Escolar Documentos
Profissional Documentos
Cultura Documentos
Series Editor
John M. Walker
School of Life Sciences
University of Hertfordshire
Hatfield, Hertfordshire, AL10 9AB, UK
Third Edition
Edited by
The field of protein NMR spectroscopy has rapidly expanded into new areas of biochemistry,
molecular biology, and cell biology research that were impossible to study as recently as 10
years ago. The potential to study macromolecular systems that were once considered too
large or too transient or too complex by using NMR spectroscopy is now being realized
with the development of innovative technologies. Standard NMR technologies are also get-
ting a facelift in part due to the pervasive nature of high-throughput approaches in bio-
chemical and biomedical research. These advances warrant a new edition of Protein NMR
Techniques that includes an authoritative but down-to-earth description of new methodolo-
gies. This edition consists of 24 chapters divided into four major categories: NMR sample
preparation, solution NMR methodologies, solid-state NMR methodologies, and data pro-
cessing. The material presented contains enough detail for use not only in specialized NMR
laboratories, but in biochemical, molecular, and cell biology research labs that have access
to high-field NMR spectrometers.
Preparing proteins for NMR spectroscopy can be a time-consuming process that may
take longer than data collection and analysis combined. Expression in bacterial cells still
remains one of the most popular ways of preparing NMR samples. Chapter 1 discusses new
methods for optimizing and increasing the production of isotope-labeled protein in bacte-
ria. However, some proteins are difficult to express in bacteria, in these cases an alternative
approach involves using yeast cells. Chapter 2 describes a methodology for producing pro-
teins in yeast that are usually secreted into the growth medium. This technique is proving
to be as robust and economic as bacterial production. One drawback to using proteins
secreted by yeast is that they may exhibit altered patterns of glycosylation and phosphoryla-
tion. To avoid this problem and achieve proper posttranslational modifications, proteins are
best produced in insect or mammalian cells. Advances in the use of these cells for producing
NMR samples are detailed in Chapters 3 and 4. Cell-free expression of proteins has become
a method of choice for high-throughput protein production especially in cases, where the
yields from in vivo overexpression are very low. Cell-free systems allow for the selective
incorporation of any isotope-labeled amino acid into a target protein with minimal scram-
bling. Chapters 5 and 6 describe the cell-free production of proteins for solution and solid-
state NMR, respectively. Finally, although well-expressed in bacterial cells, some soluble
proteins do not fold properly in sufficient quantity to permit analyses of structure, dynam-
ics, and interactions. Chapter 7 presents a methodology for expressing and purifying such
proteins in a cost-efficient manner.
The chapters on solution NMR methodologies range from the study of individual pro-
teins, large multidomain proteins, protein–ligand and protein–nucleic acid complexes
in vitro, to the study of proteins inside living cells. A strategy for studying supramolecular
systems, which has become possible due to advances in isotope labeling and NMR pulse
sequences, is described in Chapter 8. Chapter 9 presents basic protocols and the latest
improvements for measuring relaxation rates and analyzing protein dynamics. Methods to
help overcome difficulties in applying solution NMR to the study of membrane proteins are
v
vi Preface
relevant to studying proteins by using NMR. Rather than reiterating the fundamental
principles behind NMR methodologies, we have emphasized the practical aspects of experi-
mental design combined with practical advice and examples. We hope that this book will
provide both experienced NMR spectroscopists and biochemists, who are new to the field
of NMR, with enough background to successfully apply these techniques to their
research.
Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
ix
x Contents
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513
Contributors
ALAA ABDINE • CNRS and Université Paris Diderot, IBPC, Paris, France
JINWOO AHN • Department of Structural Biology, Pittsburgh Center for HIV Protein
Interactions, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA
JANET S. ANDERSON • Department of Chemistry, Union College, NY, USA
BENJAMIN BARDIAUX • NMR-supported Structural Biology, Leibnitz-Institut
für Molekulare Pharmakologie (FMP), Berlin, Germany
IAN C. BRETT • Department of Biochemistry and Cell Biology, Stony Brook University,
Stony Brook, NY, USA
CHRIS A. BROSEY • Departments of Biochemistry and Chemistry, Center for Structural
Biology, Vanderbilt University, Nashville, TN, USA
BERNHARD BRUTSCHER • Institut de Biologie Structurale – Jean-Pierre Ebel,
CNRS, CEA, UJF, UMR5075, Grenoble Cedex, France
IN-JA L. BYEON • Department of Structural Biology, Pittsburgh Center for HIV Protein
Interactions, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA
MARIE-EVE CHAGOT • Departments of Biochemistry and Chemistry,
Center for Structural Biology, Vanderbilt University, Nashville, TN, USA
WALTER J. CHAZIN • Departments of Biochemistry and Chemistry,
Center for Structural Biology, Vanderbilt University, Nashville, TN, USA
JIANGLEI CHEN • Department of Biochemistry and Molecular Biology,
School of Medicine, Wayne State University, Detroit, MI, USA
JOLYON K. CLARIDGE • Department of Biochemistry, University of Oxford, Oxford, UK
ROBERT T. CLUBB • Department of Chemistry and Biochemistry,
University of California, Los Angeles, CA, USA
DAVID COWBURN • Departments of Biochemistry and Physiology & Biophysics,
Albert Einstein College of Medicine of Yeshiva University, Bronx, NY, USA
KEVIN N. DALBY • Division of Medicinal Chemistry, University of Texas,
Austin, TX, USA; Graduate Programs, Cellular and Molecular Biology, Pharmacy,
Biomedical Engineering and Biochemistry, University of Texas, Austin, TX, USA
ARPANA DUTTA • Department of Structural Biology, University of Pittsburgh
School of Medicine, Pittsburgh, PA, USA
FABIEN FERRAGE • Département de chimie, Ecole normale supérieure et
Laboratoire des Biomolécules, CNRS UMR 7203, Paris, Cedex, France
DAVID FUSHMAN • Department of Chemistry and Biochemistry and Center for
Biomolecular Structure and Organization, University of Maryland, MD, USA
MARGARIDA GAIRÍ • Servicios Cientifico Tecnicos, Universitat de Barcelona,
Barcelona, Spain
RANAJEET GHOSE • Department of Chemistry, The City College of New York,
New York, NY, USA; Graduate Center of the City University of New York,
New York, NY, USA
xi
xii Contributors
Abstract
The Gram-negative bacterium Escherichia coli offer a means for rapid, high-yield, and economical
production of recombinant proteins. However, when preparing protein samples for NMR, high-level pro-
duction of functional isotopically labeled proteins can be quite challenging. This is especially true for the
preparation of triple-labeled protein samples in D2O (2H/13C/15N). The large expense and time-consuming
nature of triple-labeled protein production for NMR led us to revisit the current bacterial protein expres-
sion protocols. Our goal was to develop an efficient bacterial expression method for very high-level pro-
duction of triple-labeled proteins that could be routinely utilized in every NMR lab without changing
expression vectors or requiring fermentation. We developed a novel high cell-density IPTG-induction
bacterial expression method that combines tightly controlled traditional IPTG-induction expression with
the high cell-density of auto-induction expression. In addition, we optimize several key experimental
protocols and parameters to ensure that our new high cell-density bacterial expression method routinely
produces 14–25 mg of triple-labeled proteins and 15–35 mg of unlabeled proteins from 50-mL bacterial
cell cultures.
Key words: High yield protein production, Bacterial expression, Isotopic labeling, NMR
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_1, © Springer Science+Business Media, LLC 2012
1
2 V. Murray et al.
provides the cheapest way to prepare these proteins for NMR stud-
ies (3). Protein expression and purification is a routine practice in
many NMR labs, but it is not uncommon to see a drastic reduction
in protein yield when isotopically labeling the proteins, especially
when D2O must be used.
To overcome these difficulties, we have developed a novel bac-
terial expression method that combines the tightly controlled
traditional IPTG-induction bacterial expression with the high cell-
density of the auto-induction method (4). To summarize our pro-
cedure, we first determine how to make a proper starting culture,
followed by double colony selection and finally high cell-density
expression. With these optimized protocols and parameters, our
new bacterial expression method offers near gram quantity
production of triple-labeled proteins from one-liter bacterial cell
cultures, without changing expression vectors and without fermen-
tation (5). Thus, every NMR laboratory can easily apply this novel
bacterial expression method on a routine basis for the production
of a very high yield of triple-labeled proteins for their NMR structural
studies of proteins.
2. Materials
3. Methods
3.1. PAGE Sample PAGE samples come either from the cell lysate isolated immediately
Preparation after bacterial expression or from the column flow through during
protein purification.
1. For samples collected directly from bacterial cell culture:
Collect 500 μL of cells and place in a 1.7-mL microcentrifuge
tube. Spin down at 12,000 × g for 5 min at room temperature
using a microcentrifuge, discard the supernatant, and tap out
the excess on a paper towel.
2. Add 25 μL of 4× SDS loading buffer and 25 μL of water and
resuspend the pellet. Store the samples in a freezer until ready
to run a gel.
3. Before running the culture samples on a gel, place the samples
on a 90°C heat block for 30 min. Remove from the heat block;
add 50 μL of water and vortex for 30 s (see Note 6).
4. For samples collected from flow through during protein puri-
fication, mix 60 μL of column flow through with 20 μL of
6 V. Murray et al.
3.3. A Proper Starting A critical consideration for high-level bacterial expression is the
Culture preparation of a proper starting culture in a rich medium for scal-
ing up in minimal medium. The general practice in NMR labora-
tories is to grow an overnight culture using a rich medium, such as
LB, at 37°C. We observe that an overnight culture usually reaches
saturation by the next morning, and may result in plasmid instabil-
ity and loss due to several factors including basal leakage of the T7
expression system that expresses the toxic target proteins to the
host cells under this overgrowth condition (4, 6, 7). This usually
results in a poor yield of target protein. Figure 1 shows a growth
curve for E. coli BL21 (DE3) cells carrying the LCAT/pET30a
vector, in H2O-based LB, suggesting that the bacteria are in the
exponential or log phase of growth between 6 and 7.5 h at 37°C
(see Note 9). We placed particular emphasis on double colony
selection (Subheading 3.5) to ensure that a high percentage of
bacterial cells within this colony contain the DNA expression
plasmid.
1. Perform a time course of bacterial growth of a new protein
expression vector in rich (LB) medium by measuring the OD600
every 30 min for ~10–12 h in water-based rich medium and
~16–18 h in D2O-based rich medium (see Note 10). The OD600
Fig. 1. Plot of E. coli growth in 5 mL of LB medium over a 10-h period at 37°C starting with
a glycerol stock. The bacterial strain, BL-21(DE3), contains a pET30a vector expressing
the gene for lecithin:cholesterol acyltransferase (LCAT). Based on this plot, the log phase
of the culture is between an OD600 of 1 and 3.5. Note: The growth curve is vector, protein,
and bacterial strain dependent. Reproduced from Murray 2010 with permission from Cold
Spring Harbor.
8 V. Murray et al.
of the log phase and the time to reach saturation are vector,
protein, and bacterial strain-dependent; therefore, we suggest
performing this experiment before actually expressing protein
for purification and labeling (see Note 11).
2. Once the optimal OD600 of the log phase of the starting cul-
ture is determined, this will be the OD600 for all future starting
cultures using both traditional IPTG expression and high cell-
density expression methods.
3.4. Traditional IPTG This method can be scaled up or down to suit your needs. This
Method following protocol is used to make a 100-mL expression culture.
1. Prepare a 5-mL proper starting culture as described in
Subheading 3.3 in a 50-mL conical tube with holes poked in
the lid.
2. Prepare 100 mL of M9 minimal medium in a 250-mL flask.
Add 1.5 mL of the starting culture and measure the OD600.
(To check the OD600 at this point, aliquot 1 mL of the cell
culture into a cuvette and measure the OD600.) We suggest a
starting OD600 between 0.05 and 0.10 for healthy bacterial cell
growth. Place the flask in a 37°C incubator with a shaking
speed of 200 rpm.
3. Start to monitor the OD600 after 3 h (to check the OD600 at this
point, dilute 100 μL of cell culture into a cuvette containing
900 μL of distilled water and measure the OD600. The final OD600
value is ten times the spectrophotometer reading). Once the
OD600 reaches between 0.8 and 1.2 (see Note 12), remove 2 mL
of cell culture and place it in a 15-mL culture tube (for a non-
induced reference). Induce the remaining culture with 0.5 mM
IPTG. Place both the flask and the 15-mL culture tube in a 20°C
incubator overnight with a shaking speed of 200 rpm.
4. The following morning, measure the final OD600 of the cell
culture. If the protein expression is induced at an OD600 of 1,
the OD600 of the bacterial cell culture will be around 2–3, indi-
cating healthy bacterial cell growth. Harvest the cells by spin-
ning down at 10,000 × g for 10 min at 4°C using a benchtop
centrifuge. Remove the supernatant and either store the cell
pellet at −80°C or use immediately for protein purification.
5. To check protein expression levels, take 500 μL samples of
each culture (non-induced and IPTG-induced), follow the
sample preparation protocol (Subheading 3.1) and run an
SDS-PAGE (Subheading 3.2) to compare non-induced and
IPTG-induced samples.
3.5. Double Colony Since plasmid loss is encountered during bacterial expression in D2O
Selection much more frequently than in H2O, we describe a double colony
selection procedure for triple-labeling protein in D2O. Based on our
1 A Novel Bacterial Expression Method with Optimized Parameters… 9
Fig. 2. SDS-PAGE of protein expression of apoE(1–215)/pTYB1 in D2O before (a), during (b) and after (c) double colony
selection. Arrows indicate the expected protein band (~80 kDa). (a) Shows four different colonies before colony selection.
(b) Shows the results of three different colonies selected from the single colony selection (lanes 1–3) and another three
colonies selected from the double colony selection (lanes 4–6). The second colony selection was based on Colony
3 (lane 3, b). (c) Shows the results of six colonies from the double colony selection, indicating a high level of protein expres-
sion for all six colonies. Reproduced from Sivashanmugam 2009 with permission from Wiley Interscience.
5. When the OD600 of the culture reaches between 0.8 and 1 (see
Note 14), add 0.5 mM IPTG to eight of the nine cultures to
induce protein expression, clearly marking which culture is
serving as a negative control. Place the tubes in a 20°C incuba-
tor overnight with a shaking speed of 200 rpm.
6. The following morning, remove 500 μL of cell culture from
each tube and place in a microcentrifuge tube. Centrifuge at
12,000 × g for 10 min at room temperature and discard the
supernatant. Prepare SDS-PAGE samples and run a gel.
7. Choose the colony expressing the biggest protein band (using
the negative control as reference) and prepare 5 mL of LB cul-
ture in either 70 or 99% D2O and 5 μL of antibiotic containing
the colony from the master plate. Grow at 37°C until the
OD600 reaches 0.7–0.9 and spread 150 μL on an LB plate pre-
pared in 50% D2O. Invert the plate and incubate it at 37°C
overnight.
8. The following day, repeat steps 3–6 using 70 or 99% D2O for
the second round of colony selection (see Note 14). When
completed, choose the colony expressing the biggest protein
band to make a 5-mL LB culture in 100% D2O and 5 μL of
antibiotic. Grow the culture halfway through its log phase.
Add 800 μL of culture to 200 μL of 100% sterile glycerol in a
2-mL cryogenic tube with a screw top cap. Pipet up and down
to mix thoroughly and flash freeze the tubes by dipping them
in liquid nitrogen, store at −80°C. If the protein expression
level is extremely high, we suggest that you make at least 5–10
glycerol stocks of this double selected colony for future use.
3.7. Optimization Another important step for high-level protein production using
of Various Conditions high cell-density bacterial expression is to optimize the expression
conditions such as culture temperature and the induction time.
These steps are critical for the initial expression of a protein using
the high cell-density expression method. We usually use the tradi-
tional IPTG-induction method first to check if a new protein can
be expressed by using bacteria. Once the protein expression is con-
firmed by the traditional IPTG-induction method, we can opti-
mize the high cell-density expression method to produce high-yield
isotopically labeled proteins.
12 V. Murray et al.
Fig. 3. Left panel: An SDS-PAGE showing auto-induction time course of triple-labeled human apoAI expression in D2O at
room temperature. Lanes 1–7 = 24, 28, 32, 36, 40, 44, and 54 h, respectively. Right panel is a Western blot of the same
time course using anti-apoAI monoclonal antibody. Reproduced from Sivashanmugam 2009 with permission from Wiley
Interscience.
Table 1
Parameters of the time course of human apolipoprotein A–I expression
Time 24 h 28 h 32 h 36 h 40 h 44 h 54 h
3.8. Protein Protein purification depends on the fusion tag that is used, as dif-
Purification ferent tags serve different purposes. In our laboratory, we generally
use histidine tags. In this section, we describe a typical protein
purification procedure using a His-Bind Resin column.
1. Prepare 1× dilutions of all buffers. Recheck the pH to ensure
that they are still 7.9.
2. Centrifuge the cell culture at 10,000 × g for 10 min at 4°C.
Remove the supernatant and resuspend the pellets in 20 mL of
1× binding buffer. If the protein is in inclusion bodies and can
be refolded readily during dialysis, you can resuspend the
pellet in 20 mL of 1× binding buffer containing 6 M urea.
3. Lyse the cells by using either sonication or a French press.
Centrifuge the lysate at 16,000 × g for 20 min at 4°C. Collect
the supernatant and store it on ice.
4. Add 10 mL of 1× binding buffer and repeat step 2 at least twice,
and then combine all the supernatants. Depending on the pro-
tein, you may need to add additional binding buffer and repeat
step 2 3–5 times to completely extract the protein from the cells.
5. Equilibrate the affinity column with 50 mL of 1× charge buffer
(see Note 16). Remove the charge buffer and equilibrate the
column with 50 mL of 1× binding buffer. The column should
be a light blue color after equilibration.
6. Load the column with the clear lysate from step 3. The flow
rate should be ~1 mL/minute (see Note 17). Collect the flow
through and remove a 60-μL sample for SDS-PAGE.
7. Wash the column with 200 mL of 1× binding buffer, followed
by an additional 100 mL of 1× wash buffer. The flow rate of
wash buffer should also be ~1 mL/min. Elute the column with
100 mL of 1× elution buffer. Collect the last drop of elution
for SDS-PAGE to make sure that all of the protein has been
eluted from the column.
8. Perform SDS-PAGE analysis with all of the collected samples
to assess the purification.
9. Place the eluted protein into a dialysis bag and dialyze exten-
sively against water containing 20 mM (NH4)2CO3 to remove
imidazole, salts and possibly urea. After dialysis, freeze the pro-
tein sample with liquid nitrogen and lyophilize to obtain pure
triple-labeled protein powder. Run a gel to assess the purity of
the protein powder.
Table 2
Final yields of unlabeled and triple labeled proteins: high cell-density vs. traditional
IPTG method
High cell
Protein densityb (mg) IPTGb (mg) M.W. (Cal) (Da) M.W. (MS) (Da) %Dc
Triple-labeled
RAP(1–210) 20 ± 3 0.5 33,801 33,525 ± 195 ~92
RAP(91–323) 25 ± 3 0.8 36,633 36,376 ± 200 ~93
ApoE(1–183)a 18 ± 4 2 22,866 22,686 ± 116 ~89
Mouse apoAI(1–216) 15 ± 2 0.8 28,014 27,732 ± 125 ~90
Human apoAI 14 ± 1 0.6 32,814 32,401 ± 150 ~88
Unlabeled
Human apoAI 34 ± 1 1
Human apoE 17 ± 2 0.2
M.W. molecular weight
Reproduced from Sivashanmugam 2009 with permission from Wiley Interscience
a
ApoE(1–183) was expressed in 40% D2O, the rest are expressed in 99.7% D2O
b
High cell density (50-mL culture volume): high cell-density expression methods, including auto-induction and high
cell-density IPTG-induction; IPTG: the optimized traditional IPTG-induced expression. We repeated the expressions
at least three times for all proteins, the yield shown is the average ± standard deviation
c
Estimated percentage of deuteration, assuming 100% 13C and 15N-labeling. For apoE(1–183), the %D is the estimated
percentage of deuteration based on 40% D2O. For the other four proteins, the %D is the estimated percentage of deu-
teration based on 99.7% D2O
4. Notes
1. The stain can be reused multiple times. When using fresh stain,
you only need to stain gels for 15–30 min. Pour the used stain
into a separate container. When reusing stain, you may need to
stain gels longer.
16 V. Murray et al.
References
1. Swartz, J.R. (2001) Advances in Escherichia protocols for production of very high yields of
coli production of therapeutic proteins. Curr. recombinant proteins using Escherichia coli.
Opin. Biotechnol. 12, 195–201. Protein Sci. 18, 936–948.
2. Hewitt, L., and McDonnell, J.M. (2004) 6. Chen, H.C., Hwang, C.F., and Mou, D.G.
Screening and optimizing protein production in (1992) High-density Escherichia coli cultiva-
E. coli methods. Methods Mol. Biol. 278, 1–16. tion process for hyperexpression of recombi-
3. McIntosh, L.P. and Dahlquist, F.W. (1990) nant porcine growth hormone. Enzyme Microb.
Biosynthetic incorporation of 15N and 13C for Technol. 14, 321–326.
assignment and interpretation of nuclear mag- 7. Baneyx, F. (1999) Recombinant protein
netic resonance spectra of proteins. Q Rev. expression in Escherichia coli. Curr. Opin.
Biophys. 23, 1–38. Biotechnol. 10, 411–421.
4. Studier, F.W. (2005) Protein production by 8. Sambrook, J. and Russell, D. (2001) Mole-
auto-induction in high density shaking cul- cular Cloning: A Laboratory Manual (3rd ed.).
tures. Protein Expr. Purif. 41, 207–234. Cold Spring Harbor Laboratory Press, Cold
5. Sivashanmugam, A., Murray, V., Cui, C., Spring Harbor, New York (ISBN 978-
Zhang, Y., Wang, J., and Li, Q. (2009) Practical 087969577-4).
Chapter 2
Abstract
Several protein expression systems are available for the preparation of stable isotope-labeled recombinant
proteins for NMR studies. Yeast expression systems have several advantages over prokaryotic systems, such
as the widely used Escherichia coli expression system. Protein expression using the methylotrophic yeast
Pichia pastoris is commonly employed for the preparation of isotope-labeled proteins. Recently, the hemi-
ascomycete yeast Kluyveromyces lactis expression system was reported as being useful for preparing proteins
for NMR studies. Since each yeast expression system has different features, their applications have increased
in number. In this chapter, we describe procedures for the efficient production of uniformly isotope-labeled
proteins using the P. pastoris and the K. lactis yeast expression systems.
Key words: Yeast expression systems, Pichia pastoris, Kluyveromyces lactis, Stable isotope labeling,
Fed-batch fermentation, NMR
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_2, © Springer Science+Business Media, LLC 2012
19
20 T. Sugiki et al.
1.1. The P. pastoris Host cells and expression vectors are available from Invitrogen.
Expression System Based upon Invitrogen instruction manuals (7), previous reports
(8–15), and our experience, we describe here optimized cell culture
procedures for the production of isotope-labeled heterologous
proteins by P. pastoris.
P. pastoris is capable of utilizing methanol as both a carbon
source and to induce protein expression. Methanol is oxidized by alco-
hol oxidases, AOX1 and AOX2, in yeast cell peroxisomes (though
the majority of alcohol oxidase activity is attributed to AOX1). The
AOX1 promoter tightly regulates expression of the AOX1 gene,
and methanol induces AOX1 promoter activity. Thus, the AOX1
promoter is used to drive the expression of a target protein
by replacing the AOX1 gene with a cDNA encoding the desired
heterologous protein (1–6).
Invitrogen supplies several vector series that are commonly
used for protein expression in P. pastoris, including pPIC9K,
pPIC3.5K, and pPICZ. Using these vectors, a cDNA cassette con-
taining the target gene under the control of the AOX1 promoter is
inserted into the genome of P. pastoris by using homologous
recombination along with a gene coding for resistance to a drug,
such as Zeocin™ or G418, for subsequent selection of transformed
cells (7).
Target proteins are expressed either intracellularly or secreted
into the medium by P. pastoris. For secretion of target proteins,
vectors pPICZα and pPIC9K are recommended (7). The pPICZα
and pPIC9K series contain a gene that encodes a prepro signal
sequence [such as the Saccharomyces cerevisiae α-mating factor
(α-MF)] between the AOX1 promoter and the target gene (7).
Although target proteins may contain native secretion signals, in
many cases, the α-MF sequence is generally used as the sole secretion
signal. Since P. pastoris secretes a low amount of its native proteins,
a secreted target protein comprises the vast majority of the total
proteins in the culture medium, simplifying purification of the
target protein (1–7). A different set of vectors lacking secretion signal
sequences are required if the target protein is to be expressed intra-
cellularly. For cytosolic and nonglycosylated proteins, vectors pPICZ
or pPIC3.5K are recommended (7).
In this chapter, we describe methods and provide technical
advice for producing uniformly labeled [U-13C, 15N] and [U-2H, 15N]
target proteins in P. pastoris (8–14). The procedures described are
2 Isotopic Labeling of Heterologous Proteins in the Yeast… 21
1.2. The K. lactis Host cells and expression vectors are available from New England
Expression System Biolabs. Based upon instruction manuals (16), previous reports,
and our experience, we describe here optimized cell culture proce-
dures for the production of isotope-labeled heterologous proteins
by K. lactis. Heterologous recombinant proteins expressed by
K. lactis can be sequestered intracellularly or secreted into the
medium. As with P. pastoris, K. lactis secretes very low levels of native
proteins, therefore secreted target proteins comprise the vast
majority of the total protein present in the culture medium, which
simplifies purification of the target protein (16–18).
In K. lactis, the LAC4 promoter drives expression of the target
protein gene. Using the pKLAC vector, a cDNA cassette containing
the target gene under the control of the LAC4 promoter is inserted
into the genome of K. lactis by using homologous recombination
(16–18). The transcriptional activity of the LAC4 promoter is
induced by galactose, which K. lactis also utilizes as a carbon source
for cell growth. Therefore, expression of the target protein is con-
stitutively induced by cultivation in a medium containing galactose
(16–18).
The pKLAC vectors impart a fungal acetamidase gene (amdS)
to the transformants. Since only transformants that express amdS
can utilize acetamide as a sole nitrogen source, amdS is used for
selection of the desired transformants, thereby obviating the need
for expensive antibiotics, such as Zeocin™ and rendering this
auxotrophic selection method more cost-effective (16–18). In addi-
tion, auxotrophic acetamide selection enriches populations of trans-
formants in which multiple tandem copies of target cDNA are
integrated into the yeast genome (16–18). Thus, the K. lactis
protein expression system is simple, cost-effective, easily scaled-up,
and highly reproducible.
We recently established a cost-effective isotope-labeling method
utilizing the hemiascomycete yeast K. lactis (6, 19). In the most
commonly employed K. lactis expression system, 20 g/L galactose
is required as a carbon source for cell growth and as an inducer of
target protein expression (16). However, while a 20 g/L carbon
source is acceptable for uniform 15N-labeling, it is economically
infeasible for uniform 13C-labeling. Sugiki et al. reported that by
combining K. lactis strain GG799, which is characterized by weak
glucose suppression for the LAC4 promoter (16), with a fed-batch
culture method, larger amounts of protein can be expressed using a
smaller amount of glucose, thus reducing costs to a level compa-
rable to isotope-labeling by E. coli systems (6, 19). In this chapter,
we describe methods and provide technical advice for producing
uniformly labeled [U-13C, 15N] and [U-2H, 15N] target proteins in
K. lactis.
22 T. Sugiki et al.
2. Materials
The recipes for the culture media described here were derived
mainly from the manuals for the EasySelect™ Pichia Expression Kit
(Invitrogen), the Pichia Fermentation Process Guidelines (Invitrogen),
and the K. lactis Protein Expression Kit (New England Biolabs).
Autoclave sterilization is performed at 121°C for 15 min.
2.1. Uniform 13C, 1. 40% unlabeled D-glucose stock solution: Dissolve 200 g of
15
N-Labeling Using D-glucose in 1 L of water. Sterilize by aseptic filtration and store
P. pastoris at room temperature. The shelf life of this solution is approxi-
mately 1 year.
2. 100 mg/mL Stock solution of Zeocin™ (Invitrogen).
3. Yeast extract peptone dextrose sorbitol (YPDS) agar plates:
Dissolve 20 g of bacto peptone, 10 g of yeast extract, 182.2 g
of sorbitol, and 20 g of bacto agar in 950 mL of water. Sterilize
by autoclaving and cool to 50–60°C. Aseptically add 1 mL
of 100 mg/mL Zeocin™ and 50 mL of 40% glucose stock
solution, gently mix well, and dispense into sterile disporsable
Petri dishes. Store at 4°C in the dark. The shelf life of these
agar plates is 1–2 weeks.
4. 10× Yeast nitrogen base (YNB): Dissolve 34 g of YNB without
amino acids and ammonium sulfate in 1 L of water (see Note 1).
Warm it to 40–50°C to dissolve completely if needed. Sterilize
by aseptic filtration and store at 4°C. The shelf life of this solu-
tion is approximately 1 year.
5. 1 M Potassium phosphate: Mix 132 mL of 1 M K2HPO4 and
868 mL of 1 M KH2PO4. Adjust the pH to 6.0 ± 0.1 using
KOH or phosphoric acid. Sterilize by autoclaving and store at
room temperature. The shelf life of this solution is greater than
1 year.
6. 500× Biotin: Dissolve 20 mg of biotin in 100 mL of water and
warm it to 40–50°C to dissolve completely. Sterilize by aseptic
filtration and store at 4°C. The shelf life of this solution is
approximately 1 year.
7. 10% (w/v) Glycerol: Dissolve 50 mL of glycerol in 450 mL of
water. Sterilize by aseptic filtration and store at room tempera-
ture. The shelf life of this solution is over 1 year.
8. 200× PTM1 trace salts: Mix together the following ingredients
and dissolve to a final volume of 1 L in water: 6.0 g of
CuSO 4⋅5H2O, 0.08 g of NaI, 3.0 g of MgSO4⋅H2O, 0.2 g
of Na2MoO4⋅2H2O, 0.02 g of boric acid, 0.5 g of CoCl2, 20.0 g
of ZnCl2, 65.0 g of FeSO4⋅7H2O, 0.2 g of biotin, 5.0 mL of
H2SO4. Warm to 40–50°C to dissolve completely if needed.
Sterilize by aseptic filtration and store at 4°C or room temperature.
The shelf life of this solution is approximately 3 months at 4°C.
2 Isotopic Labeling of Heterologous Proteins in the Yeast… 23
2.3. Uniform 13C, 1. Yeast peptone (YP): Dissolve 20 g of bacto peptone and 10 g
15
N-Labeling Using of yeast extract in 950 mL of water. Sterilize by autoclaving
K. lactis and store at room temperature or 4°C. The shelf life of this
solution is approximately 1 year.
2 Isotopic Labeling of Heterologous Proteins in the Yeast… 25
3. Methods
13
C-methanol (0.5 % (w/v) of the total volume) to the culture
medium every 24 h during this induction phase (20) (see
Note 11). The total amount of 13C-methanol required for suf-
ficient protein expression in a cost-effective manner is approx-
imately 10–20 mL per 1 L culture medium (see Note 12).
8. Pellet the cells by centrifuging at 6,000 × g for 20 min at 4°C.
If the target protein is secreted into the culture medium, filter
the supernatant by passing it through a < 0.45-μm pore size
membrane and collect the filtrate. If needed, add appropriate
protease inhibitors to the filtrate to prevent proteolytic degra-
dation of secreted target proteins. If the target protein is
expressed in the cytoplasm of the host cells, discard the super-
natant after centrifugation and retain the cell pellet.
3.2. Uniform 2H, Deuterium isotope-labeling of target proteins is one of the most
15
N-Labeling Using important techniques for protein NMR studies, especially for analyses
P. pastoris of large molecular weight (MW > 25 K) proteins and in cross-
saturation experiments to identify intermolecular-binding sites
(6, 21–24). The P. pastoris expression system can be used to over-
express deuterium-labeled heterologous proteins.
For the efficient production of deuterated target proteins, cells
should be adapted to grow in deuterated broth medium. Adaptation
is achieved by multistage subculturing in which the deuterium
concentration is raised in a stepwise fashion. For instance, subcul-
turing of cells is performed using semideuterated (25–95% 2H2O)
medium, with subsequent culturing in fully deuterated medium
(14) (see Note 4). In the case of uniform 2H-labeling, using a
2
H-labeled carbon source (such as [2H4]methanol) only during the
induction phase is insufficient for producing fully deuterated target
proteins (a considerable amount of protons remain on methyl and
Lys δ/ε groups) (14). To achieve the nearly-complete deuteration
level required for cross-saturation experiments, a deuterium-labeled
carbon source, such as D-[2H7]glucose should also be used during
the cell growth phase (prior to the induction phase) (7, 14). In this
section, we describe a cultivation procedure for overexpressing
perdeuterated heterologous proteins in P. pastoris. Using this pro-
cedure, Ichikawa and coworkers successfully prepared a discoidin
domain of DDR2 that was more than 95% deuterated, and clearly
identified the collagen-binding site of DDR2 through transferred
cross-saturation experiments (7).
1. Prepare necessary media:
(a) Autoclave a 1-L baffled flask and a 2-L fermentation vessel,
and dry them in an oven at 60–80°C (for at least 48 h) to
completely remove residual H2O.
(b) Prepare 1.3 L of 15N-BM(2H2O) medium in a 2-L media
bottle, 10 mL of 10% unlabeled D-glucose(2H2O), and
15 mL of 10% D-[2H7]glucose(2H2O) solutions.
2 Isotopic Labeling of Heterologous Proteins in the Yeast… 29
3.3. Uniform 13C, 1. Prepare 0.5 L of sterile 13C, 15N-BMD and 1.0 L of 13C, 15N-BMD
15
N-Labeling Using feeding medium in a 2-L fermentation vessel and 1-L media
K. lactis bottle, respectively.
2. Using a sterile loop, scoop a small aliquot of K. lactis transfor-
mants from a frozen glycerol stock and streak it onto a YCB-
acetamide agar plate. Incubate the plate for 24–48 h at 30°C.
3. To produce the primary culture, inoculate a fresh, single colony
of K. lactis transformants into 5 mL of YPD medium (in a 50-mL
Corning tube) and shake at 200–250 rpm at 30°C for 48 h to
obtain a saturated biomass (OD600 > 20–30).
4. Assemble the fermentation system as described in
Subheading 3.1, step 6. Connect the media bottle containing
1 L of fresh 13C, 15N-BMD feeding medium to the inlet port of
the vessel with an appropriate length of feeding tube, and
attach a peristaltic pump at the midpoint of the feeding tube
(Fig. 1).
5. Pellet the primary culture cells by centrifuging at 3,000 × g for
5 min at 20°C and discard the supernatant. Gently resuspend
the pellet with 30 mL of fresh, sterile 13C, 15N-BMD medium
from the 2-L fermentation vessel, and pour the resuspended
cells into the remaining 13C, 15N-BMD medium in the fermen-
tation vessel. Agitate the culture medium at 600–800 rpm with
0.1–0.3 L/min feeding air at 30°C (see Notes 9, 10, and 12).
6. During the fermentation, continuously feed fresh, sterile 13C,
15
N-BMD feeding medium into the fermentation vessel at a
constant flow rate of 8.3 mL/h using the peristaltic pump (see
Note 14).
7. Pellet the cells by centrifuging at 10,000 × g for 15 min at 4°C.
If the target protein is secreted into the culture medium, filter
the supernatant by passing it through a < 0.45-μm pore size
membrane and collect the filtrate. If needed, add appropriate
protease inhibitors to the collected filtrate to prevent proteolytic
degradation of the target protein. If the target protein is expressed
in the cytoplasm of the host cells, discard the supernatant after
centrifugation, and retain the cell pellet.
Fig. 1. Overview of the fed-batch fermentation system. (a) Schematic diagram of fed-batch
fermentation. (b) Photograph of the assembled fed-batch fermentation device. a: fresh
media which is continuously fed into the fermentation vessel, b: peristaltic pump, c: fermentation
vessel, d: air flow inlet, and e: air exhaust port.
4. Notes
Acknowledgments
References
1. Cregg, J. M., Vedvick, T. S., and Raschke, W. lysozyme where phosphate ion binds using
C. (1993) Recent advances in the expression of NMR measurements. FEBS Lett. 448, 33–37.
foreign genes in Pichia pastoris. Biotechnology 12. van den Burg, H. A., de Wit, P. J., and Vervoort,
11, 905–910. J. (2001) Efficient 13C/15N double labeling of
2. Pickford, A. R., and O’Leary, J. M. (2008) the avirulence protein AVR4 in a methanol-
Isotopic labeling of recombinant proteins from utilizing strain (Mut+) of Pichia pastoris. J.
the methylotrophic yeast Pichia pastoris, in Biomol. NMR 20, 251–261.
Methods in Molecular Biology (Downing, A. K., 13. Rodriguez, E., and Krishna, N. R. (2001)
Ed.), vol. 278, pp. 17–33, Humana Press, An economical method for 13C/15N isotopic
Totowa, NJ. labeling of proteins expressed in Pichia pastoris.
3. Lin Cereghino, G. P., Cereghino, J. L., Ilgen, J. Biochem. 130, 19–22.
C., and Cregg, J. M. (2002) Production of 14. Morgan, W. D., Kragt, A., and Feeney, J. (2000)
recombinant proteins in fermenter cultures of Expression of deuterium-isotope-labelled pro-
the yeast Pichia pastoris. Curr. Opin. Biotechnol. tein in the yeast Pichia pastoris for NMR study.
13, 329–332. J. Biomol. NMR 17, 337–347.
4. Daley, R., and Hearn, M. T. (2005) Expression 15. Ichikawa, M., Osawa, M., Nishida, N.,
of heterologous proteins in Pichia pastoris : Goshima, N., Nomura, N., and Shimada, I.
a useful experimental tool in protein engi- (2007) Structural basis of the collagen-binding
neering and production. J. Mol. Recognit. 18, mode of discoidin domain receptor 2, EMBO J.
119–138. 26, 4168–4176.
5. Cos, O., Ramόn, R., Montesinos, J. L., and 16. New England Biolabs, Inc. K. lactis Protein
Valero F. (2006) Operational strategies, moni- Expression Kit: Instruction manual. Available at
toring and control of heterologous protein http://www.neb.com/nebecomm/ManualFiles/
production in the methylotrophic yeast Pichia manualE1000.pdf.
pastoris under different promoters: A review. 17. Colussi, P. A., and Taron, C. H. (2005)
Microbial Cell Factories 5, 17–36. Kluyveromyces lactis LAC4 promoter variants
6. Takahashi, H., and Shimada, I. (2010) Production that lack function in bacteria but retain full
of isotopically labeled heterologous proteins in function in K. lactis, Appl. Environ. Microbiol.
non-E. coli prokaryotic cells. J. Biomol. NMR 71, 7092–7098.
46, 3–10. 18. Read, J. D., Colussi, P. A., Ganatra, M. B., and
7. Invitrogen Corporation. EasySelect™ Pichia Taron, C. H. (2007) Acetamide selection of
Expression Kit: A manual of methods for expres- Kluyveromyces lactis cells transformed with an
sion of recombinant proteins using pPICZ and integrative vector leads to high-frequency for-
pPICZα in Pichia pastoris, Version I. Available mation of multicopy strains. Appl. Environ.
at http://tools.invitrogen.com/content/sfs/man- Microbiol. 73, 5088–5096.
uals/easyselect_man.pdf. 19. Sugiki, T., Shimada, I., and Takahashi, H.
8. Laroche, Y., Strome, V., De Meutter, J., (2008) Stable isotope labeling of protein by
Messens, J., and Lauwereys, M. (1994) High- Kluyveromyces lactis for NMR study. J. Biomol.
level secretion and very efficient isotopic labeling NMR 42, 159–162.
of tick anticoagulant peptide (TAP) expressed 20. Invitrogen Corporation. Pichia Fermentation
in the methylotrophic yeast, Pichia pastoris. Process Guidelines. Available at http://toolsja.
Bio/Technol 12, 1119–1124. invitrogen.com/content/sfs/manuals/pichia-
9. Denton, H., Smith, M., Husi, H., Uhrin, D., ferm_prot.pdf.
Barlow, P. N., Batt, C. A. , and Sawyer, L. (1998) 21. Takahashi, H., Nakanishi, T., Kami, K., Arata,
Isotopically labeled bovine β-lactoglobulin for Y., and Shimada, I. (2000) A novel NMR
NMR studies expressed in Pichia pastoris. method for determining the interface of large
Protein Expr. Purif. 14, 97–103. protein-protein complexes. Nat. Struct. Biol. 7,
10. Wood, M. J., and Komives, E. A. (1999) 220–223.
Production of large quantities of isotopically 22. Nakanishi, T., Miyazawa, M., Sakakura, M.,
labeled protein in Pichia pastoris by fermenta- Terasawa, H., Takahashi, H., and Shimada, I.
tion.. J. Biomol. NMR 13, 149–159. (2002) Determination of the interface of a large
11. Mine, S., Ueda, T., Hashimoto, Y., Tanaka, Y., protein complex by transferred cross-saturation
and Imoto, T. (1999) High-level expression of measurements. J. Mol. Biol. 318, 245–249.
uniformly 15N-labeled hen lysozyme in Pichia 23. Takahashi, H., Miyazawa, M., Ina, Y., Fukunishi,
pastoris and identification of the site in hen Y., Mizukoshi, Y., Nakamura, H., and Shimada,
36 T. Sugiki et al.
I. (2006) Utilization of methyl proton reso- Pichia pastoris fermentation. Biotechnol. Bioeng
nances in cross-saturation measurement for 72, 1–11.
determining the interfaces of large protein-pro- 33. Cai, M., Huang, Y., Sakaguchi, K., Clore, G.
tein complexes. J. Biomol. NMR 34, 167–177. M., Gronenborn, A. M., and Craigie, R. (1998)
24. Shimada, I., Ueda, T., Matsumoto, M., Sakakura, An efficient and cost-effective isotope labeling
M., Osawa, M., Takeuchi, K., Nishida, N., and protocol for proteins expressed in Escherichia
Takahashi, H. (2008) Cross-saturation and coli. J. Biomol. NMR 11, 97–102.
transferred cross-saturation experiments. Prog. 34. Zhang, Y., Liu, R., and Wu, X. (2007) The proteolytic
Nucl. Magn. Reson. Spectrosc. 54, 123–140. systems and heterologous proteins degradation
25. Koch, V., Rüffer, H.-M., Schügerl, K., in the methylotrophic yeast Pichia pastoris.
Innertsberger, E., Menzel, H., and Weis, J. Annal. Microbiol. 57, 553–560.
(1995) Effect of antifoam agents on the medium 35. Li, Z., Xiong, F., Lin, Q., d’Anjou, M., Daugulis,
and microbial cell properties and process per- A. J., Yang, D. S., and Hew, C. L. (2001) Low-
formance in small and large reactors. Process temperature increases the yield of biological
Biochem. 30, 435–446. active herring antifreeze protein in Pichia pastoris.
26. Macauley-Patrick, S., Fazenda, M. L., McNeil, Protein Expr. Purif. 21, 438–445.
B., and Harvey, L. M. (2005) Heterologous 36. Jahic, M., Gustavsson, M., Jansen, A.-K., Martinelle,
protein production using the Pichia pastoris M., and Enfors, S.-O. (2003) Analysis and
expression system. Yeast 22, 249–270. control of proteolysis of a fusion protein in
27. Shapiro, R. I., Wen, D., Levesque, M., Pichia pastoris fed-batch processes. J. Biotechnol.
Hronowski, X., Gill, A., Garber, E. A., Galdes, 102, 45–53.
A., Strauch, K. L., and Taylor, F. R. (2003) 37. Jahic, M., Wallberg, F., Bollok, M., Garcia, P.,
Expression of sonic hedgehog-Fc fusion protein and Enfors, S.-O. (2003) Temperature limited
in Pichia pastoris. Identification and control of fed-batch technique for control of proteolysis
post-translational, chemical, and proteolytic in Pichia pastoris bioreactor cultures. Microbial
modifications. Protein Expr. Purif. 29, 272–283. Cell Factories 2, 6–17.
28. Files, D., Ogawa, M., Scaman, C. H., and 38. Surribas, A., Stahn, R., Montesinos, J. L., Enfors,
Baldwin, S. A. (2001) A Pichia pastoris fermen- S.-O., Valero, F., and Jahic, M. (2007) Production
tation process for producing high-levels of of Rhizopus oryzae lipase from Pichia pastoris
recombinant human cystatin-C. Enzyme using alternative operational strategies. J. Biotechnol.
Microbial Technol. 29, 335–340. 130, 291–299.
29. Zhang, W., Hywood Potter K. J., Plantz, B. A., 39. Shi, X., Karkut T., Chamankhah, M., Alting-
Schlegel, V. L., Smith, L. A., and Meagher, M. Mees, M., Hemmingsen, S. M., and Hegedus,
M. (2003) Pichia pastoris fermentation with D. (2003) Optimal conditions for the expres-
mixed-feeds of glycerol and methanol: growth sion of a single-chain antibody (scFv) gene
kinetics and production improvement. J. Ind. in Pichia pastoris. Protein Expr. Purif. 28,
Microbiol. Biotechnol. 30, 210–215. 321–330.
30. Xie, J., Zhang, L., Ye, Q., Zhou, Q., Xin, L., 40. Yao, X. Q., Zhao, H. L., Xue, C., Zhang, W.,
Du, P., and Gan, R. (2003) Angiostatin pro- Xiong, X. H., Wang, Z. W., Li, X. Y., and Liu,
duction in cultivation of recombinant Pichia Z. M. (2009) Degradation of HSA-
pastoris fed with mixed carbon sources. AX15(R13K) when expressed in Pichia pas-
Biotechnol. Lett. 25, 173–177. toris can be reduced via the disruption of
31. McGrew, J. T., Leiske, D., Dell, B., Klinke, R., YPS1 gene in this yeast. J. Biotechnol. 139,
Krasts, D., Wee, S. F., Abbott, N., Armitage, 131–136.
R., and Harrington, K. (1997) Expression of 41. Madduri, K., Badger, M., Li, Z.-S., Xu, X.,
trimeric CD40 ligand in Pichia pastoris: use of a Thornburgh, S., Evans, S., and Dhadialla, T. S.
rapid method to detect high-level expressing (2009) Development of stable isotope and sele-
transformants. Gene 187, 193–200. nomethionine labeling methods for proteins
32. d’Anjou, M. C., and Daugulis, A. J. (2001) A expressed in Pseudomonas fluorescens. Protein
rational approach to improving productivity in Expr. Purif. 65, 57–65.
Chapter 3
Abstract
Recent years have seen remarkable progress in applying nuclear magnetic resonance (NMR) spectroscopy
to proteins that have traditionally been difficult to study due to issues with folding, posttranslational
modification, and expression levels or combinations thereof. In particular, insect cells have proved useful
in allowing large quantities of isotope-labeled, functional proteins to be obtained and purified to homoge-
neity, allowing study of their structures and dynamics by using NMR. Here, we provide protocols that
have proven successful in such endeavors.
Key words: Isotope labeling, Baculovirus, Nuclear magnetic resonance, Recombinant protein
expression, Insect cells
1. Introduction:
Baculovirus-Insect
Cell Expression
System Isotope labeling of proteins represents an important and often
required tool for the application of nuclear magnetic resonance
(NMR) spectroscopy to investigate the structure and dynamics of
proteins. So far, the great majority of isotope-labeled proteins have
been expressed in Escherichia coli (1) because of the ease of cloning
and expressing proteins at low cost. When possible, protein pro-
duction should be performed in a prokaryotic system (i.e., E. coli),
since this strategy is the most cost-effective and allows the most
flexibility. However, human or complex proteins often cannot be
expressed in E. coli in an active, correctly folded, posttranslationally
modified form (glycosylated, phosphorylated, etc.). The capacity
of E. coli for protein folding and forming disulfide bonds is not
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_3, © Springer Science+Business Media, LLC 2012
37
38 K. Saxena et al.
the insertion of the gene of interest into the viral genome at the
polh locus, resulting in the production of a recombinant virus
genome and allowing the powerful polyhedron promoter to drive
protein expression of the foreign gene. Since the efficiency of
homologous recombination is quite low, identification, isolation,
and selection of recombinant virus were traditionally achieved by
labor-intensive, technically demanding plaque assays. Due to the
deletion of the polyhedrin gene, the recombinant plaque has a
more clearly distinct morphology than the parental virus contain-
ing the polh gene. Subsequently, additional rounds of plaque
screening are required to separate the desired recombinant virus
from the parental wild-type virus. However, discriminating between
polyhedron-positive and -negative plaques and isolating recombi-
nant virus turned out to be a serious problem for many investiga-
tors who used the BvE for the production of recombinant proteins.
Nowadays, these technical issues and the time-consuming plaque
purification processes are eliminated. In the next section, some of
the key developments in the BvE are presented and discussed.
1.2. Commercially Generally, the baculovirus genome is considered too large to insert
Available Baculovirus a foreign gene directly. In most applications of the BvE, the gene
Expression Systems of interest is therefore cloned into a transfer vector, which contains
sequences that flank the polyhedron gene in the baculovirus
genome. The virus genome and the transfer vector are cotrans-
fected into the insect cells and the gene of interest inserts into the
virus genome via homologous recombination (see Fig. 1) under
the control of the strong late viral polyhedrin promoter (35). Since
a mixture of recombinant and original parental virus is produced
after the initial replication, time-consuming plaque purification
and isolation are required before protein expression can proceed.
BacVector (Merck Biosciences), Baculo-Gold and pBacPAK (BD
Biosciences), and Bac-N-Blue (Invitrogen) are commercially avail-
able BvEs that use homologous recombination to integrate foreign
genes into the virus genome.
New developments in generating recombinant virus by using
site-specific transpositions (Bac-to-Bac or BaculoDirect, Invitrogen)
or progress in recombination methodology with an engineered
baculovirus containing a lethal mutation in an essential gene (open
reading frame (ORF1629), flashBAC from Oxford Expression
Technologies, or BacMagic from EMD Chemicals, Novagen) have
facilitated the use of BvE for a larger user community (28).
Generally, these improvement strategies can be classified into transfer
plasmid modifications and parental baculovirus genome modifica-
tion (28). The flashBAC and BacMagic are the most promising
BvEs so far, since the efficiency of recombination in both systems
is 100%. Therefore, these BvEs overcome the requirement of time-
consuming plaque assays and protein expression can be started
directly after one or two rounds of virus amplification. This technology
42 K. Saxena et al.
1.3. Insect Cell Lines The main insect cell lines used for cotransfections and baculovirus
amplification are Spodoptera frugiperda Sf9 or Sf21 (derivatives of
the fall armyworm). Trichoplusia ni BTI 5B1-4 (36) (High Five™)
cells are generally used for the production of secreted recombinant
proteins and not for virus production because of the increased pos-
sibility of generating virus mutants (37). Due to the high-mannose
and paucimannose types of glycosylation that are obtained in insect
cells, no therapeutic protein is currently produced using this sys-
tem as this would compromise in vivo bioactivity and potentially
induce allergenic reactions. Engineering insect cells with glycosyl-
transferases allows the production of proteins with mammalian-
type sugars (38).
1.4. Expression of Due to the high costs of incorporating stable isotopes into insect
Labeled Recombinant cells, it is recommended that the recombinant protein expression
Protein in the is optimized using the chosen BvE before starting with labeled
Baculovirus fermentation. There are only a few studies reporting the incorpo-
Expression System ration of stable isotopes into proteins expressed by baculovirus-
mediated insect cell expression (39–47).
In contrast to the initial trials of amino acid-type selective
labeling of proteins in user-defined insect cell media, nowadays
there are commercial media (BioExpress-2000, Cambridge Isotope
Laboratories, CIL) available for the different labeling applications.
Expression of uniformly 13C- 15N-labeled Abelson Kinase domain
(13C–15N BioExpress-2000) was the first example of backbone
NMR resonance assignments of a recombinant protein expressed
using the BvE (48). So far, most uniform labeling protein work in
insect cells is only performed by industrial research groups due to
the extraordinary costs of the required media. It is not possible to
cultivate insect cells in minimal medium, since this host requires
essential amino acids for its growth. Reports of selective amino
acid isotope labeling in BvE are more frequently cited in the litera-
ture, since this approach is easy and fast and does not require
expensive medium for labeling. Even in the absence of a backbone
assignment of the target protein, structural information can be
deduced from the selective labeling approach based on an existing
X-ray structure of the protein. From a practical aspect, it should be
considered that there are essential and nonessential amino
acids (alanine, cysteine, glutamic acid, glutamine, aspartic acid,
and asparagine) in insect cells whose content in the medium
44 K. Saxena et al.
NH4
Alanine Phenylalanine
Valine
Fig. 2. A schematic presentation of amino acid metabolism in E. coli and Sf9 with respect to 15N: The black arrows symbolize
pathways present in both expression hosts. Pathways that only exist in E. coli are shown in grey. The strength of the arrows
reflects the intensity of the conversion. Picture modified from Bruggert et al. (42).
2. Materials
3. Methods
3.1. Cotransfection For efficient transfection, high-quality DNA of the transfer plasmid
of Insect Cells is prepared using commercially available plasmid DNA purification
kits (Qiagen, Novagen) (see Note 4). Additionally, it is recom-
mended to use fresh, rapidly proliferating cells (see Note 5) for
transfection experiments and to have positive and negative trans-
fection controls (see Note 6).
1. For each cotransfection, prepare one 35-mm2 plate. Seed the
dishes with insect cells at least 1 h before use. Use 1 × 106 cells/
dish for Sf9 cells and 1.5 × 106 cells/dish for Sf21 cells in 2 mL
of SF900 II insect cell culture medium. Allow the cells to
attach by incubating at 28°C for 20 min.
2. During the 1-h incubation period, prepare a DNA–liposome
complex cotransfection mix of BacMagic DNA and Insect
GeneJuice® transfection reagent for each transfection. Assemble
the following components, in the order listed, in a sterile tube
(bijoux) (see Note 7):
1 mL serum-free, antibiotic-free SF900 II insect cell culture
medium
5 μL Insect GeneJuice
5 μL BacMagic DNA (100 ng total)
5 μL Transfer vector DNA (500 ng total)
1.015 mL Total volume
3. Incubate at room temperature for 15–30 min to allow the
DNA–liposome complexes to form.
4. Remove the culture medium from the 35-mm2 dishes of cells
using a sterile pipette, ensuring that the cell monolayer is not
disrupted (see Note 8).
5. Immediately after the medium has been removed from the
cells, add the 1 mL of the DNA–liposome complex dropwise
to the center of each dish (see Note 9) and incubate in a plastic
sandwich box at 28°C for a minimum of 5 h or overnight.
6. After the incubation period, add another 1 mL of SF900 II
insect cell culture medium to each dish and continue the
incubation for 5 days in total.
3 Isotope Labeling in Insect Cells 47
3.3. Analysis Before proceeding with use of the virus in subsequent protein
of Recombinant expression experiments, a plaque assay to determine the accurate
Protein Expression titer is strongly recommended by the manufacturer of the flash-
BAC and BacMagic technology (see Note 16). This allows the
calculation of MOI and ensures cross-referencing and reproduc-
ibility in the following experiments. However, for analysis of protein
expression, it is not absolutely required to know the titer of the
recombinant virus (see Note 17). Pilot expression analysis can also
be performed by infecting cells with different amounts of P1 virus
and monitoring the expression of the recombinant protein.
According to the protein expression literature, it is recommended
that exponentially growing cells should be infected at a high MOI
to ensure that all cells are infected simultaneously and that the cul-
ture is synchronous. However, in our lab, various P1 viruses are
directly added (0.5–10%) to the insect cells for determining the
best ratio for optimal recombinant protein expression without
knowing the exact titer of the different viruses. The best time to
48 K. Saxena et al.
Fig. 3. Protein overexpression and purification using insect cells. SDS PAGE (left ), Western Blot (right ): Protein samples
loaded are BC (before Ni-NTA chromatography), FT (Flowthrough), M (Molecular weight marker), W (Wash), E (Elution).
3 Isotope Labeling in Insect Cells 49
3.4. Production While fermenters and bioreactors are now routinely used to produce
of Isotopically recombinant proteins in insect cells, good yields can be achieved in
Labeled Protein shake cultures at a fraction of the cost. Large, disposable shake
flasks can be obtained that hold up to 1.5 L of insect cell culture
on a shaking platform in a warm room or incubator. The most suit-
able culture protocol for isotope labeling of a recombinant protein
in insect cells involves an initial growth phase for the insect cells in
an unlabeled growth/expression medium and a subsequent expres-
sion phase, where cells are centrifuged and resuspended into the
labeling medium prior to infection with the recombinant baculo-
virus (see Note 22). This centrifugation step allows much higher
expression levels of the labeled protein compared to one with
growth and expression in labeling medium without the change in
medium. Recently, isotope-labeled expression media for insect cells
have become commercially available (BioExpress-2000, CIL), but
the source and the concentration of the amino acids in these media
are not published. However, for site-specific labeling of amino
acids in proteins expressed in insect cells, adding the isotope-
labeled amino acid of interest to a standard insect cell medium
before infection (see Note 23) also works well and the cost for this
labeling strategy is comparable to that when using E. coli.
The following protocol is adapted from the Application Note:
Efficient uniform isotope labeling of proteins expressed in
Baculovirus-infected insect cells using BioExpress® 2000 (Insect cell)
medium by André Strauss, Gabriele Fendrich, and Wolfgang Jahnke.
Prior to performing isotope labeling of a protein, optimize the
culture and BV-infection conditions in unlabeled medium (e.g.,
BioExpress 2000-U or SF900 II for expression of the protein).
1. Cultivate several 100 mL cultures of Sf9 cells, adapted to
growth in serum-free SF900 II medium, in 500-mL Erlenmeyer
flasks for 3 days at 27°C with shaking at 90 rpm.
2. Prepare the uniform isotope-labeling medium (see Note 24).
3. When the final cell density of the culture has reached 1.5 × 106
cells/mL (~3 days), sterile centrifuge the cells at 400 × g for
20 min at 20°C.
4. Resuspend the pelleted cells in 100-mL portions in labeled
BioExpress® 2000-CN medium and transfer them to fresh
500-mL Erlenmeyer flasks.
5. Add the recombinant virus stock at a titer of 0.5–2 × 108 pfu/mL
to an MOI = 1–2, according to optimized conditions.
6. Grow the 100 mL cultures of baculovirus-infected Sf9 cells for
3 days post infection in labeled BioExpress® 2000-CN medium
at 27°C, with shaking at 90 rpm (see Note 25).
7. Harvest the cells expressing the labeled recombinant protein
by centrifuging at 400 × g for 20 min at 20°C (see Note 26);
resuspend the pelleted cells in 20 mL of PBS containing
50 K. Saxena et al.
4. Notes
11. The expected titer of this initial viral seed stock is generally
about 1 × 107 pfu/mL. In contrast to virus-free cells, virus-
infected cells appear grainy with enlarged nuclei and do not
form a confluent monolayer.
12. It is important that the cells are healthy and in the log phase of
growth to ensure that virus replication occurs efficiently to
generate high-titer stocks of virus for subsequent use in expres-
sion studies suitable for transfection into cells.
13. Virus-infected cells have an increased need for oxygen, and
therefore the contents of the flasks should be shaken at quite
high speeds to maximize aeration. The surface area-to-volume
ratio should also be as large as possible for maximum gas
exchange. Do not overfill the flasks.
14. Under a phase-contrast inverted microscope, cells infected
with virus appear grainy when compared to healthy cells. The
infected cells become uniformly rounded and enlarged, with
distinct enlarged nuclei.
15. The virus stock can be stored in the dark at 4°C for 6–12 months,
although the titer begins to drop after 3–4 months. Titer the
virus before use and reamplify if necessary. The addition of
2–5% serum when using serum-free medium can be helpful in
avoiding a drop in titer. Virus may be frozen at −80°C for lon-
ger periods of time. Avoid multiple freeze–thaw cycles.
16. Protocols for plaque assays can be found in the manuals of
Novagen and Oxford Expression Technologies.
17. After cotransfecting and harvesting the seed stock of virus for
further amplification, it is possible to harvest the remaining
cells from the dish and prepare these for SDS-PAGE/Western
blotting. This gives a quick check for gene expression.
18. The most commonly used times are at 24, 48, 72, and 96 h
post infection (hpi). Some proteins may be very stable and
accumulate to high levels by 96 hpi, and others may start to
degrade and thus need to be harvested much earlier.
19. High Five cell lines (Invitrogen) often increase the yield of
secreted recombinant proteins.
20. For secreted protein expression, the supernatant has to be
analyzed.
21. Standard molecular biology procedures like SDS-PAGE and
Western-blotting are precisely described in Molecular Cloning:
A Laboratory Manual) (49).
22. A similar cost-reducing labeling strategy in E. coli was also
described for the production of isotope-labeled protein by
Marley and coworkers (50).
52 K. Saxena et al.
References
1. Goto, N. K., and Kay, L. E. (2000) New 9. Hu, Y. C. (2008) Baculoviral vectors for gene
developments in isotope labeling strategies for delivery: a review. Curr. Gene Ther. 8, 54–65.
protein solution NMR spectroscopy. Curr. 10. Kost, T. A., Condreay, J. P., Ames, R. S., Rees,
Opin. Struct. Biol. 10, 585–592. S., and Romanos, M. A. (2007) Implementation
2. Schein, C. H. (1991) Optimizing protein fold- of BacMam virus gene delivery technology in a
ing to the native state in bacteria. Curr Opin. drug discovery setting. Drug Discov. Today 12,
Biotechnol. 2, 746–750. 396–403.
3. Qing, G., Ma, L. C., Khorchid, A., Swapna, G. 11. Hitchman, R. B., Possee, R. D., and King, L.
V., Mal, T. K., Takayama, M. M., Xia, B., A. (2009) Baculovirus expression systems for
Phadtare, S., Ke, H., Acton, T., Montelione, recombinant protein production in insect cells.
G. T., Ikura, M., and Inouye, M. (2004) Cold- Recent Pat. Biotechnol. 3, 46–54.
shock induced high-yield protein production 12. Lam, J. S., Huang, H., and Levitz, S. M.
in Escherichia coli. Nat. Biotechnol. 22, (2007) Effect of differential N-linked and
877–882. O-linked mannosylation on recognition of fun-
4. Young, J. C., Agashe, V. R., Siegers, K., and gal antigens by dendritic cells. PLoS One 2,
Hartl, F. U. (2004) Pathways of chaperone- e1009.
mediated protein folding in the cytosol. Nat. 13. Dasgupta, S., Navarrete, A. M., Bayry, J.,
Rev. Mol. Cell. Biol. 5, 781–791. Delignat, S., Wootla, B., Andre, S., Christophe,
5. Esposito, D., and Chatterjee, D. K. (2006) O., Nascimbeni, M., Jacquemin, M., Martinez-
Enhancement of soluble protein expression Pomares, L., Geijtenbeek, T. B., Moris, A.,
through the use of fusion tags. Curr. Opin. Saint-Remy, J. M., Kazatchkine, M. D., Kaveri,
Biotechnol. 17, 353–358. S. V., and Lacroix-Desmazes, S. (2007) A role
6. Andersen, C. L., Matthey-Dupraz, A., for exposed mannosylations in presentation of
Missiakas, D., and Raina, S. (1997) A new human therapeutic self-proteins to CD4+ T
Escherichia coli gene, dsbG, encodes a periplas- lymphocytes. Proc. Natl. Acad. Sci. U.S.A 104,
mic protein involved in disulphide bond forma- 8965–8970.
tion, required for recycling DsbA/DsbB and 14. Moore, K. L. (2003) The biology and enzy-
DsbC redox proteins. Mol. Microbiol. 26, mology of protein tyrosine O-sulfation. J. Biol.
121–132. Chem. 278, 24243–24246.
7. Bardwell, J. C. (1994) Building bridges: disul- 15. Daly, R., and Hearn, M. T. (2005) Expression
phide bond formation in the cell. Mol. Microbiol. of heterologous proteins in Pichia pastoris:
14, 199–205. a useful experimental tool in protein engineer-
8. Pickford, A. R., and O’Leary, J. M. (2004) ing and production. J. Mol. Recognit. 18,
Isotopic labeling of recombinant proteins from 119–138.
the methylotrophic yeast Pichia pastoris. 16. Hamilton, S. R., and Gerngross, T. U. (2007)
Methods Mol. Biol. 278, 17–33. Glycosylation engineering in yeast: the advent
3 Isotope Labeling in Insect Cells 53
of fully humanized yeast. Curr. Opin. Biotechnol. 29. O´Reilly, D. R., Miller, L., Luckow, V.A. (1992)
18, 387–392. Baculovirus Expression Vectors - a Laboratory
17. Chisholm, V., Chen, C. Y., Simpson, N. J., and Manual. WH Freeman, New York.
Hitzeman, R. A. (1990) Molecular and genetic 30. van Regenmortel, M. H., Mayo, M. A.,
approach to enhancing protein secretion. Fauquet, C. M., and Maniloff, J. (2000) Virus
Methods Enzymol. 185, 471–482. nomenclature: consensus versus chaos. Arch.
18. Dorner, A. J., and Kaufman, R. J. (1990) Virol. 145, 2227–2232.
Analysis of synthesis, processing, and secretion 31. Hunt, I. (2005) From gene to protein: a review
of proteins expressed in mammalian cells. of new and enabling technologies for multi-
Methods Enzymol. 185, 577–596. parallel protein expression. Protein Expr. Purif.
19. Moir, D. T. (1989) Yeast mutants with increased 40, 1–22.
secretion efficiency. Biotechnology 13, 32. Carbonell, L. F., Klowden, M. J., and Miller, L.
215–231. K. (1985) Baculovirus-mediated expression of
20. Biemans, R., Thines, D., Rutgers, T., De Wilde, bacterial genes in dipteran and mammalian
M., and Cabezon, T. (1991) The large surface cells. J. Virol. 56, 153–160.
protein of hepatitis B virus is retained in the 33. Smith, G. E., Summers, M. D., and Fraser, M.
yeast endoplasmic reticulum and provokes its J. (1983) Production of human beta interferon
unique enlargement. DNA Cell. Biol. 10, in insect cells infected with a baculovirus
191–200. expression vector. Molecular Cell. Biol. 3,
21. Gennaro, D. E., Hoffstein, S. T., Marks, G., 2156–2165.
Ramos, L., Oka, M. S., Reff, M. E., Hart, T. 34. Ayres, M. D., Howard, S. C., Kuzio, J., Lopez-
K., and Bugelski, P. J. (1991) Quantitative Ferber, M., and Possee, R. D. (1994) The
immunocytochemical staining for recombinant complete DNA sequence of Autographa cali-
tissue-type plasminogen activator in transfected fornica nuclear polyhedrosis virus. Virology
Chinese hamster ovary cells. Proc. Soc. Exp. 202, 586–605.
Biol. Med. 198, 591–598. 35. Kitts, P. A., and Possee, R. D. (1993) A method
22. Shuster, J. R. (1991) Gene expression in yeast: for producing recombinant baculovirus expres-
protein secretion. Curr. Opin. Biotechnol. 2, sion vectors at high frequency. Biotechniques
685–690. 14, 810–817.
23. Mollaaghababa, R., Davidson, F. F., Kaiser, C., 36. Kost, T. A., Condreay, J. P., and Jarvis, D. L.
and Khorana, H. G. (1996) Structure and (2005) Baculovirus as versatile vectors for pro-
function in rhodopsin: expression of functional tein expression in insect and mammalian cells.
mammalian opsin in Saccharomyces cerevisiae. Nat. Biotechnol. 23, 567–575.
Proc. Natl. Acad. Sci. U.S.A 93, 37. Friesen, P. D., and Nissen, M. S. (1990) Gene
11482–11486. organization and transcription of TED, a lepi-
24. Robinson, A. S., Hines, V., and Wittrup, K. D. dopteran retrotransposon integrated within the
(1994) Protein disulfide isomerase overexpres- baculovirus genome. Mol. Cell. Biol. 10,
sion increases secretion of foreign proteins in 3067–3077.
Saccharomyces cerevisiae. Biotechnology (N Y) 38. Harrison, R. L., and Jarvis, D. L. (2006)
12, 381–384. Protein N-glycosylation in the baculovirus-
25. Shusta, E. V., Raines, R. T., Pluckthun, A., and insect cell expression system and engineering of
Wittrup, K. D. (1998) Increasing the secretory insect cells to produce “mammalianized”
capacity of Saccharomyces cerevisiae for produc- recombinant glycoproteins. Adv. Virus Res. 68,
tion of single-chain antibody fragments. Nat. 159–191.
Biotechnol. 16, 773–777. 39. DeLange, F., Klaassen, C. H., Wallace-Williams,
26. Butz, J. A., Niebauer, R. T., and Robinson, A. S. E., Bovee-Geurts, P. H., Liu, X. M., DeGrip,
S. (2003) Co-expression of molecular chaper- W. J., and Rothschild, K. J. (1998) Tyrosine
ones does not improve the heterologous structural changes detected during the photo-
expression of mammalian G-protein coupled activation of rhodopsin. J. Biol. Chem. 273,
receptor expression in yeast. Biotechnol. Bioeng. 23735–23739.
84, 292–304. 40. Creemers, A. F., Klaassen, C. H., Bovee-Geurts,
27. Durocher, Y., and Butler, M. (2009) Expression P. H., Kelle, R., Kragl, U., Raap, J., de Grip, W.
systems for therapeutic glycoprotein produc- J., Lugtenburg, J., and de Groot, H. J. (1999)
tion. Curr. Opin. Biotechnol. 20, 700–707. Solid state 15N NMR evidence for a complex
28. Jarvis, D. L. (2009) Baculovirus-insect cell Schiff base counterion in the visual G-protein-
expression systems. Methods Enzymol. 463, coupled receptor rhodopsin. Biochemistry 38,
191–222. 7195–7199.
54 K. Saxena et al.
41. Bellizzi, J. J., Widom, J., Kemp, C. W., and 46. Stockman, B. J., Kothe, M., Kohls, D., Weibley,
Clardy, J. (1999) Producing selenomethionine- L., Connolly, B. J., Sheils, A. L., Cao, Q.,
labeled proteins with a baculovirus expression Cheng, A. C., Yang, L., Kamath, A. V., Ding,
vector system. Structure 7, R263-267. Y. H., and Charlton, M. E. (2009) Identification
42. Bruggert, M., Rehm, T., Shanker, S., of allosteric PIF-pocket ligands for PDK1 using
Georgescu, J., and Holak, T. A. (2003) A novel NMR-based fragment screening and 1H-15N
medium for expression of proteins selectively TROSY experiments. Chem. Biol. Drug Des.
labeled with 15N-amino acids in Spodoptera fru- 73, 179–188.
giperda (Sf9) insect cells. J. Biomol. NMR 25, 47. Jahnke, W., Grotzfeld, R. M., Pelle, X., Strauss,
335–348. A., Fendrich, G., Cowan-Jacob, S. W., Cotesta,
43. Strauss, A., Bitsch, F., Cutting, B., Fendrich, S., Fabbro, D., Furet, P., Mestan, J., and
G., Graff, P., Liebetanz, J., Zurini, M., and Marzinzik, A. L. (2010) Binding or bending:
Jahnke, W. (2003) Amino-acid-type selective distinction of allosteric Abl kinase agonists from
isotope labeling of proteins expressed in antagonists by an NMR-based conformational
Baculovirus-infected insect cells useful for assay. J. Am. Chem. Soc. 132, 7043–7048.
NMR studies. J. Biomol. NMR 26, 367–372. 48. Vajpai, N., Strauss, A., Fendrich, G., Cowan-
44. Strauss, A., Bitsch, F., Fendrich, G., Graff, P., Jacob, S. W., Manley, P. W., Jahnke, W., and
Knecht, R., Meyhack, B., and Jahnke, W. Grzesiek, S. (2008) Backbone NMR resonance
(2005) Efficient uniform isotope labeling of assignment of the Abelson kinase domain in
Abl kinase expressed in Baculovirus-infected complex with imatinib. Biomol. NMR Assign.
insect cells. J. Biomol. NMR 31, 343–349. 2, 41–42.
45. Betz, M., Vogtherr, M., Schieborr, U., Elshorst, 49. Maniatis, T. (1982) Molecular cloning : a labo-
B., Grimme, S., Pescatore, B., Langer, T., ratory manual / T. Maniatis, E.F. Fritsch, J.
Saxena, K., and Schwalbe, H. (2008) Chemical Sambrook, Cold Spring Harbor Laboratory,
Biology of Kinases Studied by NMR Cold Spring Harbor, N.Y.
Spectroscopy, In Chemical Biology (Prof. Dr. 50. Marley, J., Lu, M., and Bracken, C. (2001)
Stuart L. Schreiber, P. D. T. M. K. P. D. G., A method for efficient isotopic labeling of recom-
Ed.), pp 852–890. binant proteins. J. Biomol. NMR 20, 71–75.
Chapter 4
Abstract
Isotope labeling of proteins represents an important and often required tool for the application of nuclear
magnetic resonance (NMR) spectroscopy to investigate the structure and dynamics of proteins. Mammalian
expression systems have conventionally been considered to be too weak and inefficient for protein expression.
However, recent advances have significantly improved the expression levels of these systems. Here, we
provide an overview of some of the recent developments in expression strategies for mammalian expression
systems in view of NMR investigations.
Key words: Isotope labeling, Nuclear magnetic resonance, Recombinant protein expression, Human
embryonic kidney cells
1. Introduction:
Mammalian Cell
Expression System
Mammalian cell expression systems are being increasingly used to
1.1. Principle express proteins for structural biology. This is evident from the
of Mammalian increase in the number of structures available in the Protein Data
Cell-Mediated Bank (PDB) (1) of proteins purified from such sources. The main
Protein Expression reason for switching to these higher eukaryotic expression systems
is the ability to produce biologically active cell surface receptors
and secreted glycoproteins. The functionality of these proteins is
linked to the requirement for posttranslational modifications, such
as glycosylation and disulfide bond formation, which can often only
be satisfied by using mammalian systems and not other eukaryotic
systems, such as yeast and insect cells.
Protein expression in mammalian cells involves transfection
with a plasmid carrying the gene of interest under the control of a
mammalian promoter. The mammalian gene can be introduced
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_4, © Springer Science+Business Media, LLC 2012
55
56 A. Dutta et al.
1.2. Mammalian Mammalian cell lines used for transfection with virus-based vectors
Cell Lines carrying the gene of interest usually already carry vectors encoding
viral early genes corresponding to the origin of replication used by
the new vector. Some of these early genes are SV40 T antigen,
EBNA-1, E1, and E2 which bind to SV40, EBV, and BPV ori,
respectively. The early genes act in trans to initiate replication by
binding to the ori and they also act as enhancers, thereby increas-
ing the copy number of viral vector and the expression levels of
the protein of interest (13). Since very high copy numbers of the
viral genes can lead to host cell death, regulation of viral replication
can be imposed by transfecting mammalian cells with a different
vector carrying the early genes (13). Mammalian cell lines that are
widely used for protein expression are CHO and Human Embryonic
Kidney 293 (HEK293) cells. Both are suitable for use in adherent
and suspension cultures. The advantage of using CHO cells is the
availability of various auxotrophs that can be used as selection
markers for transfection. An example of such an auxotroph is
dihydrofolate reductase (DHFR) deficient cells which are triple
auxotrophs for hypoxanthine, glycine, and thymidine (14).
Transfection of foreign genes along with DHFR genes in these
cells allows for the selection of clones in a medium devoid of the
above nutrients. Another advantage of this system is that it helps
amplify the foreign gene when DHFR deficient cells are grown in
the presence of methotrexate, which blocks DHFR activity. This
causes the transfected cells to deal with low DHFR activity by
amplifying the copy number of DHFR, thus amplifying the copy
number of the transfected gene of interest. The disadvantages of
the CHO cell line is that these cells do not carry all of the sugar
transferring enzymes (15) that are present in human cell lines. This
may lead to the production of functionally nonrelevant proteins.
Further, some of the posttranslational modifications of human
proteins are not appropriately carried out in these cells (16). It has
also been seen that expression levels are usually lower in CHO cells
than in HEK293 cells (12, 17). Therefore, the human cell line,
HEK293, is used for proteins that require functionally sensitive
posttranslational modifications.
58 A. Dutta et al.
Fig. 1. Dephased (red line) and nondephased (blue line) 15N detected 13C/15N CP REDOR spectra of selectively labeled
rhodopsin in DOPC lipid bilayers at 220 K. (a) a,e-15N-Trp; (b) 13C¢-Leu/a,e-15N-Trp; (c) 13U-Thr/a,e-15N-Trp; (d) 13C¢-Cys/a,e-
15
N-Trp; (e) 13C¢-Pro/a,e-15N-Trp; (f) 13C¢-Gly/a,e-15N-Trp. Figure reproduced with permission from (29).
4 Isotope Labeling in Mammalian Cells 61
2. Materials
2.1. Expression 1. Plasmid containing the gene of interest (see Note 1).
of Isotope-Labeled 2. HEK293 cells.
Recombinant Proteins
3. Dulbecco’s modified Eagle medium (DMEM F-12).
4. Blasticidin (500 mg/mL stock solution prepared in DMEM
F-12).
5. Geneticin, G418 (100 mg/mL stock solution prepared in
DMEM F-12).
6. 0.05% Trypsin–EDTA (Gibco).
7. Penicillin–streptomycin (PS): 100 U/mL of each.
8. Fetal bovine serum (FBS).
9. Phosphate buffer saline (PBS): Autoclaved.
10. Tetracycline: 200 mg/mL (100× stock).
11. Sodium butyrate: 500 mM (100× stock), filtered.
12. Complete medium: DMEM F-12, 10% FBS, 1% PS.
13. Induction medium: Complete medium containing 2 mg/mL
tetracycline and 5 mM sodium butyrate.
14. Selection medium: Complete medium containing 5 mg/mL
blasticidin and 1, 2, or 3 mg/mL of G418.
15. Cryo medium: Complete medium containing 10% DMSO.
16. 2.5 M CaCl2.
17. BES: 50 mM N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic
acid, pH 7.2, 250 mM NaCl, 1.5 mM Na2HPO4, 1 M NaOH
used to adjust pH.
18. 15- and 10-cm2 Tissue culture plates.
19. 6- and 24-Well cell culture plates.
20. Sterile forceps.
21. Cloning rings.
22. Vacuum grease.
23. Dodecyl maltoside.
Table 1
A. Dutta et al.
3. Methods
3.1. Expression Following is a protocol for generating an inducible stable cell line
of Isotope-Labeled of HEK293 (30–32).
Recombinant Proteins
1. Wake up HEK293 cells from cryostocks and maintain them in
15 cm2 plates in 25 mL of complete medium containing 5 mg/mL
blasticidin.
2. Split the cells into a 10-cm2 dish at 80% confluency.
3. Split the cells 1:10 or 1:8 (~1–2 million cells per plate) into com-
plete medium containing 5 mg/mL blasticidin the day before
transfection.
4. On the day of transfection, the cells should be 30–40% confluent.
To transfect the cells, prepare the following cocktail, adding
components in the order given (volumes are for one 10-cm2
plate): In a falcon tube, mix 30 mg of plasmid DNA, 50 mL of
2.5 M CaCl2, 500 mL of BES, and sterile water to 1 mL.
Incubate the mixture for exactly 1 min and gently add it to the
cells. After 1 h, verify the efficiency of transfection by the presence
of calcium phosphate precipitate which appears as fine sand
particles between cells.
5. Incubate the plates at 35°C and 3% CO2 for 19 h.
64 A. Dutta et al.
14. Change the selection medium on the cells every 2–3 days.
15. As soon as the cells in the higher density well reach confluence,
induce with induction medium.
16. Harvest the cells 48-h post induction.
(a) After harvesting, solubilize the cells in a detergent suitable
for the membrane protein of interest (e.g., 1% dodecyl
maltoside).
(b) Centrifuge the solubilized cells at 126,000 × g for 20 min
at room temperature and collect the supernatant.
17. Check for levels of expression by performing a Western dot
blot on serial dilutions of the supernatant in PBS, containing
1% of the detergent used for solubilization, versus protein sam-
ples of known concentrations (see Note 4).
18. When the cells from the low density well reach confluence,
split 1:5 into two 10-cm2 cell culture dishes. Add 0.5 mL of
0.05% trypsin–EDTA to a 6-well plate, and then add 1.5 mL
of selection medium. Add 1 mL of this cell suspension to each
of the two 10-cm2 plates.
19. After the cells reach confluence on the 10-cm2 plate, prepare
three 1-mL glycerol stocks from each plate. To prepare glyc-
erol stocks:
(a) Wash 70–90% confluent plates twice with PBS and
trypsinize with 1 mL of 0.05% trypsin–EDTA for 1 min.
(b) Add 10 mL of complete medium (no selection) and col-
lect the cells in a 15-mL falcon tube.
(c) Centrifuge the cells at 600 × g for 10 min at 4°C.
(d) Aspirate the medium and gently resuspend the cells in
3 mL of cryo medium.
(e) Transfer the cells to cryo vials in 1-mL aliquots and place
the tubes in cryo boxes. Incubate at −20°C for 1 h, and then
−80°C overnight.
(f) On the following day, move the tubes to liquid nitrogen
storage tanks.
3.2. Specific Isotope After estimating the yield from above (Subheading 3.1) and iden-
Labeling of Proteins tifying the highest yield clone, the following procedure is used to
prepare a protein sample, in suspension culture, in which particular
amino acids are specifically isotope labeled (30–32).
1. Grow the highest yielding stable cell line clone in complete
medium.
2. Split the cells 1:5 into 10-cm2 dishes after ~3 days using 15 mL
of selection medium.
66 A. Dutta et al.
4. Conclusions
5. Notes
References
1. Nettleship, J. E., Assenberg, R., Diprose, J. M., Rosser, M. P., MacRobbie, J., Olsen, C. L., and
Rahman-Huq, N., and Owens, R. J. (2010) Cobb, R. R. (2006) High levels of protein
Recent advances in the production of proteins expression using different mammalian CMV
in insect and mammalian cells for structural promoters in several cell lines. Protein Expr.
biology. J. Struct. Biol. 172, 55–65. Purif. 45, 115–124.
2. Aricescu, A. R., Lu, W., and Jones, E. Y. (2006) 13. Van Craenenbroeck, K., Vanhoenacker, P., and
A time- and cost-efficient system for high- Haegeman, G. (2000) Episomal vectors for
level protein production in mammalian gene expression in mammalian cells. Eur. J.
cells. Acta. Crystallogr. D Biol. Crystallogr. 62, Biochem. 267, 5665–5678.
1243–1250. 14. Urlaub, G., and Chasin, L. A. (1980) Isolation
3. Lee, J. E., Fusco, M. L., and Ollmann Saphire, of Chinese hamster cell mutants deficient in
E. (2009) An efficient platform for screening dihydrofolate reductase activity. Proc. Natl.
expression and crystallization of glycoproteins Acad. Sci. U.S.A. 77, 4216–4220.
produced in human cells. Nat. Protoc. 4, 15. Grabenhorst, E., Schlenke, P., Pohl, S., Nimtz,
592–604. M., and Conradt, H. S. (1999) Genetic engi-
4. Nettleship, J. E., Rahman-Huq, N., and Owens, neering of recombinant glycoproteins and the
R. J. (2009) The production of glycoproteins glycosylation pathway in mammalian host cells.
by transient expression in mammalian cells. Glycoconj. J. 16, 81–97.
Methods Mol. Biol. 498, 245–263. 16. Suttie, J. W. (1986) Report of Workshop on
5. Reeves, P. J., Callewaert, N., Contreras, R., and expression of vitamin K-dependent proteins in bac-
Khorana, H. G. (2002) Structure and function terial and mammalian cells. Madison, Wisconsin,
in rhodopsin: high-level expression of rho- USA, April 1986, Thromb. Res. 44, 129–134.
dopsin with restricted and homogeneous 17. Schlaeger, E. J., Kitas, E. A., and Dorn, A.
N-glycosylation by a tetracycline-inducible (2003) SEAP expression in transiently trans-
N-acetylglucosaminyltransferase I-negative fected mammalian cells grown in serum-free
HEK293S stable mammalian cell line. Proc. suspension culture. Cytotechnology 42, 47–55.
Natl. Acad. Sci. U.S.A. 99, 13419–13424.
18. Takahashi, H., and Shimada, I. (2010)
6. Sclimenti, C. R., and Calos, M. P. (1998) Epstein- Production of isotopically labeled heterologous
Barr virus vectors for gene expression and trans- proteins in non-E. coli prokaryotic and eukary-
fer. Curr. Opin. Biotechnol. 9, 476–479. otic cells. J. Biomol. NMR 46, 3–10.
7. Sambrook, J., Rodgers, L., White, J., and 19. Hansen, A. P., Petros, A. M., Mazar, A. P.,
Gething, M. J. (1985) Lines of BPV-transformed Pederson, T. M., Rueter, A., and Fesik, S. W.
murine cells that constitutively express influenza (1992) A practical method for uniform isotopic
virus hemagglutinin. EMBO J. 4, 91–103. labeling of recombinant proteins in mammalian
8. Grossi, M. P., Caputo, A., Rimessi, P., Chiccoli, cells. Biochemistry 31, 12713–12718.
L., Balboni, P. G., and Barbanti-Brodano, G. 20. Lustbader, J. W., Birken, S., Pollak, S., Pound,
(1988) New BK virus episomal vector for com- A., Chait, B. T., Mirza, U. A., Ramnarain, S.,
plementary DNA expression in human cells. Canfield, R. E., and Brown, J. M. (1996)
Arch. Virol. 102, 275–283. Expression of human chorionic gonadotropin
9. Piechaczek, C., Fetzer, C., Baiker, A., Bode, J., uniformly labeled with NMR isotopes in
and Lipps, H. J. (1999) A vector based on the Chinese hamster ovary cells: an advance toward
SV40 origin of replication and chromosomal rapid determination of glycoprotein structures.
S/MARs replicates episomally in CHO cells. J. Biomol. NMR 7, 295–304.
Nucleic Acids Res. 27, 426–428. 21. Shindo, K., Masuda, K., Takahashi, H., Arata, Y.,
10. Sambrook, J., Fritsch, E. F., and Maniatis, T. and Shimada, I. (2000) Backbone 1H, 13C, and 15N
(1989) Molecular Cloning: A Laboratory Manual, resonance assignments of the anti-dansyl antibody
CSH Laboratory Press, Cold Spring Harbor, NY. Fv fragment. J. Biomol. NMR 17, 357–358.
11. Doerfler, W., Schubbert, R., Heller, H., 22. Archer, S. J., Bax, A., Roberts, A. B., Sporn, M.
Kammer, C., Hilger-Eversheim, K., Knoblauch, B., Ogawa, Y., Piez, K. A., Weatherbee, J. A.,
M., and Remus, R. (1997) Integration of for- Tsang, M. L., Lucas, R., Zheng, B. L., and
eign DNA and its consequences in mammalian et al. (1993) Transforming growth factor beta 1:
systems. Trends Biotechnol. 15, 297–301. NMR signal assignments of the recombinant
12. Xia, W., Bringmann, P., McClary, J., Jones, P. protein expressed and isotopically enriched
P., Manzana, W., Zhu, Y., Wang, S., Liu, Y., using Chinese hamster ovary cells. Biochemistry
Harvey, S., Madlansacay, M. R., McLean, K., 32, 1152–1163.
4 Isotope Labeling in Mammalian Cells 69
23. Werner, K., Richter, C., Klein-Seetharaman, J., 28. Han, M., and Smith, S. O. (1995) High-
and Schwalbe, H. (2008) Isotope labeling of resolution structural studies of the retinal–
mammalian GPCRs in HEK293 cells and char- Glu113 interaction in rhodopsin. Biophys. Chem.
acterization of the C-terminus of bovine rho- 56, 23–29.
dopsin by high resolution liquid NMR spec- 29. Werner, K., Lehner, I., Dhiman, H. K., Richter,
troscopy. J. Biomol. NMR 40, 49–53. C., Glaubitz, C., Schwalbe, H., Klein-
24. Arata, Y., Kato, K., Takahashi, H., and Shimada, Seetharaman, J., and Khorana, H. G. (2007)
I. (1994) Nuclear magnetic resonance study of Combined solid state and solution NMR stud-
antibodies: a multinuclear approach. Methods ies of alpha,epsilon-15N labeled bovine rhodop-
Enzymol. 239, 440–464. sin. J. Biomol. NMR 37, 303–312.
25. Klein-Seetharaman, J., Reeves, P. J., Loewen, 30. Reeves, P. J., Kim, J. M., and Khorana, H. G.
M. C., Getmanova, E. V., Chung, J., Schwalbe, (2002) Structure and function in rhodopsin: a
H., Wright, P. E., and Khorana, H. G. (2002) tetracycline-inducible system in stable mamma-
Solution NMR spectroscopy of (alpha -15N) lian cell lines for high-level expression of opsin
lysine-labeled rhodopsin: The single peak mutants. Proc. Natl. Acad. Sci. U.S.A. 99,
observed in both conventional and TROSY- 13413–13418.
type HSQC spectra is ascribed to Lys-339 in 31. Reeves, P. J., Callewaert, N., Contreras, R., and
the carboxyl-terminal peptide sequence. Proc. Khorana, H. G. (2002) Structure and function
Natl. Acad. Sci. U.S.A. 99, 3452–3457. in rhodopsin: high-level expression of rhodop-
26. Klein-Seetharaman, J., Yanamala, N. V., Javeed, sin with restricted and homogeneous
F., Reeves, P. J., Getmanova, E. V., Loewen, M. N-glycosylation by a tetracycline-inducible
C., Schwalbe, H., and Khorana, H. G. (2004) N-acetylglucosaminyltransferase I-negative
Differential dynamics in the G protein-coupled HEK293S stable mammalian cell line 452. Proc.
receptor rhodopsin revealed by solution NMR. Natl. Acad. Sci. U.S.A. 99, 13419–13424.
Proc. Natl. Acad. Sci. U.S.A. 101, 3409–3413. 32. Reeves, P. J., Thurmond, R. L., and Khorana,
27. Patel, A. B., Crocker, E., Reeves, P. J., H. G. (1996) Structure and function in rho-
Getmanova, E. V., Eilers, M., Khorana, H. G., dopsin: high level expression of a synthetic
and Smith, S. O. (2005) Changes in interhelical bovine opsin gene and its mutants in stable
hydrogen bonding upon rhodopsin activation. mammalian cell lines. Proc. Natl. Acad. Sci.
J. Mol. Biol. 347, 803–812. U.S.A. 93, 11487–11492.
Chapter 5
Abstract
The cell-free expression system using an Escherichia coli extract is a practical method for producing
isotope-labeled proteins. The advantage of the cell-free system over cellular expression is that any isotope-
labeled amino acid can be incorporated into the target protein with minimal scrambling, thus providing
opportunities for advanced isotope labeling of proteins. We have modified the standard protocol for E. coli
cell-free expression to cope with two problems specific to NMR sample preparation. First, endogenous
amino acids present in the E. coli S30 extract lead to dilution of the added isotope. To minimize the content
of the remaining amino acids, a gel filtration step is included in the preparation of the E. coli extract.
Second, proteins produced by the cell-free system are not necessarily homogeneous due to incomplete
processing of the N-terminal formyl-methionine residue, which complicates NMR spectra. Therefore, the
protein of interest is engineered to contain a cleavable N-terminal histidine-tag, which generates a homo-
geneous protein after the digestion of the tag. Here, we describe the protocol for modified E. coli cell-free
expression.
Key words: Cell-free synthesis, S30 extract from E. coli, NMR, Stable isotope labeling
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_5, © Springer Science+Business Media, LLC 2012
71
72 M. Takeda and M. Kainosho
2. Materials
2.1. Preparation 1. RNase-free Milli-Q PF Plus water (Millipore) (see Note 1).
of E. coli S30 Extract 2. LB medium: Dissolve 20 g of Luria Broth powder in 1 L of
water.
3. Incomplete rich medium: Dissolve 5.6 g of KH2PO4, 28.9 g of
K2HPO4, 1 g of Bacto yeast extract, and 1.5 mg of thiamine
hydrochloride in 1 L of water, and autoclave the solution. Cool
to room temperature and add 50 mL of 40% (w/v) D-glucose
and 10 mL of 100 mM Mg(OAc)2.
4. BL21 Star (DE3) cells (Invitrogen) (see Note 2).
5. French press (see Note 3).
6. S30 buffer: Combine 10 mL of 1 M Tris–OAc, pH 8.2, 10 mL
of 1.4 M Mg(OAc) 2, and 10 mL of 6 M KOAc in a final
volume of 1 L of RNase-free water, atuoclave, and add 1 mL
of 1 M DTT.
7. Gel filtration apparatus: Pack Sephadex G25 resin into a
2.5 × 20-cm chromatography column and set the column verti-
cally at 4°C. After the resin settles, attach a funnel to the top of
the column and equilibrate with 500 mL of S30 buffer (see
Note 4).
8. Dialysis tubing (molecular weight cutoff: 6–8 kDa).
9. Polyethylene glycol (PEG)-8000: Avg. MW 8,000 Da.
10. 2-mercaptoethanol.
11. Diethyl pyrocarbonate (DEPC).
12. Centrifugation tube: Wash the tubes with DEPC water and
autoclave to completely remove RNase.
13. 40% (w/v) D-glucose.
14. 100 mM Mg(OAc)2.
15. 1 M Tris–OAc, pH 8.2.
16. 1.4 M Mg(OAc)2.
17. 6 M KOAc.
18. 1 M DTT.
74 M. Takeda and M. Kainosho
3. Methods
Fig. 1. Flowchart of the cell-free strategy (reproduced from ref. 14 with permission from
Elsevier Science).
5 Cell-Free Protein Production for NMR Studies 77
a 105
110
Q3E2
T5 Q3E2
Q3E2 Q3E2
T5 T70
115
D2 D2 T70
15N (p.p.m.)
R74 R74
I9 Q3
I9 Q3 120
K77
D78 K77
D78
M71
A73
K13 A73 M71
K13
125
A1 L4 L4
M0
130
11 10 9 8 7 6
1H (p.p.m.)
b G113
105
G33
N137D2 N137D2
G40 G59 Q135E2 Q135E2
T62
G96 G132
V55
G23
Q8E2 Q41E2 N53D2
110
T146 Q41E2
Q8E2
N53D2 N111D2
T5 N111D2 Q143E2
G98 Q143E2 Q3E2 N42D2
N42D2 Q3E2
G25 G134 T29 T26 Q49E2
S17 Q49E2
D58
T44 D95 D22 M145
G61
T117 T110 T79 N60D2
T70
N97D2
Q135
E127
R90
F19
N60D2 115
T28 E54
N42 M109 Y99
M72 D93 N97D2
15N (p.p.m.)
D131
D129 F92
S81 E139 D2E67 D20 R74
R106
N53 Q49 N60 T34
Y138 A103
I125 M76 I52 Q143
Q3 V91
F65 M36 Q41 K75
F89 L69 R126 E87 H107 E47
E119
F16 R37 E84 S38V108 L112 M51 E11
A128
V142 D122 M124
E140
L32 E7
N97
I9 E83D50 M144
E123 Q8 E104 120
E6 E120 L39
E45 F12 D24 L48 E114K77 E14 K30
L18
D133 A46
A10 E31
D118 L105 V121
R86 D78 D56
M71
A102 F68 I85 A88 L4 V35
A73
K13
E82 A15 N111
A1 D80
K21
S101 I63
K115 K148 125
V136 F141 L116 K94
I130 A147
D64
I27
I100
N137
A57
130
11 10 9 8 7 6
Fig. 2. (1H, 15N) HSQC experiments of (U-13C, 15N) CaM. (a) CaM synthesized by the E. coli
cell-free system. (b) CaM synthesized with the N-terminal tag following its removal by
thrombin digestion. The extra peaks in (a) and the peaks in (b) are labeled with their
assignments (reproduced from ref. 14 with permission from Elsevier Science).
5 Cell-Free Protein Production for NMR Studies 79
3.2. Preparation of T7 1. Using a glycerol stock of E. coli strain BL21 (DE3) trans-
RNA Polymerase formed with a plasmid encoding T7 RNA polymerase, inocu-
late 10 mL of LB medium containing the appropriate antibiotic
in a 50-mL tube and then incubate with shaking at 37°C
overnight.
2. Inoculate 1 L of M9 medium supplemented with tryptone
peptone, containing the appropriate antibiotic, with the 10 mL
culture from step 1 into a 2-L flask that has been prewarmed
to 37°C. Incubate the cells with shaking at 37°C until the culture
reaches an OD600 of 0.5.
3. Induce the expression of T7 RNA polymerase by adding IPTG
to a final concentration of 0.5 mM. Incubate the cells with
shaking for 8 h at 37°C and then centrifuge the culture at
5,000 × g for 10 min at 4°C. Store the cell pellet at −80°C.
80 M. Takeda and M. Kainosho
Fig. 3. 13C-filtered 1H-NMR spectra of CaM. (a) 15N-labeled CaM synthesized with the
conventional extract. (b) 2H-, 13C-, 15N-labeled CaM synthesized with the conventional S30
extract. (c) 2H-, 13C-, 15N-labeled CaM synthesized with the improved, dialyzed S30 extract.
All of the spectra are adjusted for the intensities of the amide region. Since the peaks of the
protons attached to 13C are filtered, only the protons attached to 12C give rise to resonances
in the aliphatic region (reproduced from ref. 14 with permission from Elsevier Science).
3.3. Cell-Free 1. Wear sanitary gloves (see Note 22) and thaw the LM mixture
Synthesis Reaction and the S30 extract (see Note 23).
2. Prepare the reaction and dialysis solutions (Subheading 2.3).
In the case of a small-scale reaction for evaluating expression
levels, the typical volumes of the reaction and dialysis solu-
tions are 0.5 and 2.0 mL, respectively. For large production
quantities, each volume is scaled up while maintaining the
volume ratio of the reaction solution to the dialysis solution
(see Note 24).
3. Transfer the dialysis solution to an RNase-free vessel. Tie one
end of the dialysis tube firmly. Transfer the reaction solution to
the dialysis tube, and tie off the other end. Fold the tubing 4–6
times and place it into the vessel such that the dialysis tube is
completely submerged in the dialysis solution.
82 M. Takeda and M. Kainosho
4. Notes
Acknowledgments
References
1. Clemens, M.J., and Prujin, G.J. (1999) Protein protein structure determinations. Nature 440,
synthesis in Eukaryotic Cell-free systems. Oxford 52–57.
University Press, New York, pp. 129–165. 12. Takeda, M., Ikeya, T., Güntert, P., and
2. Kramer, G., Kudlicki, W., and Hardesty, B. Kainosho, M. (2007) Automated structure
(1999) Cell-free Coupled Transcription- determination of proteins with the SAIL-FLYA
Translation systems from Escherichia coli. Oxford NMR method. Nature Protocol 2, 2896–2902.
University Press, New York, pp. 129–165. 13. Kainosho, M., and Güntert, P. (2010) SAIL-
3. Zubay, G. (1973) In vitro synthesis of protein Stereo-array isotope labeling. Q. Rev. Biophys.
in microbial systems. Ann. Rev. Gen. 7, 267–287. 7, 1–54.
4. Spirin, A.S., Barano, V.I., Ryabova, L.A., 14. Torizawa, T., Shimizu, M., Taoka, M., Miyano,
Ovodov, S.Y., and Alakhov, Y.B. (1988) A con- H., and Kainosho, M. (2004) Efficient produc-
tinuous cell-free translation system capable of tion of isotopically labeled proteins by cell-free
producing polypeptides in high yield. Science synthesis: A practical protocol. J. Biomol. NMR
242, 1162–1164. 30, 311–325.
5. Kim, D.M., Kigawa, T., Choi, C.Y., and 15. Pratt, J.M. (1984) Transcription and Trans-
Yokoyama, S. (1996) A highly efficient cell-free lation: A Practical Approach, IRL Press,
protein synthesis system from Escherichia coli. New York, pp. 179–209.
Eur. J. Biochem. 239, 881–886. 16. Davanloo, P., Rosenberg, A.H., Dunn, J.J.,
6. Kim, D.M., and Swartz, J.R. (2000) Prolonging and Studier, F.W. (1984) Cloning and expres-
cell-free protein synthesis by selective reagent sion of the gene for bacteriophage T7 RNA
additions. Biotechnol. Prog. 16, 385–390. polymerase. Proc. Natl. Acad. Sci. U.S.A. 81,
7. Kigawa, T., Yabuki, T., Yoshida, Y., Tsutsui, 2035–2039.
M., Ito, Y., Shibata, T., and Yokoyama, S. 17. Zawadzki, V., and Gross, H.J. (1991) Rapid
(1999) Cell-free production and stable-isotope and simple purification of T7 RNA polymerase.
labeling of milligram quantities of proteins. Nucl. Acid Res. 19,1948.
FEBS Lett. 442,15-19. 18. Grodberg, J., and Dunn, J.J. (1988) ompT
8. Madin, K., Sawasaki, T., Ogasawara, T., and encodes the Escherichia coli outer membrane
Endo, Y. (2000) A highly efficient and robust protease that cleaves T7 RNA polymerase dur-
cell-free protein synthesis system prepared from ing purification. J. Bacteriol. 170, 1245–1253.
wheat embryos: plants apparently contain a sui- 19. Huang, Y.H., Leblanc, P., Apostolou, V.,
cide system directed at ribosomes. Proc. Natl. Stewart, B., and Moreland, R.B. (1995)
Acad. Sci. U.S.A. 97, 559–564. Comparison of Milli-Q PF Plus water to
9. Shimizu, Y., Inoue, A., Tomari, Y., Suzuki, T., DEPC-treated water in the preparation and
Yokogawa, T., Nishikawa, K., and Ueda, T. (2001) analysis of RNA. Biotechniques 19, 656–661.
Cell-free translation reconstituted with purified 20. Zawada, J., and Swartz, J.R. (2006) Effects of
components. Nat. Biotechnol. 19, 751–755. growth rate on cell extract performance in cell-
10. Kigawa, T., Muto, Y., and Yokoyama, S. (1995) free protein synthesis. Biotechnol. Bioeng. 94,
Cell-free synthesis and amino acid-selective 618–624.
stable isotope labeling of proteins for NMR 21. Liu, D.V., Zawada J.F., and Swartz, J.R.
analysis. J. Biomol. NMR 6, 129–134. (2005) Streamlining Escherichia coli S30
11. Kainosho, M., Torizawa, T., Iwashita, Y., extract preparation for economical cell-
Terauchi, T., Ono. A,M., and Güntert, P. free protein synthesis. Biotechnol. Prog. 21,
(2006) Optimal isotope labelling for NMR 460–465.
Chapter 6
Abstract
Although cell-free expression is a relative newcomer to the biochemical toolbox, it has already been
reviewed extensively, even in the more specialized cases such as membrane protein expression, nanolipo-
protein particles, and applications to crystallography and nuclear magnetic resonance (NMR). Solid-state
NMR is also a newcomer to the structural biology toolbox, with its own specificities in terms of sample
preparation. Cell-free expression and solid-state NMR are a promising combination that has already proven
useful for the structural study of membrane proteins in their native environment, the hydrated lipid bilayer.
We describe below several protocols for preparing MscL, a mechanosensitive membrane channel, using
cell-free expression destined for a solid-state NMR study. These protocols are flexible and can easily be
applied to other membrane proteins, with minor adjustments.
Key words: In vitro synthesis, Integral membrane proteins, Membrane protein reconstitution,
Liposomes, Solid-state NMR
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_6, © Springer Science+Business Media, LLC 2012
85
86 A. Abdine et al.
2. Materials
2.1. Cell-Free All solutions are prepared with autoclaved nuclease-free Milli-Q
Expression of the water (see Note 1).
Mechanosensitive
1. Thermoregulated shaker for the cell-free reaction vessel.
Channel MscL in
Detergent Micelles 2. Nuclease-free 50-mL tubes.
3. Nuclease-free 1.5-mL tubes.
4. Nuclease-free glass pipettes.
5. Nuclease-free pipette tips (0–10, 10–200, 200–1,000 μL, and
1–5 mL), autoclaved at 121°C for 20 min.
6. pIVEX-2.3-mscL plasmid (3). Encodes the E. coli MscL
C-terminally fused to a His6 tag under the control of the T7
promoter. Dissolve in pure nuclease-free water and store ali-
quoted at −20°C, at 0.5 μg/μL.
7. RTS 9000 cell-free expression kit (Roche/5Prime). Contains
lyophilized E. coli lysate, reaction mix, feeding mix, reconstitu-
tion buffer, continuous exchange reaction vessel, and a syringe
(see Fig. 1, Notes 2–5). Store at −80°C.
88 A. Abdine et al.
Fig. 1. Roche/5Prime RTS 9000 reaction vessel. The reaction compartment (10 mL) and the
feeding compartment (100 mL) are each accessible through two screws. The reaction
compartment contains the cell lysate, plasmid DNA, detergent or preformed liposomes and
is where the coupled transcription/translation reaction takes place. The separate feeding
compartment provides additional ions, energy substrates, nucleotides and amino acids,
through a semipermeable membrane (MW cutoff 10 kDa). Simultaneously, by-products
that may inhibit the reaction are diluted through the same membrane into the feeding
compartment. This continuous exchange allows cell-free expression to last for up to 24 h.
Table 1
Preparation of amino acid solutions for MscL cell-free expression
in a volume V = 110 mL (see Subheading 3.1 and Fig. 1)
2.2. Cell-Free All solutions are prepared with autoclaved nuclease-free Milli-Q
Expression of the water (see Note 1).
Mechanosensitive
1. Thermoregulated shaker for the cell-free reaction vessel.
Channel MscL in
Liposomes 2. Mini-Extruder with two 1-mL gastight microsyringes and
0.1-μm cutoff polycarbonate membranes.
3. Nuclease-free 50-mL tubes.
4. Nuclease-free 1.5-mL tubes.
5. Nuclease-free glass pipettes.
6. Nuclease-free pipette tips (0–10, 10–200, 200–1000 µL and
1–5 mL). Autoclave at 121°C for 20 min.
6 Cell-Free Membrane Protein Expression for Solid-State NMR 91
2.3. Sample All solutions described should be prepared using Milli-Q water and
Characterization stored at room temperature unless stated otherwise.
1. FA diluted solution: 37% formaldehyde in water, store at 4°C.
2. Dimethyl sulfoxide (DMSO).
3. 1 M NaOH: To adjust pH.
4. 1 M HCl: To adjust pH.
5. 1 M Tris base, pH 8.
6. SDS running buffer: 100 mM Tris base, 100 mM HEPES,
0.1% SDS, adjust to pH 8 using NaOH or HCl.
7. Staining buffer: 0.5% Coomassie Brilliant Blue R-250 dye, 50%
ethanol and 10% acetic acid.
8. Destaining buffer: 20% ethanol and 10% acetic acid.
9. 2× SDS loading buffer: 0.2 M Tris base, pH 8, 8% SDS,
40% glycerol, 0.4% bromophenol blue, 0.4 M DTT, adjust pH
8 using 1 M NaOH or 1 M HCl.
10. 1× SDS loading buffer: dilute 1 mL of 2× SDS loading buffer
with 1 mL of water.
11. DSS solution: 50 mM disuccinimidylsuberate in DMSO.
12. HEPES–KCl buffer (see Subheading 2.1).
13. Polyvinylidene fluoride (PVDF) membranes.
14. TBS-Tween buffer 1: 50 mM Tris base, pH 7.4, 150 mM
NaCl, 0.05% Tween-20.
15. Monoclonal anti-histidineperoxidase-conjugate antibody solu-
tion (for the dot blot): Dilute 2,000× from the stock solution
with TBS-Tween buffer 1, prepare fresh.
16. Diaminobenzidine solution: Dissolve one tablet of diamin-
obenzidine and one tablet of urea hydrogen peroxide in 5 mL
of water, prepare fresh.
17. Western-blotting detection kit.
18. Nonfat dry milk (powder).
19. Chemiluminescence reagents (H2O2 and luminol), store at 4°C.
20. Towbin buffer: 25 mM Tris base, 192 mM glycine, 20% iso-
propanol, and 0.1% SDS. Store without isopropanol at 4°C,
add isopropanol just before use.
21. Ponceau S solution: 0.1% Ponceau S, 5% acetic acid.
22. TBS-Tween buffer 2: TBS-Tween buffer 1 with 5% (w/v)
nonfat dry milk, freshly made.
23. Mouse anti-histidine antibody solution (for the Western-blot):
Dilute 5,000× from the stock solution with TBS-Tween buffer 1,
prepare fresh.
24. Anti-mouseperoxidase-coupled antibody solution (for the
Western-blot): Dilute 10,000× from the stock solution with
TBS-tween buffer 1, prepare fresh.
6 Cell-Free Membrane Protein Expression for Solid-State NMR 93
2.4. NMR Sample Four millimeter diameter rotors for high-resolution magic-angle
Preparation spinning solid-state NMR (see Fig. 2).
Fig. 2. Four millimeter diameter rotors for high-resolution magic-angle spinning solid-
state NMR. The sample (approximately 50 μL) is contained in the center of the zirconium
rotor, thanks to the Teflon insert bottom and top, and is kept tightly in place by the top
screw and the Kel-F cap.
94 A. Abdine et al.
3. Methods
3.1. Cell-Free The method below describes the production of an MscL sample
Expression for solid-state NMR, using a cell-free protein expression system in
of the Mechano- detergent micelles. Although any labeling scheme could be per-
sensitive Channel formed, the sample described below was labeled on isoleucines and
MscL in Detergent threonines, using 13C/15N-labeled amino acids, and was subse-
Micelles quently used in solid-state NMR experiments (14).
3.1.1. Day 1: Cell-Free 1. Calculate the amount of amino acids to be incorporated during
Protein Expression in protein expression. The concentration of each amino acid in
Detergent Micelles the protein cell-free expression system (V = 110 mL, see Fig. 1)
is usually ~1 mM. For a given protein, each amino acid con-
centration depends on its occurrence, n, in the protein
sequence. In our case, we have obtained good results with a
final concentration of each amino acid equal to n/10, expressed
in mM. The corresponding weight of each amino acid, of
molecular weight MWi, is therefore (n × V × MWi/10),
expressed in mg. The corresponding volume of each amino
acid solution of concentration Ci, where Ci is 168 or 140 μM
(Subheading 2.1), is (n × V/10)/Ci, expressed in mL. The
weights and final volumes calculated for this protocol are sum-
marized in Table 1.
2. Thaw the DTT, RTS reconstitution buffer, and MscL plasmid
at room temperature.
3. Thaw the other components of the RTS kit (E. coli lysate, reac-
tion mix, and feeding mix) on ice.
4. Prepare the unlabeled amino acid solutions (see Subheading 2.1).
5. Prepare the labeled amino acid solutions (see Subheading 2.1).
6. Add the labeled amino acid solution to the unlabeled amino
acid solution.
7. Add 3 mL of 40 mM DTT, sonicate (80 W for 5 min or until
clear), and store on ice.
8. Reconstitute the lyophilized E. coli lysate in 5.2 mL of recon-
stitution buffer. Shake the bottle gently (see Note 8).
9. Reconstitute the lyophilized reaction mix in 2.2 mL of recon-
stitution buffer. Shake the bottle gently (see Note 8).
6 Cell-Free Membrane Protein Expression for Solid-State NMR 95
3.1.2. Day 2: Purification 1. After 22 h, stop the cell-free protein expression. Remove a 20 μL
of Detergent/Protein aliquot of the reaction run-on solution for gel electrophoresis.
Micelles 2. Remove the reaction run-on mix using a pipette, add 100 μL
of the AEBSF solution to prevent protease digestion, and stir
at room temperature for 15 min (see Note 9).
3. Dilute the sample to a final volume of 50 mL using FPLC
buffer A1 and centrifuge at 10,000 × g for 15 min at 4°C.
4. Connect the chelating column to the FPLC system (see Note 10).
Pass 4–5 column volumes (CV) of distilled water through
the column, to wash away the ethanol solution in which it has
been kept.
5. Degas the FPLC buffers for 10 min in an ultrasonic bath. Wash
the pump inlets with the degassed solutions. Wash the system
with FPLC buffer A1 until the UV and conductivity baselines
are stable.
96 A. Abdine et al.
3.1.3. Day 3: Sample 1. Dialyze a second time against 1 L of 1× dialysis buffer for 2 h
Characterization and at 4°C. Remove a 20 μL aliquot for gel electrophoresis.
Reconstitution into 2. Assess the cell-free expression level, FPLC purification effi-
Liposomes ciency, and oligomeric state by SDS-PAGE (see Subheadings
3.3.1 and 3.3.3).
6 Cell-Free Membrane Protein Expression for Solid-State NMR 97
3.1.4. Day 4: Preparation 1. Add another 300 mg of wet polystyrene beads to each tube.
of the Sample for Incubate for another 2 h, at 4°C.
Solid-State NMR 2. After 2 h, set the tubes vertically for 5 min and discard the
pellet and the beads.
3. Divide the supernatant into four tubes and centrifuge at
100,000 × g for 30 min at 4°C. Remove a 20 μL aliquot of one
supernatant for gel electrophoresis, discard the rest and resus-
pend the pellets with 1 mL of HEPES–KCl buffer for each
pellet. Split the new resuspended pellets into two tubes and
centrifuge at 100,000 × g for 30 min at 4°C. Repeat one more
time so that the entire sample fits into a single tube. Remove a
final 20 μL aliquot of the supernatant for gel electrophoresis.
Dry the final pellet under argon and store it at −20°C.
4. Assess the reconstitution by SDS-PAGE (see Subheading 3.3.1),
and by proteoliposome density characterization on a sucrose gra-
dient (see Subheading 3.3.9). Also, estimate the lipid and water
content (see Subheadings 3.3.8 and 3.3.10, and Note 15).
5. The average quantity of protein obtained with the RTS 9000
kit is 10 mg in 10 mL of reaction mix, which is generally suf-
ficient to prepare two or three NMR samples. Transfer ~30 mg
of pelleted sample to the NMR rotor (see Subheading 3.4).
6. Store the remaining pellet at −20°C.
98 A. Abdine et al.
3.2. Cell-Free The method below describes the production of an MscL sample
Expression of the for solid-state NMR, using a cell-free protein expression system
Mechanosensitive directly into lipid vesicles. Although any labeling scheme could be
Channel MscL performed, the sample described below was labeled on arginines,
in Liposomes isoleucines, methionines, phenylalanines and prolines, using
13
C/15N-labeled amino acids, and was subsequently used in solid-
state NMR experiments (19). Lipids used were synthetic DOPC,
but other lipids or complex mixtures can also be used, such as
asolectin (18) or nanodiscs (17).
3.2.1. Day 1: Cell-Free 1. Thaw the DTT solution, RTS reconstitution buffer and MscL
Protein Expression plasmid at room temperature.
2. Thaw the other components of the RTS kit (E. coli lysate, reac-
tion mix, and feeding mix) on ice.
3. Prepare the DOPC liposomes (Subheading 2.2).
4. Prepare the HEPES solutions (Subheading 2.1).
5. Prepare the unlabeled amino acid solutions (Subheading 2.1).
6. Prepare the labeled amino acid solutions (Subheading 2.1).
7. Add the labeled amino acid solutions to the unlabeled amino
acid solutions.
8. Add 3 mL of 40 mM DTT, sonicate (80 W for 5 min or until
clear), and store on ice.
9. Reconstitute the lyophilized E. coli lysate in 2.7 mL of recon-
stitution buffer. Shake the bottle gently (see Note 8).
10. Add the liposome solution (2.5 mL) to the E. coli lysate once
the latter is completely dissolved.
11. Reconstitute the lyophilized reaction mix in 2.2 mL of recon-
stitution buffer. Shake the bottle gently (see Note 8).
12. Reconstitute the lyophilized feeding mix in 80 mL of reconsti-
tution buffer. Shake the bottle gently (see Note 8).
13. Prepare the feeding solution by adding 26 mL of the reconsti-
tuted amino acid solution (mixture of unlabeled and labeled
amino acid solutions) and 3 mL of 40 mM DTT to the feeding
mix solution (80 mL).
14. Prepare the reaction solution by adding the reconstituted reac-
tion mix (2.2 mL), 2.7 mL of the reconstituted amino acid solu-
tion and 0.3 mL of 40 mM DTT to the reconstituted E. coli
lysate with the liposomes (5.2 mL). Take a 20 μL aliquot of this
reaction solution for later comparison on gel electrophoresis.
15. Add 300 μL of the MscL plasmid (~150 μg) to the reaction
solution.
16. Open both screws of the reaction compartment and fill the
reaction compartment with the reaction solution using a nucle-
ase-free pipette (see Fig. 1). Remove any air bubbles by tapping
the vessel lightly.
6 Cell-Free Membrane Protein Expression for Solid-State NMR 99
17. Open both screws of the feeding compartment and fill the
feeding compartment with the feeding solution, using the
provided syringe (see Fig. 1). Remove any air bubbles by
tapping the vessel lightly.
18. Insert the reaction vessel into a thermoregulated shaker.
19. Set the shaking speed to 800 rpm.
20. Set the temperature to 30°C.
21. Incubate the reaction for 22 h.
3.2.2. Day 2: Preparation 1. After 22 h, stop the cell-free protein expression. Remove a
of the Sample for 20 μL aliquot of the reaction run-on solution for gel
Solid-State NMR electrophoresis.
2. Remove the reaction run-on mix using a pipette, add 100 μL
of the AEBSF solution to prevent protease digestion, and stir
at room temperature for 15 min (see Note 9).
3. Split the sample into six tubes and centrifuge at 100,000 × g for
30 min at 4°C. Remove a 20 μL aliquot of one supernatant for
gel electrophoresis, discard the rest and resuspend each pellet
with 1 mL of HEPES–KCl buffer. Split the resuspended pellets
into three tubes and centrifuge at 100,000 × g for 30 min at
4°C. Repeat one more time so that the entire sample fits into a
single tube. Remove a final 20 μL aliquot of the supernatant
for gel electrophoresis. Dry the final pellet under argon and
store it at −20°C.
4. Assess cell-free expression efficiency and oligomeric state by
SDS-PAGE (see Subheadings 3.3.2 and 3.3.3). At this point,
the expressed protein is generally almost pure.
5. If the protein was expressed with a polyhistidine tag, after SDS-
PAGE, wash the gel in water and perform a Western-blot
transfer (see Subheading 3.3.5).
6. Estimate the protein and lipid concentration using the BCA
and Rouser methods respectively (see Subheadings 3.3.7 and
3.3.8, and Note 13).
7. Characterize the proteoliposome density on a sucrose gradient
(see Subheading 3.3.9), and estimate the water content (see
Subheading 3.3.10 and Note 15).
8. The average quantity of protein obtained with the RTS 9000
kit is 10 mg in 10 mL of reaction mix, which is generally suf-
ficient to prepare two or three NMR samples. About 30 mg of
pelleted sample is transferred into the NMR rotor (see
Subheading 3.4).
9. Store the remaining pellet at −20°C.
100 A. Abdine et al.
3.3.3. Electrophoresis If the protein is an oligomer (like MscL), assess its integrity by
for Oligomeric State cross-linking the protein with different free primary amine target-
Characterization ing reagents, like formaldehyde (FA) or disuccinimidylsuberate
(DSS) (3, 25). NB: the protein buffer must not contain primary
amines (such as in Tris).
1. Mix 5 μg of purified MscL, either in detergent or liposomes, in
10 μL of HEPES–KCl buffer and 0.54 μL of the FA diluted
solution.
2. Mix 5 μg of purified MscL, either in detergent or liposomes, in
10 μL of HEPES–KCl buffer and 0.2 μL of 50 mM DSS.
3. Incubate both samples, without shaking, for 30 min at room
temperature.
6 Cell-Free Membrane Protein Expression for Solid-State NMR 101
Fig. 3. Coomassie blue stained SDS-PAGE of MscL: (1) purified in detergent, (2) cross-
linked with formaldehyde, (3) cross-linked with disuccinimidylsuberate. Formaldehyde
generates five protein bands on the gel, corresponding to various oligomeric forms, from
monomers to pentamers, while disuccinimidylsuberate generates mostly pentamers, con-
firming the pentameric nature of Escherichia coli MscL (3, 25).
3.3.4. Dot Blot 1. Deposit 2 μL of each fraction of purified protein directly onto
a PVDF membrane.
2. Allow the protein solution to air-dry.
3. Prepare the TBS-Tween buffers 1 and 2.
4. Prepare the monocolonal antibody solution.
5. When the protein spots are completely dry, incubate the mem-
brane in the TBS-Tween buffer 2 for 30 min.
6. Wash the membrane twice for 5 min with the TBS-Tween
buffer 1.
102 A. Abdine et al.
3.3.6. Protein Protein concentration is measured using the BCA method (26, 27),
Quantification using the test tube procedure of the Micro BCA Protein Assay Kit:
in Detergents
1. Prepare eight BSA standards by diluting the stock solution
with the solubilization buffer (the final BSA concentration
should fall between 0.5 and 20 μg/mL). Make three replicates
of each dilution. Introduce 500 μL of each into 1.5-mL tubes
(24 tubes in total) and store at room temperature.
2. Dilute the unknown protein sample with the solubilization
buffer to three different concentrations that are expected to be
between 1 and 20 μg/mL. Make three replicates of each dilu-
tion. Transfer 500 μL of each sample into 1.5-mL tubes (nine
tubes in total).
3. Add 500 μL of Micro BCA Working Reagent to each of the
1.5-mL tubes (24 BSA standards and nine unknown proteins)
and incubate for 60 min at 60°C. Cool to room temperature.
4. Measure the absorbance at 562 nm.
5. Plot a standard curve based on the absorbance of the BSA
samples.
6. Deduce the protein concentration of each unknown protein
sample.
3.3.7. Protein Lipids need to be removed from the sample by precipitating the
Quantification proteins using cold acetone, followed by a centrifugation to sepa-
in Proteoliposomes rate the protein pellet from the supernatant containing the lipids:
1. Prepare the 24 BSA standard tubes (Subheading 3.3.6, step 1).
2. Dilute the unknown protein sample with the solubilization
buffer to three different concentrations that are expected to be
between 1 and 20 μg/mL. Make three replicates of each dilu-
tion. Transfer 500 μL of each sample into 1.5-mL tubes (nine
tubes in total).
3. Add 1 mL of cold acetone to each tube of unknown protein.
Vortex the tubes and incubate for 60 min at −20°C. Centrifuge
at 10,000 × g for 10 min at room temperature. Discard the
supernatant. Incubate the tubes for 30 min at room tempera-
ture to allow for acetone evaporation. Add 500 μL of the solu-
bilization buffer, and vortex the tubes again.
4. Add 500 μL of Micro BCA Working Reagent to each of the
1.5-mL tubes (24 BSA standards and nine unknown proteins)
and incubate for 60 min at 60°C. Cool to room temperature.
5. Measure the absorbance at 562 nm.
6. Plot a standard curve based on the absorbance of the BSA
samples.
7. Deduce the protein concentration of each unknown protein
sample.
104 A. Abdine et al.
3.3.8. Lipid Quantification The phospholipid content in the proteoliposome sample is assessed
in Proteoliposomes by the Rouser method, which measures the phosphate concentra-
tion in the sample (28). Full eye, face, and skin protection is
required for this method.
1. Set a heating block at 180°C.
2. Wash the glass tubes in nitric acid solution before use and dry
them in an oven.
3. Prepare five phosphate standard samples by diluting the phos-
phate stock solution into five samples containing 1–5 μg of
KH2PO4 per tube. Classically, 5 μg of KH2PO4 give an absor-
bance of 0.9 at 800 nm. Store at 4°C.
4. Prepare the ascorbic acid solution (Subheading 2.3).
5. Collect three proteoliposomes volumes containing approxi-
mately 1–5 μg of lipids. Transfer the samples into clean glass
tubes.
6. Add 0.65 mL of the perchloric acid solution and place the
tubes in the heated block at 180°C for 30 min of digestion,
until the yellow color has disappeared.
7. Add 0.65 mL of the perchloric acid solution per phosphate
standard tube (digestion is not necessary).
8. Put all the tubes on ice, and set the heated block at 100°C
(alternatively, a boiling water bath can be used).
9. Once cooled, add to the tubes the following: 3.3 mL of water,
0.5 mL of the ammonium molybdate solution, and 0.5 mL of the
ascorbic acid solution. Agitate on a vortex after each addition.
10. Put the tubes in the heated block at 100°C for 5 min.
11. Put the tubes on ice for 5 min.
12. Read the absorbance of the cooled samples at 800 nm.
13. Plot a standard curve based on the absorbance of the phos-
phate standard samples. The exact concentration in the phos-
pholipid samples is calculated based on the standard curve.
3.3.9. Proteoliposome A discontinuous sucrose flotation gradient analysis (21, 29) can be
Density Characterization performed to check that the membrane protein is correctly recon-
stituted into proteoliposomes. The proteoliposomes can be layered
at the bottom (see below) or at the top (see Note 16) of the sucrose
layers.
1. After MscL reconstitution, add 120 mg of sucrose to 0.2 mL of
the proteoliposome suspension (containing approximately 96 μg
of protein) and mix gently by pipetting, until the sucrose has
completely dissolved. The final volume is approximately
0.265 mL and the final sucrose concentration is 45%. Adjust the
final volume to 0.4 mL with the buffered 45% sucrose solution.
6 Cell-Free Membrane Protein Expression for Solid-State NMR 105
3.4. NMR Sample 1. Weigh the empty rotor, cap, insert bottom, top, and screw
Preparation (see Fig. 2).
2. Introduce the insert bottom into the rotor.
3. Introduce a couple of mg of the pelleted sample into the rotor
on the tip of a spatula. Place the rotor into a 1.5-mL tube and
centrifuge it at 10,000 × g for 1 min at room temperature.
Repeat this until approximately 30 mg of sample has been
introduced into the rotor.
4. Place the insert top into the rotor, without the top screw. Clean
the upper part the rotor with a precision wiper before capping
tightly. At this point, the rotor containing sample should be
stored at −20°C until the NMR experiment is performed.
5. Introduce the rotor into the magic-angle spinning NMR probe
and into the magnet of the NMR spectrometer. Spin the rotor
at ~10 kHz for 15 min.
106 A. Abdine et al.
6. Extract the rotor and open it. Clean the upper part of the rotor
again, in case a drop of water has come out. Place the top
screw, tighten it, and then cap the rotor tightly. Weigh the full
rotor to deduce the final sample mass. A typical sample consists
of 3 mg proteins, 12 mg lipids and 15 mg water.
4. Notes
Acknowledgments
References
1. Lyford, L. K., and Rosenberg, R. L. (1999) 4. Klammt, C., Lohr, F., Schafer, B., Haase, W.,
Cell-free expression and functional reconstitu- Dötsch, V., Ruterjans, H., Glaubitz, C., and
tion of Homo-oligomeric α7 Nicotinic Bernhard, F. (2004) High level cell-free expres-
Acetylcholine Receptors into Planar Lipid sion and specific labeling of integral membrane
Bilayers. J. Biol. Chem. 274, 25675–25681. proteins. Eur. J. Biochem. 271, 568–580.
2. Elbaz, Y., Steiner-Mordoch, S., Danieli, T., 5. Wagner, S., Bader, M. L., Drew, D., and de Gier,
and Schuldiner, S. (2004) In vitro synthesis of J. W. (2006) Rationalizing membrane protein
fully functional EmrE, a multidrug transporter, overexpression. Trends Biotechnol. 24, 364–371.
and study of its oligomeric state. Proc. Natl. 6. Eshaghi, S. (2009) High-throughput expression
Acad. Sci. U.S.A. 101, 1519–1524. and detergent screening of integral membrane
3. Berrier, C., Park, K. H., Abes, S., Bibonne, A., proteins. Methods Mol. Biol. 498, 265–271.
Betton, J. M., and Ghazi, A. (2004) Cell-free 7. Tate, C. G. (2010) Practical considerations of
synthesis of a functional ion channel in the membrane protein instability during purifica-
absence of a membrane and in the presence of tion and crystallisation. Methods Mol. Biol. 601,
detergent. Biochemistry 43, 12585–12591. 187–203.
6 Cell-Free Membrane Protein Expression for Solid-State NMR 109
8. Breyton, C., Pucci, B., and Popot, J.-L. (2010) 19. Abdine, A., Verhoeven, M. A., and
Amphipols and fluorinated surfactants: Two Warschawski, D. E. (2011) Cell-free expres-
alternatives to detergents for studying mem- sion and labeling strategies for a new decade
brane proteins in vitro. Methods Mol. Biol. 601, in solid-state NMR. New Biotechnol. 28,
219–245. 272–276.
9. White, S. H. (2010) Membrane Proteins of 20. He, M. (2008) Cell-free protein synthesis:
Known Structure. University of California at applications in proteomics and biotechnology.
Irvine. http://blanco.biomol.uci.edu/ New Biotechnol. 25, 126–132.
Membrane_Proteins_xtal.html Accessed 25 21. Rigaud, J.-L., and Lévy, D. (2003)
July 2011. Reconstitution of Membrane Proteins into
10. Warschawski, D. E. (2010) Membrane Proteins of Liposomes. Methods Enzymol. 372, 65–86.
Known Structure Determined by NMR. Drorlist. 22. Triba, M. N., Zoonens, M., Popot, J.-L.,
http://www.drorlist.com/nmr/MPNMR.html Devaux, P. F., and Warschawski, D. E. (2006)
Accessed 25 July 2011. Reconstitution and alignment by a magnetic
11. Schneider, B., Junge, F., Shirokov, V. A., Durst, field of a β-barrel membrane protein in bicelles.
F., Schwarz, D., Dötsch, V., and Bernhard, F. Eur. Biophys. J. 35, 268–275.
(2010) Membrane protein expression in cell- 23. Leitz, A. J., Bayburt, T. H., Barnakov, A. N.,
free systems. Methods Mol. Biol. 601, 165–186. Springer, B. A., and Sligar, S. G. (2006)
12. Sobhanifar, S., Reckel, S., Junge, F., Schwarz, Functional reconstitution of β2-adrenergic
D., Kai, L., Karbyshev, M., Löhr, F., Bernhard, receptors utilizing self-assembling Nanodisc
F., and Dötsch, V. (2010) Cell-free expression technology. BioTechniques 40, 601–612.
and stable isotope labelling strategies for mem- 24. De Grip, W. J., Van Oostrum, J., and Bovee-
brane proteins. J. Biomol. NMR 46, 33–43. Geurts, P. H. M. (1998) Selective detergent-
13. Lehner, I., Basting, D., Meyer, B., Haase, W., extraction from mixed detergent/lipid/protein
Manolikas, T., Kaiser, C., Karas, M., and micelles, using cyclodextrin inclusion com-
Glaubitz, C. (2008) The key residue for sub- pounds: a novel generic approach for the prep-
strate transport (Glu14) in the EmrE dimer is aration of proteoliposomes. Biochem. J. 330,
asymmetric. J. Biol. Chem. 283, 3281–3288. 667–674.
14. Abdine, A., Verhoeven, M. A., Park, K.-H., 25. Sukharev, S. I., Schroeder, M. J., and McCaslin,
Ghazi, A., Guittet, E., Berrier, C., Van D. R. (1999) Stoichiometry of the Large
Heijenoort, C., and Warschawski, D. E. (2010) Conductance Bacterial Mechanosensitive
Structural study of the membrane protein Channel of E. coli. A Biochemical Study. J.
MscL using cell-free expression and solid-state Membrane Biol. 171, 183–193.
NMR. J. Magn. Reson. 204, 155–159. 26. Smith, P. K., Krohn, R. I., Hermanson, G. T.,
15. Kalmbach, R., Chizhov, I., Schumacher, M. Mallia, A. K., Gartner, F. H., Provenzano, M.
C., Friedrich, T., Bamberg, E., and Engelhard, D., Fujimoto, E. K., Goeke, N. M., Olson, B.
M. (2007) Functional cell-free synthesis of a J., and Klenk, D. C. (1985) Measurement of
seven helix membrane protein: in situ insertion protein using bicinchoninic acid. Anal.
of bacteriorhodopsin into liposomes. J. Mol. Biochem. 150, 76–85.
Biol. 371, 639–648. 27. Wiechelman, K. J., Braun, R. D., and
16. Marques, B., Liguori, L., Paclet, M. H., Villegas- Fitzpatrick, J. D. (1988) Investigation of the
Mendéz, A., Rothe, R., Morel, F., Lenormand, bicinchoninic acid protein assay: identification
J.-L. (2007) Liposome-mediated cellular deliv- of the groups responsible for color formation.
ery of active gp91(phox). PLoS One 2, e856. Anal. Biochem. 175, 231–237.
17. Katzen, F., Fletcher, J. E., Yang, J. P., Kang, D., 28. Rouser, G., Fkeischer, S., and Yamamoto, A.
Peterson, T. C., Cappuccio, J. A., Blanchette, (1970) Two dimensional thin layer chromato-
C. D., Sulchek, T., Chromy, B. A., Hoeprich, P. graphic separation of polar lipids and determi-
D., Coleman, M. A., and Kudlicki, W. (2008) nation of phospholipids by phosphorus analysis
Insertion of membrane proteins into discoidal of spots. Lipids 5, 494–496.
membranes using a cell-free protein expression 29. Laird, D. M., Eble, K. S., and Cunningham, C.
approach. J. Proteome Res. 7, 3535–3542. C. (1986) Reconstitution of mitochondrial
18. Berrier, C., Guilvout, I., Bayan, N., Park, F0.F1-ATPase with phosphatidylcholine using
K.-H., Mesneau, A., Chami, M., Pugsley, A. P., the nonionic detergent, octylglucoside. J. Biol.
and Ghazi, A. (2011) Coupled cell-free synthe- Chem. 261, 14844–14850.
sis and lipid vesicle insertion of a functional oli- 30. Etezady-Esfarjani, T., Hiller, S., Villalba, C.,
gomeric channel MscL - MscL does not need and Wüthrich, K. (2007) Cell-free protein syn-
the insertase YidC for insertion in vitro. thesis of perdeuterated proteins for NMR
Biochim. Biophys. Acta 1808, 41–46. studies. J. Biomol. NMR 39, 229–238.
Chapter 7
Abstract
NMR analyses of the structure, dynamics, and interactions of the Src family kinases (SFKs) have been
hindered by the limited ability to obtain sufficient amounts of properly folded, soluble protein from bacterial
expression systems, to allow these studies to be performed in an economically viable manner. In this chapter,
we detail our attempts to overcome these difficulties using the catalytic domain (SrcCD) of c-Src, the pro-
totypical SFK, as an illustrative example. We describe in detail two general methods to express and purify
SrcCD from Escherichia coli expression systems in both fully active wild-type and kinase-deficient mutant
forms, allowing the efficient and cost-effective labeling by NMR-active isotopes for solution NMR studies.
Key words: Protein tyrosine kinases, Src-family kinases, Escherichia coli expression systems
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_7, © Springer Science+Business Media, LLC 2012
111
112 A. Piserchio et al.
Fig. 1. (a) Domain arrangement in inactive full-length c-Src (PDB ID: 2SRC). The residues comprising the linker between
the SH2 and the catalytic domain (CD), which engages the SH3 domain, is shown in ball-and-stick representation, as are
the residues that comprise the C-terminal tail. The regulatory tyrosine residues in the activation loop (Y416, unphosphory-
lated in inactive c-Src) and the C-terminal tail (Y527, phosphorylated in inactive c-Src) are also shown in ball-and-stick
representation. The ATP-analog, ANP (phosphoaminophosphonic acid-adenylate), bound to the ATP-binding site, is also
shown in ball-and-stick representation. (b) Expanded view of SrcCD with the N-lobe and C-lobes indicated. Important
regulatory elements that include the activation loop, Gly-rich loop, the HRD (that precedes the catalytic loop) and DFG
motifs are also indicated. Y416 is shown in a ball-and-stick representation and the αC helix is indicated by a dashed
ellipse. The ligand ANP is also shown bound to the ATP-binding pocket.
7 Expression and Purification of Src-family Kinases for Solution NMR Studies 113
2. Expression
of the Catalytic
Domain of c-Src The first strategy that we utilized for the bacterial expression of
in Escherichia coli c-Src involved (1) the addition of an N-terminal MBP fusion tag to
enhance both solubility and folding, (2) the coexpression of GroEL
and GroES chaperones to further facilitate the accumulation of
properly folded protein, an idea originally proposed by Cole (46),
and (3) a decrease in the growth temperature after induction to
15°C, thus reducing the rapid accumulation of large quantities of
proteins and macromolecular crowding effects that tend to
facilitate aggregation.
MBP is widely used to enhance the production of fused proteins
through a mechanism that is far from clear. Several possible mecha-
nisms have been proposed including the formation of large vesicular
aggregates with the misfolded or unfolded proteins contained
inside (47) or a general chaperoning effect (48). Our initial
attempts to produce the catalytic domain of the so-called “kinase-
dead mutant” bearing a Lys → Met mutation (SrcCDK295M) (49) in
M9 minimal medium showed that in fact, just the introduction of
a N-terminal MBP fusion, while not eliminating the presence of
protein in the insoluble inclusion bodies, drastically increased the
amount of soluble material. However, a majority of the protein
produced in this fashion found in solution consisted of soluble
aggregates, presumably formed by misfolded material. Observation
of similar phenomena has added credence to the speculation that
MBP fusions can form micellar-like systems, formed by unfolded
material in the core, and by the soluble MBP domain on the sur-
face (47). Though a relatively small amount of properly folded
SrcCDK295M was obtained by these preparations, it allowed us to
collect the first 15N-HSQC spectra of SrcCDK295M. However, the
quantity of soluble, properly folded SrcCD obtained using this
procedure was not sufficient for detailed NMR analyses. This was
because a large number of experiments, collected over long periods
of time, are required for the resonance assignment process, neces-
sitating multiple NMR samples produced using expensive labeling
supplies. Nevertheless, the observation that material suitable for
NMR experiments could be obtained from an E. coli expression
system was indeed encouraging.
The largest improvement in the yield of properly folded
SrcCDK295M came when the N-terminal MBP fusion system was
transformed into E. coli cells that overexpress GroEL and GroES
chaperones. Cole and coworkers have shown that overexpression
of these bacterial chaperones allow the purification of small amounts
of SrcCDK295M from E. coli (49). We found that a more substantial
improvement, also achievable in the minimal medium used for
NMR sample preparation, can be obtained by integrating the MBP
fusion and the GroELS systems. Furthermore, the same approach
116 A. Piserchio et al.
Fig. 2. Production and purification of SrcCDK295M. (a) SDS-PAGE gel (20%) showing different steps in the purification
process: lane 1 – GroEL and GroES (expressed alone), lane 2 – cell lysate supernatant, lane 3 – cell lysate pellet, lane 4 –
metal affinity column flow through, lane 5 – metal affinity column eluate, lane 6 – product after thrombin cleavage
(MBP – higher molecular weight and SrcCDK295M – lower molecular weight), lane 7 – SrcCDK295M after Q-column purification.
(b) Effects of controlled phosphorylation and dephosphorylation of SrcCDK295M on the Q column elution profiles. Dashed line:
sample preincubated with catalytic amounts of commercial wild-type c-Src and Mg+2/ATP. Dark solid line: sample preincu-
bated with alkaline phosphatase (nonspecific phosphatase). Light solid line: sample preincubated with Lyp phosphatase
(specific for phosphorylated Y416). The identity of pSrcCDK295M was confirmed by immunoblots using anti-pTyr antibodies
and MS/MS, following trypsin digestion.
Fig. 3. (a) SDS-PAGE gel (20%) showing the various steps in the purification of wild-type SrcCD (His6-tagged): lane 1 – cell
lysate supernatant, lane 2 – cell lysate pellet, lane 3 – metal affinity column flow-through (unbound material), lane 4 –
metal affinity column eluate, lane 5 – cSrc after purification using gel filtration. The positions of Cdc37, MBP-Lyp and SrcCD
in lanes 3–5, respectively, are highlighted by boxes. Samples in lanes 4 and 5 were concentrated by a factor of 10. (b) Size
exclusion column profile of the eluate from the metal affinity column purification. (c) Immunoblots of phosphorylated
wild-type SrcCD (preincubated with 2 mM ATP and 10 mM Mg+2) with anti-pTyr antibody. All immunoblots were visualized
using the Bio-Rad ChemiDoc system. Lane 1 – phosphorylated SrcCD, lane 2 – phosphorylated SrcCD after incubation
with alkaline phosphatase.
3. Materials
3.1. Expression 1. H-MBP-3C vector (58): a kind gift from Dr. Kaushik Dutta,
Vectors and New York Structural Biology Center (see Note 1).
Competent Cells 2. pACYCDuet vector: for simultaneous expression of two
proteins (see Note 2): SrcCD and Cdc37 should be cloned
into this vector following the protocol described below.
3. 0.1-cm gap sterile electroporation cuvette for bacteria.
4. Electroporator (see Note 3).
5. pREP4-groELS electrocompetent BL21 (DE3) cells
(see Note 4).
6. BL21 (DE3) T1 competent cells.
14
17. N ammonium chloride.
18. 0.22-μm Filters.
4. Methods
4.1. Expression and The detailed protocol provided below is for the bacterial expres-
Purification of the sion and purification of the catalytic domain of the kinase deficient
Catalytic Domain of construct (SrcCDK295M) or the wild-type domain in the presence of
c-Src as an N-terminal a high-affinity inhibitor.
MBP Fusion
4.1.2. Protocol for the 1. Lyse cells using a French Press (1,100 psi), followed by
Purification of SrcCD from centrifugation at 16,000 g for 20 min at 4°C. Transfer the
the MBP Fusion Protein supernatant to a centrifuge tube.
2. Check the pH of the supernatant (see Note 13), if it is below
7.5, adjust by adding small amounts of solid tris-hydroxym-
ethyl-aminomethane (Tris).
3. Transfer preequilibrated cobalt-resin into the centrifuge tube
containing the lysate supernatant, and incubate for about 1 h
under rotation or slight agitation at 4°C.
4. Pour the lysate and beads into an empty column. Elute the
lysate and wash the column with four bed volumes of lysis
buffer and two bed volumes of lysis buffer containing 3 mM
imidazole (3 μL Imidazole buffer/mL of lysis buffer).
5. Elute the protein with four bed volumes of the lysis buffer
containing 300 mM imidazole (combine three volumes of
Imidazole buffer with seven volumes of lysis buffer).
6. Add 20 μL of EDTA solution/mL of eluate (final concentra-
tion: 10 mM; see Note 14) to the eluate prior to dialysis and
then dialyze against cleavage buffer at 4°C.
7. Determine the concentration of the fusion protein by measuring
the OD280 using 119,100 M-1-cm-1 as the extinction coefficient.
8. Add 5 units of α-thrombin/mg of fusion protein to the solu-
tion, immediately add β-mercaptoethanol to a final concentra-
tion of 4 mM. Cleave overnight at 4°C.
9. Check for completion of the cleavage reaction by running an
SDS–PAGE gel. Add more thrombin if necessary. When the
cleavage is complete, add solid DTT and the protease inhibi-
tor, AEBSF (4-(2-aminoethyl) benzenesulfonyl fluoride) to
final concentrations of 5 mM and 100 μM, respectively. Add
AP23464, a high-affinity kinase inhibitor, to the cleavage buf-
fer to a final concentration of 100 μM for wild-type SrcCD.
10. Dilute the completed reaction fivefold into ion exchange chro-
matography buffer A, and load the resulting sample onto a
Q column (ion-exchange) at a flow-rate of 4 mL/min (the volume
at this stage is likely to be large, possibly more than 100 mL;
see Note 15).
11. Elute the cleaved kinase by running a salt gradient from 0 to
300 mM NaCl, using ion exchange chromatography buffer B,
in 120 min using a flow-rate of 1 mL/min. Unphosphorylated
SrcCDK295M appears at around 15.4 mS/cm and SrcCDK295M
phosphorylated at Tyr416 appears at around 16.9 mS/cm. Wild-
type SrcCD bound with the kinase inhibitor, AP23464, appears
at around 20.4 mS/cm. MBP elutes at around 8–10 mS/cm,
multiple peaks may appear due to nonspecific phosphorylation
(see Note 16).
124 A. Piserchio et al.
4.2. Expression We provide below a step-by-step protocol for the expression and
and Purification purification of wild-type SrcCD without the use of an N-terminal
of the Fully Active, MBP tag. The SrcCD obtained using this protocol is fully active
Wild-Type Catalytic and does not require a kinase inhibitor to allow purification to
Domain of c-Src homogeneity. To implement this protocol, first SrcCD and Cdc37
have to be cloned into a pACYCDuet vector (Chloramphenicol
resistant) using standard procedures. SrcCD should be inserted in
the first cloning site using the EcoRI and SalI restriction enzymes
(resulting in a N-terminal His6-tag for SrcCD, see Note 17), and
Cdc37 in the second site using XhoI and NdeI (see Note 18).
SrcCD expression was also tested with and without the concomi-
tant coexpression of a third protein, the tyrosine phosphatase, Lyp
(61) (a kind gift from Dr. Ronald Seidel, Albert Einstein College
of Medicine). Lyp selectively dephosphorylates the activating Tyr
residue in SFKs (specifically Tyr394 in Lck), and was obtained in a
H-MBP-3C Ampicillin resistant vector (see Note 1).
1. If using electrocompetent cells, transfer ~20 ng of pACYC-
Duet vector into the Eppendorf tube containing BL21(DE3)
T1 electrocompetent cells. Transfer the contents of the
Eppendorf tube into a prechilled 0.1 cm gap electroporation
cuvette and incubate for 2 min on ice (see Note 19). If also
overexpressing the phosphatase, then transfer both vectors
(pACYC and H-MBP-3C, 20 ng each) into the cuvette at the
same time (see Note 20).
2. Follow steps 4–14 of Subheading 4.1.1 but use Chloram-
phenicol (35 mg/L) and Ampicillin only if overexpressing the
phosphatase. This applies to step 4 (Chloramphenicol, or
Ampicillin and Chloramphenicol double resistant plates) and
step 6 (Chloramphenicol, or Ampicillin and Chloramphenicol
added to the M9 medium).
3. Purify the protein following steps 1–6 of Subheading 4.1.2.
Note that the final purified protein retains the His6-tag. If the
tag needs to be removed, a thrombin cleavage site could be
inserted at step 1 in the SrcCD forward primer (described in
Note 10). Then steps 7–9 from Subheading 4.1.2. would also
apply. Note that, in our construct, the MBP-phosphatase
fusion protein has an N-terminal His6 tag and coelutes with
the kinase.
4. After elution, dialyze the protein(s) against dialysis buffer.
5. Concentrate the protein(s) down to 2 mL and inject it onto a
size exclusion column equilibrated with size exclusion
chromatography buffer. Note that the phosphatase elutes in
the void (presumably due to some level of self-association)
while the kinase elutes consistently at the volume expected for
its molecular weight, and are both readily isolated.
7 Expression and Purification of Src-family Kinases for Solution NMR Studies 125
5. Conclusions
6. Notes
12. All bacterial growth postinduction (after the initial half an hour
at 37°C, see Note 11) was carried out at a low temperature
(15°C). It is well known that low temperatures decrease the
degradation rates of unstable proteins; in the case of SrcCD,
we observed that low temperature growth is essential for the
production of soluble, properly folded SrcCD. We speculate
that SrcCD folding is a slow process that is in competition with
the formation of aggregates both soluble and insoluble. In our
protocol both the N-terminal MBP fusion and the GroEL/
GroES chaperones (or the Cdc37 cochaperone) provide assis-
tance in stabilizing the nascent chain and protect it from
aggregation. Lower temperatures likely assist this process by
reducing the rates of protein synthesis and aggregation.
13. We noticed that after long induction times at low temperature,
E. coli cells release significant amounts of acid molecules into
the lysis buffer. This is, in some instances, sufficient to over-
come the buffering power of the solution leading to a pH
below 7.0, compromising efficient binding of the His6 tag to
the metal affinity column.
14. EDTA is introduced to chelate Co2+ cations. Metal leaks from
the resin can be caused by mechanical stress and by high
Imidazole concentrations. If not chelated by EDTA, Co2+ can
be difficult to eliminate by dialysis, since it can bind the His6-tag
as well as the metal binding pockets in SrcCD if left in the sam-
ple. Co2+ forms a reddish precipitate once a substantial amount
of reducing agent has been added (interfering with the remain-
ing purification process), while the Co2+ remaining in solution,
being paramagnetic, degrades the quality of the NMR spectra.
15. We normally do not load more than 15 mg of fusion protein
into the ion exchange column at a given time.
16. Unlike the catalytically compromised mutant, wild-type SrcCD
does not elute once loaded on the Q column. Presumably the
wild-type protein is less stable under conditions of low ionic
strength used for sample loading, and precipitates on the
column. We found however, that this construct could also be
successfully loaded and eluted using the protocol described
above by adding a high-affinity kinase inhibitor (59) after
cleaving the fusion protein.
17. For SrcCD (human), we used the following primers:
Forward: GGTGGTGGAATTCGTCCAAGCCGCAGA
CTCAGGG.
Reverse: GGTGGTGGTCGACCTAGAGGTTCTCCCC
GGGCTGGTA.
128 A. Piserchio et al.
Acknowledgments
References
1. Hunter, T. (2000) Signaling – 2000 and 17. Roux, P. P., and Blenis, J. (2004) ERK and p38
beyond. Cell 100, 113–127. MAPK-activated protein kinases: a family of
2. Johnson, L. N. (2009) The regulation of pro- protein kinases with diverse biological functions.
tein phosphorylation. Biochem. Soc. Trans. 37, Microbiol. Mol. Biol. Rev. 68, 320–344.
627–641. 18. Chen, Z., Gibson, T. B., Robinson, F., Silvestro,
3. Tarrant, M. K., and Cole, P. A. (2009) The L., Pearson, G., Xu, B., Wright, A., Vanderbilt,
chemical biology of protein phosphorylation. C., and Cobb, M. H. (2001) MAP kinases.
Annu. Rev. Biochem. 78, 797–825. Chem. Rev. 101, 2449–2476.
4. Ostman, A., and Bohmer, F. D. (2001) 19. Hubbard, S. R., and Till, J. H. (2000) Protein
Regulation of receptor tyrosine kinase signal- tyrosine kinase structure and function. Annu.
ing by protein tyrosine phosphatases. Trends Rev. Biochem. 69, 373–398.
Cell Biol. 11, 258–266. 20. Shi, Z., Resing, K. A., and Ahn, N. G. (2006)
5. Shi, Y. (2009) Serine/threonine phosphatases: Networks for the allosteric control of protein
mechanism through structure. Cell 139, kinases. Curr. Opin. Struct. Biol. 16, 686–692.
468–484. 21. Fedorov, O., Muller, S., and Knapp, S. (2010)
6. Kolibaba, K. S., and Druker, B. J. (1997) The (un)targeted cancer kinome. Nature Chem.
Protein tyrosine kinases and cancer. Biochim. Biol. 6, 166–169.
Biophys. Acta 1333, F217–248. 22. Grueneberg, D. A., Degot, S., Pearlberg, J., Li,
7. Blume-Jensen, P., and Hunter, T. (2001) W., Davies, J. E., Baldwin, A., Endege, W.,
Oncogenic kinase signalling. Nature 411, Doench, J., Sawyer, J., Hu, Y., Boyce, F., Xian,
355–365. J., Munger, K., and Harlow, E. (2008) Kinase
8. Chong, P. K., Lee, H., Kong, J. W., Loh, M. requirements in human cells: I. Comparing
C., Wong, C. H., and Lim, Y. P. (2008) kinase requirements across various cell types.
Phosphoproteomics, oncogenic signaling and Proc. Natl. Acad. Sci. U.S.A. 105,
cancer research. Proteomics 8, 4370–4382. 16472–16477.
9. Pawson, T., and Kofler, M. (2009) Kinome sig- 23. Baldwin, A., Li, W., Grace, M., Pearlberg, J.,
naling through regulated protein–protein Harlow, E., Munger, K., and Grueneberg, D.
interactions in normal and cancer cells. Curr. A. (2008) Kinase requirements in human cells:
Opin. Cell Biol. 21, 147–153. II. Genetic interaction screens identify kinase
requirements following HPV16 E7 expression
10. Johnson, L. N. (2009) Protein kinase inhibi- in cancer cells. Proc. Natl. Acad. Sci. U.S.A.
tors: contributions from structure to clinical 105, 16478–16483.
compounds. Q. Rev. Biophys. 42, 1–40.
24. Bommi-Reddy, A., Almeciga, I., Sawyer, J.,
11. Noble, M. E., Endicott, J. A., and Johnson, L. Geisen, C., Li, W., Harlow, E., Kaelin, W. G.,
N. (2004) Protein kinase inhibitors: insights Jr., and Grueneberg, D. A. (2008) Kinase
into drug design from structure. Science 303, requirements in human cells: III. Altered kinase
1800–1805. requirements in VHL−/− cancer cells detected
12. Nichols, G. L. (2003) Tyrosine kinase inhibi- in a pilot synthetic lethal screen. Proc. Natl.
tors as cancer therapy. Cancer Invest. 21, Acad. Sci. U.S.A. 105, 16484–16489.
758–771. 25. Grueneberg, D. A., Li, W., Davies, J. E.,
13. Shawver, L. K., Slamon, D., and Ullrich, A. Sawyer, J., Pearlberg, J., and Harlow, E. (2008)
(2002) Smart drugs: tyrosine kinase inhibitors Kinase requirements in human cells: IV.
in cancer therapy. Cancer Cell 1, 117–123. Differential kinase requirements in cervical and
14. Nolen, B., Taylor, S., and Ghosh, G. (2004) renal human tumor cell lines. Proc. Natl. Acad.
Regulation of protein kinases; controlling activ- Sci. U.S.A. 105, 16490–16495.
ity through activation segment conformation. 26. Wiesner, S., Wybenga-Groot, L. E., Warner,
Mol. Cell 15, 661–675. N., Lin, H., Pawson, T., Forman-Kay, J. D.,
15. Kornev, A. P., and Taylor, S. S. (2010) Defining and Sicheri, F. (2006) A change in conforma-
the conserved internal architecture of a protein tional dynamics underlies the activation of Eph
kinase. Biochim. Biophys. Acta 1804, receptor tyrosine kinases. EMBO J. 25,
440–444. 4686–4696.
16. Parsons, S. J., and Parsons, J. T. (2004) Src 27. Vajpai, N., Strauss, A., Fendrich, G., Cowan-
family kinases, key regulators of signal trans- Jacob, S. W., Manley, P. W., Grzesiek, S., and
duction. Oncogene 23, 7906–7909. Jahnke, W. (2008) Solution conformations and
130 A. Piserchio et al.
dynamics of Abl kinase-inhibitor complexes 40. Sicheri, F., Moarefi, I., and Kuriyan, J. (1997)
determined by NMR substantiate the different Crystal structure of the Src family tyrosine
binding modes of imatinib/nilotinib and dasa- kinase Hck. Nature 385, 602–609.
tinib. J. Biol. Chem. 283, 18292–18302. 41. Weijland, A., Neubauer, G., Courtneidge, S.
28. Piserchio, A., Ghose, R., and Cowburn, D. A., Mann, M., Wierenga, R. K., and Superti-
(2009) Optimized bacterial expression and Furga, G. (1996) The purification and charac-
purification of the c-Src catalytic domain for terization of the catalytic domain of Src
solution NMR studies. J. Biomol. NMR 44, expressed in Schizosaccharomyces pombe.
87–93. Comparison of unphosphorylated and tyrosine
29. Piserchio, A., Warthaka, M., Devkota, A. K., phosphorylated species. Eur. J. Biochem. 240,
Kaoud, T. S., Lee, S., Abramczyk, O., Ren, P., 756–764.
Dalby, K. N., and Ghose R. (2011) Solution 42. Feder, D., and Bishop, J. M. (1990) Purification
NMR insights into docking interactions involv- and enzymatic characterization of pp60c-src
ing inactive ERK2. Biochemistry 50, from human platelets. J. Biol. Chem. 265,
3660–3672. 8205–8211.
30. Masterson, L. R., Mascioni, A., Traaseth, N. J., 43. Seeliger, M. A., Young, M., Henderson, M. N.,
Taylor, S. S., and Veglia, G. (2008) Allosteric Pellicena, P., King, D. S., Falick, A. M., and
cooperativity in protein kinase A. Proc. Natl. Kuriyan, J. (2005) High yield bacterial expres-
Acad. Sci. U.S.A. 105, 506–511. sion of active c-Abl and c-Src tyrosine kinases.
31. Masterson, L. R., Cheng, C., Yu, T., Tonelli, Protein Sci. 14, 3135–3139.
M., Kornev, A., Taylor, S. S., and Veglia, G. 44. Stover, D. R., Liebetanz, J., and Lydon, N. B.
(2010) Dynamics connect substrate recogni- (1994) Cdc2-mediated modulation of pp60c-
tion to catalysis in protein kinase A. Nature src activity. J. Biol. Chem. 269,
Chem. Biol. 6, 821–828. 26885–26889.
32. Liu, D., Xu, R., and Cowburn, D. (2009) 45. Pellecchia, M., Bertini, I., Cowburn, D., Dalvit,
Segmental isotopic labeling of proteins for C., Giralt, E., Jahnke, W., James, T. L.,
nuclear magnetic resonance. Meth. Enzymol. Homans, S. W., Kessler, H., Luchinat, C.,
462, 151–175. Meyer, B., Oschkinat, H., Peng, J., Schwalbe,
33. Boggon, T. J., and Eck, M. J. (2004) Structure H., and Siegal, G. (2008) Perspectives on
and regulation of Src family kinases. Oncogene NMR in drug discovery: a technique comes of
23, 7918–7927. age. Nature Rev. Drug Discov. 7, 738–745.
34. Martin, G. S. (2001) The hunting of the Src. 46. Cole, P. A. (1996) Chaperone-assisted protein
Nature Rev. Mol. Cell Biol. 2, 467–475. expression. Structure 4, 239–242.
35. Alvarez, R. H., Kantarjian, H. M., and Cortes, 47. Nomine, Y., Ristriani, T., Laurent, C., Lefevre,
J. E. (2006) The role of Src in solid and hema- J. F., Weiss, E., and Trave, G. (2001) Formation
tologic malignancies: development of new- of soluble inclusion bodies by HPV E6 onco-
generation Src inhibitors. Cancer 107, protein fused to maltose-binding protein.
1918–1929. Protein Expr. Purif. 23, 22–32.
36. Russello, S. V., and Shore, S. K. (2003) Src in 48. Kapust, R. B., and Waugh, D. S. (1999)
human carcinogenesis. Front. Biosci. 8, Escherichia coli maltose-binding protein is
s1068–1073. uncommonly effective at promoting the solu-
37. Trevino, J. G., Summy, J. M., and Gallick, G. E. bility of polypeptides to which it is fused.
(2006) Src inhibitors as potential therapeutic Protein Sci. 8, 1668–1674.
agents for human cancers. Mini Rev. Med. 49. Wang, D., Huang, X. Y., and Cole, P. A. (2001)
Chem. 6, 681–687. Molecular determinants for Csk-catalyzed
38. Grgurevich, S., Mikhael, A., and McVicar, D. W. tyrosine phosphorylation of the Src tail.
(1999) The Csk homologous kinase, Chk, Biochemistry 40, 2004–2010.
binds tyrosine phosphorylated paxillin in 50. Ozkirimli, E., and Post, C. B. (2006) Src kinase
human blastic T cells. Biochem. Biophys. Res. activation: a switched electrostatic network.
Commun. 256, 668–675. Protein Sci. 15, 1051–1062.
39. Zrihan-Licht, S., Deng, B., Yarden, Y., McShan, 51. Cowan-Jacob, S. W. (2006) Structural biology
G., Keydar, I., and Avraham, H. (1998) Csk of protein tyrosine kinases. Cell Mol. Life Sci.
homologous kinase, a novel signaling molecule, 63, 2608–2625.
directly associates with the activated ErbB-2 52. Welch, W. J., and Feramisco, J. R. (1982)
receptor in breast cancer cells and inhibits their Purification of the major mammalian heat
proliferation. J. Biol. Chem. 273, 4065–4072. shock proteins. J. Biol. Chem. 257,
14949–14959.
7 Expression and Purification of Src-family Kinases for Solution NMR Studies 131
53. Hahn, J. S. (2009) The Hsp90 chaperone selective trisubstituted purine-based com-
machinery: from structure to drug develop- pounds. Chem. Biol. Drug Des. 67, 46–57.
ment. BMB Rep. 42, 623–630. 60. Fiaux, J., Bertelsen, E. B., Horwich, A. L., and
54. Zuehlke, A., and Johnson, J. L. (2010) Hsp90 Wuthrich, K. (2004) Uniform and residue-spe-
and co-chaperones twist the functions of diverse cific 15N-labeling of proteins on a highly deu-
client proteins. Biopolymers 93, 211–217. terated background. J. Biomol. NMR 29,
55. Caplan, A. J., Mandal, A. K., and Theodoraki, 289–297.
M. A. (2007) Molecular chaperones and pro- 61. Wu, J., Katrekar, A., Honigberg, L. A., Smith, A.
tein kinase quality control. Trends Cell Biol. 17, M., Conn, M. T., Tang, J., Jeffery, D., Mortara,
87–92. K., Sampang, J., Williams, S. R., Buggy, J., and
56. Karnitz, L. M., and Felts, S. J. (2007) Cdc37 Clark, J. M. (2006) Identification of substrates of
regulation of the kinome: when to hold ‘em and human protein-tyrosine phosphatase PTPN22.
when to fold ‘em. Science STKE 2007, pe22. J. Biol. Chem. 281, 11002–11010.
57. Whitelaw, M. L., Hutchison, K., and Perdew, G. 62. Burz, D. S., Dutta, K., Cowburn, D., and
H. (1991) A 50-kDa cytosolic protein complexed Shekhtman, A. (2006) In-cell NMR for pro-
with the 90-kDa heat shock protein (hsp90) is tein–protein interactions (STINT-NMR).
the same protein complexed with pp60v-Src Nature Protoc. 1, 146–152.
hsp90 in cells transformed by the Rous sarcoma 63. Selenko, P., and Wagner, G. (2006) NMR
virus. J. Biol. Chem. 266, 16436–16440. mapping of protein interactions in living cells.
58. Alexandrov, A., Dutta, K., and Pascal, S. M. Nature Meth. 3, 80–81.
(2001) MBP fusion protein with a viral pro- 64. Sakakibara, D., Sasaki, A., Ikeya, T., Hamatsu,
tease cleavage site: one-step cleavage/purifica- J., Hanashima, T., Mishima, M., Yoshimasu,
tion of insoluble proteins. Biotechniques 30, M., Hayashi, N., Mikawa, T., Walchli, M.,
1194–1198. Smith, B. O., Shirakawa, M., Guntert, P., and
59. Dalgarno, D., Stehle, T., Narula, S., Schelling, Ito, Y. (2009) Protein structure determination
P., van Schravendijk, M. R., Adams, S., Andrade, in living cells by in-cell NMR spectroscopy.
L., Keats, J., Ram, M., Jin, L., Grossman, T., Nature 458, 102–105.
MacNeil, I., Metcalf, C., 3rd, Shakespeare, W., 65. Campos-Olivas, R., Marenchino, M., Scapozza,
Wang, Y., Keenan, T., Sundaramoorthi, R., L. and Gervasio, F. L. (2011) Backbone assign-
Bohacek, R., Weigele, M., and Sawyer, T. ment of the tyrosine kinase Src catalytic domain
(2006) Structural basis of Src tyrosine kinase in complex with imatinib. Biomol. NMR Asgn.
inhibition with a new class of potent and 5, 221–224.
Chapter 8
Abstract
Over the recent years, there has been increased interest in applying NMR spectroscopy for the characterization
of proteins and protein complexes of large molecular weight. The combination of multidimensional NMR,
novel pulse sequences allowing for the selection of slowly relaxing coherence pathways, and the development
of a range of labeling techniques has enabled high-resolution NMR analyses of supramolecular systems of
even megadalton size. Here, we describe how NMR can be used to obtain structural information in large
systems by using as an example the recent structure determination of SecA ATPase (204 kDa) in complex
with a signal peptide.
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_8, © Springer Science+Business Media, LLC 2012
133
134 S.-R. Tzeng et al.
2. Materials
3. Methods
3.1. Protein Labeling 1. Pick a freshly transformed colony (see Note 1) of BL21(DE3)
for Methyl Detection cells and inoculate a 1–2 mL culture of L-broth in D2O containing
0.1% of glucose at 37°C until cells reach an OD600 of ~0.7–0.8.
2. Centrifuge the cells at 1,200 × g for 20 min at room tempera-
ture and resuspend them in 5–10 mL of sterile M9 medium in
D2O (M9/D2O) in a sterile flask to a starting OD600 of ~0.05.
Incubate the culture in a shaking incubator (220–250 rpm) at
37°C until it reaches an OD600 of ~0.6.
136 S.-R. Tzeng et al.
a b
12 Ile 20
14 Met 21
16 Ala
22
(p.p.m.)
(p.p.m.)
18
Leu, Val 23
13C
13C
20
24
22
25
24
26
26
27
2.0 1.6 1.2 0.8 0.4 0.0 1.2 1.0 0.8 0.6 0.4 0.2 0.0
1H (p.p.m.)
Fig. 1. (a) 1H-13C HMQC of [U-2H,12C], Ala-, Leu-, Met-, Val-, Ile-δ1-[13CH3] labeled protein. The methyl groups of all five
amino acids can be labeled with no scrambling. (b) 1H-13C HMQC of the same protein as in (a, black) but prepared using
10% BIOEXPRESS (green). The Leu methyl groups are completely suppressed, whereas the Val methyl groups are only
minimally affected.
8 NMR Studies of Large Protein Systems 137
3.2. NMR Assignment To assign a large protein, such as SecA (204 kDa; 901 residues per
subunit), a domain-parsing strategy is followed.
1. Isolate and characterize, by using NMR, virtually all domains
of the full-length protein and a number of fragments compris-
ing contiguous domains (see Note 10). The size of the isolated
domains and fragments should be such that backbone and
side-chain assignment is feasible using standard approaches.
2. Prepare Ala, Ile, Met, Leu, and Val methyl-labeled samples for
the full-length protein and the domains thereof (see Note 11).
Record methyl-TROSY for all of the samples, overlay, and
compare the spectra of the individual domains against the
spectra of the longer fragments and full-length protein. If good
resonance correspondence among domains, fragments, and
the full-length protein can be demonstrated, then assignment
is in principle transferable.
3. Record standard triple-resonance NMR experiments for isolated
domains and obtain backbone assignments. Record the three-
dimensional spectra required for the assignment of the methyl
groups (see Note 12).
4. Transfer the methyl assignments obtained for the isolated
domains to the larger fragments and finally to the full-length
protein by visually inspecting the methyl-TROSY spectra. Only
the assignment of the obvious and well-dispersed resonances
can be safely transferred this way.
5. Record 13C HMQC-NOESY-HMQC spectra (see Note 13) for
the methyl-labeled samples. Use the NOE patterns to confirm
and extend the assignment transfer from the domains to the
full-length protein. If a crystal structure is available, it can be
used to determine the distances between the methyl groups
and assist with the assignment.
6. Prepare site-directed mutations to assign ambiguous resonances
and further extend and confirm the assignments (see Note 14).
3.4. Structure 1. Determine the interface between the ligand and the large protein
Determination using differential line broadening (9). The residues affected by
complex formation can be used as ambiguous restraints.
2. If the ligand is a flexible peptide, use transferred NOESY (10)
to determine the structure of the peptide in the complex.
Determine the structure of the complex by using a CNS-based
software, such as HADDOCK (11) or Xplor-NIH (12). Use
the crystal structure of the large protein to define the starting
conformation, and both unambiguous and ambiguous
restraints obtained from NOE, PRE, line broadening, and
chemical shift perturbation experiments.
4. Notes
References
1. Ruschak, A. M., and Kay, L. E. (2010) Methyl limited nuclear overhauser effect data.
groups as probes of supra-molecular structure, Biochemistry 39, 5355–5365.
dynamics and function. J. Biomol. NMR 46, 8. Tang, C., Schwieters, C., and Clore, G. (2007)
75–87. Open-to-closed transition in apo maltose-binding
2. Goto, N., Gardner, K., Mueller, G., Willis, R., protein observed by paramagnetic NMR. Nature
and Kay, L. (1999) A robust and cost-effective 449, 1078–1082.
method for the production of Val, Leu, Ile (delta 9. Takeuchi, K., and Wagner, G. (2006) NMR
1) methyl-protonated 15N-, 13C-, 2H-labeled studies of protein interactions. Curr. Opin.
proteins. J. Biomol. NMR 13, 369–374. Struct. Biol. 16, 109–117.
3. Tugarinov, V., Kanelis, V., and Kay, L. E. (2006) 10. Post, C. (2003) Exchange-transferred NOE
Isotope labeling strategies for the study of high- spectroscopy and bound ligand structure
molecular-weight proteins by solution NMR determination. Curr. Opin. Struct. Biol. 13,
spectroscopy. Nat. Protoc. 1, 749–754. 581–588.
4. Tugarinov, V., Hwang, P., Ollerenshaw, J., and 11. de Vries, S. J., van Dijk, M., and Bonvin, A. M.
Kay, L. (2003) Cross-correlated relaxation (2010) The HADDOCK web server for
enhanced 1H-13C NMR spectroscopy of methyl data-driven biomolecular docking. Nat. Protoc.
groups in very high molecular weight proteins 5, 883–897.
and protein complexes. J. Am. Chem. Soc. 125, 12. Schwieters, C. D., Kuszewski, J. J., Tjandra, N.,
10420–10428. and Clore, G. M. (2003) The Xplor-NIH
5. Sprangers, R., and Kay, L. E. (2007) Quantitative NMR molecular structure determination package.
dynamics and binding studies of the 20S protea- J. Magn. Reson. 160, 65–73.
some by NMR. Nature 445, 618–622. 13. Isaacson, R., Simpson, P., Liu, M., Cota, E.,
6. Gelis, I., Bonvin, A., Keramisanou, D., Koukaki, Zhang, X., Freemont, P., and Matthews, S.
M., Gouridis, G., Karamanou, S., Economou, (2007) A new labeling method for methyl
A., and Kalodimos, C. G. (2007) Structural transverse relaxation-optimized spectroscopy
basis for signal-sequence recognition by the NMR spectra of alanine residues. J. Am. Chem.
translocase motor SecA as determined by NMR. Soc. 129, 15428–15429.
Cell 131, 756–769. 14. Tugarinov, V., and Kay, L. (2003) Ile, Leu, and
7. Battiste, J., and Wagner, G. (2000) Utilization Val methyl assignments of the 723-residue
of site-directed spin labeling and high-resolution malate synthase G using a new labeling strategy
heteronuclear nuclear magnetic resonance for and novel NMR methods. J. Am. Chem. Soc.
global fold determination of large proteins with 125, 13868–13878.
Chapter 9
Abstract
Nitrogen-15 relaxation is the most ubiquitous source of information about protein (backbone) dynamics
used by NMR spectroscopists. It provides the general characteristics of hydrodynamics as well as internal
motions on subnanosecond, micro- and millisecond timescales of a biomolecule. Here, we present a full
protocol to perform and analyze a series of experiments to measure the 15N longitudinal relaxation rate,
the 15N transverse relaxation rate under an echo train or a single echo, the 15N–1H dipolar cross-relaxation
rate, as well as the longitudinal and transverse cross-relaxation rates due to the cross-correlation of the
nitrogen-15 chemical shift anisotropy and the dipolar coupling with the adjacent proton. These rates can
be employed to carry out model-free analyses and can be used to quantify accurately the contribution of
chemical exchange to transverse relaxation.
Key words: Nuclear magnetic resonance, Protein dynamics, Relaxation rates, Nitrogen-15, Longitudinal
relaxation, Transverse relaxation, Cross-relaxation, Cross-correlated relaxation, Chemical exchange
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_9, © Springer Science+Business Media, LLC 2012
141
142 F. Ferrage
1.1. Theory We provide here the minimal theoretical background that is necessary
to define the various terms that are used in this review. For a
detailed understanding of relaxation theory, the reader should refer
to some of the many reviews that have been published on the
subject (16–19). Relaxation is the irreversible process through which
a spin system evolves toward a steady state. Different elements of the
density operator evolve with different auto-relaxation rates, and
they convert into one another with well-defined cross-relaxation
rates. The molecular processes underlying relaxation are the fluc-
tuations of high-amplitude orientation-dependent interactions,
such as dipole–dipole couplings and the anisotropy of the chemical
shift. The amplitude of these interactions is known so that one
can extract the characteristics of motions from relaxation rates.
Relaxation rates do not depend directly on the correlation function
of interaction Hamiltonians but on their Fourier transform: the
spectral density function J(w). Here, we describe a protocol to measure
the following relaxation rates for 15N nuclei in a 15N–1H pair.
The longitudinal auto-relaxation rate, R1:
Rex = p A pB Dw 2 t ex , (7)
2. Materials
3. Methods
3.2. Auto-Relaxation The sequences presented in Fig. 1a, b should be employed for this
Rates Measurements experiment. There is no fully satisfactory combination of water-flip
back schemes and suppression of CSA/DD cross-correlated cross-
3.2.1. Longitudinal
relaxation pathways during the relaxation delay. One solution is to
Relaxation Rate
saturate the water resonance in each scan and use a long recovery
delay between scans. To saturate the water resonance, drop the
water-flip back pulse at the end of the first INEPT in Fig. 1a and
use a strong gradient G2. In that case, the two shaped pulses in
Fig. 1b can be substituted either by composite pulse decoupling
(see Fig. 1c) or a series of proton 180° pulses (see Fig. 1d). Here,
we use a scheme that can be used when no amide resonance is
lying too close to the water resonance. It is one of the standard
sequences available on Bruker spectrometers.
1. Set pulse durations and amplitudes according to the spectrom-
eter specifications, the calibration and the description of the
pulse sequence in Fig. 1a, b. Set delays according to the pulse
sequence. Calibrate proton shape pulses according to the
spectrometer-specific protocol.
2. Set the recovery delay between scans to a large value, usually
2 s; 3 s can be used if spectrometer time is not a problem.
3. For a well-behaved protein, start with eight scans (16–32 for
low-concentration samples). Set the relaxation delay to a low
value (we use 20 ms). Run the first 1D experiment. Phase and
save this spectrum.
4. Set the relaxation delay to a value close to the expected average
T1. At 300 K, this is 500 ms for a globular 100-residue protein
and 700 ms for a 150-residue protein. Acquire a new spectrum.
9 Protein Dynamics by 15N Nuclear Magnetic Relaxation 147
Fig. 1. Pulse sequences used for measuring 15N auto-relaxation rates. (a) General scheme. (b) Relaxation sequence for
measuring the longitudinal relaxation rate R1; (c) relaxation sequence for measuring the transverse relaxation rate under a
single echo R2echo; (d) relaxation sequence for the measurement of the transverse relaxation rate under a CPMG echo train
R2CPMG. All narrow (filled) and (wide ) open rectangles represent 90° and 180° pulses, respectively. Pulse phases are along
the x-axis of the rotating frame unless otherwise mentioned. Proton composite pulse decoupling during the delay t1 was
performed with a GARP scheme and a radio frequency field (rf) amplitude of 1 kHz. The 13C channel pulse was a 500 ms
smoothed CHIRP pulse (37), with a sweep of 60 kHz on a 600 MHz spectrometer; the carrier at the center of the pulse was
110 ppm. Composite-pulse decoupling during acquisition was performed on the 15N channel with a GARP scheme (38) and
an rf amplitude of 1,090 Hz. The delay ta is 2.56 ms; the delay tb can be adjusted around 5 ms. The phase cycles were:
f1 = {x, −x }; f3 = {x, x, −x, −x }; f4 = {x, x, −x, −x }; f5 = {−y, −y, y, y }; facq = {x, −x, −x, x }. When relaxation block (b) is used,
f2 = { y, y, y, y, −y, −y, −y, −y } and facq = {x, −x, −x, x, −x, x, x, −x }. The amplitude profile of the pulsed field gradient was
a sine bell shape. Their durations and peak amplitudes over the x, y, and z orientations (when triple axis gradients are
available) were, respectively: G1; 600 ms, 9.5 G/cm, 9.5 G/cm, 0; G2; 1 ms, 0, 0, 30 G/cm; G3; 600 ms, 15 G/cm, –15 G/cm,
0; G4; 1 ms, 0, 0, 40 G/cm; G5; 1 ms, 0, 0, 8.1 G/cm. Coherence selection was achieved by inverting the amplitude of the
gradient G4 and phase f1. (b) The carrier is placed at 8.2 ppm during the relaxation block; gray bell-shaped pulses are
1.6 ms Q3 Gaussian cascade pulses at 600 MHz (39) (see Note 9). (c) WALTZ-16 decoupling should be used for 1H decoupling
during the relaxation block (40). See text for how to choose the relaxation delays. (d) Gray rectangles are 180° pulses,
depending on the spectrometer and probe, they should be either at high power or less, but should not be longer than
100 ms. t should be set to 500 ms.
148 F. Ferrage
( )
exp -R1av t max = 0.3 Û R1av = - ln 0.3 / t max (10)
( )
and exp -R1avt j = I j . (11)
So that
3.2.2. Transverse The sequence presented in Fig. 1b should be employed for this
Relaxation Rates experiment. Only relaxation delays that lead to full cycles of the pro-
with a Single Echo ton composite pulse decoupling (CPD) should be used, which
modifies slightly the protocol from the measurement of longitudi-
nal relaxation rates.
150 F. Ferrage
Fig. 2. Example of a longitudinal relaxation decay curve as shown by Grace during the
Curvefit procedure.
3.2.3. Transverse For the sake of consistency, this protocol is described here. However,
Relaxation Rates with this is the most challenging experiment as heating from high
a CPMG Scheme radiofrequency fields may lead to bubble formation and sample
degradation. It is strongly advised to run this experiment as the
last one of the series.
1. Most parameters are identical to those used in longitudinal
relaxation measurements. Set the power level for 15N 180° pulses
during the CPMG echo train. This experiment is one of the
most demanding that can be run on a high-resolution probe.
The maximum power (i.e., shortest nitrogen-15 pulses) that can
be employed during a CPMG echo train is usually provided to
users for each probe on each spectrometer. If no value is
recommended, ask the person in charge of the spectrometer
maintenance. If you are in charge of the spectrometer mainte-
nance, ask the manufacturer. Often, particularly on cryogenic
probes, these pulses should not be applied at full power.
For proper accuracy of the experiment, 180° pulses should be no
longer than 100 ms.
2. In addition to Fig. 1d, include a temperature-compensation
loop at the beginning of the sequence. At the end of the
recycling delay, include a series of far off-resonance (200 kHz
works fine) 15N 180° pulses at the amplitude of the CPMG
echo train. Do not include 1H 180° pulses as these would alter the
1
H longitudinal polarization. The duration of this train should
be such that the total number of 15N 180° pulses is constant
whatever the relaxation delay is. An additional delay, equal to
the relaxation delay, should also be added so that the total
duration for the recovery of proton polarization is constant.
3. Follow steps 2–7 of the protocol for longitudinal relaxation
(Subheading 3.2.1). Zero can be used as the shortest relax-
ation delay.
4. Use a long recovery delay (preferably 2 s) and a large number
of dummy scans (512) to let the temperature control system
reach equilibrium before the first experiment.
5. Follow steps 9–16 of the R1 analysis (Subheading 3.2.1) to obtain
transverse relaxation rates under a CPMG train, R2CPMG.
3.3. 15N–{ 1H} Nuclear The sequence presented in Fig. 3 should be used for this experiment.
Overhauser Effect It displays a series of improvements introduced recently (25–27).
Measurements The two experiments (under effective proton saturation and at
equilibrium) have to be run in an interleaved manner.
1. Set up the saturation scheme (Fig. 3b). The saturation element
is symmetric, 1H pulses have a 180° flip angle, the carrier is placed
in the amide region (8.2 ppm), the interpulse delay tNOE should
be a multiple of 1/JNH, where JNH is the 1H–15N one-bond scalar
coupling constant. In globular proteins at low pH, tNOE = 22 ms
152 F. Ferrage
Fig. 3. Pulse sequence used for recording steady-state 15N–{1H} nuclear Overhauser effects. For each measurement, reference
and steady-state experiments have to be recorded in an interleaved manner. In the reference experiment, one should run
the part of the pulse program displayed in box (a). At the end of the recovery delay TNOE = 10 s (or more), the proton carrier
is placed on resonance with the water signal and a very selective water-flip back pulse is applied (3 ms sinc shaped or
longer). To record steady-state experiments, the boxed sequence in (a) is substituted by the scheme shown in (b) for the
effective saturation of amide proton resonances. After an optional delay T¢NOE = 2 s for stable detection of the lock signal,
the proton carrier is placed in the center of the amide region (at 8.2 ppm) as shown by the arrow labeled by N. The motif
[delay tNOE/2 – 180° pulse – delay tNOE/2] is repeated nNOE times. The interpulse delay, tNOE, is typically 22 ms (11 ms may
also be used, see text). The rf amplitude for the pulses should be 7.5 kHz at a 500 MHz Larmor frequency and 9 kHz at
600 MHz. A gradient G1 is applied at the end of the last tNOE/2 delay to suppress all transverse components of the proton
polarization. The carrier was moved on-resonance with the water signal as indicated by the W arrow. The number of cycles,
nNOE, was set so that the total duration for effective saturation was 4 s. All narrow (filled ) and wide (open) rectangles represent 90°
and 180° pulses, respectively. Pulse phases are along the x-axis of the rotating frame unless otherwise mentioned. Proton
composite pulse decoupling during the delay t1 was performed with a GARP scheme and an rf amplitude of 1 kHz.
Composite-pulse decoupling during acquisition was performed on the 15N channel with a GARP scheme (38) and an rf
amplitude of 1,090 Hz. The delay ta is 2.56 ms. The phase cycles were: f1 = {y, –y}; f2 = {x, x, –x, –x}; f3 = {x, x, –x, –x };
f4 = {−y, –y, y, y }; facq = {x, –x, –x, x }. The amplitude profile of the pulsed field gradient was a sine bell shape. The durations
and peak amplitudes over the x, y, and z orientations were, respectively: G1; 600 ms, 15 G/cm, 15 G/cm, 0; G2; 1 ms, 0, 0,
25 G/cm; G3; 1 ms, 0, 0, 40 G/cm; G4; 1 ms, 0, 0, 8.1 G/cm. Coherence selection was achieved by inverting the amplitude
of the gradient G3 and phase f4.
3.4. Measurements These experiments are less commonly run in spite of their great
of CSA/DD Cross- utility. The set up turns out to be straightforward after more typi-
Correlated Cross- cal relaxation experiments have been run.
Relaxation Rates 1. Set pulse durations and amplitudes according to the spectrom-
eter specifications, the calibration and the description of the pulse
sequence, which is shown in Fig. 4. Set delays according to
the pulse sequence. Calibrate proton shape pulses according
to the spectrometer-specific protocol. Most of this part has
already been set up in the preceding experiments.
2. For both experiments, the minimum number of scans imposed
by the phase cycle is higher, leading to very long experimental
times when high resolution data needs to be collected (large
number of t1 points). On the other hand, shorter recycle
delays can be used, typically between 1 and 1.5 s with little loss
of sensitivity and no loss of accuracy.
3. These experiments use symmetrical reconversion (15, 28). This
means that all four relaxation pathways in a two-operator space
are detected. These four experiments have to be run in an
154 F. Ferrage
Fig. 4. Pulses sequences for measuring CSA/DD cross-correlated cross-relaxation rates. (a) General scheme; (b) relaxation
block for measuring the transverse cross-relaxation rate; (c) relaxation block for measuring the longitudinal cross-relaxation
rate. Only specific elements are detailed here. The delay d is equal to the shortest value of t1 so that the first effective value
of t1 is zero. The phase cycles were: f1 = { y, –y }; f2 = {x, x, x, x, –x, –x, –x, –x}; f3 = {x, x, x, x, x, x, x, x, –x, –x, –x, –x, –x,
–x, –x, –x}; f4 = {x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, –x, –x, –x, –x, –x, –x, –x, –x, –x, –x, –x, –x, –x, –x, –x, –x}; f5 = {x};
f6 = {x, x, –x, –x}; f¢6 = {x, x, y, y}; facq = {x, –x, –x, x, –x, x, x, –x, –x, x, x, –x, x, –x, –x, x, –x, x, x, –x, x, –x, –x, x, x, –x, –x,
x, –x, x, x, –x}. Gradient durations and peak amplitudes over the x, y, and z orientations were, respectively: G1; 600 ms, 15 G/
cm, 15 G/cm, 0; G2; 600 ms, 6.5 G/cm, 0, 0; G3; 2 ms, 0, 0, 40 G/cm; G4; 1 ms, –9.5 G/cm, 9.5 G/cm, 0 G/cm; G5; 600 ms,
3.5 G/cm, 0 G/cm, 14.5 G/cm; G6; 1 ms, –35 G/cm, –35 G/cm, –35 G/cm; G7; 1 ms, 35 G/cm, 35 G/cm, 35 G/cm. Frequency
sign discrimination was performed using States-TPPI.
9 Protein Dynamics by 15N Nuclear Magnetic Relaxation 155
3.4.1. Transverse 1. Set up the relaxation delay(s). The limiting factor for sensitivity
Cross-Correlated is the intensity in the cross-relaxation experiments (pathways II
Cross-Relaxation Rates and III in Fig. 4a). This intensity is maximum when the relax-
ation delay is equal to the auto-relaxation time. The optimal
relaxation delay typically lies between 30 and 80 ms. Explore
this interval in 20 ms steps. The intensity versus time curve is
usually flat around the maximum. When two relaxation delays
appear to give the same maximum intensity, first check a relax-
ation delay at the midpoint between the two. If the intensity is
better, keep this delay. If the intensity is the same, choose the
shortest delay so the intensity in the auto-relaxation experi-
ments will be better.
2. This experiment can be run with a single relaxation delay. If time
permits, repeat the optimal relaxation delay and run another delay
with sufficient intensity on the cross-relaxation experiments.
3. Follow steps 9–14 of the longitudinal relaxation protocol
(Subheading 3.2.1) to obtain the intensities of the four interleaved
experiments. The following quantity should be computed:
(
S (4T ) = tanh dNt 4T . ) (14)
4. If only one relaxation delay is recorded, simply invert Eq. 14.
If several time points are recorded, compile the values of S and
the corresponding errors in a text file and follow steps 15 and
16 of the longitudinal relaxation protocol (Subheading 3.2.1).
The header of each N.in file should set the hyperbolic tangent
mode for the fit to evaluate dtN .
3.4.2. Longitudinal 1. Set up the relaxation delay(s). The constraints are similar to
Cross-Correlated those for the transverse cross-relaxation rate (Subheading 3.4.1).
Cross-Relaxation Rates A difference is that the cross-relaxation experiments are now
number I and IV because of the conversion between longitudinal
polarization and two-spin order in the middle of the cross-
relaxation delay. The optimal delay should be found between 80
and 200 ms. However, particularly in nondeuterated proteins,
there is a strong advantage to using shorter delays. If not,
one has to calculate a correction for the rates that takes into
account the effects of proton–proton cross-relaxation (15). In that
case, at least two well-separated time points should be recorded.
2. This experiment can be run with a single relaxation delay if it is
short and/or the protein is deuterated. To make sure that the
correction is small, it is advisable to run two experiments with
different delays.
156 F. Ferrage
From these2 ratios, fit the value of D. In most cases, evaluate the
factor 32d1N DT 3 / 3 , it should be much smaller than 1 so that
D can be obtained from a simple exponential fit.
5. Compute for each time point:
3.5. Analysis and The analysis of this dataset closely follows the one presented by
Interpretation of 15N Kroenke et al. (14). The use of more accurate measurements of
Relaxation Rates nuclear Overhauser effects and longitudinal CSA/DD cross-correlated
3.5.1. Identification 1. Compute the transverse relaxation rate expected in the absence
of Chemical Exchange of exchange R20, for each backbone 15N nucleus:
R20 = (R1 - 1.25s NH )dtN / dlN - 1.08s NH . (17)
3.5.2. A Note on the The ensemble of rates can be used as the input for a model-free
Model-Free Analysis of 15N analysis (29). A series of software are available: modelfree (30); fast
Relaxation Rates modelfree (31); tensor2 (32); and dynamics (33) are good examples.
I do not describe here the use of such software. The only point that
discussed is the choice of the ensemble of relaxation rates that
should be employed. The above-mentioned softwares are designed
to use the longitudinal relaxation rate R1, the transverse relaxation
158 F. Ferrage
rate R2CPMG, and the NOE ratio as inputs. In most cases, these rates
should be used. One of the most difficult tasks of the analysis is the
detection of an exchange contribution to transverse relaxation
rates R2CPMG. When a nonzero contribution RexCPMG is identified,
the rates are at best fit with a simpler model for fast motions. When
a protein shows significant contributions of chemical exchange to
the relaxation of many 15N nuclei, this can be detrimental to the
overall quality of the analysis of hydrodynamic properties and fast
local motions. The analysis described herein provides exchange-
free transverse relaxation rates R20. If, and only if, the ensemble of
RexCPMG rates shows a very flat baseline around zero (see for instance
ref. 15), the R20 rates can be employed instead of the R2CPMG rates
in the model-free analysis.
4. Notes
Acknowledgments
References
1. Mittermaier, A., and Kay, L. E. (2006) Review – An Analytical Approach and Its Application to
New tools provide new insights in NMR studies of Cooperative Ca2+ Binding by Calbindin Dgk. J.
protein dynamics. Science 312, 224–228. Am. Chem. Soc. 115, 9832–9833.
2. Palmer, A. G. (2004) NMR characterization 6. Frederick, K. K., Marlow, M. S., Valentine, K. G.,
of the dynamics of biomacromolecules. Chem. and Wand, A. J. (2007) Conformational entropy
Rev. 104, 3623–3640. in molecular recognition by proteins. Nature
3. Massi, F., Wang, C. Y., and Palmer, A. G. (2006) 448, 325–329.
Solution NMR and computer simulation stud- 7. Kuszewski, J., Gronenborn, A. M., and Clore, G.
ies of active site loop motion in triosephos- M. (1999) Improving the packing and accuracy
phate isomerase. Biochemistry 45, of NMR structures with a pseudopotential for
10787–10794. the radius of gyration. J. Am. Chem. Soc. 121,
4. Tjandra, N., Feller, S. E., Pastor, R. W., and Bax, 2337–2338.
A. (1995) Rotational Diffusion Anisotropy of 8. Ryabov, Y., and Fushman, D. (2007) Structural
Human Ubiquitin from N-15 NMR Relaxation. Assembly of Multidomain Proteins and Protein
J. Am. Chem. Soc. 117, 12562–12566. Complexes Guided by the Overall Rotational
5. Akke, M., Brüschweiler, R., and Palmer III, A. G. Diffusion Tensor. J. Am. Chem. Soc. 129,
(1993) NMR Order Parameters and Free Energy: 7894–7902.
162 F. Ferrage
9. Ryabov, Y. E., and Fushman, D. (2007) A model chemical exchange in N-15-labeled proteins.
of interdomain mobility in a multidomain pro- Magn. Reson. Chem. 41, 866–876.
tein. J. Am. Chem. Soc. 129, 3315–3327. 22. Kempf, J. G., and Loria, J. P. (2004) Measurement
10. Palmer, A. G., and Massi, F. (2006) Characte- of Intermediate Exchange Phenomena. Meth.
rization of the dynamics of biomacromolecules Mol. Biol. 278, 185–231.
using rotating-frame spin relaxation NMR spec- 23. Findeisen, M., Brand, T., and Berger, S. (2007)
troscopy. Chem. Rev. 106, 1700–1719. A H-1-NMR thermometer suitable for cryo-
11. Palmer, A. G. (2004) NMR characterization of probes. Magn. Reson. Chem. 45, 175–178.
the dynamics of biomacromolecules. Chem. Rev. 24. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu,
104, 3623–3640. G., Pfeifer, J., and Bax, A. (1995) NMRPipe: a
12. Vallurupalli, P., Hansen, D. F., and Kay, L. E. Multidimensional Spectral Processing System
(2008) Structures of invisible, excited protein Based on UNIX Pipes. J. Biomol. NMR 6,
states by relaxation dispersion NMR spec- 277–293.
troscopy. Proc. Natl. Acad. Sci. U.S.A. 105, 25. Ferrage, F., Piserchio, A., Cowburn, D., and
11766–11771. Ghose, R. (2008) On the measurement of
13. Kay, L. E., Torchia, D. A., and Bax, A. (1989) N-15-{H-1} nuclear Overhauser effects. J. Magn.
Backbone Dynamics of Proteins as Studied by Reson. 192, 302–313.
N-15 Inverse Detected Heteronuclear NMR- 26. Ferrage, F., Cowburn, D., and Ghose, R. (2009)
Spectroscopy – Application to Staphylococcal Accurate Sampling of High-Frequency Motions
Nuclease. Biochemistry 28, 8972–8979. in Proteins by Steady-State 15N-{1H} Nuclear
14. Kroenke, C. D., Loria, J. P., Lee, L. K., Rance, Overhauser Effect Measurements in the Presence
M., and Palmer III, A. G. (1998) Longitudinal of Cross-Correlated Relaxation. J. Am. Chem.
and Transverse H-1-N-15 Dipolar N-15 Chemical Soc. 131, 6048–6049.
Shift Anisotropy Relaxation Interference: 27. Ferrage, F., Reichel, A., Battacharya, S., Cowburn,
Unambiguous Determination of Rotational D., and Ghose, R. (2010) On the measurement
Diffusion Tensors and Chemical Exchange of N-15-{H-1} nuclear Overhauser effects. 2.
Effects in Biological Macromolecules. J. Am. Effects of the saturation scheme and water signal
Chem. Soc. 120, 7905–7915. suppression. J. Magn. Reson. 207, 294–303.
15. Pelupessy, P., Ferrage, F., and Bodenhausen, G. 28. Pelupessy, P., Espallargas, G. M., and
(2007) Accurate Measurement of Longitudinal Bodenhausen, G. (2003) Symmetrical recon-
Cross-Relaxation Rates in Nuclear Magnetic version: measuring cross-correlation rates
Resonance. J. Chem. Phys. 126, 134508. with enhanced accuracy. J. Magn. Reson. 161,
16. Korzhnev, D. M., Billeter, M., Arseniev, A. S., 258–264.
and Orekhov, V. Y. (2001) NMR Studies of 29. Lipari, G., and Szabo, A. (1982) Model-Free
Brownian Tumbling and Internal Motions in Approach to the Interpretation of Nuclear
Proteins. Prog. Nucl. Magn. Reson. Spectrosc. 38, Magnetic Resonance Relaxation In Macromo-
197–266. lecules 1. Theory and Range of Validity. J. Am.
17. Luginbuhl, P., and Wuthrich, K. (2002) Semi- Chem. Soc. 104, 4546–4559.
classical nuclear spin relaxation theory revisited 30. Mandel, A. M., Akke, M., and Palmer III, A. G.
for use with biological macromolecules. Prog. (1995) Backbone Dynamics of Escherichia coli
Nucl. Magn. Reson. Spectrosc. 40, 199–247. Ribonuclease HI : Correlations with Structure
18. Nicholas, M. P., Eryilmaz, E., Ferrage, F., and Function in an Active Enzyme. J. Mol. Biol.
Cowburn, D., and Ghose, R. (2010) Nuclear 246, 144–163.
spin relaxation in isotropic and anisotropic media, 31. Cole, R., and Loria, J. P. (2003) FAST-Modelfree:
Prog. Nucl. Magn. Reson. Spectrosc. 57, 111–158. A program for rapid automated analysis of solu-
19. Cavanagh, J., Fairbrother, W. J., Palmer III, A. tion NMR spin-relaxation data. J. Biomol. NMR
G., Rance, M., and Skelton, N. J. (2006) Protein 26, 203–13.
NMR Spectroscopy: Principles and practice, 32. Dosset, P., Hus, J. C., Blackledge, M., and
Academic Press, San Diego. Marion, D. (2000) Efficient analysis of macro-
20. Wang, L. C., Pang, Y. X., Holder, T., Brender, J. molecular rotational diffusion from heteronu-
R., Kurochkin, A. V., and Zuiderweg, E. R. P. clear relaxation data. J. Biomol. NMR 16,
(2001) Functional dynamics in the active site of 23–28.
the ribonuclease binase. Proc. Natl. Acad. Sci. 33. Fushman, D., Cahill, S., and Cowburn, D.
U.S.A. 98, 7684–7689. (1997) The main chain dynamics of the
21. Wang, C. Y., and Palmer, A. G. (2003) Solution dynamin pleckstrin homology (PH) domain
NMR methods for quantitative identification of in solution: analysis of 15N relaxation with
9 Protein Dynamics by 15N Nuclear Magnetic Relaxation 163
monomer/dimer equilibration. J. Mol. Biol. 37. Bohlen, J. M., and Bodenhausen, G. (1993)
266, 173–194. Experimental Aspects of Chirp NMR-
34. Butterwick, J. A., Loria, J. P., Astrof, N. S., Spectroscopy. J. Magn. Reson. A 102, 293–301.
Kroenke, C. D., Cole, R., Rance, M., and Palmer, 38. Shaka, A. J., Barker, P. B., and Freeman, R.
A. G. (2004) Multiple time scale backbone (1985) Computer-Optimized Decoupling
dynamics of homologous thermophilic and Scheme for Wideband Applications and Low-
mesophilic ribonuclease HI enzymes. J. Mol. Biol. Level Operation. J. Magn. Reson. 64,
339, 855–871. 547–552.
35. Sarkar, R., Moskau, D., Ferrage, F., Vasos, P. R., 39. Emsley, L., and Bodenhausen, G. (1990)
and Bodenhausen, G. (2008) Single or triple Gaussian Pulse Cascades-New Analytical
gradients? J. Magn. Reson. 193, 110–118. Functions for Rectangular Selective Inversion
36. Muhandiram, D. R., Yamazaki, T., Sykes, B. and In-phase Excitation in NMR. Chem. Phys.
D., and Kay, L. E. (1995) Measurement of 2H Lett. 165, 469–476.
T1ro Relaxation Times in Uniformly 40. Shaka, A. J., Keeler, J., Frenkiel, T., and Freeman,
13
C-Labeled and Fractionally 2H-Labeled R. (1983) An Improved Sequence for Broad
Proteins in Solution. J. Am. Chem. Soc. 117, Band Decoupling – WALTZ-16. J. Magn. Reson.
11536–11544. 52, 335–38.
Chapter 10
Abstract
Advances in solution nuclear magnetic resonance (NMR) methodology that enable studies of very large
proteins have also paved the way for studies of membrane proteins that behave like large proteins due to
the added weight of surfactants. Solution NMR has been used to determine the high-resolution structures
of several small, membrane proteins dissolved in detergent micelles and small bicelles. However, the usual
difficulties with membrane proteins in producing, purifying, and stabilizing the proteins away from native
membranes remain, requiring intensive screening efforts. Low levels of heterologous expression can be the
most detrimental aspect to studying membrane proteins. This is exacerbated for NMR studies because of
the costs of isotopically enriched media. Thus, solution NMR studies have tended to focus on relatively
small, membrane proteins that can be expressed into inclusion bodies and refolded. Here, we describe
the methods used to produce, purify, and refold the proton channel M2 into detergent micelles, and the
procedures used to determine chemical shift assignments and the atomic level structure of the closed form
of the homotetrameric channel.
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_10, © Springer Science+Business Media, LLC 2012
165
166 J.K. Claridge and J.R. Schnell
Fig. 1. An overview of the homotetrameric structure of the high pH, closed M2 proton
channel. Alpha helices are shown as cylinders. An idealized schematic of the DHPC micelle
is shown as a single layer of detergent molecules with the hydrocarbon chains coating the
transmembrane domain. The positioning of the micelle is based on the absence of cross
peaks in amide backbone strips at the resonance frequency of water in an 15N-edited
NOESY (24).
168 J.K. Claridge and J.R. Schnell
2. Materials
2.4. NMR Chemical ● All NMR experiments are conducted on a 14.1 T spectrometer
Shift Assignments equipped with a cryogenic probe.
and Restraint ● Processing and preliminary analysis of data, peak fitting, and
Measurements extraction of intensities for determining scalar bond couplings
are performed using NMRPipe (25).
● Resonance assignments and quantitation of NOE cross-peak
intensities are performed using CARA (26).
● Prediction of backbone dihedral angles from chemical shifts is
performed using TALOS (27).
● Cylindrically shaped polyacrylamide gel: A solution containing
4.5% acrylamide concentration with an acrylamide/bisacryl-
amide molar ratio of 40 is cast in a 6-mm cylindrical vessel.
● A gel press kit (New Era Enterprises, Inc.) is used to push the
cylindrical gel (6 mm in diameter) into an open-ended 4.2-mm
inner diameter NMR tube.
3. Methods
Fig. 2. A generalized flowchart showing the production of tetrameric M2 samples from trpLE fusions expressed in
Escherichia coli.
10 Bacterial Production and Solution NMR Studies of a Viral… 171
3.1. Expression 1. Transform the plasmid into E. coli BL21(DE3) pLysS cells,
of the TrpLE-M2 plate on LB agar containing kanamycin and chloramphenicol,
Fusion Protein and incubate overnight at 37°C.
2. Inoculate cultures of 200–500 mL Luria–Bertani (LB) media
with individual colonies of freshly transformed bacteria and
grow overnight at 37°C with moderate shaking (150 rpm).
3. Centrifuge the cultures at 2,000 × g for 25 min at 4°C and resus-
pend into 40 mL of chilled M9 medium. Add the resuspended
cells to the large-scale M9 cultures (2–4 L) such that the OD600
is 0.2–0.3 relative to water. Grow each liter of culture in a 2.5-L
baffled flask at 37°C with moderate shaking (150 rpm).
4. When the OD600 reaches 0.6–0.7, induce expression of the
trpLE-M2 fusion by adding IPTG from a stock of 1 M IPTG
to a final concentration of 1 mM. Grow overnight. The final
OD600 is typically between 1.2 and 1.4 (see Note 2).
5. Analyze protein expression levels by using SDS-PAGE (12%
Bis–Tris gel). Spin down cell quantities equivalent to 250 mL
of OD 600 = 0.6 cell culture at 5,000 × g for 5 min at room
temperature and redissolve the pellet into 40 mL of 8 M urea,
20 mL of 4× LDS sample buffer, 5 mL of 1 M DTT, and 5 mL
of 10% SDS. Load 20 mL of sample per lane and run the gel
using MES SDS running buffer.
3.2. Fusion Protein 1. Harvest the cells by centrifugation at 5,000 × g for 30 min at
Purification, Cleavage, 4°C. Resuspend the cell pellets in lysis buffer using a Dounce
and Preparation homogenizer, and sonicate on ice for a total of 3 min with a
of Pure M2 20% duty cycle. Spin down the inclusion bodies and cell debris
at 10,000 × g for 25 min at 4°C. Solubilize the pellets in 40 mL
of lysis buffer and spin down at 10,000 × g for 25 min at 4°C to
purify away additional water-soluble contaminants (see Note 3).
2. Dissolve the water-insoluble matter in 50 mL (per 1 L culture)
of guanidine buffer using a Dounce homogenizer. Pellet undis-
solved matter, which includes nucleic acids, by centrifugation
at 100,000 × g for 1.5 h at 4°C. Add the supernatant to 2 mL
of Ni-NTA agarose beads (see Note 4) preequilibrated with
guanidine buffer. After a 1-h incubation at 4°C with gentle
rotation, pour the slurry into a gravity column and wash with
80 mL of guanidine buffer. Elute bound trpLE-M2 fusion
172 J.K. Claridge and J.R. Schnell
3.3. Reconstitution 1. Dissolve purified peptide (1.2 mg) into reconstitution buffer
of M2 into Detergent (see Note 7) at a concentration of 250 mM (final volume is
Micelles 960 mL), split into three 3.5 kDa MWCO dialysis cups, and
dialyze for 12 h against 2 L of NMR buffer with slow stirring
and one buffer change at 10 h.
2. Concentrate the sample to ~0.7 mM monomer using a cen-
trifugal concentrator with a 5 kDa MWCO.
3. Add rimantadine to 10 mM. Rimantadine binds at four equiva-
lent sites near the gate on the lipid-facing side of the channel
and stabilizes the closed conformation of the pore. Because of
the poor water solubility of rimantadine, it is added from a
0.3 M stock that contains 80 mM DHPC (Subheading 2.3).
4. Add deuterium oxide for the magnetic field lock to a concen-
3.4. NMR Chemical tration of 5% in three steps.
Shift Assignments
and Restraint
1. To achieve nearly complete sequence-specific backbone chemi-
cal shift assignment of 1HN, 15N, 13C¢, 13Ca, and 13Cb, use
Measurements
10 Bacterial Production and Solution NMR Studies of a Viral… 173
Fig. 3. Projection of the 1H,1H plane of 15N-separated NOESY spectra on an M2 sample containing (a) fully protonated or
(b) fully deuterated DHPC detergent showing the loss of information due to spin diffusion when the hydrocarbon chain of DHPC
is protonated (see Note 9). NOE mixing times were 90 and 110 ms, respectively.
174 J.K. Claridge and J.R. Schnell
Fig. 4. Final, lowest-energy structure of the high pH M2 channel showing the position of
methyl groups throughout the transmembrane domain, including the helix–helix interfaces.
Helices are shown as ribbons and methyl protons as filled circles. Structural elements in
adjacent helices are shaded darker or lighter.
4. Notes
Acknowledgments
References
1. LeMaster, D. M. (1994) Isotope labeling in 13. Marsden, R. L., Lee, D., Maibaum, M., Yeats,
solution protein assignment and structural C., and Orengo, C. A. (2006) Comprehensive
analysis. Prog. Nucl. Magn. Reson. Spectrosc. 26, genome analysis of 203 genomes provides struc-
371–419. tural genomics with new insights into protein
2. Pervushin, K., Riek, R., Wider, G., and family space. Nucleic Acids Res. 34, 1066–1080.
Wuthrich, K. (1997) Attenuated T2 relaxation 14. Staley, J. P., and Kim, P. S. (1994) Formation of
by mutual cancellation of dipole-dipole cou- a native-like subdomain in a partially folded
pling and chemical shift anisotropy indicates an intermediate of bovine pancreatic trypsin inhib-
avenue to NMR structures of very large bio- itor. Protein Sci. 3, 1822–1832.
logical macromolecules in solution. Proc. Natl. 15. Ma, C., Marassi, F. M., Jones, D. H., Straus, S.
Acad. Sci. U.S.A 94, 12366–12371. K., Bour, S., Strebel, K., Schubert, U., Oblatt-
3. Ruschak, A. M., and Kay, L. E. (2010) Methyl Montal, M., Montal, M., and Opella, S. J.
groups as probes of supra-molecular structure, (2002) Expression, purification, and activities
dynamics and function. J. Biomol. NMR 46, of full-length and truncated versions of the
75–87. integral membrane protein Vpu from HIV-1.
4. Gautier, A., Mott, H. R., Bostock, M. J., Protein Sci. 11, 546–557.
Kirkpatrick, J. P., and Nietlispach, D. (2010) 16. Zheng, H., Zhao, J., Wang, S., Lin, C. M.,
Structure determination of the seven-helix Chen, T., Jones, D. H., Ma, C., Opella, S., and
transmembrane receptor sensory rhodopsin II Xie, X. Q. (2005) Biosynthesis and purification
by solution NMR spectroscopy. Nat. Struct. of a hydrophobic peptide from transmembrane
Mol. Biol. 17, 768–774. domains of G-protein-coupled CB2 receptor.
5. Imai, S., Osawa, M., Takeuchi, K., and Shimada, J. Pept. Res. 65, 450–458.
I. (2010) Structural basis underlying the dual 17. Diefenderfer, C., Lee, J., Mlyanarski, S., Guo, Y.,
gate properties of KcsA. Proc. Natl. Acad. Sci. and Glover, K. J. (2009) Reliable expression
U.S.A. 107, 6216–6221. and purification of highly insoluble transmem-
6. Bruggert, M., Rehm, T., Shanker, S., Georgescu, brane domains. Anal. Biochem. 384, 274–278.
J., and Holak, T. A. (2003) A novel medium for 18. Call, M. E., Schnell, J. R., Xu, C., Lutz, R. A.,
expression of proteins selectively labeled with Chou, J. J., and Wucherpfennig, K. W. (2006)
15
N-amino acids in Spodoptera frugiperda (Sf9) The structure of the zetazeta transmembrane
insect cells. J. Biomol. NMR 25, 335–348. dimer reveals features essential for its assembly
7. Strauss, A., Bitsch, F., Cutting, B., Fendrich, with the T cell receptor. Cell 127, 355–368.
G., Graff, P., Liebetanz, J., Zurini, M., and 19. Chong, Y. H., Ball, J. M., Issel, C. J., Montelaro,
Jahnke, W. (2003) Amino-acid-type selective isotope R. C., and Rushlow, K. E. (1991) Analysis of
labeling of proteins expressed in Baculovirus- equine humoral immune responses to the trans-
infected insect cells useful for NMR studies. membrane envelope glycoprotein (gp45) of
J. Biomol. NMR 26, 367–372. equine infectious anemia virus. J. Virol. 65,
8. Makino, S., Goren, M. A., Fox, B. G., and Markley, 1013–1018.
J. L. Cell-free protein synthesis technology in 20. Smith, J. G., Mothes, W., Blacklow, S. C., and
NMR high-throughput structure determination. Cunningham, J. M. (2004) The mature avian
Methods Mol. Biol. 607, 127–147. leukosis virus subgroup A envelope glycoprotein
9. Kim, H. J., Howell, S. C., Van Horn, W. D., Jeon, is metastable, and refolding induced by the syn-
Y. H., and Sanders, C. R. (2009) Recent Advances ergistic effects of receptor binding and low pH is
in the Application of Solution NMR Spectroscopy coupled to infection. J. Virol. 78, 1403–1410.
to Multi-Span Integral Membrane Proteins. Prog. 21. Bertrand, K., Squires, C., and Yanofsky, C.
Nucl. Magn. Reson. Spectrosc. 55, 335–360. (1976) Transcription termination in vivo in the
10. Warchawski, D. (2010) Membrane proteins of leader region of the tryptophan operon of
known structure determined by NMR. http:// Escherichia coli. J. Mol. Biol. 103, 319–337.
www.drorlist.com/nmr/MPNMR.html. 22. Miozzari, G. F., and Yanofsky, C. (1978) Translation
11. Chill, J. H., Louis, J. M., Miller, C., and Bax, of the leader region of the Escherichia coli trypto-
A. (2006) NMR study of the tetrameric KcsA phan operon. J. Bacteriol. 133, 1457–1466.
potassium channel in detergent micelles. Protein 23. Kleid, D. G., Yansura, D., Small, B., Dowbenko,
Sci. 15, 684–698. D., Moore, D. M., Grubman, M. J., McKercher,
12. Liu, Y., Engelman, D. M., and Gerstein, M. P. D., Morgan, D. O., Robertson, B. H., and
(2002) Genomic analysis of membrane protein Bachrach, H. L. (1981) Cloned viral protein
families: abundance and conserved motifs. vaccine for foot-and-mouth disease: responses
Genome Biol. 3, research0054. in cattle and swine. Science 214, 1125–1129.
10 Bacterial Production and Solution NMR Studies of a Viral… 179
24. Schnell, J. R., and Chou, J. J. (2008) Structure peptide as revealed by three-bond carbon-carbon
and mechanism of the M2 proton channel of couplings and 13C chemical shifts. J. Biomol.
influenza A virus. Nature 451, 591–595. NMR 7, 256–260.
25. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., 37. Hu, J.-S., Grzesiek, S., and Bax, A. (1997) Chi1
Pfeifer, J., and Bax, A. (1995) NMRPipe: a multi- angle information from a simple two-dimensional
dimensional spectral processing system based on NMR experiment which identifies trans 3JNCg
UNIX pipes. J. Biomol. NMR 6, 277–293. couplings in isotopically enriched proteins.
26. Keller, R. (2004) The Computer Aided Resonance J. Biomol. NMR 9, 323–328.
Assignment Tutorial, First Edition ed., Cantina- 38. Hu, J.-S., Grzesiek, S., and Bax, A. (1997)
Verlag, Goldau, Switzerland. http://cara.nmr- Two-dimensional NMR methods for determining
software.org/downloads/3-85600-112-3.pdf. c1 angles of aromatic residues in proteins from
27. Cornilescu, G., Delaglio, F., and Bax, A. (1999) three-bond JC¢Cg and JNCgcouplings. J. Am.
Protein backbone angle restraints from searching Chem. Soc. 119, 1803–1804.
a database for chemical shift and sequence 39. Tycko, R., Blanco, F. J., and Ishii, Y. (2000)
homology. J. Biomol. NMR 13, 289–302. Alignment of biopolymers in strained gels:
28. Schwieters, C. D., Kuszewski, J., Tjandra, N., A new way to create detectable dipole-dipole
and Clore, G. M. (2002) The Xplor-NIH NMR couplings in high-resolution biomolecular
molecular structure determination package. J. NMR. J. Am. Chem. Soc. 122, 9340–9341.
Magn. Reson. 160, 66–74. 40. Sass, H. J., Musco, G., Stahl, S. J., Wingfield, P.
29. Zweckstetter, M., Bax, A. (2000) Prediction of T., and Grzesiek, S. (2000) Solution NMR of
sterically induced alignment in a dilute liquid proteins within polyacrylamide gels: diffusional
crystalline phase: aid to protein structure deter- properties and residual alignment by mechanical
mination by NMR. J. Am. Chem. Soc. 122, stress or embedding of oriented purple mem-
3791–3792. branes. J. Biomol. NMR 18, 303–309.
30. Salzmann, M., Wider, G., Pervushin, K., and 41. Chou, J. J., Gaemers, S., Howder, B., Louis, J.
Wuthrich, K. (1999) Improved sensitivity and M., Bax, A. (2001) A simple apparatus for gen-
coherence selection for [N-15,H-1]-TROSY erating stretched polyacrylamide gels, yielding
elements in triple resonance experiments. J. uniform alignment of proteins and detergent
Biomol. NMR 15, 181–184. micelles. J. Biomol. NMR 21, 377–382.
31. Kay, L. E., Ikura, M., Tschudin, R., and Bax, A. 42. Weigelt, J. (1998) Single Scan, Sensitivity- and
(1990) Three-dimensional triple resonance Gradient-Enhanced TROSY for Multidimen-
NMR spectroscopy of isotopically enriched sional NMR Experiments. J. Am. Chem. Soc. 120,
proteins. J. Magn. Reson. 89, 496–514. 10778–10779.
32. Neri, D., Szyperski, T., Otting, G., Senn, H., 43. Cornilescu, G., Marquardt, J. L., Ottiger, M.,
and Wuthrich, K. (1989) Stereospecific nuclear and Bax, A. (1998) Validation of protein struc-
magnetic resonance assignments of the methyl ture from anisotropic carbonyl chemical shifts
groups of valine and leucine in the DNA- in a dilute liquid crystalline phase. J. Am. Chem.
binding domain of the 434 repressor by biosyn- Soc. 120, 6836–6837.
thetically directed fractional 13C labeling. 44. Lakomek, N. A., Lange, O. F., Walter, K. F.,
Biochemistry 28, 7510–7516. Fares, C., Egger, D., Lunkenheimer, P., Meiler,
33. Grzesiek, S., Vuister, G. W., and Bax, A. (1993) J., Grubmuller, H., Becker, S., de Groot, B. L.,
A simple and sensitive experiment for measure- and Griesinger, C. (2008) Residual dipolar cou-
ment of JCC couplings between backbone car- plings as a tool to study molecular recognition of
bonyl and methyl carbons in isotopically enriched ubiquitin. Biochem. Soc. Trans. 36, 1433–1437.
proteins. J. Biomol. NMR 3, 487–493. 45. Loria, J. P., Rance, M., and Palmer, A. G.
34. Vuister, G. W., Wang, A. C., and Bax, A. (1993) (1999) A relaxation-compensated Carr-Purcell-
Measurement of three-bond nitrogen-carbon J Meiboom-Gill sequence for characterizing
couplings in proteins uniformly enriched in 15N chemical exchange by NMR spectroscopy. J.
and 13C. J. Am. Chem. Soc. 115, 5334–5335. Am. Chem. Soc. 121, 2331–2332.
35. Bax, A., Vuister, G. W., Grzesiek, S., Delaglio, F., 46. Allerhand, A., and Thiele, E. (1966) Analysis of
Wang, A. C., Tschudin, R., and Zhu, G. (1994) Carr-Purcell Spin-Echo NMR Experiments on
Measurement of homo- and heteronuclear Multiple-Spin Systems. II. The Effect of Chemical
J couplings from quantitative J correlation. Exchange. J. Chem. Phys. 45, 902–916.
Methods Enzymol. 239, 79–105. 47. Golovanov, A. P., Hautbergue, G. M., Wilson,
36. MacKenzie, K. R., Prestegard, J. H., and S. A., and Lian, L. Y. (2004) A simple method
Engelman, D. M. (1996) Leucine side-chain for improving protein solubility and long-term
rotamers in a glycophorin A transmembrane stability. J. Am. Chem. Soc. 126, 8933–8939.
Chapter 11
Abstract
The integrity and propagation of the genome depend upon the fidelity of DNA processing events, such as
replication, damage recognition, and repair. Requisite to the numerous biochemical tasks required for
DNA processing is the generation and manipulation of single-stranded DNA (ssDNA). As the primary
eukaryotic ssDNA-binding protein, Replication Protein A (RPA) protects ssDNA templates from stray
nuclease cleavage and untimely reannealment. More importantly, RPA also serves as a platform for orga-
nizing access to ssDNA for readout of the genetic code, recognition of aberrations in DNA, and processing
by enzymes. We have proposed that RPA’s ability to adapt to such a broad spectrum of multiprotein
machinery arises in part from its modular organization and interdomain flexibility. While requisite for
function, RPA’s modular flexibility has presented many challenges to providing a detailed characterization
of the dynamic architecture of the full-length protein. To enable the study of RPA’s interdomain dynamics
and responses to ssDNA binding by biophysical methods including NMR spectroscopy, we have success-
fully produced recombinant full-length RPA in milligram quantities at natural abundance and enriched
with NMR-active isotopes.
Key words: Replication Protein A, DNA processing, Protein modularity, Isotopic labeling,
Recombinant expression, Protein purification, NMR spectroscopy
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_11, © Springer Science+Business Media, LLC 2012
181
182 C.A. Brosey et al.
the trimeric core of the protein (70C, 32D, 14), from which
emanate the flexibly linked N-terminal domains of RPA70 (70N,
70A, 70B), as well as the disordered N-terminal and structured
C-terminal domains of RPA32 (32N and 32C, respectively).
Binding of ssDNA is facilitated by domains 70A, 70B, 70C, and
32D, which together occupy an occluded site size of 30 nucle-
otides (1). Interactions with other DNA-processing proteins are
primarily mediated by domains 70N and 32C, and the principal
DNA-binding domains 70A and 70B (1, 2).
As a universal participant in DNA processing, RPA must inter-
act with a wide array of structurally unique multiprotein complexes.
The flexible, modular organization of the protein is thought to be
critical for enabling such structural adaptability (3). Although
high-resolution X-ray or NMR structures of all individual RPA
domains are available (4–8), the dynamic interdomain organiza-
tion of full-length RPA and the accompanying structural altera-
tions imposed by DNA processing have not been extensively
characterized. The full-length protein’s intrinsic flexibility poses
several challenges to study by X-ray diffraction; and at 116 kDa,
RPA falls outside the size limit of conventional NMR methods
(30–40 kDa). Application of advanced NMR approaches, however,
namely, deuterium labeling and TROSY- or CRINEPT-based tech-
niques, has allowed this size limitation to be extended to proteins
in excess of 100 kDa (9–12). This, combined with the discrete
distribution of molecular mass among RPA domains (50 kDa for
the trimer core and 10–14 kDa for the remaining domains), makes
feasible characterization of the full-length protein by NMR (13).
Here, we describe the production of full-length RPA by recom-
binant expression in Escherichia coli and subsequent purification of
the protein by a series of FPLC steps. The protocols provided
include those required for preparation of 2H-, 15N-enriched RPA
for study by NMR spectroscopy.
2. Materials
2.2. Cell Expression 1. RPA pET15b BL21(DE3) pLysS LB plate (Subheading 2.1).
Testing 2. Sterile 10-mL test culture tubes.
3. LB medium: 10 g/L tryptone, 10 g/L NaCl, 5 g/L yeast
extract dissolved in Milli-Q water and autoclaved at 121°C for
15 min.
4. 1,000× Antibiotic stocks (Subheading 2.1).
5. 1 M IPTG: Sterilize by filtration at 0.2 μm and store at −20°C.
6. 2× SDS loading buffer: 100 mM Tris–HCl, pH 6.8, 4% (w/v)
SDS (electrophoresis grade), 0.2% (w/v) bromophenol blue,
20% (w/v) glycerol, 200 mM β-mercaptoethanol (βME,
added fresh) (14).
7. 8 M urea.
8. Precast 4–12% Bis-Tris SDS-PAGE gel (Invitrogen).
9. 1× MES SDS running buffer (Invitrogen).
10. 1× Prestained molecular weight standards.
11. SimplyBlue SafeStain.
2.4. Starter Cultures 1. RPA pET15B BL21(DE3) pLysS LB plate (Subheading 2.1).
2.4.1. LB Medium 2. 250 mL LB starter culture (Subheading 2.3.1).
2.4.2. Minimal Medium 1. RPA pET15B BL21(DE3) pLysS LB plate (Subheading 2.1).
and Deuterated Minimal 2. Sterile 10-mL test culture tubes.
Medium
3. LB medium (Subheading 2.2).
4. 1,000× Antibiotic stocks (Subheading 2.1).
5. 250 mL Minimal medium starter culture (Subheading 2.3.2).
2.6. RPA Purification 1. Lysis buffer: Dissolve two complete EDTA-free protease inhibitor
cocktail tablets (Roche) in 80 mL of Ni-NTA buffer A
2.6.1. Cell Lysis
(Subheading 2.6.2) in a 150-mL glass beaker on ice immedi-
ately prior to use.
2. 100-mL Glass homogenizer.
3. Sonic dismembrator.
4. 25-mm diameter, 0.45-μm syringe filter.
2.6.4. Superdex 200 Gel 1. Centrifugal concentrators (15 mL, 30 kDa MWCO).
Filtration Chromatography 2. 0.22-μm centrifugal spin filters.
3. Refrigerated Äkta FPLC purification system and accessories.
4. Gel filtration buffer: 20 mM HEPES, pH 7.5, 100 mM NaCl,
5 mM βME, 10 μM ZnCl2, 200 mM arginine; adjusted to tar-
get pH with concentrated HCl, filtered at 0.45 μm under
vacuum, and stored at 4°C (see Notes 6 and 7).
5. Superdex 200 HR 10/30 column (GE Healthcare).
3. Methods
3.1. Cell Ensuring robust antibiotic selection of the RPA vector on solid
Transformation medium is vital to enabling the success of subsequent liquid
cultures (see Note 2).
1. Thaw 100 μL of BL21(DE3) pLyS competent cells on ice, gen-
tly combine with 100 ng of RPA pET15b vector, and incubate
for 30 min on ice.
2. Heat shock the cells at 42°C for 45 s and incubate on ice for
2 min. Add 900 μL of sterile SOC recovery medium.
3. Incubate the cells for 1 h at 37°C and 200–230 rpm, then
centrifuge the cells at 16,100 × g for 1 min at room tempera-
ture (see Note 8). Remove 900 μL of the clarified SOC medium
and gently resuspend the cells in the remaining medium prior
to plating on an LB medium plate.
4. Incubate the plates overnight at 37°C.
3.2. Cell Expression Expression testing allows confirmation of the expression capability
Testing of the transformed bacterial colonies prior to scaling up protein
production. The testing also allows for selection of colonies with
the most robust expression.
11 Preparation of the Modular Multi-Domain Protein RPA… 187
3.3. Preparation This medium serves for the production of unlabeled RPA.
of Culture Media Preparation should include a 250-mL starter culture to accommo-
date a 6-L large-scale culture.
3.3.1. LB Medium
1. Dissolve LB components (Subheading 2.2) in Milli-Q water in
a 500-mL baffled Erlenmeyer flask (starter culture) or 2.8-L
baffled Fernbach flasks (large-scale culture), autoclave at 121°C
for 15 min, and cool to 50–60°C.
2. Add ampicillin and chloramphenicol at 1:1,000 dilution imme-
diately prior to inoculation.
188 C.A. Brosey et al.
3.3.2. Minimal Medium This medium serves for the production of 15N-enriched RPA.
Preparation should include a 250 mL starter culture to accommo-
date a 6 L large-scale culture. A 250 mL minimal medium culture
also serves as an adaptation culture for production of deuterated
protein.
1. Dilute 10× M9 salts in Milli-Q water to 1× in a 500-mL baffled
Erlenmeyer flask (starter culture) or 2.8-L baffled Fernbach
flasks (large-scale culture), autoclave at 121°C for 15 min, and
cool to 50–60°C.
2. Add the following components immediately prior to inocu-
lation: 0.5 g/L of 15NH4Cl (see Note 11), 2 mL/L of 1 M
MgSO4, 100 μL/L of 1 M CaCl2, 10 mL/L of 20% glucose,
1 mL/L of 1 M thiamine hydrochloride, and antibiotic stocks
at 1:1,000 dilution.
3.3.3. Deuterated Minimal This medium serves for six 1 L large-scale production cultures and
Medium is prepared immediately prior to inoculation after the success of the
250 mL minimal medium adaptation culture has been ascertained.
1. Dry autoclave six 2.8-L Fernbach flasks with baffles and allow
to dry thoroughly overnight.
2. Dissolve all dry components in 6 L of 99% D2O (see Notes 12
and 13) and immediately sterilize the medium in 1-L batches
by using sterile vacuum filtration systems (i.e., 1 L/unit). This
apparatus filters the medium directly into a sterile 1-L bottle.
Chloramphenicol is not included in the medium at this stage to
ease the metabolic burden on the cells.
3. Carefully transfer each 1 L of sterile deuterated minimal
medium to a dry, sterile Fernbach flask (see Note 14).
3.5. Large-Scale Cell 1. Prepare six 1 L cultures of rich LB medium, minimal medium, or
Culture and deuterated minimal medium as described above (Subheading 3.3)
Overexpression and inoculate each with 30 mL (40 mL for deuterated minimal
medium) of the corresponding overnight starter culture.
2. Grow the cultures at 37°C, 200–230 rpm, until an A600 of
0.6–0.7 is reached (see Note 15).
3. Allow the cultures to equilibrate for half an hour with agitation
at 18°C (or room temperature for deuterated minimal medium)
prior to induction. Collect a preinduction SDS-PAGE sample
as described above (Subheading 3.2) and induce the cells with
1 mM IPTG. Allow cells to express overnight (approximately
16–18 h).
4. The A600 at the end of the expression period should be 1.8–2.0
for LB medium cultures and 0.9–1.0 for standard and deuterated
minimal media cultures. Collect postinduction SDS-PAGE
samples from the cultures as described above (Subheading 3.2).
Harvest the cultures by centrifuging at 10,000 × g for 20 min
at 4°C.
5. Decant the supernatant and reserve the spent deuterated media
for recycling (15). Spent LB or minimal media may be decon-
taminated by the addition of bleach or a 1% Terg-a-zyme solu-
tion for 30 min, and then discarded. If purification does not
follow immediately, transfer the pellets to sterile 50-mL conical
tubes and freeze at −80°C. Run pre- and postinduction SDS-
PAGE samples as in Subheading 3.2 to confirm the presence
of RPA expression.
3.6. RPA Purification Purification of RPA involves three primary steps: Ni-NTA affinity,
heparin, and size-exclusion chromatography. For best results, the
protocol should be completed over the course of 2 days, where
Ni-NTA and heparin chromatography steps are accomplished the
first day and the final gel filtration step is carried out on the second
day. If necessary, the Ni-NTA and heparin steps may be divided
into two separate days and protein fractions from each purification
kept at 4°C overnight. Ideally, though, the time from cell lysis to
the final gel filtration exchange should be kept to a minimum.
3.6.1. Cell Lysis 1. If cells have been frozen at −80°C, thaw the pellets by
submerging the 50-mL conical tubes in cool water. Meanwhile,
prepare and chill the lysis buffer and pre-chill the 100-mL glass
homogenizer on ice (see Note 16).
2. Transfer all 6 L of RPA cell pellets into the 100-mL homoge-
nizer, rinse the 50-mL conical tubes with ice-cold lysis buffer,
and add the rinse and any remaining lysis buffer to the
homogenizer.
190 C.A. Brosey et al.
3.6.3. Desalting Exchange The charged DNA-binding clefts of RPA render it sensitive to the
and Heparin absence of ambient salt. Effective binding of RPA to the heparin
Chromatography matrix, however, requires a low salt content in the loading buffer.
Direct dialysis into the loading buffer (heparin buffer A) usually
provokes extensive precipitation of RPA. Buffer exchange by desalt-
ing, however, allows for the rapid and successful transfer of RPA
into the loading buffer with minimal aggregation. Once the series
of FPLC desalting runs are complete, it is imperative to load the
exchanged protein directly onto the heparin column to restore a
stabilizing ionic environment. As before, loading, washing, and
11 Preparation of the Modular Multi-Domain Protein RPA… 191
3.6.4. Superdex 200 Gel Gel filtration provides a final polishing step for the purification
Filtration Chromatography and ensures the removal of any trace RPA70 or low-molecular-weight
contaminants. As before, loading and elution are implemented
automatically using a pre-programmed Äkta FPLC method.
1. Equilibrate the Superdex 200 HR 10/30 column with 1.5 CVs
of filtered Milli-Q water and 1.5 CVs of gel filtration buffer.
2. Pre-rinse two centrifugal concentrators (15 mL, 30 kDa
MWCO) with Milli-Q by centrifuging at 3,700 × g for 10 min
at 4°C. Concentrate the heparin RPA pool to ~300–500 μL
(see Note 19).
3. Filter the concentrate by using a 0.22-μm centrifugal spin filter
in a refrigerated (4°C) centrifuge and load onto the Superdex
200 HR 10/30 column at 0.3 mL/min, collecting 0.5-mL
fractions for 1.5 CVs.
192 C.A. Brosey et al.
4. As before, assess the presence of RPA from the A280 trace and SDS-
PAGE of relevant fractions (sampling 5 μL of each fraction).
5. Before the final RPA fractions are pooled, acquire final A280
and A260 measurements by UV-Vis spectrophotometry to
ensure that the selected fractions are DNA-free, as determined
by A260/A280 ratios of 0.64 or less.
4. Notes
1. The tricistronic RPA pET15b vector was a gift from the lab of
Alexey Bochkarev. 6×-His tags with thrombin cleavage sites
precede the RPA70 and RPA14 subunits. The order of subunit
open reading frames (ORFs) is as follows: RPA70, RPA14,
RPA32.
2. Using freshly prepared ampicillin stock in the LB medium is
important for ensuring robust RPA transformation. Even
though the choice of the BL21(DE3) pLysS cell line is designed
to circumvent leaky expression, even small amounts of nonin-
duced RPA can potentially result in resistant cells with less than
robust expression.
3. Antibiotic stocks may be aliquoted and stored at −20°C for
future use. For long-term storage of ampicillin stocks, storage
at −80°C is recommended.
4. The 70C domain of RPA contains a zinc-binding motif, for
which ZnCl2 is included in the purification buffers.
5. The most effective imidazole concentrations in the Ni-NTA
buffers will depend on how recently the Ni-NTA resin has been
charged. For freshly charged resin, the imidazole concentration
11 Preparation of the Modular Multi-Domain Protein RPA… 193
that all buffers are kept ice cold and that heating from other
steps of the lysis (sonication, centrifugation) is kept to a
minimum.
17. The ice-water bath serves as a heat sink during sonication of
the lysate.
18. Initializing the FPLC system (pump washing, cleaning super-
loops, setting up fraction collectors), as well as equilibration of
the Ni-NTA column, should occur prior to or concurrently
with cell lysis to ensure that the clarified, filtered lysate can be
loaded directly into the system as soon as it is available.
19. Centrifugal concentrators are usually spun in 10–15-min incre-
ments and carefully mixed with each addition of the Ni-NTA
pool to prevent buildup and aggregation of RPA at the base of
the concentrator.
20. The desalting resolution of the HiPrep 26/10 Desalting
column (GE Healthcare) is 10 mL; that is, the column can
effectively exchange 10 mL of injected sample into the tar-
get buffer without contamination from the original buffer.
To avoid desalting too concentrated a volume of RPA and
triggering aggregation, the Ni-NTA pool is processed in
three sequential 10-mL batches.
21. As with the Ni-NTA step, initializing the FPLC system and
equilibrating the desalting and heparin columns should occur
prior to or concurrently with concentrating the RPA Ni-NTA
pool to allow for immediate loading once the target volume
is reached.
Acknowledgments
The authors would like to thank Dr. Dalyir Pretto and Susan Meyn.
This work was supported by the National Institutes of Health
operating grant R01 GM65484 and graduate training grant T32
GM08320.
References
1. Wold, M. S. (1997) Replication protein A: repair, and recombination. J. Biol. Chem. 279,
A heterotrimeric, single-stranded DNA-binding 30915–30918.
protein required for eukaryotic DNA metabo- 4. Bochkarev, A., Pfuetzner, R. A., Edwards, A. M.,
lism. Annu. Rev. Biochem. 66, 61–92. and Frappier, L. (1997) Structure of the single-
2. Fanning, E., Klimovich, V., and Nager, A. R. stranded-DNA-binding domain of replication
(2006) A dynamic model for replication pro- protein A bound to DNA. Nature 385,
tein A (RPA) function in DNA processing 176–181.
pathways. Nuc. Acids Res. 34, 4216–4137. 5. Jacobs, D. M., Lipton, A. S., Isern, N. G.,
3. Stauffer, M. E., and Chazin, W. J. (2004) Daughdrill, G. W., Lowry, D. F., Gomes, X.,
Structural mechanisms of DNA replication, and Wold, M. S. (1999) Human replication
11 Preparation of the Modular Multi-Domain Protein RPA… 195
protein A: Global fold of the N-terminal RPA- 10. Riek, R., Wider, G., Pervushin, K., and
70 domain reveals a basic cleft and flexible Wuthrich, K. (1999) Polarization transfer by
C-terminal linker. J. Biomol. NMR 14, cross-correlated relaxation in solution NMR
321–331. with very large molecules. Proc. Natl. Acad.
6. Bochkarev, A., Bochkareva, E., Frappier, L., Sci. U.S.A. 96, 4918–4923.
and Edwards, A. M. (1999) The crystal struc- 11. Riek, R., Pervushin, K., and Wuthrich, K.
ture of the complex of replication protein A (2000) TROSY and CRINEPT: NMR with
subunits RPA32 and RPA14 reveals a mecha- large molecular and supramolecular structures
nism for single-stranded DNA binding. EMBO J. in solution. Trends Biochem. Sci. 25,
18, 4498–4504. 462–468.
7. Mer, G., Bochkarev, A., Gupta, R., Bochkareva, 12. Tugarinov, V., Hwang, P. M., and Kay, L. E.
E., Frappier, L., Ingles, C. J., Edwards, A. M., (2004) Nuclear magnetic resonance spectros-
and Chazin, W. J. (2000) Structural basis for copy of high-molecular-weight proteins. Annu.
the recognition of DNA repair proteins UNG2, Rev. Biochem. 73, 107–146.
XPA, and RAD52 by replication factor A. Cell
103, 449–456. 13. Brosey, C. A., Chagot, M. E., Ehrhardt, M.,
Pretto, D. I., Weiner, B. E., and Chazin, W. J.
8. Bochkareva, E., Korolev, S., Lees-Miller, S. P., (2009) NMR analysis of the architecture and
and Bochkarev, A. (2002) Structure of the RPA functional remodeling of a modular multi-
trimerization core and its role in the multistep domain protein, RPA. J. Am. Chem. Soc. 131,
DNA-binding mechanism of RPA. EMBO J. 6346–6347.
21, 1855–1863.
9. Pervushin, K., Riek, R., Wider, G., and 14. Sambrook J and Russell D. (2001) Molecular
Wuthrich, K. (1997) Attenuated T2 relaxation cloning: A laboratory manual, vol. 3, 3rd ed.
by mutual cancellation of dipole-dipole cou- Cold Spring Harbor Laboratory Press,
pling and chemical shift anisotropy indicates an New York.
avenue to NMR structures of very large bio- 15. Li, M. X., Corson, D. C., and Sykes, B. D.
logical macromolecules in solution. Proc. Natl. (2002) Structure determination by NMR iso-
Acad. Sci. U.S.A. 94, 12366–12371. tope labeling. Meth. Mol. Biol. 173, 255–265.
Chapter 12
Abstract
This chapter describes the preparation of NMR quantities of RNA purified to single-nucleotide resolution
for protein–RNA interaction studies. The protocol is easily modified to make nucleotide-specific isotopically
labeled RNAs or uniformly labeled RNA fragments for ligation to generate segmentally labeled RNAs.
Key words: In vitro transcription, Single-nucleotide resolution, RNA synthesis, Protein–RNA interactions,
RNA purification, Isotopic labeling
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_12, © Springer Science+Business Media, LLC 2012
197
198 C.A. Theimer et al.
Fig. 1. RNA synthesis and purification. (a) An RNA synthesis and purification flowchart, including the general steps from
initial template design to NMR data collection. (b) DNA template design for RNA transcription. The RNA product is presented
on the annealed DNA template/T7 Top promoter duplex in lower case letters. The DNA sequences of the T7 RNA polymerase
promoter sequence and DNA template for transcription are presented in upper case letters. The two nucleotides at the 5¢
end of the template, which should have 2¢-OMe substitutions, are indicated by asterisks. (c) A representative 20% (19:1)
acrylamide:bisacrylamide gel showing the observed bands from a typical analytical transcription experiment compared to
the DNA template.
12 NMR Studies of Protein–RNA Interactions 199
2. Materials
2.1. DNA Template 1. Template DNA: Complementary DNA template for RNA
Design for RNA transcription, 0.1–1.0 mM in water, store at −20°C, Fig. 1b
Synthesis (see Note 1).
2. T7 Top DNA: Purified T7 RNA polymerase promoter DNA
(coding strand, 5¢-TAATACGACTCACTATA-3¢), 1 mM in
water, store at −20°C.
2.2. RNA Synthesis 1. 50 mM Annealed DNA template: T7 Top DNA stock solution
in water, store at −20°C.
2. 10× Transcription buffer: 400 mM Tris–HCl, pH 8.0, 10 mM
spermidine, 0.1% (w/v) Triton X-100; store at room temperature.
3. 100 mM dithiothreitol (DTT): Store in 1–2-mL aliquots
at −20°C.
4. 100 mM ATP, CTP, GTP, and UTP: Dissolve nucleotides in
water and adjust to pH 8.0 using 1M NaOH; store in 1–2-mL
aliquots at −20°C (see Note 2).
5. 1M MgCl2: Store at room temperature.
6. T7 RNA polymerase/ribonuclease (RNAse) inhibitor mixture:
T7 RNA polymerase in solution, dialyzed into 30 mM HEPES
(adjusted to pH 7.5 with 1M NaOH), 0.1M potassium glutamate,
0.25 mM EDTA, 0.05% Tween-20, 1 mM DTT, 200 mM
NaCl. Following dialysis, RNAse inhibitor (see Note 3) and
glycerol (50% (w/v) final concentration) are added. Store in
1-mL aliquots at −80°C.
7. Gel running buffer (1× TBE): 90 mM Tris base, 90 mM boric
acid, and 2 mM EDTA (using 0.5M EDTA, pH 8.0, stock
solution). Typically made as a 5× or 10× concentrated stock
and diluted as needed; store at room temperature.
8. Denaturing acrylamide gel solution: 20% acrylamide/bisacryl-
amide (19:1), 1× TBE, and 7.8 M urea; store at 4°C (see Note 4).
9. 10% (w/v) Ammonium persulfate solution (APS): Store at 4°C.
10. Denaturing gel loading buffer: 80% formamide, 10 mM EDTA
(using 0.5M EDTA, pH 8.0, stock solution), 0.025% (w/v)
xylene cyanol, and 0.025% (w/v) bromophenol blue; store at
room temperature.
11. Toluidine blue stain: 0.25% (w/v) in water; prepare as needed,
and store at room temperature.
12. 100% Ethanol; store at −20°C.
13. 0.5M EDTA stock solution: 0.5M EDTA, adjust to pH 8.0
using solid NaOH pellets.
14. N,N,N´,N´-tetramethyl-ethane-1,2-diamine (TEMED).
200 C.A. Theimer et al.
2.3. RNA Purification 1. Denaturing gel loading buffer, denaturing acrylamide gel solu-
tion, APS, and gel running solutions are identical to those
described in Subheading 2.2.
2. Polyester-backed, silica-based, fluorescent thin-layer chroma-
tography (TLC) plates.
3. Low-salt buffer: 10 mM monosodium phosphate/disodium
phosphate buffer, pH 7.6 (adjusted with 1M NaOH or 1M
HCl if the pH is off by more than 0.2 pH units), 1 mM EDTA,
and 200 mM KCl; store at room temperature (see Note 5).
4. High-salt buffer: 10 mM monosodium phosphate/disodium
phosphate buffer, pH 7.6 (adjusted with 1M NaOH or 1M
HCl if the pH is off by more than 0.2 pH units), 1 mM EDTA,
and 1.5M KCl; store at room temperature (see Note 5).
5. HiTrap-Q anion exchange column (5 mL prepacked).
6. 20% (w/v) Ethanol.
7. Amicon-pressurized stirred cell and 1,000 and/or 3,000
molecular weight cutoff (MWCO) membranes.
8. NMR buffer for preliminary RNA experiments or for RNA–
protein interaction studies (typically, 10–20 mM sodium phos-
phate buffer, pH 6.0–7.0, 0–200 mM KCl).
9. 70% Ethanol.
10. 100% Ethanol.
11. Electroelution chamber.
2.4. Preliminary RNA 1. 99.99% D2O: Store in sealed ampoules at room temperature.
Analysis by NMR 2. 4M potassium chloride: Store at room temperature.
3. Acid and base solutions for pH optimization: Usually, 0.1 and
1M hydrochloric acid and 0.1 and 1M potassium (or sodium)
hydroxide; store at room temperature.
4. Sodium azide.
5. Standard and Shigemi NMR tubes.
3. Methods
3.1. DNA Template Commonly, a partially double-stranded DNA template is used for
Design for RNA in vitro transcription of large quantities of RNA, since only the
Synthesis promoter region of the template DNA must be double stranded
for T7 RNA polymerase to bind and initiate transcription (Fig. 1b)
(16). Standard PCR can be used to make the DNA template com-
pletely double stranded if it improves transcription efficiency and
the overall yield for a particular template, although we do not often
find this to be necessary. The biggest issue for synthesizing RNA
by using in vitro transcription is RNA product heterogeneity due
to in vitro transcription artifacts (Fig. 1c). It has been demonstrated
that, in vitro, T7 RNA polymerase can produce off-target products
as a result of 5¢-heterogeneity, 3¢-heterogeneity, and RNA-
templated RNA addition (16–19). T7 RNA polymerase only syn-
thesizes RNA products that contain at least one 5¢-G nucleotide.
Increased transcription efficiency is observed with two or three
5¢-G nucleotides, but multiple G nucleotides at the 5¢-end of the
RNA has been shown to cause 5¢-heterogeneity (additional G
nucleotides) in the RNA product (18). To strike a balance between
transcription efficiency and artifact generation, we prefer, when
possible, to start sequences with no more than two G nucleotides in a
row. 3¢-heterogeneity appears to be a result of the runoff transcrip-
tion mechanism and often results in the nontemplated addition of
one or two nucleotides (N + 1 and, less often, N + 2 products) (16).
This problem can be diminished or overcome completely by hav-
ing the DNA template synthesized with 2¢-methoxyl groups on the
two terminal 5¢ nucleotides (20, 21), and we strongly recommend that
template DNAs be synthesized with this modification; Fig. 1b.
Typically, we see RNA-templated RNA addition products that
vary in size from ~10–20 nucleotides longer than the expected
transcribed product. Since this phenomenon is sequence dependent,
some sequences display no extraneous long products and some
sequences make large amounts of long products (17, 19). A study
has been performed on DNA sequence mutations that reduce or
abolish this behavior (22); when sequence alteration is not feasible,
there are transcription conditions that can help to reduce this problem
(19). In addition, strategies to eliminate 5¢- and 3¢-heterogeneity
in RNA samples using ribozyme-based cleavages are also performed,
as described in detail elsewhere (13). Finally, mass spectrometry is
an excellent technique to check that the final RNA product is the
correct length and has the correct sequence composition.
The DNA template is complementary to the sequence of the
RNA and the top strand of the T7 promoter sequence (Fig. 1b).
Both DNA sequences for RNA transcription can either be ordered
202 C.A. Theimer et al.
3.2. RNA Synthesis The transcription reaction is essentially the same for making selec-
tively or uniformly 13C-,15N-isotopically labeled RNA as it is for
unlabeled RNA. All nucleotides are prepared and stored separately
so that one or more unlabeled nucleotides can be replaced with
13
C-,15N-isotopically labeled nucleotides for making selectively
labeled RNA samples or all four unlabeled nucleotides can be
replaced with 13C-,15N-isotopically labeled nucleotides for uniform
labeling. The primary differences are as follows: (1) since labeled
nucleotides are prohibitively expensive, we typically run labeled
transcriptions at a concentration of 2 mM for each nucleotide and
(2) we often explore additional transcription optimization condi-
tions (additional transcription components), such as adding inorganic
pyrophosphatase (IPP) or polyethylene glycol (avg. MW 8,000),
for improved transcription efficiency (23).
Before performing large-scale transcription reactions (10–
50 mL), it is necessary to determine the optimum transcription
conditions for every new DNA template (Fig. 1c). The conditions
that need to be considered include NTP concentration (2–4 mM),
12 NMR Studies of Protein–RNA Interactions 203
size can be used. Rinse the glass plates and spacers with water,
followed by ethanol, and wipe dry immediately prior to use.
Although very small gels can be made using standard protein
SDS-gel apparatus, we do not advise this as the resolution on
such gels is generally too low for single-nucleotide resolution
and thus not satisfactory for this purpose.
10. Prepare a 0.75-mm-thick 20% gel by mixing 50 mL of denaturing
acrylamide gel solution with 500 mL of 10% APS. To this solu-
tion, mix in 50 mL of TEMED and immediately pour the gel.
These gels are a single layer and the comb (10, 14, or 20 well)
is inserted immediately after the entire gel is poured. The gel
takes approximately 30 min to polymerize. The percentage of
acrylamide:bisacrylamide solution used depends on the size of
the RNA product that is run on the gel. We typically use 20%
for RNA transcripts up to 40–50 nucleotides long, 15% for
RNA transcripts between 45 and 70 nucleotides, and 10–12%
for longer RNA transcripts.
11. When the gel is polymerized, remove the comb and carefully
rinse the wells with water by using a 30-mL syringe equipped
with a 22-gauge needle. Place the gel in the apparatus and fill
the buffer chambers with gel running buffer. The gel should be
pre-run at 150 V (or 15 W) for 15–20 min and the wells rinsed
again with gel running buffer prior to loading the samples.
12. Once the samples are loaded, run the gel for 2–3 h at 150 V or
until the bromophenol blue dye front (dark blue) is within
~3 cm of the bottom of the gel.
13. Stain the gel with toluidine blue stain on an orbital shaker for
15 min. Destain the gel with water (multiple exchanges) until
you observe good contrast between the dark blue nucleic acid
bands and the background of the gel. The optimal transcription
conditions are chosen based on the test conditions that produce
the highest intensity of the RNA product band.
14. Once the optimal solution conditions have been identified, the
large-scale transcription reaction can be performed. Generally,
we perform a 30 mL transcription reaction in a sterile, disposable,
blue-capped 50-mL centrifuge tube. We find that this is a large
enough transcription volume to generate a reasonable quantity
of RNA for NMR samples (150–400 nmol of RNA) for most
RNA sequences, although (rarely) some RNA sequences transcribe
much better or much worse than this. The large-scale reaction
is a direct scale-up of the small-scale reaction. The only differences
are the volumes of reagents used and that it is usually not necessary
to centrifuge the solution after vortex mixing. The T7 RNA
polymerase/RNAse inhibitor mixture is still the last ingredient
added before incubation and the tube is swirled gently after
adding the polymerase.
12 NMR Studies of Protein–RNA Interactions 205
15. Incubate the reactions for 4–8 h at 37°C. The reactions can be
incubated overnight, although this can be risky if there is any
possibility of RNAse contamination or if the RNA sequence
is prone to undesirable side reactions, like RNA-primed
RNA addition by T7 RNA polymerase, as described above (see
Note 10).
16. Centrifuge the reaction solution at 4,000 × g for 5 min at 4°C
and decant the supernatant, which contains the RNA product,
into a sterile 250-mL centrifuge bottle. There is a large pellet
from precipitated inorganic pyrophosphate; this does not contain
any RNA and can be discarded.
17. Add 1/10th the volume of 0.5M EDTA, pH 8.0 (1 mL of
EDTA for every 10 mL of transcription), swirl to mix, and then
add 2.5–3 times the volume of cold 100% ethanol (~75–90 mL of
ethanol for a 30 mL transcription) to the transcription solution.
18. Place the centrifuge bottle containing the RNA product at −20°C
overnight to precipitate (see Note 11).
3.3. RNA Purification Denaturing PAGE is our primary means of purifying large quantities
of RNA to single-nucleotide resolution for NMR. Although we have
not experimented recently with currently available preparative-scale
HPLC columns, in the past we found that HPLC purification would
not purify RNAs larger than ~30 nucleotides to single-nucleotide
resolution on a preparative scale. In addition, while we frequently
purify small quantities of RNA using native PAGE, the native gels
must be run in the cold room (4°C), typically take a long time to
run, and rarely yield single-nucleotide resolution when the single-band
products are checked on analytical denaturing gels for purity.
1. Centrifuge the bottle at 14,000 × g for 45 min at 4°C.
2. Very gently, decant the supernatant, as soon as the centrifuge
stops running (see Note 12). There should be a visible white
pellet on the wall of the bottle. The size of the pellet is not a
direct reflection of product yield, since the majority of the size
of the pellet is due to salt precipitation.
3. If the pellet is very large, put 20–50 mL of cold 70% (w/v) etha-
nol in the bottle very gently to wash away excess salt. Centrifuge
with the pellet located against the outer wall at 14,000 × g for
30 min at 4°C, and again decant the supernatant immediately
(see Note 13).
4. Dry the pellet by placing the capless centrifuge bottle underneath
the hood, angled so that the residual ethanol is not lying
directly on top of the pellet. Evaporating off all residual ethanol
typically takes about 2 h if the supernatant was properly poured
off without disturbing the pellet.
206 C.A. Theimer et al.
12. Remove the gel from the apparatus, carefully remove one glass
plate, and cover the gel with plastic wrap. Flip the gel over
(plastic wrap on the bottom) onto a fluorescent TLC plate and
gently remove the top glass plate. When a handheld short-
wavelength UV light (254 nm) is shone onto the gel, the plate
glows green and the RNA band is visible as a grey to black
shadow on the fluorescent green background (UV shadowing).
Carefully cut the product bands out of the gel with a clean
razor blade and transfer them to a sterile, disposable 50-mL
centrifuge tube and discard the rest of the gel. Repeat for all of
the gels (exposure to UV light should be minimized to avoid
UV damage of RNA).
13. Elute the RNA product by using a 4-trap Elutrap electroelution
chamber (Whatman), with 1× TBE as the running buffer. Gel slices
should be cut into small pieces but not crushed, and placed
into the gel holding chamber between the BT1 and BT2
membranes (see Note 16).
14. Run the elution at 150 V and collect the RNA sample from the
trap at 2–3-h intervals, typically for 9 h. For longer RNAs, we
frequently reduce the voltage to 50 V and run overnight to
ensure complete elution of the product from the gel slices. The
progress of the elution can be tracked by calculating the amount
of RNA in the eluant at each time point, based on UV absor-
bance at 260 nm. Time points should be stored at −20°C until
all of the RNA has been collected, and the elution is finished.
15. Thaw and pool all of the eluted RNA fractions. Load this material
onto an HiTrap-Q anion-exchange column hooked up to a
peristaltic pump (flow rate 3–5 mL/min) and pre-equilibrated
with low-salt buffer. Wash the column with ten-column
volumes (50 mL) or more of low-salt buffer. Elute the RNA
with high-salt buffer and collect 4-mL fractions. The purified
RNA is typically found entirely in the second and third fractions.
The column must always be thoroughly rinsed with low- and
high-salt buffers between samples and immediately stored in
20% ethanol when not in use to prevent contamination.
16. Pool the two RNA containing fractions (8 mL) in a 50-mL cen-
trifuge tube, add 24 mL of cold 100% ethanol, and store the
RNA overnight at −20°C to precipitate.
17. Remove the centrifuge tube directly from the freezer, make a
weight-matched balance tube, and centrifuge at 15,000 × g for
45 min at 4°C. Very gently, decant the supernatant as soon as
the centrifuge stops running. At this point, it is possible to
again perform a 70% ethanol wash as described above (step 3).
18. Dry the pellet by placing the capless centrifuge tube under-
neath the hood, angled so that the residual ethanol is not lying
directly on top of the pellet. Evaporating off all residual ethanol
typically takes about 2 h.
208 C.A. Theimer et al.
3.4. Preliminary RNA Once an initial unlabeled RNA sample is made, a basic set of NMR
Analysis by NMR experiments is run, including 1D 1H NMR experiments collected
in 95% H2O/5% D2O (5–10°C) for pH and salt titrations, NOESYs
collected in 95% H2O/5% D2O (5–10°C) and D2O (20–30°C),
and a TOCSY collected in D2O (20–30°C), to assess the properties
of the RNA in solution at NMR concentrations. Optimal salt and
pH conditions are assessed by the number, intensity, and line
widths of the detectable imino proton resonances compared to the
number of imino protons which are expected to be protected from
rapid exchange due to hydrogen bonding in Watson-Crick and
noncanonical base pairs (Fig. 2). Generally, we find that the optimal
pH for RNA samples falls between 6.0 and 7.0 and the optimal salt
conditions vary from 0 to 200 mM monovalent salt, depending
on the ability of the sequence to form alternative conformations
(see Note 18). The optimal solution conditions are typically dictated
by the protein of interest, but it is important to be aware of the
expected behavior of the RNA under the appropriate solution
conditions.
Degradation products and alternative conformations (including
unwanted dimerization) can be identified both from additional imino
proton resonances in the 1D and 2D spectra collected in 95% H2O/5%
D2O (Fig. 2). In addition, the 1D 1H spectrum helps identify
any small-molecule contaminants in the RNA sample. The most
common contaminants in RNA samples (and their causes) are as
follows: acrylamide and/or urea (insufficient volume of low-salt
wash of the anion-exchange column before eluting the RNA product),
EDTA and/or ethanol (too few buffer exchanges during desalting
and concentration), and ethanol and/or glycerol (insufficient soaking
and rinsing of the Amicon membrane prior to installing in the
12 NMR Studies of Protein–RNA Interactions 209
Fig. 2. 1D and 2D imino proton NMR spectra of RNA. The secondary structure of a hairpin RNA and the imino proton region
of 1D and 2D NOESY 1H spectra (500 MHz) of the RNA in 95% H2O/5% D2O at 10°C demonstrate the presence of sharp
imino proton resonances. A clear sequential walk through the stem base pairs is indicated for the expected hairpin stem.
The smaller imino proton resonances and NOE peaks are due to an alternative conformation in solution.
stirred cell). EDTA can be eliminated from the high- and low-salt
anion-exchange buffers if it is a recurring contaminant in the final
samples. Ethanol contamination can easily be removed by freezing
the sample in liquid nitrogen, lyophilizing the sample to dryness,
and resuspending in 95% H2O/5% D2O (repeated as necessary).
Glycerol typically requires exhaustive dialysis or buffer exchanges
in the Amicon-stirred cell to be completely removed.
1. For titrations, the RNA sample was previously concentrated
into water during the last steps of RNA purification
(Subheading 3.3, steps 19–21). Add 5% (w/v) 99.99% D2O to
the RNA sample (usually, ~0.1–0.3 mM RNA for solution con-
dition optimization) and adjust to the starting pH (~pH 5.5)
using acid and base solutions (see Note 19).
2. Transfer the sample to a standard NMR tube, insert the NMR
tube into the spectrophotometer, and allow the sample to
equilibrate to 10°C for 5 min before the instrument is set up
for data collection.
3. Collect a baseline 1D imino proton spectrum (using 1,1 echo
water suppression) for a low-salt, low-pH (~5.5–6.0), RNA
sample at 10°C; usually, 128 scans are sufficient (see Note 20).
210 C.A. Theimer et al.
3.5. RNA–Protein This section focuses on preparing the protein for investigating RNA–
Interactions by NMR protein interactions. Protein purifications are well-established and
isotopic labeling of proteins has been reviewed extensively (12, 24,
25). Other than the standard issues and concerns for preparing
NMR quality protein samples, there are two issues that also need
to be considered. First, RNAses can bind to or copurify with the
protein of interest and may potentially degrade the RNA during
complex formation. Second, high-affinity cellular RNA can copurify
with the protein. Usually, several different purification steps, such
as affinity tag purification, followed by ion-exchange and size-exclu-
sion chromatography, are sufficient to ensure complete removal of
RNAses (26). If it is also necessary to remove RNA bound to protein,
this can be done through treatment with 0.1–0.5% PEI (27) to
ensure removal of all RNA bound to the protein of interest.
Uniformly 15N-labeled protein is typically obtained by overex-
pression from E. coli in M9 minimal medium and purified using
affinity columns for either His6- or GST-tagged proteins. This step
is often followed by size-exclusion chromatography. Depending on
the purity of the protein, as judged by Coomassie staining of SDS
12 NMR Studies of Protein–RNA Interactions 211
4. Notes
1. All water for RNA studies must be distilled deionized 18.0 MW-cm
quality from a purification system that is properly maintained
(cartridges replaced as suggested by manufacturer). Large bottles
of autoclaved water are kept on hand in the laboratory for
solution making, and all solutions that come into contact with
RNA are sterile filtered using Nalgene sterile filter units with a
0.2-mm pore size prior to storage. Unless otherwise noted, all
solutions should be handled in this manner and “water” in the
12 NMR Studies of Protein–RNA Interactions 213
RNA structures and are often needed for larger RNA structures
to form, completely small RNAs typically fold without magne-
sium, and the NMR samples last longer (degrade less quickly)
if magnesium is excluded from the NMR buffer, whenever
possible.
19. It is very important that, when the pH of RNA samples in
water is being adjusted, you are careful not to overshoot the
desired pH. Alternately adding acid and base to reach the
appropriate pH results unnecessarily increases the salt concen-
tration, which also affects the intensity of the imino proton
resonance peaks.
20. It is important to ensure that the sweep width is wide enough
not to miss any unusual imino proton chemical shifts (proto-
nated cytosine residues, for example). Therefore, we routinely
collect out to 18 ppm initially, although almost all imino proton
peaks are found between ~9 and 15 ppm.
21. PEI can be very difficult to remove from a protein preparation,
as it cannot be dialyzed out effectively. Ammonium sulfate pre-
cipitation is the standard method of separating the protein of
interest from the PEI, with the protein precipitating and the
PEI remaining in the supernatant solution. The precise percent-
age of ammonium sulfate necessary is protein dependent and
the optimal percentage to be used should be tested on small
samples of the protein to ensure maximal protein recovery.
22. Examining the fingerprint region of the free protein spectrum:
If the peaks are well-dispersed with minimal overlap in the center
region of the spectrum, this is an indication that the protein is
well-folded. Each peak in this spectrum represents a unique
amide peak from the protein. If the protein is not well-folded,
it is still worth testing the titration with RNA, as it is possible
that the RNA may induce conformational change of the protein
upon binding.
23. It is best to start with a lower end of the range of protein concen-
trations since formation of RNA:RNA multimers can be an issue
at high concentrations. This titration also gives some idea of
the stoichiometry of RNA:protein complex. If a large amount
of RNA (high RNA:protein ratio) must be added to observe
chemical shift perturbations, this may be an indication that
measuring NOEs between the RNA and protein is difficult.
24. One must keep in mind that it is possible to get shifted reso-
nances (on the protein or the RNA) not only in the binding
site, but also in other regions due to allosteric modulation or,
more likely, changes in the pH or monovalent salt concentra-
tion of the solution. This is why it is particularly important to
try to have both the RNA and protein prepared such that their
final solution conditions are identical.
12 NMR Studies of Protein–RNA Interactions 217
References
1. Autexier, C., and Triki, I. (1999) Tetrahymena (2008) Multiple segmental and selective iso-
telomerase ribonucleoprotein RNA–protein tope labeling of large RNA for NMR structural
interactions. Nucl. Acids Res. 27, 2227–2234. studies. Nucl. Acids Res. 36, e89.
2. Bachand, F., Triki, I., and Autexier, C. (2001) 15. Peterson, R. D., Theimer, C. A., Wu, H., and
Human telomerase RNA–protein interactions. Feigon, J. (2004) New applications of 2D
Nucl. Acids Res. 29, 3385–3393. filtered/edited NOESY for assignment and
3. Greider, C. W., and Blackburn, E. H. (1987) structure elucidation of RNA and RNA–protein
The telomere terminal transferase of Tetrahymena complexes. J. Biomol. NMR 28, 59–67.
is a ribonucleoprotein enzyme with two kinds of 16. Milligan, J. F., Groebe, D. R., Witherell, G. W.,
primer specificity. Cell 51, 887–898. and Uhlenbeck, O. C. (1987) Oligoribo-
4. Staley, J. P., and Woolford, J. L., Jr. (2009) nucleotide synthesis using T7 RNA polymerase
Assembly of ribosomes and spliceosomes: com- and synthetic DNA templates. Nucl. Acids Res.
plex ribonucleoprotein machines. Curr. Opin. 15, 8783–8798.
Cell Biol. 21, 109–118. 17. Cazenave, C., and Uhlenbeck, O. C. (1994)
5. Ban, N., Nissen, P., Hansen, J., Moore, P. B., RNA template-directed RNA synthesis by T7
and Steitz, T. A. (2000) The complete atomic RNA polymerase. Proc. Natl. Acad. Sci. U.S.A.
structure of the large ribosomal subunit at 2.4 91, 6972–6976.
A resolution. Science 289, 905–920. 18. Pleiss, J. A., Derrick, M. L., and Uhlenbeck, O.
6. Cech, T. R. (2000) Structural biology. The ribo- C. (1998) T7 RNA polymerase produces 5¢ end
some is a ribozyme, Science 289, 878–879. heterogeneity during in vitro transcription from
7. Schluenzen, F., Tocilj, A., Zarivach, R., Harms, certain templates. RNA 4, 1313–1317.
J., Gluehmann, M., Janell, D., Bashan, A., 19. Triana-Alonso, F. J., Dabrowski, M., Wadzack, J.,
Bartels, H., Agmon, I., Franceschi, F., and and Nierhaus, K. H. (1995) Self-coded 3¢-exten-
Yonath, A. (2000) Structure of functionally sion of run-off transcripts produces aberrant prod-
activated small ribosomal subunit at 3.3 ang- ucts during in vitro transcription with T7 RNA
stroms resolution. Cell 102, 615–623. polymerase. J. Biol. Chem. 270, 6298–6307.
8. Wimberly, B. T., Brodersen, D. E., Clemons, 20. Kao, C., Rudisser, S., and Zheng, M. (2001) A
W. M., Jr., Morgan-Warren, R. J., Carter, A. P., simple and efficient method to transcribe RNAs
Vonrhein, C., Hartsch, T., and Ramakrishnan, with reduced 3¢ heterogeneity. Methods 23,
V. (2000) Structure of the 30S ribosomal sub- 201–205.
unit. Nature 407, 327–339. 21. Kao, C., Zheng, M., and Rudisser, S. (1999)
9. Ule, J. (2008) Ribonucleoprotein complexes in A simple and efficient method to reduce non-
neurologic diseases. Current Opinion in templated nucleotide addition at the 3 terminus
Neurobiology 18, 516–523. of RNAs transcribed by T7 RNA polymerase.
10. Whittaker, J. W. (2007) Selective isotopic labeling RNA 5, 1268–1272.
of recombinant proteins using amino acid 22. Nacheva, G. A., and Berzal-Herranz, A. (2003)
auxotroph strains. Methods. Mol. Biol. 389, Preventing nondesired RNA-primed RNA
175–188. extension catalyzed by T7 RNA polymerase.
11. Cowburn, D., Shekhtman, A., Xu, R., Ottesen, Eur. J. Biochem./FEBS 270, 1458–1465.
J. J., and Muir, T. W. (2004) Segmental isoto- 23. Cunningham, P. R., and Ofengand, J. (1990)
pic labeling for structural biological applica- Use of inorganic pyrophosphatase to improve
tions of NMR. Methods Mol. Biol. 278, 47–56. the yield of in vitro transcription reactions cata-
12. Liu, D., Xu, R., and Cowburn, D. (2009) lyzed by T7 RNA polymerase. BioTechniques 9,
Segmental isotopic labeling of proteins for 713–714.
nuclear magnetic resonance. Methods Enzymol. 24. Gardner, K. H., and Kay, L. E. (1998) The use
462, 151–175. of 2H, 13C, 15N multidimensional NMR to
13. Lu, K., Miyazaki, Y., and Summers, M. F. (2010) study the structure and dynamics of proteins.
Isotope labeling strategies for NMR studies of Ann. Rev. Biophys. Biomol. Struct. 27, 357–406.
RNA. J. Biomol. NMR 46, 113–125. 25. Marley, J., Lu, M., and Bracken, C. (2001)
14. Nelissen, F. H., van Gammeren, A. J., Tessari, M., A method for efficient isotopic labeling of recom-
Girard, F. C., Heus, H. A., and Wijmenga, S. S. binant proteins. J. Biomol. NMR 20, 71–75.
218 C.A. Theimer et al.
26. Wu, H., Finger, L. D., and Feigon, J. (2005) HIV-1 Rev. Protein Expression and Purification
Structure determination of protein/RNA 63, 112–119.
complexes by NMR. Methods Enzymol. 394, 28. Khanna, M., Wu, H., Johansson, C., Caizergues-
525–545. Ferrer, M., and Feigon, J. (2006) Structural
27. Marenchino, M., Armbruster, D. W., and study of the H/ACA snoRNP components
Hennig, M. (2009) Rapid and efficient purifi- Nop10p and the 3¢ hairpin of U65 snoRNA.
cation of RNA-binding proteins: application to RNA 12, 40–52.
Chapter 13
Abstract
This chapter describes the methods to form and optimize samples of protein–DNA complexes that are
suitable for detailed structure and dynamics studies by NMR spectroscopy.
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_13, © Springer Science+Business Media, LLC 2012
219
220 M.D. Sam and R.T. Clubb
Fig. 1. Flowchart showing the procedures used to form and optimize protein–DNA
complexes for NMR studies.
2. Materials
3. Methods
3.1. Binding Affinity Biochemical assays should be available to rapidly estimate the affinity
Measurements of wild-type and mutant proteins for different DNA molecules.
A variety of methods can be employed to measure binding, such as
the EMSA, isothermal titration calorimetry (ITC), fluorescence
anisotropy, surface plasmon resonance (SPR; e.g., Biacore), and
fluorescence quenching (if the protein contains an appropriately
positioned tryptophan). However, we favor the EMSA because it is
robust and simple to perform (23). This procedure has been
described in detail previously (24) and is outlined below.
1. Mix the following components: 12 mL of 2× binding buffer,
3 mL of 1 mg/mL BSA, 2 mL of 0.5 mg/mL Poly dI/dC (com-
petitor DNA) for a total of 17 mL.
2. Add to this mixture the protein stock solution and the
appropriate amount of H2O to achieve a final volume of 23 mL
(see Subheading 2.1 and Note 2).
3. Incubate on ice for 20 min.
4. Add 1 mL of 32P-labeled DNA probe (~4,000 cpm/mL).
5. Incubate on ice for 20 min.
13 Preparation and Optimization of Protein–DNA… 225
Invert the gel onto FLEX TLC plates with the saran wrap on
top of the TLC plates. Remove the remaining gel plate and in
a darkroom, use a handheld UV lamp (254 nm) to locate the
ssDNA. Quickly excise the DNA using a razor blade. Cut the
excised gel containing the desired ssDNA into ½-in. pieces to
ensure efficient DNA electroelution. The purpose of the
FLEX TLC plate is to enhance the DNA signals during
exposure to UV lights.
6. Assemble an Elutrap™ electroelution device (Whatman).
Membranes for trapping DNA are BT1 (14 nucleotide cutoff)
and BT2 (cellulose acetate membrane). Alternatively, BT1 can be
replaced with a low-molecular-weight cutoff dialysis membrane.
7. Elute the DNA in 1× TBE. Run the Elutrap at 150 V for
8–10 h (remove eluted DNA from the Elutrap two to three
times during this period).
8. Dialyze extensively with dialysis buffer to remove all denatur-
ing reagents from the ssDNA.
9. Determine the DNA concentration from a UV absorbance
reading at 260 nm (A260). The extinction coefficient may be
calculated online using a program provided by IDT (http://
www.idtdna.com/analyzer/Applications/OligoAnalyzer/).
3.4. Preparation of the Our group follows a conservative approach when forming protein–
Protein–DNA Complex DNA complexes to minimize sample loss due to precipitation
(17–22) (see Note 10). In this procedure, dilute concentrations of
13 Preparation and Optimization of Protein–DNA… 227
3.5. Assessing the 1. To assess the quality of the DNA spectrum, record 1D 1H
Quality of the Protein– spectra using a 1331 pulse adjusted to maximally excite the
DNA Complex imino protons (25).
2. Compare the spectra of the complex and the isolated dsDNA
molecule to assess protein binding.
3. If sufficient material is present, use a similar excitation scheme
to record a 2D 1H NOESY spectrum, which reveals whether
the appropriate imino–imino cross peaks are present. Ascertain
the quality of the protein spectrum by comparing the 1H–15N
HSQC spectra of the free and DNA-bound forms of the pro-
tein. The spectrum of the complex should be well-resolved,
exhibit uniform signal intensities, and differ substantially from
the NMR spectrum of the DNA-free protein (Fig. 2).
In favorable cases, it may be possible to observe intermo-
lecular NOEs using protein–DNA complexes containing only
15
N-labeled protein. Depending on the structure of the complex,
intermolecular NOE cross peaks are sometimes observed in
the 2D 1H NOESY spectrum between the imino protons of
Fig. 2. The 1H–15N HSQC spectrum of the Int-DNA complex, which exhibits good dispersion
and uniform line widths. Data for this complex were suitable to determine a high-
resolution NMR structure (17–22).
13 Preparation and Optimization of Protein–DNA… 229
the DNA and protein protons that resonate upfield of 1.2 ppm.
This is because the most upfield resonances in the 1H spectrum
of DNA are the thymine H5 methyl groups (1.2 and 1.6 ppm).
Intermolecular NOEs between the protein amide and DNA
imino protons can sometimes also be seen in the 3D 15N-edited
NOESY spectrum of the 15N-labeled complex. However, a full
assessment of whether a complex is suitable for structure
determination by NMR requires the acquisition of the appro-
priate 2D and 3D edited and filtered NOESY experiments
using samples in which the protein is labeled with 13C and 15N.
4. Notes
Acknowledgments
We thank Dr. Evgeny Fadeev for making Fig. 2. This work was
supported by a grant from the National Institutes of Health to
R.T.C. (R01 AI52217).
References
1. Pabo, C. O., and Sauer, R. T. (1992) solution structure of an Antennapedia homeodo-
Transcription Factors-Structural Families and main-DNA complex. J. Mol. Biol. 234,
Principles of DNA Recognition. Annu. Rev. 1084–1093.
Biochem. 61, 1053–95. 7. Omichinski, J. G., Clore, G. M., Schaad, O.,
2. Garvie, C. W., and Wolberger, C. (2001) Felsenfeld, G., Trainor, C., Appella, E., Stahl,
Recognition of specific DNA sequences. Mol. S. J., and Gronenborn, A. M. (1993) NMR
Cell. 8, 937–946. structure of a specific DNA complex of
3. Nadassy, K., Wodak, S. J., and Janin, J. (1999) Zn-containing DNA binding domain of GATA-1.
Structural features of protein-nucleic acid rec- Science 261, 438–446.
ognition sites. Biochemistry 38, 1999–2017. 8. Kalodimos, C. G., Biris, N., Bonvin, A. M.,
4. Jen-Jacobson, L. (1997) Protein-DNA recog- Levandoski, M. M., Guennuegues, M., Boelens,
nition complexes: conservation of structure R., and Kaptein, R. (2004) Structure and flex-
and binding energy in the transition state. ibility adaptation in nonspecific and specific
Biopolymers 44, 153–180. protein-DNA complexes. Science 305,
5. Boehr, D. D., Nussinov, R., and Wright, P. E. 386–389.
(2009) The role of dynamic conformational 9. Qian, Y. Q., Otting, G., and Wuthrich, K.
ensembles in biomolecular recognition. Nat. (1993) NMR detection of hydration water in
Chem. Biol. 5, 789–796. the intermolecular interface of a protein-DNA
6. Billeter, M., Qian, Y. Q., Otting, G., Muller, M., complex. J. Am. Chem. Soc. 115, 1189–1190.
Gehring, W., and Wuthrich, K. (1993) 10. Iwahara, J., and Clore, G. M. (2006) Direct
Determination of the nuclear magnetic resonance observation of enhanced translocation of a
232 M.D. Sam and R.T. Clubb
homeodomain between DNA cognate sites by 19. Fadeev, E. A., Sam, M. D., and Clubb, R. T.
NMR exchange spectroscopy. J. Am. Chem. (2009) NMR structure of the amino-terminal
Soc. 128, 404–405. domain of the lambda integrase protein in
11. Tjandra, N., and Bax, A. (1997) Direct mea- complex with DNA: immobilization of a
surement of distances and angles in biomolecules flexible tail facilitates beta-sheet recognition
by NMR in a dilute liquid crystalline medium. of the major groove. J. Mol. Biol. 388,
Science 278, 1111–1114. 682–690.
12. Clore, G. M., and Iwahara, J. (2009) Theory, 20. Sam, M. D., Cascio, D., Johnson, R. C., and
practice, and applications of paramagnetic Clubb, R. T. (2004) Crystal structure of the
relaxation enhancement for the characteriza- excisionase-DNA complex from bacteriophage
tion of transient low-population states of bio- lambda. J. Mol. Biol. 338, 229–240.
logical macromolecules and their complexes. 21. Wojciak, J. M., Connolly, K. M., and Clubb, R.
Chem. Rev. 109, 4108–4139. T. (1999) NMR structure of the Tn916 inte-
13. Pervushin, K., Riek, R., Wider, G., and grase-DNA complex. Nature Struct. Biol. 6,
Wuthrich, K. (1997) Attenuated T2 relaxation 366–373.
by mutual cancellation of dipole-dipole cou- 22. Wojciak, J. M., Iwahara, J., and Clubb, R. T.
pling and chemical shift anisotropy indicates an (2001) The Mu repressor-DNA complex con-
avenue to NMR structures of very large bio- tains an immobilized “wing” within the minor
logical macromolecules in solution. Proc. Natl. groove. Nature Struct. Biol. 8, 84–90.
Acad. Sci. USA 94, 12366–12371. 23. Buratowski, S., and Chodosh, L. A. (2001)
14. Cavanagh, J., Fairbrother, W. J., Palmer, A. G. Mobility shift DNA-binding assay using gel
I., Rance, M., and Skelton, N. J. (2006) Protein electrophoresis. Curr. Protoc. Mol. Biol.,
NMR Spectroscopy: Principles & Practice (2nd Chapter 12, Unit 12 2.
ed.), Academic Press, San Diego. 24. Taylor, J. D., Ackroyd, A. J., and Halford, S. E.
15. Iwahara, J., Wojciak, J. M., and Clubb, R. T. (1994) The gel shift assay for the analysis of
(2001) Improved NMR spectra of a protein- DNA-protein interactions, in DNA-protein
DNA complex through rational mutagenesis interactions, principles and protocols (Kneale, G.
and the application of a sensitivity optimized G., Ed.), Humana Press, Totowa, NJ.
isotope-filtered NOESY experiment. J. Biomol. 25. Hore, P. J. (1983) A new method for water
NMR 19, 231–241. suppression in the proton NMR spectra of
16. Sambrook, J., Fritsch, E. F., and Maniatis, T. aqueous solutions. J. Magn. Reson. 54,
(1989) Molecular Cloning, A laboratory man- 539–542.
ual, 2nd ed., Cold Spring Harbor Laboratory 26. Louis, J. M., Martin, R. G., Clore, G. M., and
Press. Gronenborn, A. M. (1998) Preparation of uni-
17. Abbani, M. A., Papagiannis, C. V., Sam, M. D., formly isotope-labeled DNA oligonucleotides
Cascio, D., Johnson, R. C., and Clubb, R. T. for NMR spectroscopy. J. Biol. Chem. 273,
(2007) Structure of the cooperative Xis-DNA 2374–2378.
complex reveals a micronucleoprotein filament 27. Xiong, A. S., Yao, Q. H., Peng, R. H., Duan,
that regulates phage lambda intasome assem- H., Li, X., Fan, H. Q., Cheng, Z. M., and Li,
bly. Proc. Natl. Acad. Sci. USA 104, Y. (2006) PCR-based accurate synthesis of
2109–2114. long DNA sequences. Nat Protoc. 1,
18. Iwahara, J., Iwahara, M., Daughdrill, G. W., 791–797.
Ford, J., and Clubb, R. T. (2002) The struc- 28. Andrus, A., and Kuimelis, R. G. (2001) Analysis
ture of the Dead ringer-DNA complex reveals and purification of synthetic nucleic acids using
how AT-rich interaction domains (ARIDs) rec- HPLC. Curr. Protoc. Nucleic Acid Chem.,
ognize DNA. EMBO J. 21, 1197–1209. Chapter 10, Unit 10 5.
Chapter 14
Abstract
Nuclear magnetic resonance (NMR) has evolved into a powerful tool for characterizing protein–ligand
interactions in solution under near physiological conditions. It is now frequently harnessed to assess the
affinity and specificity of interactions; to identify binding epitopes on proteins and ligands; and to charac-
terize the structural rearrangements induced by binding.
The first section of this chapter provides a general overview of the NMR study of protein–ligand
interactions. The section is divided according to two main categories of experiments: those based on
observing protein signals and those based on observing ligand signals. The next section explains two case
studies performed in the authors’ laboratory. The first of these deals with the interaction between vascular
endothelial growth factor and a peptidic ligand, and includes a detailed protocol of chemical shift
perturbation experiments. The second one reports on the interaction between prolyl oligopeptidase and a
small molecule as monitored by ligand saturation transfer difference (STD), and illustrates how NMR can
be used to confirm binding and to identify the binding epitope of a ligand.
Key words: Protein–ligand interactions, Chemical shift perturbation, Saturation transfer difference,
NMR, Vascular endothelial growth factor, Prolyl oligopeptidase, Protein observed experiments,
Ligand observed experiments
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_14, © Springer Science+Business Media, LLC 2012
233
234 M. Goldflam et al.
1.1. Protein Observed Although in some cases monodimensional (1D) 1H-NMR experi-
Experiments ments have been used to characterize the protein–ligand binding
by following the 1H chemical shifts of specific residues in the pro-
tein, most experiments on protein observation entail bidimensional
(2D) NMR. Conventional 1D-1H spectra typically cannot resolve
the individual proton signals of the protein. This limitation can be
overcome by distributing the information along two dimensions
and by employing heteronuclear spectroscopy (i.e., studying mag-
netically active nuclei other than protons that are present in pro-
teins, such as 15N and 13C). Since the natural abundance of 15N and
13
C (0.37% and 1.1%, respectively) is too low for NMR experi-
ments, the protein to be studied must be isotopically labeled,
usually, through expression in E. coli. Several efficient labeling
schemes are available, and choosing the right one can greatly simplify
NMR studies of proteins.
The most widely used labeling method is uniform labeling with
15
N. For proteins expressed recombinantly in E. coli, 15N (in the
form of an ammonium salt) is added to the expression media. This
simple modification provides near quantitative isotopic labeling of
the protein: all backbone amides as well the nitrogen containing
side chains are labeled with this magnetically active nucleus.
Heteronuclear 1H-15N correlation NMR experiments that allow
236 M. Goldflam et al.
Fig. 1. (a) Basic pulse sequence for the 1H-15N HSQC experiment. The narrow and wide
bars depict 90° and 180° pulses, respectively. The delays (t ; equal to 1/[41JHN]) allow
magnetization evolution to be transferred between coupled nuclei. 15N magnetization
evolves for t1 and 1H magnetization is directly detected during t2. Double Fourier transfor-
mation along t1 and t2 generates a 2D correlation spectrum with frequencies F1 and F2,
respectively, as shown in (b) Every signal in the spectrum corresponds to one NH group in
the protein and gives information on the chemical shifts of an amide nitrogen (F1) and an
amide proton (F2) that are directly coupled through the coupling constant 1JHN.
1.2. Ligand Observed All ligand observation experiments are based on the difference in
Experiments NMR parameters between the bound and free states of the ligand.
The changes in nuclear Overhauser effects (NOEs) when ligands
bind to receptor proteins are especially interesting (22). Ligands
with molecular weight lower than 1,000 U exhibit short correla-
tion times (τc) and show only weak positive NOEs, very small neg-
ative NOEs, or no NOEs at all, depending on the magnetic field
strength and the molecular weight. Proteins, due to their size,
show large τc, large negative NOEs, and highly efficient spin
diffusion. Upon binding, the ligand forms a high molecular weight
complex with the protein; consequently its properties change,
especially its NOE behavior, with the appearance of strong nega-
tive NOEs, usually called transferred NOEs (trNOEs). The differ-
ence in the properties between the bound and free ligand is as
large as the difference in molecular weight between the ligand and
the protein–ligand complex. Since these differences have a direct
impact on the observable NMR parameters of the ligand, several
experiments can be used to detect and characterize the binding
event. Most ligand observed methods are based on one of the
following: assessment of changes in conventional NMR parameters
240 M. Goldflam et al.
Fig. 2. Basic pulse sequence for an STD experiment. Selective saturation via a train of N
selective 90° pulses separated by a delay δ is performed during block A. Protein signals
are suppressed during an R2 relaxation filtering delay (block B). After a module for water
suppression (block C), the 1H signal is detected during the FID (block D).
whereby
I STD
STD Ampl. = e × h = e × (2)
I0
STDAmpl. corresponds to the STD amplification factor (28) and is a
correction for total ligand concentration to the STD effect; ISTD is
the intensity of an individual proton in the STD spectrum, and I0,
intensity of the same proton in the reference spectrum; e is the
ligand excess; h is the fraction of ISTD from I0; STDmax is the maximal
STD intensity achievable with long saturation times and corre-
sponds to the STD intensity in the absence of T1 bias.
Finally, STD experiments can be used to determine dissociation
constants if a titration curve is recorded with varying ligand
concentrations at the same saturation time (30). To do this, the
244 M. Goldflam et al.
1.3. Ligand Versus Ligand-based and protein-based approaches have distinct advantages
Protein Observed and disadvantages. The former yield information about the strength
Experiments of the interaction, the binding epitope, and the conformation of
the ligand. They can be used to simultaneously screen several
compounds for their ability to bind to a protein of interest. Ligand-
based experiments have simple requirements. First, they do not
require isotopically labeled protein. Secondly, there is no upper
limit for protein size, but the difference between protein and ligand
has to be substantial enough to result in differential relaxation
behaviors. Contrariwise, protein observation experiments are cur-
rently only feasible for proteins weighing ca. 40 kDa or less.
Moreover, they demand considerable amounts of isotopically
labeled protein, which must be stable at high concentrations for
long periods of time. When signal assignment is available, protein-
based experiments may provide more information. Most impor-
tantly, they can be used to identify one or several binding sites of
the ligand on the protein and indicate zones of structural rear-
rangement. Furthermore, in these experiments, formation of ligand
aggregates cannot be misinterpreted as an interaction. However,
the data in protein observation experiments are easiest to interpret
when the binding site has been saturated. In the case of low affinity
ligands, this point may be beyond the limit of solubility. In conclu-
sion, there are cases for which one approach is better suited than
the other. Nevertheless, the full power of NMR to characterize
protein–ligand interactions can only be exploited if both approaches
are combined. Several examples from academic and industrial drug
discovery projects are testament to the great success of a combined
approach (7, 8, 33, 34).
2. Materials
3. Methods
3.1. Protein-Based The following case study describes NMR monitoring of the inter-
Study on the Binding action between VEGF and the peptidic ligand P-7i, presented here
of VEGF to the Peptidic as a representative example of a CSP experiment (3). The 23-kDa
Ligand P-7i VEGF11–109 construct used is a truncated version of VEGF121 which
exhibits excellent solubility and stability. Moreover, it is readily
labeled with 15N and is a symmetric homodimer, making it highly
suited to protein-based NMR experiments. Additionally, the signal
assignment for this construct is available (35). P-7i is a 19 amino
acid-long analog of v107 that was discovered by phage display
(40). It differs from v107 by a single mutation: the Ile-7 is D rather
than L, which translates to a reduced affinity for VEGF (252 μM)
compared to the wild type (1.0 μM). Although P-7i is relatively
large (MW = ca. 2.3 kDa) for NMR studies of protein–ligand
binding, it was selected as an example because it exhibits fast and
intermediate exchange behavior and is amenable to CSP studies.
3.1.1. Sample Preparation 1. Starting from a stock solution of P-7i in water, prepare six
aliquots (two at each of three concentrations) with a total
amount of ligand corresponding to 50, 100, or 200 μM in a
volume of 160 μL (19, 37, or 75 μg, respectively). Freeze the
aliquots in 1.5-mL Eppendorf tubes and lyophilize (Fig. 3).
14 NMR Studies of Protein–Ligand Interactions 247
3.1.2. Recording of CSP 1. Equilibrate the VEGF sample inside the NMR spectrometer
NMR Spectra for 15 min at 318 K before starting the NMR spectra acquisi-
tion. This temperature is chosen because the available signal
assignment was performed at this temperature (35). Execute
standard NMR procedures: tune the probe, shim the magnetic
field, calibrate the length of the 90° pulses for 1H and 15N, and
optimize the water suppression.
2. Record a 1H-15N HSQC spectrum using the FAST-HSQC
experiment (41), which uses pulsed field gradients and a
WATERGATE (42) module to efficiently suppress the water
signal. For this, 2,048 × 256 complex points with a total of
8 transients per increment are used. The total experiment time
is ca. 40 min.
3. Once the spectrum is obtained, remove the sample from the
spectrometer and transfer it to the first Eppendorf tube contain-
ing the lyophilized ligand (see Fig. 3). Add 0.5% of DMSO-d6
to ensure ligand solubility (see Note 6), vortex the sample and
centrifuge (5,000 RCF, 1 min, RT), and transfer all of the liquid
to the previously used NMR tube. Introduce the resulting sam-
ple, containing the protein and the ligand at the first titration
concentration, into the spectrometer. Equilibrate the sample at
318 K for 15 min, re-shim the magnetic field and record a new
spectrum using the same conditions described in step 2.
4. Repeat step 3 until spectra are recorded for each titration point.
By dissolving the lyophilized ligand in the sample incremen-
tally, the concentration of P-7i increases stepwise over the
course of the titration from 0 to 50 μM, 100, 200 , 300, 500,
and finally, 700 μM (Fig. 3).
Fig. 3. Stepwise addition of P-7i to VEGF: overview of lyophilized ligand aliquots, the respective concentration increments
and the total concentration over the course of titration. After the reference experiment on the sample containing only protein,
the ligand concentration was increased stepwise by transferring the protein sample to Eppendorf tubes with the denoted
amounts of lyophilized P-7i prior to acquisition of the next NMR spectrum. A total of seven spectra were acquired by
repeating this procedure: one reference spectrum, plus one spectrum at each of the six titration points.
248 M. Goldflam et al.
3.1.3. Data Analysis 1. To process the HSQC spectra, increase the number of points
in the indirect dimension (F1) from 256 to 512 by linear
prediction and then zero fill to 1,024 points to yield a
2,048 × 1,024 matrix. Adjust the phase correction manually
and apply a squared sine weighting function in both dimen-
sions. Process all the spectra acquired in the titration experi-
ment identically using Topspin 2.0. Figure 4 shows the seven
superimposed spectra acquired from the titration of VEGF
with P-7i.
2. Determination of CSP requires peak picking and subsequent
assignment of the peaks to the corresponding residues. Use
the program Cara to do this for the first and last spectra of
the titration (see Note 7).
3. Extract the relevant data for mapping the binding site by
calculating the distance between the position of the reference
Fig. 4. (a) Seven superimposed 1H-15N HSQC spectra of a 100-μM sample of uniformly-15N labeled VEGF11–109 at 318 K
titrated with 0–700 μM of P-7i (600 MHz with cryoprobe). (b) Zoom of Lys48 shifts (reproduced from ref. 3 with kind per-
mission from Wiley).
14 NMR Studies of Protein–Ligand Interactions 249
⎛ 2 ⎛ Δδ N ⎞ 2 ⎞
Δδ NH = ⎜ Δδ H +⎜ (4)
⎝ ⎝ 5 ⎠⎟ ⎟⎠
Fig. 5. Histogram of the P-7i induced CSP of the backbone amides Δd for every residue of
VEGF11–109 observed in the 1H-15N HSQC experiment. The histogram was calculated for the
shifts between the reference spectrum (VEGF alone) and the spectrum of the sample with
the highest ligand concentration (VEGF plus 700 μM of P-7i). The lower (dashed ) horizon-
tal line represents the mean shift, and the upper (dotted ) horizontal line, the cutoff for
significant changes (mean shift plus one standard derivation).
250 M. Goldflam et al.
[L ]− [L]
Δδ NH = F × 0
[P ] 0
(5)
K + [L ]+ 2 [P ]− (K + [L ]) + 4 [P ](K − [L ]+ [P ])
2
D 0 0 D 0 0 D 0 0
=F×
2 [P ] 0
3.2. Ligand-Based The following case study was chosen to give a step-by-step expla-
Study on Binding nation of an STD experiment. The example is based on work
of POP to the Ligand performed in the authors’ laboratory to characterize the interac-
Baicalin tion between the protease POP and the flavonoid baicalin (43).
14 NMR Studies of Protein–Ligand Interactions 251
Fig. 7. (a) Relative amount of bound ligand plotted against the total ligand concentration
and fitted to a model of two independent identical binding sites. Analysis is based on
Lys48 peak displacement in the spectra of the VEGF-(P-7i) complex. (b) Results of fitting
for the four selected residues (reproduced from ref. 3 with kind permission from Wiley).
3.2.1. Optimization 1. Equilibrate the sample containing only POP in the NMR
of STD Parameters spectrometer to 308 K (see Note 11) for 15 min. Follow the
and Confirmation of standard procedure: tune the probe, shim the magnetic field,
Baicalin as a POP Ligand calibrate the length of the 90° pulse, and optimize the water
suppression (see Note 12).
2. To optimize the protein saturation, record a 1H-spectrum of
POP to identify promising signals in the aliphatic region.
The aliphatic region of POP shows a signal at ca. 0.9 ppm
that decreases in intensity in the up-field direction without
occurrence of any new maxima. Therefore, the closer to this
value the protein is irradiated, the greater the saturation (see
Note 13).
3. To determine if protein irradiation is affecting any of the
ligand’s signals, acquire several STD spectra of a sample of
baicalin alone, using values of 0, −1, and −2 ppm for the
on-resonance frequency. No STD signals are observed in any
case. This negative control experiment confirms that no
direct irradiation of baicalin or baicalin-aggregates occurs.
252 M. Goldflam et al.
Fig. 8. (a) 1H STD spectrum of baicalin (500 μM) in the presence of POP (10 μM) recorded at 600 MHz and 308 K.
The protein signals were suppressed by applying a spin lock filter. (b) 1H-reference spectrum of baicalin. *Signals arising
from sample impurities (reproduced from ref. 43 with kind permission from Elsevier).
14 NMR Studies of Protein–Ligand Interactions 253
Fig. 9. (a) Saturation transfer difference amplification for individual protons at different saturation times. The data were
acquired with a ninefold excess of baicalin over POP (20 μM). (b) Build-up curves obtained for individual protons in baicalin.
The key to the nomenclature for the individual protons is provided in Fig. 10 (adjusted from ref. 43 with kind permission
from Elsevier).
Fig. 10. The STD data were fitted to a monoexponential equation, from which the STDmax and the saturation rate constant
(ksat) were obtained. The initial slope directly correlates to the proximity of the corresponding proton to the protein and is
the product of STDmax and ksat. The relative STDs were calculated by setting the proton with the greatest STD effect to 100%
(adjusted from ref. 43 with kind permission from Elsevier).
4. Notes
References
1. Wiseman, T., Williston, S., Brandts, J. F., and 11. Tugarinov, V., and Kay, L. E. (2005) Methyl
Lin, L. N. (1989) Rapid measurement of groups as probes of structure and dynamics in
binding constants and heats of binding using a NMR studies of high-molecular-weight
new titration calorimeter. Anal. Biochem. 179, proteins. Chembiochem. 6, 1567–1577.
131–137. 12. Bodenhausen, G., and Ruben, D. J. (1980)
2. Baldwin, M. A. (2005) Mass spectrometers for Natural abundance nitrogen-15 NMR by
the analysis of biomolecules. Methods Enzymol. enhanced heteronuclear spectroscopy. Chemical
402, 3–48. Physics Letters 69, 185–189.
3. Dyachenko, A., Goldflam, M., Vilaseca, M., 13. Kay, L., Keifer, P., and Saarinen, T. (1992) Pure
and Giralt, E. (2010) Molecular recognition at absorption gradient enhanced heteronuclear
protein surface in solution and gas phase: Five single quantum correlation spectroscopy with
VEGF peptidic ligands show inverse affinity improved sensitivity. J. Am. Chem. Soc. 114,
when studied by NMR and CID-MS. 10663–10665.
Biopolymers 94, 689–700. 14. Morris, G. A., and Freeman, R. (1979)
4. Englebienne, P., Hoonacker, A. V., and Verhas, Enhancement of nuclear magnetic resonance
M. (2003) Surface plasmon resonance: princi- signals by polarization transfer. J. Am. Chem.
ples, methods and applications in biomedical Soc. 101, 760–762.
sciences. Spectroscopy 17, 255–273. 15. Gang, Z., and Price, W. S. Solvent signal
5. Fielding, L. (2003) NMR methods for the suppression in NMR. Prog. Nucl. Magn. Reson.
determination of protein-ligand dissociation Spectrosc. 56, 267–288.
constants. Curr. Top. Med. Chem. 3, 39–53. 16. Shuker, S. B., Hajduk, P. J., Meadows, R. P.,
6. Carlomagno, T. (2005) Ligand-target interac- and Fesik, S. W. (1996) Discovering High-
tions: what can we learn from NMR? Annu. Affinity Ligands for Proteins: SAR by NMR.
Rev. Biophys. Biomo.l Struct. 34, 245–266. Science 274, 1531–1534.
7. Lepre, C. A., Moore, J. M., and Peng, J. W. 17. Smrcka, A. V., Kichik, N., Tarrago, T.,
(2004) Theory and applications of NMR-based Burroughs, M., Park, M. S., Itoga, N. K.,
screening in pharmaceutical research. Chem. Stern, H. A., Willardson, B. M., and Giralt, E.
Rev. 104, 3641–3676. (2010) NMR analysis of G-protein betagamma
8. Dalvit, C. (2009) NMR methods in fragment subunit complexes reveals a dynamic G(alpha)-
screening: theory and a comparison with other Gbetagamma subunit interface and multiple
biophysical techniques. Drug Discov. Today 14, protein recognition modes. Proc. Natl. Acad.
1051–1057. Sci. USA 107, 639–644.
9. Tarrago, T., Claasen, B., Kichik, N., Rodriguez- 18. Pellecchia, M. (2005) Solution nuclear mag-
Mias, R. A., Gairi, M., and Giralt, E. (2009) A netic resonance spectroscopy techniques for
cost-effective labeling strategy for the NMR probing intermolecular interactions. Chem.
study of large proteins: selective 15 N-labeling Biol. 12, 961–971.
of the tryptophan side chains of prolyl oligo- 19. Reibarkh, M., Malia, T. J., and Wagner, G.
peptidase. Chembiochem. 10, 2736–2739. (2006) NMR distinction of single- and
10. Bogan, A. A., and Thorn, K. S. (1998) Anatomy multiple-mode binding of small-molecule
of hot spots in protein interfaces. J. Mol. Biol. protein ligands. J. Am. Chem. Soc. 128,
280, 1–9. 2160–2161.
258 M. Goldflam et al.
20. Medek, A., Hajduk, P., Mack, J., and Fesik, S. affinity to proteins via magnetization transfer
(2000) The Use of Differential Chemical Shifts from bulk water. J. Biomol. NMR 18,
for Determining the Binding Site Location and 65–68.
Orientation of Protein-Bound Ligands. J. Am. 32. Dalvit, C., Fogliatto, G., Stewart, A., Veronesi,
Chem. Soc. 122, 1241–1242. M., and Stockman, B. (2001) WaterLOGSY as
21. Krishnamoorthy, J., Yu, V. C., and Mok, Y. K. a method for primary NMR screening: Practical
(2010) Auto-FACE: an NMR based binding aspects and range of applicability. J. Biomol.
site mapping program for fast chemical NMR 21, 349–359.
exchange protein-ligand systems. PLoS One 5, 33. Pellecchia, M., Sem, D. S., and Wuthrich, K.
e8943. (2002) NMR in drug discovery. Nat. Rev.
22. Neuhaus, D., and Williamson, M. P. (2000) Drug Discov. 1, 211–219.
The Nuclear Overhauser Effect in Structural 34. Oltersdorf, T., Elmore, S. W., Shoemaker, A.
and Conformational Analysis, 2nd Edition ed., R., Armstrong, R. C., Augeri, D. J., Belli, B.
Wiley, New York. A., Bruncko, M., Deckwerth, T. L., Dinges, J.,
23. Meiboom, S., and Gill, D. (1958) Modified Hajduk, P. J., Joseph, M. K., Kitada, S.,
Spin-Echo Method for Measuring Nuclear Korsmeyer, S. J., Kunzer, A. R., Letai, A., Li,
Relaxation Times, Review of Scientific C., Mitten, M. J., Nettesheim, D. G., Ng, S.,
Instruments 29, 688–691. Nimmer, P. M., O’Connor, J. M., Oleksijew,
24. Ni, F., and Scheraga, H. A. (1994) Use of the A., Petros, A. M., Reed, J. C., Shen, W., Tahir,
Transferred Nuclear Overhauser Effect To S. K., Thompson, C. B., Tomaselli, K. J., Wang,
Determine the Conformations of Ligands Bound B., Wendt, M. D., Zhang, H., Fesik, S. W., and
to Proteins. Accts. Chem. Res. 27, 257–264. Rosenberg, S. H. (2005) An inhibitor of Bcl-2
25. Moriz, M., and Bernd, M. (1999) family proteins induces regression of solid
Characterization of Ligand Binding by Saturation tumours. Nature 435, 677–681.
Transfer Difference NMR Spectroscopy. Angew. 35. Fairbrother, W. J., Champe, M. A., Christinger,
Chem. Int. Ed. Engl. 38, 1784–1788. H. W., Keyt, B. A., and Starovasnik, M. A.
26. Klein, J., Meinecke, R., Mayer, M., and Meyer, (1997) 1H, 13C, and 15N backbone assign-
B. (1999) Detecting Binding Affinity to ment and secondary structure of the receptor-
Immobilized Receptor Proteins in Compound binding domain of vascular endothelial growth
Libraries by HR-MAS STD NMR. J. Am. factor. Protein Sci. 6, 2250–2260.
Chem. Soc. 121, 5336–5337. 36. Bruker Corporation, (2007) Topspin 2.0,
27. Groves, P., Kover, K. E., Andre, S., http://www.bruker-biospin.com/nmr_soft-
Bandorowicz-Pikula, J., Batta, G., Bruix, M., ware.html.
Buchet, R., Canales, A., Canada, F. J., Gabius, 37. Keller, R. (2004) The Computer Aided
H. J., Laurents, D. V., Naranjo, J. R., Resonance Assignment Tutorial, 1st edition ed.,
Palczewska, M., Pikula, S., Rial, E., Strzelecka- CANTINA Verlag.
Kiliszek, A., and Jimenez-Barbero, J. (2007) 38. Origin Corporation, (2007) Origin 8.0,
Temperature dependence of ligand-protein http://www.originlab.com/.
complex formation as reflected by saturation 39. Chemical computing group, (2009) http://
transfer difference NMR experiments. Magn. www.chemcomp.com/index.htm.
Reson. Chem. 45, 745–748. 40. Pan, B., Li, B., Russell, S. J., Tom, J. Y.,
28. Mayer, M., and Meyer, B. (2001) Group Cochran, A. G., and Fairbrother, W. J. (2002)
epitope mapping by saturation transfer differ- Solution structure of a phage-derived peptide
ence NMR to identify segments of a ligand in antagonist in complex with vascular endothe-
direct contact with a protein receptor. J. Am. lial growth factor. J. Mol. Biol. 316,
Chem. Soc. 123, 6108–6117. 769–787.
29. Mayer, M., and James, T. L. (2004) NMR- 41. Mori, S., Abeygunawardana, C., Johnson, M.
based characterization of phenothiazines as a O., and van Zijl, P. C. (1995) Improved sensi-
RNA binding scaffold. J. Am. Chem. Soc. 126, tivity of HSQC spectra of exchanging protons
4453–4460. at short interscan delays using a new fast
30. Meyer, B., and Peters, T. (2003) NMR HSQC (FHSQC) detection scheme that
spectroscopy techniques for screening and avoids water saturation. J. Magn. Reson. B
identifying ligand binding to protein receptors. 108, 94–98.
Angew. Chem. Int. Ed. Engl. 42, 864–890. 42. Piotto, M., Saudek, V., and Sklenář, V. (1992)
31. Dalvit, C., Pevarello, P., Tato, M., Veronesi, Gradient-tailored excitation for single-quantum
M., Vulpetti, A., and Sundstrom, M. (2000) NMR spectroscopy of aqueous solutions.
Identification of compounds with binding J. Biomol. NMR 2, 661–665.
14 NMR Studies of Protein–Ligand Interactions 259
43. Tarrago, T., Kichik, N., Claasen, B., Prades, R., 45. Schumann, F. H., Riepl, H., Maurer, T.,
Teixido, M., and Giralt, E. (2008) Baicalin, a Gronwald, W., Neidig, K. P., and Kalbitzer, H. R.
prodrug able to reach the CNS, is a prolyl (2007) Combined chemical shift changes and
oligopeptidase inhibitor. Bioorg. Med. Chem. amino acid specific chemical shift mapping of
16, 7516–7524. protein-protein interactions. J. Biomol. NMR
44. Voehler, M. W., Collier, G., Young, J. K., 39, 275–289.
Stone, M. P., and Germann, M. W. (2006) 46. Hwang, T. L., and Shaka, A. J. (1995) Water
Performance of cryogenic probes as a function suppression that works. Excitation sculpting
of ionic strength and sample tube geometry. using arbitrary waveforms and pulsed field gra-
J. Magn. Reson. 183, 102–109. dients. J. Magn. Reson. 112, 275–279.
Chapter 15
Abstract
A living cell is a complex system that contains many biological macromolecules and small molecules necessary
for survival, in a relatively small volume. It is within this crowded and complex cellular environment that
proteins function making in-cell studies of protein structure and binding interactions an exciting and
important area of study. Nuclear magnetic resonance (NMR) spectroscopy is a particularly attractive
method for in-cell studies of proteins since it provides atomic-level data noninvasively in solution. In addition,
NMR has recently undergone significant advances in instrumentation to increase sensitivity and in methods
development to reduce data acquisition times for multidimensional experiments. Thus, NMR spectroscopy
lends itself to studying proteins within a living cell, and recently “in-cell NMR” studies have been reported
from several laboratories. To date, this technique has been successfully applied in Escherichia coli (E. coli),
Xenopus laevis (X. laevis) oocytes, and HeLa host cells. Demonstrated applications include protein assignment
as well as de novo 3D protein structure determination. The most common use, however, is to probe binding
interactions and structural modifications directly from proton nitrogen correlation spectra. E. coli is the
most extensively used cell type thus far and this chapter is largely confined to reviewing recent literature
and describing methods and detailed protocols for in-cell NMR studies in this bacterial cell.
Key words: In-cell NMR spectroscopy, Protein NMR spectroscopy, Escherichia coli, Fast NMR
spectroscopy
1. Introduction
1.1. In-Cell NMR Within a cell, there are many different proteins, other biological
Spectroscopy: macromolecules, and small molecules that must function properly
The Technique for a cell to survive. The simplest organisms are estimated to utilize
and Information a few hundred types of small molecules and genomically encode up
It Reveals to 1,000 different proteins (1, 2), with the human genome estimated
to encode 10–100-fold more (3). The large number and diversity
of small molecules and biological macromolecules form a cellular
environment, where these molecules function, that is very crowded
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_15, © Springer Science+Business Media, LLC 2012
261
262 K.E. Robinson et al.
1.2. In-Cell NMR In-cell NMR spectroscopy within E. coli has been used to determine
Spectroscopy: the 3-D backbone assignment of GB1 (8), determine the 3-D
Current Achievements de novo structure of the gene product TTHA1718 (9, 10), as well
and Challenges as to study protein–protein interactions (7), protein–DNA interac-
tions (11), binding events (12), and identify potential new drugs
(13). These studies have all been performed using 15N and/or 13C
isotopic labeling of proteins under 20 kDa. Both uniform labeling
and selective methyl-group labeling strategies have been applied
(14, 15). Labeling schemes typically use isotope 15N enrichment for
2-D fingerprint data collection and also incorporate uniform 13C
labeling when collecting 3-D data for assignment and structure
determination. The 1H–13C HSQC fingerprint can also be useful,
but is often complicated by significant background signals that arise
from metabolites within the cell incorporating the 13C isotope (15).
However, it has been demonstrated that utilizing specific 13C methyl-
group labeling can be beneficial for observing larger proteins that
15 In-Cell NMR Spectroscopy in Escherichia coli 263
110.0
115.0
120.0
N (ppm)
125.0
130.0
135.0
Fig. 1. 1H–15N HSQC spectrum from E. coli cells grown in isotopically labeled medium without protein overexpression.
The signals detected here are referred to as the background signal for in-cell NMR spectroscopy.
1.3. In-Cell NMR The studies cited above illustrate that in-cell NMR utilizing E. coli
Spectroscopy: enables the study of proteins within a physiologically relevant envi-
Looking Ahead ronment. However, the technique requires that the protein of
interest is expressed to intracellular concentrations sufficient for
NMR detection that are greater than the concentrations of most
cellular proteins. Thus, overexpression generates a labeled protein
in the cellular environment but is not strictly in vivo. Ideally, the
protein concentration should be tightly controlled to understand
what effects a high concentration may cause within a living cell as
well as potentially lower the concentration while remaining within
the detection limits of NMR. It has been demonstrated that protein
expression levels for in-cell NMR can be controlled with tunable
promoters, such as the arabinose promoter (7). External delivery
of isotopically labeled protein can provide another strategy to control
concentrations of labeled proteins in cells; however, to date, this
has only been reported using Xenopus laevis oocytes or HeLa cells
(39, 40) (Fig. 2). The overexpression protocol is a useful strategy
because the protein of interest is never exposed to a noncellular
environment; however, E. coli lack much of the complexity found
within eukaryotic organisms as they are not capable of performing
most posttranslational modifications and do not have organelles.
Therefore, it is also of interest to develop complementary methods
to overexpress proteins within a eukaryotic organism, such as yeast
or insect cells, so that the protein of interest is always exposed to
268 K.E. Robinson et al.
Target protein
Cell-penetrating
peptide
Injection needle
E. coli
HeLa
Fig. 2. Methods for in-cell NMR spectroscopy. E. coli cells are used for overexpressing
protein within the cell of interest. Proteins are difficult to overexpress in eukaroptic cells;
therefore, currently, two protocols are used. For X. laevis oocytes, physical injection has
been used while for HeLa cells cell-penetrating peptides have been used.
2. Materials
2.1. Protein Expression 1. E. coli BL21 DE3 cells containing the expression plasmid (see
Note 1).
2. Luria-Broth (LB) Medium: 10 g NaCl, 10 g Bacto-Tryptone,
5 g yeast extract in 1 L dIH2O. Autoclave to sterilize and store
at 25°C.
3. LB agar selection plates: LB medium with 20 g Bacto-Agar in
1 L dIH2O. Autoclave to sterilize. Add the desired antibiotic
after the medium has cooled to 50°C. Store plates at 4°C.
4. 5× M9 Salts: 64 g Na2HPO4, 15 g KH2PO4, 2.5 g NaCl, 5.0 g
NH4Cl in 1 L dIH2O. Autoclave to sterilize. To incorporate
15
N, use 15NH4Cl; final pH is 7.2. Store at 25°C.
5. 1,000× trace metal mix: 2 mM H3BO3, 2 mM CuSO4, 2 mM
CoCl2, 10 mM MnCl2, 2 mM NiSO4, 2 mM (NH4)6Mo7O24.
Filter sterilize and store at 25°C.
6. M9 minimal medium: 1× M9 salts, 2 mM MgSO4, 1 mM FeCl3,
25 mM ZnSO4, 100 mM CaCl2, 1× trace metal mix, 0.0005%
thiamine, 0.3% (w/v) glucose (see Note 5). To incorporate
13
C, use 0.3% (w/v) 13C glucose.
15 In-Cell NMR Spectroscopy in Escherichia coli 269
Table 1
Induction Agent Composition
Table 2
Selection Agent Composition
3. Methods
3.1. Protein Expression The typical strategy for studying a protein using in-cell NMR
for Preparing an spectroscopy entails incorporating 15N uniform isotope labeling
In-Cell NMR Sample and collecting a 2-D 1H–15N HSQC. Uniform 13C labeling for the
1
H–13C HSQC experiment is rarely done as the background signal
detected is quite high. However, selective labeling schemes to
incorporate 13C isotope into methyl groups can be beneficial,
particularly for larger proteins, and significantly reduce the back-
ground signal detected as very few metabolites incorporate the 13C
label introduced by this strategy. An alternative to labeling methyl
groups with 13C isotope is to use 19F trifluoromethyl groups, but
this procedure has limitations both in protein production and the
spectral information produced.
If the target protein yields a successful in-cell HSQC spectrum,
then higher dimensional experiments and/or binding studies can be
designed and implemented. Binding studies require tight regulation
over the delivery and concentration of the type of binding partner
to be studied. For protein–protein interaction studies, it is desirable
to be able to detect both proteins individually and to have independent
control over expressing or delivering both proteins to the same cell
(see Note 6). The experimental design depends on the goal of the
experiment; however, all in-cell NMR experiments have a similar
basic protocol as described below (Fig. 3).
1. Using a frozen glycerol stock or single LB selection plate colony,
inoculate 5 mL of LB medium with the appropriate antibiotic
(Table 2). Incubate the culture for 12–16 h at 37°C in a shaking
incubator. The container used for all cultures should be at least
three to four times the volume of the medium volume to ensure
proper aeration. We find it convenient to start this growth the
afternoon before expression.
2. Start two 50 mL cultures in LB medium containing the appro-
priate antibiotic (Table 2) by inoculating from the overnight
culture to a starting OD600 » 0.05 (see Note 7). Two cultures
are grown to have a sample that can be used for setting NMR
shims and other parameters on the spectrometer without losing
valuable data collection time.
15 In-Cell NMR Spectroscopy in Escherichia coli 271
al
ign
Centrifuge Centrifuge tS
tec
Switch media 20% Slurry De
No Lyse
Sig cells
Grow Induce na
l
3.2. Cell Viability Assay 1. Using a marker, draw partitions on the outside of an LB agar
(41) (see Note 3) selective plate to section it as shown in Fig. 4.
2. Perform two to three serial 1:100 dilutions, followed by three
serial 1:10 dilutions into fresh LB medium (Fig. 4). To reduce
error, adjust each dilution to a final volume of 1 mL. Vortex
each dilution prior to preparing the next dilution. This is done
to ensure that cells are evenly suspended.
3. Plate 10 mL drops of each dilution on the previously marked
selective plate, one drop per section (Fig. 4). Vortex the sample
in between each drop. Allow drops to thoroughly dry on the
plate and incubate at 37°C for 12–16 h.
4. Count the number of colonies per section. To calculate cell
viability, identify the section(s) containing 3–30 single colonies,
count the number of colonies for each 10 mL drop (these are
typically called colony-forming units (CFU)), and divide by the
volume (V) plated (in mL) times the dilution factor (D) (10−x):
(CFU/(V × D)). A healthy bacterial sample with an OD600 of
~1 should give ~109 CFU/mL. It is important to keep in mind
that cell viability altered by just one dilution indicates that 90%
of the viable cells have died off.
11
4
11
4 11
10
6 6
10 10
8
8 9
102 104 106 107 108 109 8
Serial Dilutions 9
9
NMR Tube Selective Plate
3.3. Freeze–Thaw 1. Prepare the freeze–thaw lysis buffer immediately before use
Lysis Protocol (see (see Note 9).
Note 8) 2. Resuspend one of each of the two samples gathered from each
point of the protein induction (i.e., before and after induction)
in 50 mL of lysis buffer.
3. Place the samples in a dry ice/ethanol bath for 5 min. Completely
thaw at room temperature. Repeat the freeze–thaw cycle two
more times. Be careful when thawing as some proteins are
temperature sensitive and may be aggregated by high or low
temperatures (see Note 7).
4. Spin the samples at 16,000 × g for 10 min at 25°C. Transfer the
supernatant to a fresh Eppendorf tube; this is the “cleared lysate”
sample. The pellet left behind is the “insoluble” sample.
5. Mix 50 mL of cleared lysate with 50 mL of 2× loading dye.
Resuspend the insoluble pellet using 25 mL of 2× loading dye and
25 mL of dIH2O. Boil the samples for 10 min.
6. Resuspend the cell pellets that were not used in step 2 in
25 mL of 2× loading dye and 25 mL of dIH2O. Boil the samples
for 10 min.
7. Run an SDS-PAGE gel to analyze expression levels and determine
the solubility of the protein (42). Solubility is determined by
what sample the target protein appears in. If the target protein
appears in the cleared lysate sample, then the protein is soluble.
If it appears in the insoluble pellet, then it is insoluble.
8. This method can also be applied to an in-cell NMR sample if
there was no protein signal detected during the experiment.
The volume of lysis buffer should be equal to the volume of
the NMR sample. Follow the freeze–thaw protocol through
step 4, then place the cleared lysate into an NMR tube, and
examine it for a protein signal.
4. Notes
Acknowledgments
The authors wish to thank Dr. Ronald A. Venters and Dr. Brian E.
Coggins for their useful discussions and comments on this
manuscript.
References
1. Dobson, C. M. (2004) Chemical space and 9. Sakakibara, D., Sasaki, A., Ikeya, T., Hamatsu,
biology. Nature 432, 824–828. J., Hanashima, T., Mishima, M., Yoshimasu,
2. Goto, S., Okuno, Y., Hattori, M., Nishioka, T., M., Hayashi, N., Mikawa, T., Walchli, M.,
and Kanehisa, M. (2002) LIGAND: database of Smith, B. O., Shirakawa, M., Guntert, P., and
chemical compounds and reactions in biological Ito, Y. (2009) Protein structure determination
pathways. Nucl. Acids Res. 30, 402–404. in living cells by in-cell NMR spectroscopy.
3. Lander, E. S., et al. (2001) Initial sequencing Nature 458, 102–105.
and analysis of the human genome. Nature 10. Ikeya, T., Sasaki, A., Sakakibara, D., Shigemitsu,
409, 860–921. Y., Hamatsu, J., Hanashima, T., Mishima, M.,
4. Goodsell, D. S. (1991) Inside a living cell. Yoshimasu, M., Hayashi, N., Mikawa, T.,
Trends Biochem. Sci. 16, 203–206. Nietlispach, D., Walchli, M., Smith, B. O.,
5. Ellis, R. J., and Minton, A. P. (2003) Cell biology: Shirakawa, M., Guntert, P., and Ito, Y. (2010)
join the crowd. Nature 425, 27–28. NMR protein structure determination in living
6. Hall, D., and Minton, A. P. (2003) Macromolecular E. coli cells using nonlinear sampling. Nat.
crowding: qualitative and semiquantitative Protoc. 5, 1051–1060.
successes, quantitative challenges. Biochim. Biophys. 11. Augustus, A. M., Reardon, P. N., and Spicer, L.
Acta. 1649, 127–139. D. (2009) MetJ repressor interactions with
7. Burz, D. S., Dutta, K., Cowburn, D., and DNA probed by in-cell NMR. Proc. Natl. Acad.
Shekhtman, A. (2006) Mapping structural Sci. USA 106, 5065–5069.
interactions using in-cell NMR spectroscopy 12. Hubbard, J. A., MacLachlan, L. K., King, G. W.,
(STINT-NMR). Nat. Methods 3, 91–93. Jones, J. J., and Fosberry, A. P. (2003) Nuclear
8. Reardon, P. N., and Spicer, L. D. (2005) magnetic resonance spectroscopy reveals the
Multidimensional NMR spectroscopy for protein functional state of the signalling protein CheY
characterization and assignment inside cells. in vivo in Escherichia coli. Mol. Microbiol. 49,
J. Am. Chem. Soc. 127, 10848–10849. 1191–1200.
276 K.E. Robinson et al.
13. Xie, J., Thapa, R., Reverdatto, S., Burz, D. S., 25. Mandelshtam, V. A., Taylor, H. S., and Shaka,
and Shekhtman, A. (2009) Screening of small A. J. (1998) Application of the filter diagonalization
molecule interactor library by using in-cell method to one- and two-dimensional NMR
NMR spectroscopy (SMILI-NMR). J. Med. spectra. J. Magn. Reson. 133, 304–312.
Chem. 52, 3516–3522. 26. Kupce, E., and Freeman, R. (2003) Fast multi-
14. Serber, Z., Keatinge-Clay, A. T., Ledwidge, R., dimensional NMR of proteins. J. Biomol. NMR
Kelly, A. E., Miller, S. M., and Dotsch, V. 25, 349–354.
(2001) High-resolution macromolecular NMR 27. Schanda, P., Kupce, E., and Brutscher, B.
spectroscopy inside living cells. J. Am. Chem. (2005) SOFAST-HMQC experiments for
Soc. 123, 2446–2447. recording two-dimensional heteronuclear corre-
15. Serber, Z., Straub, W., Corsini, L., Nomura, A. lation spectra of proteins within a few seconds.
M., Shimba, N., Craik, C. S., Ortiz de J. Biomol. NMR 33, 199–211.
Montellano, P., and Dotsch, V. (2004) Methyl 28. Kupce, E., and Freeman, R. (2003) Projection-
groups as probes for proteins and complexes in reconstruction of three-dimensional NMR
in-cell NMR experiments. J. Am. Chem. Soc. spectra. J. Am. Chem. Soc. 125, 13958–13959.
126, 7119–7125. 29. Coggins, B. E., Venters, R. A., and Zhou, P.
16. Tugarinov, V., and Kay, L. E. (2003) Ile, Leu, (2004) Generalized reconstruction of n-D
and Val methyl assignments of the 723-residue NMR spectra from multiple projections: appli-
malate synthase G using a new labeling strategy cation to the 5-D HACACONH spectrum of
and novel NMR methods. J. Am. Chem. Soc. protein G B1 domain. J. Am. Chem. Soc. 126,
125, 13868–13878. 1000–1001.
17. Goto, N. K., Gardner, K. H., Mueller, G. A., 30. Coggins, B. E., and Zhou, P. (2007) Sampling
Willis, R. C., and Kay, L. E. (1999) A robust of the NMR time domain along concentric
and cost-effective method for the production of rings. J. Magn. Reson. 184, 207–221.
Val, Leu, Ile (delta 1) methyl-protonated 15N-, 31. Barna, J. C. J., Laue, E. D., Mayger, M. R.,
13
C-, 2H-labeled proteins. J. Biomol. NMR 13, Skilling, J., and Worrall, S. J. P. (1987)
369–374. Exponential Sampling, an alternative method
18. Li, C., Wang, G. F., Wang, Y., Creager-Allen, for sampling in two-dimensional NMR experi-
R., Lutz, E. A., Scronce, H., Slade, K. M., Ruf, ments. J. Magn. Reson. 73, 69–77.
R. A., Mehl, R. A., and Pielak, G. J. (2010) 32. Hiller, S., Fiorito, F., Wuthrich, K., and Wider, G.
Protein (19)F NMR in Escherichia coli. J. Am. (2005) Automated projection spectroscopy
Chem. Soc. 132, 321–327. (APSY). Proc. Natl. Acad. Sci. USA 102,
19. Serber, Z., Ledwidge, R., Miller, S. M., and 10876–10881.
Dotsch, V. (2001) Evaluation of parameters 33. Eghbalnia, H. R., Bahrami, A., Tonelli, M.,
critical to observing proteins inside living Hallenga, K., and Markley, J. L. (2005) High-
Escherichia coli by in-cell NMR spectroscopy. resolution iterative frequency identification
J. Am. Chem. Soc. 123, 8895–8901. for NMR as a general strategy for multidimen-
20. Li, C., Charlton, L. M., Lakkavaram, A., Seagle, sional data collection. J. Am. Chem. Soc. 127,
C., Wang, G., Young, G. B., Macdonald, J. M., 12528–12536.
and Pielak, G. J. (2008) Differential dynamical 34. Kupce, E., and Freeman, R. (2004) Projection-
effects of macromolecular crowding on an reconstruction technique for speeding up
intrinsically disordered protein and a globular multidimensional NMR spectroscopy. J. Am.
protein: implications for in-cell NMR spectros- Chem. Soc. 126, 6429–6440.
copy. J. Am. Chem. Soc. 130, 6310–6311. 35. Venters, R. A., Coggins, B. E., Kojetin, D.,
21. Barnes, C. O., and Pielak, G. J. (2010) In-cell Cavanagh, J., and Zhou, P. (2005) (4,2)D
protein NMR and protein leakage, Proteins Projection – reconstruction experiments for
79, 347–351. protein backbone assignment: application to
22. Burz, D. S., and Shekhtman, A. (2008) In-cell human carbonic anhydrase II and calbindin
biochemistry using NMR spectroscopy. PLoS D(28 K). J. Am. Chem. Soc. 127, 8785–8795.
One 3, e2571. 36. Coggins, B. E., Venters, R. A., and Zhou, P.
23. Dedmon, M. M., Patel, C. N., Young, G. B., (2005) Filtered backprojection for the recon-
and Pielak, G. J. (2002) FlgM gains structure struction of a high-resolution (4,2)D CH3-NH
in living cells. Proc. Natl. Acad. Sci. USA 99, NOESY spectrum on a 29 kDa protein. J. Am.
12681–12684. Chem. Soc. 127, 11562–11563.
24. McNulty, B. C., Young, G. B., and Pielak, G. J. 37. Kupce, E., and Freeman, R. (2003) Recon-
(2006) Macromolecular crowding in the Esche- struction of the three-dimensional NMR spectrum
richia coli periplasm maintains alpha-synuclein of a protein from a set of plane projections.
disorder. J. Mol. Biol. 355, 893–897. J. Biomol. NMR 27, 383–387.
15 In-Cell NMR Spectroscopy in Escherichia coli 277
38. Coggins, B. E., and Zhou, P. (2006) Polar Fourier High-resolution multi-dimensional NMR
transforms of radially sampled NMR data. J. Magn. spectroscopy of proteins in human cells. Nature
Reson. 182, 84–95. 458, 106–109.
39. Selenko, P., Serber, Z., Gadea, B., Ruderman, J., 41. Beckman, J. S., and Siedow, J. N. (1985)
and Wagner, G. (2006) Quantitative NMR Bactericidal agents generated by the peroxidase-
analysis of the protein G B1 domain in Xenopus catalyzed oxidation of para-hydroquinones.
laevis egg extracts and intact oocytes. Proc. J. Biol. Chem. 260, 14604–14609.
Natl. Acad. Sci. USA 103, 11904–11909. 42. Sambrook, J., and Russell, D. W., (Eds.) (2001)
40. Inomata, K., Ohno, A., Tochio, H., Isogai, S., Molecular Cloning: A Laboratory Manual, Vol.
Tenno, T., Nakase, I., Takeuchi, T., Futaki, S., 3, Cold Spring Harbor Laboratory Press, Cold
Ito, Y., Hiroaki, H., and Shirakawa, M. (2009) Spring Harbor.
Chapter 16
Abstract
Perdeuteration and back substitution of exchangeable protons in microcrystalline proteins, in combination
with recrystallization from D2O-containing buffers, significantly reduce 1H, 1H dipolar interactions. This
way, amide proton line widths on the order of 20 Hz are obtained. Aliphatic protons are accessible either
via specifically protonated precursors or by using low amounts of H2O in the bacterial growth medium.
The labeling scheme enables characterization of structure and dynamics in the solid-state without dipolar
truncation artifacts.
Key words: Magic angle spinning solid-state NMR, Perdeuteration, 2H labeling, Microcrystalline
proteins, 15N relaxation, Order parameters, Protein dynamics
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_16, © Springer Science+Business Media, LLC 2012
279
280 B. Reif
2. Correlation
Spectroscopy
To record proton-detected experiments, solvent suppression
becomes a major issue. Water suppression can be achieved using
pulsed field gradients (79), cross-polarization (CP) periods as spin
locks to purge unwanted solvent magnetization (82), or a combina-
tion of both (84). In experiments that are carried out with proteins
that are recrystallized from buffers containing 100% H2O, the 1H
line width of most of the resonances is typically on the order of
150–250 Hz and 80–150 Hz in the absence and presence of
homonuclear decoupling, respectively (85). We showed recently
that ultrahigh-resolution 1H spectra are obtained in MAS solid-
state NMR if the respective perdeuterated protein is recrystallized
from a buffer containing 90% D2O (86) (Fig. 2). The resulting 1H
line width is on the order of 17–35 Hz for MAS spinning frequen-
cies in the range of 8–24 kHz. This approach enables 1H-detected
2D 1H, 15N correlation spectroscopy without the need for homo-
and heteronuclear dipolar decoupling.
Similarly, high-resolution spectra can be recorded for methyl
protons in perdeuterated peptides and proteins (Fig. 3a) (87). The
bacteria that overexpress the SH3 domain are grown in a medium
containing glucose that is only ~97% enriched in deuterium. The
likelihood that a proton gets incorporated into a methyl group is,
therefore, on the order of 10%. The canonical line widths in the 1H
282 B. Reif
Fig. 1. Proton density in the α-spectrin SH3 domain upon deuteration. (a) Protonated sample. (b) Sample recrystallized
from 100% H2O: Only exchangeable protons of the protein and protons of refined hydration water are displayed. (c) Sample
recrystallized from 10% H2O and 90% D2O. On average, every molecule contains only 21 protons, assuming that there are
53 hydration water molecules as found in the X-ray structure (PDB: 1U06) (83). (d) Upon deuteration, 1H, 1H dipolar interactions
are strongly attenuated due to the chemical dilution of the proton spins in the sample.
a G51
Y13
110 V53
V9 R21 V58 S19
V23
N Chemical shift [ppm]
A56 Y57
115 N35
G28 Q50 D40 T24
D14
F52 Y15
K18 E45
120 I30 K59 T32
M25
K27 R49 L8
V44 K39 E17
L10 K43 W41 E22
125 S36 K26
15
L61
K60 V46 L34 Q16
We41 A11
L31 L12 D62
A55
130 L33
20
G51
G28
10 We41
1H
A56
L61
0
0 20 40 60 80 100 120
Rotor Period [•s]
Fig. 2. (a) 1H-detected 1H, 15N correlation recorded with a perdeuterated α-spectrin SH3 sample that was recrystallized
from a buffer containing 90% D2O. (b) Amide proton line widths as a function of MAS rotation frequencies for selected residues.
Reproduced by permission of Wiley from Chevelkov et al. (86).
a
I30δ
12
14
T37 L34δ1
L8δ1
V58γ2
T32 L61δ1 L8δ2
24 V9γ2
L33δ2
L10δ2
L12δ2 L33δ1
26 L34δ2
L31 L10δ1
28
2.5 2.0 1.5 1.0 0.5 0.0
1
H Chemical Shift (ppm)
b S36β P54δ
A55
6 S36β 3
DQ Chemical Shift (ppm)
T24β
8 P54α A56 G51 4
T32β T37α
K39
K60 A11
Y15
G28
10 P20α 5
T37β
2Hα
2 α
T32α
H
V23
V53
I30
T24α
12 V44 6
F52 V58
70 60 50
13Cα Chemical Shift (ppm)
Fig. 3. (a) 1H, 13C correlation recorded for a [U – 2H, 13C, 15N]-labeled sample of the α-spectrin
SH3 domain. The experiment makes use of the residual protonation of the precursors
employed during protein biosynthesis. Spectra in gray and black are recorded using INEPT
and CP, respectively, for magnetization transfer. Cross peaks highlighted with circles are
only visible in the INEPT version of the experiment. Reproduced by permission of Elsevier
from Agarwal et al. (87). (b) Cα spectral region of the 13C-detected 2H-DQ, 13C correlation
experiment applied to the SH3 domain. Dα resonances are as narrow as 16 Hz (A56) in the
2
H DQ dimension. Reproduced by permission of ACS from Agarwal et al. (93).
16 Deuterated Peptides and Proteins: Structure and Dynamics Studies… 285
S36-Cβ
T24-Cβ T32-Cβ S19-Cβ
5 H2 O
T37-Cβ T37-Cα
1H Chemical Shift (ppm)
6 T37-OH
1J
T24-HN HN
7 T32-OH T32-Cα
8
T37-HN
T24-OH
9
L33-HN
N38-HN
S36-HN
10
74 72 70 68 66 64 62
13C Chemical Shift (ppm)
Fig. 4. Threonine spectral region of a 13C-detected 1H, 13C correlation recorded for a [U – 2H,
13
C, 15N]-labeled sample of the α-spectrin SH3 domain. The spectrum was recorded at
5°C. Peaks are split into doublets in the 1H indirect dimension due to evolution of the 1JNH
scalar coupling. Reproduced by permission of ACS from Agarwal et al. (95).
286 B. Reif
a b
βRL = βMA ΔνFWHM
= 10 Hz ΔνFWHM
2H-DQ
= 10 Hz βRL =
54.71° , βMA
βRL = 54.79°
24 kHz
10 kHz
60 30 0 –30 –60
250 200 150 100 50 0 –50 2H Chemical Shift (Hz)
15
N Frequency [Hz]
Fig. 5. Effect of the MAS spinning axis on the 15N (a) and 2H (b) line width. (a) The simulation of the 15N spectrum (15N–1H
spin pair) assumes an 15N chemical shift anisotropy of 100 ppm and a dipolar and scalar coupling to a directly bonded
proton of 10 kHz and −95 Hz, respectively. The external magnetic field strength was set to 14.1 T, corresponding to a proton
Larmor frequency of 600 MHz. Simulations are shown for a MAS rotation frequency of 10 kHz (dashed) and 24 kHz (solid).
( )
Mis-setting the spinning angle from the “magic angle” ( b MA = b RL = arctan 2 ≈ 54.73561) reintroduces the sum
and difference anisotropy for the upfield and downfield components, respectively. Reproduced by permission of ACS from
Chevelkov et al. (97). (b) Simulation of the 2H DQ (top) and DQ (bottom) spectrum. The quadrupolar coupling was assumed
to be 100 kHz, setting h = 0.1. In the simulation, the Euler angle b RL, which describes the angle between the principal axes
of the rotor fixed frame and the laboratory coordinate system, was varied as indicated. All simulations were carried out
using the program SIMPSON (98). Reproduced by permission of ACS from Agarwal et al. (93).
3. Characterization
of Dynamics
in the Solid State
In solution-state NMR, the relaxation properties of an NMR
observable are largely determined by the overall tumbling of
molecule in the solvent. Local structural fluctuations, which are
often of greater interest than the characterization of the overall
correlation time of a molecule, are therefore difficult to access. The
situation is different when microcrystalline proteins are considered.
In MAS solid-state NMR experiments, overall tumbling is absent
and relaxation is mostly driven by local structural fluctuations.
Therefore, the 15N-T1 relaxation time of an amide nitrogen in the
protein backbone can fluctuate by several orders of magnitude
(104–107). Figure 6 shows the 15N-T1 relaxation times that were
288 B. Reif
a b
6
5 6
5 15
4 4 N 1 (500 MHz)
3 15
3 N 1 (600 MHz)
2 2
– T1 [sec]
10 10
7 7
6 6
5 5
4 4
15N
3 3
2
2
1
1
15N T1 (600 MHz) 7
7 15N T (900 MHz) 6
6 1 5
5 4
10 20 30 40 50 60 10 20 30 40 50 60
Residue Number Residue Number
Fig. 6. 15N T1 relaxation times of the α-spectrin SH3 domain recorded in the solid state (a) and in solution (b). Reproduced
by permission of ACS from Chevelkov et al. (108).
with
2
⎛ γ γ h⎞ (2)
d = ⎜ H N3 ⎟ ≡ ω HN
2 2
⎝ 2π r ⎠NH
( )
2
c = ⎡⎣ γ N B0 σ|| − σ ⊥ ⎤⎦ ≡ ωN2 ·Δσ 2
2
16 Deuterated Peptides and Proteins: Structure and Dynamics Studies… 289
with ⎛m ⎞g g
DHN = ⎜ 0 ⎟ H 3 N ,
⎝ 4π ⎠ rNH
⎡⎛ 1 ⎞ ⎛1 ⎞⎤
dN + DHN H z = dN + DHN ⎢⎜ + H z ⎟ − ⎜ − H z ⎟ ⎥
⎣ ⎝ 2 ⎠ ⎝ 2 ⎠⎦
= d N + DHN ⎡⎣H a ⎤⎦ − DHN ⎡⎣H b ⎤⎦ (4)
⎧⎪ d N + DHN ⎡⎣H a ⎤⎦ ; upfield component
=⎨
⎪⎩dN − DHN ⎡⎣H ⎤⎦ ; downfield component
b
Effective
4°C 10 °C 17 °C 24 °C Temperature
D62
~30 Hz 21.4 Hz
Fig. 7. 15N columns extracted from a 2D 1H, 15N correlation experiment that was recorded without 1H decoupling in the 15N
evolution period. The MAS rotation frequency was kept constant at 13 kHz. Contributions from a static correlation between
the 15N CSA tensor and the 1H, 15N dipole as a possible source of the effect should, therefore, be constant and independent
of a change in temperature. The anisotropy of the multiplet intensities increases at lower temperatures indicating that
the correlation time implied by dynamics is decreased. The effects are, in particular, pronounced for D62 for which the
upfield component cannot be detected when the sample temperature is adjusted to 4°C. Reproduced by permission of ACS
from Chevelkov et al. (97).
292 B. Reif
and its directly bonded 1H (9, 131–133). Again, the dipolar interaction
can be determined more reliably if perdeuterated proteins are
employed since residual 1H, 1H dipolar interactions can be neglected
as a possible source of error (134). Dipolar couplings, extracted
from CPPI-type experiments (135–137), are almost unaffected by
the radio frequency inhomogeneities of the probe, and thus enable
a more accurate determination of the absolute value of the coupling.
A comparison of solid-state and solution-state order parameters
allows identification of slow motional processes that are normally
not easily observed in solution (110).
To analyze motion quantitatively, the spectral density functions
in Eqs. 1 and 5 need to be expressed explicitly. This is not trivial,
since the exact form of the spectral density function depends on the
underlying motional model (111). In the framework of extended
model-free formalism (138, 139), the spectral density functions
Jm(w) are expressed as Lorentzian functions that depend on two
correlation times, ts and tF, and two order parameters, SS and SF,
referring to slow and fast motional processes, respectively:
(
J (w ) = 1 − S F2 ) 1 + tw t
F
2 2 (
+ S F2 1 − SS2 ) 1 + tw t
S
2 2
.
(6)
F S
⎢ R
(
χ = ⎨∑ ⎢ expt R1,theo
i − R1,expt
i ) ⎥⎦ ⎣ η
(
⎥ + ⎢ expt η − η
theo
⎥ ⎬.
⎦ ⎪⎭
(7)
)
⎩⎪ ⎣
i 1,i
a b
S S2 S S2
τS (ns) τS (ns)
c d
SS2
SS2
τS (ns)
τS (ns)
Fig. 8. Rms difference plots between experimental and theoretical data as a function of SS2 and τS for the residue Q16 in
α-spectrin SH3. For the best fit, we obtain τF = 22 ps, and SF2 = 0.819. Data included in the fit are: (a) 15N T1 measured at 14.1 T
and 21.1 T. (b) 15N T1 measured at 14.1 T and hDD/CSA. (c) 15N T1 measured at 21.1 T and hDD/CSA. (d) 15N T1 measured at 14.1 T, 15N T1
measured at 21.1 T and h DD/CSA. Reproduced by permission of Springer from Chevelkov et al. (140).
steepness of the minimum, but leaves the best fit for τS and SS2 approxi-
mately unaltered. This is in agreement with previous findings
(106, 109).
We expect that this kind of analysis will become more and more
important in the future to characterize the dynamics of membrane
proteins and amyloid fibrils. However, for soluble/crystalline proteins,
solid-state NMR might also be the method of choice to quantitate
dynamic processes. Overall tumbling, which is the major source of
relaxation in solution, is absent in the solid state. Local motional
processes are directly reflected in the respective relaxation rates,
and the quantitative characterization of dynamics should thus be
more accurate.
In addition to backbone dynamics, information on side-chain
dynamics is obtained from analysis of the 2H Pake tensor. In the
past, specific deuterium labeling was used to investigate the dynamics
of various crystalline and amorphous solids, like liquid crystals (141),
polymers (3, 142, 143), biomembranes (144, 145), membrane
proteins (146–148), and enzymes (149). If the increment in the
294 B. Reif
2
H dimension of a multidimensional experiment is chosen to be small
enough, the resulting spinning sideband manifold can be employed
to extract the anisotropy and asymmetry parameters for the 2H
quadrupolar tensor in uniformly perdeuterated proteins (102, 150).
The (scaled) anisotropy and asymmetry yield the order parameter
and give direct information on the implicated motional model,
respectively. The quadrupolar interaction, which dominates the
deuterium spectral line shape, is very sensitive to molecular motion
over a large kinetic window (3, 151). This method should apply
to motional processes that are faster compared to the size of the
quadrupolar interaction (ca. 165 kHz of a sp3 hybridized carbon
(152)). Intermediate motions (10−4 to 10−7 s) result in line shape
distortions due to anisotropic 2H T2 relaxation (142, 153, 154).
In crystalline proteins, this scenario is difficult to observe due to low
signal-to-noise ratios. Faster processes can, in principle, be analyzed
by measuring 2H T1 relaxation times. The anisotropy of the spin-
lattice relaxation time T1 was used in the past to study fast molecular
motion (10−8 to 10−12 s) (151). Uniformly deuterated spin systems,
similar to the case of 15N–2H T1 relaxation times, also suffer from
(2H, 2H) spin diffusion. The measured 2H R1 rates are generally
averaged because of cross talk among the 2H spins (155). An alternative
route to assess fast side-chain dynamics involves the incorporation
of selective 13C labels into the methyl groups, making use of selectively
isotopically enriched amino acid precursors. This can be achieved
by employing α-ketoisovalerate ((12CD3) (13CHD2)-CD-CO-COO−)
in protein biosynthesis, which yields efficient labeling of one methyl
group (−CD2H) in valine and leucine residues (89, 90). Interestingly,
the resulting side-chain 13C-T1 relaxation times match those found
in solution, demonstrating that motional processes in the solid
state and in solution are highly similar (89, 155). This similarity
opens the door for future characterization of biomolecular dynamics
in which MAS solid-state NMR might play a major role given
the fact that relaxation parameters are independent of molecular
tumbling. Thus, local structural fluctuations are accessible with
much higher precision in the solid-state compared to solution-state
NMR experiments.
References
1. Pauli, J., Van Rossum, B.-J., Förster, H., De 3. Schmidt-Rohr, K., and Spiess, H. W. (1994)
Groot, H. J. M., and Oschkinat, H. (2000) Multidimensional Solid-State NMR and
Sample Optimization and Identification of Polymers, Academic Press, London.
Signal Patterns of Amino Acid Side Chains in 4. LeMaster, D. M., and Kushlan, D. M. (1996)
2D-RFDR Spectra of the α-Spectrin SH3 Dynamical Mapping of E. coli Thioredoxin via
Domain. J. Magn. Reson. 143, 411–416. 13
C NMR Relaxation Analysis. J. Am. Chem.
2. McDermott, A., Polenova, T., Böckmann, A., Soc. 118, 9255–9264.
Zilm, K. W., Paulsen, E. K., Martin, R. W., and 5. Stringer, J. A., Bronnimann, C. E., Mullen,
Montelione, G. T. (2000) Partial Assignments C. G., Zhou, D. H. H., Stellfox, S. A., Li, Y.,
for uniformly (13C,15N)-enriched BPTI in the Williams, E. H., and Rienstra, C. M. (2005)
solid state. J. Biomol. NMR 16, 209–219. Reduction of RF-induced sample heating with
16 Deuterated Peptides and Proteins: Structure and Dynamics Studies… 295
solid-state NMR spectroscopy. J. Am. Chem. 36. Tycko, R. (2006) Molecular structure of
Soc. 127, 12965–12974. amyloid fibrils: insights from solid-state NMR.
26. Hiller, M., Krabben, L., Vinothkumar, K. R., Quart. Rev. Biophys. 39, 1–55.
Castellani, F., Van Rossum, B., Kühlbrandt, 37. Paravastu, A. K., Qahwash, I., Leapman, R.
W., and Oschkinat, H. (2005) Solid-State D., Meredith, S. C., and Tycko, R. (2009)
Magic-Angle Spinning NMR of Outer- Seeded growth of beta-amyloid fibrils from
Membrane Protein G from Escherichia coli. Alzheimer’s brain-derived fibrils produces a
ChemBioChem. 6, 1679–1684. distinct fibril structure. Proc. Natl Acad. Sci.
27. Agarwal, V., Fink, U., Schuldiner, S., and USA 106, 7443–7448.
Reif, B. (2007) MAS Solid-State NMR 38. Jaroniec, C. P., MacPhee, C. E., Astrof, N. S.,
Studies on the Multidrug Transporer EmrE. Dobson, C. M., and Griffin, R. G. (2002)
BBA- Biomembranes 1768, 3036–3043. Molecular conformation of a peptide fragment
28. Etzkorn, M., Martell, S., Andronesi, O. C., of transthyretin in an amyloid fibril. Proc.
Seidel, K., Engelhard, M., and Baldus, M. (2007) Natl. Acad. Sci. USA 99, 16748–16753.
Secondary structure, dynamics, and topology 39. Jaroniec, C. P., MacPhee, C. E., Bajaj, V. S.,
of a seven-helix receptor in native membranes, McMahon, M. T., Dobson, C. M., and Griffin,
studied by solid-state NMR spectroscopy. R. G. (2004) High-resolution molecular structure
Angew. Chem. Int. Edt. 46, 459–462 of a peptide in an amyloid fibril determined
29. Shi, L., Lake, E. M. R., Ahmed, M. A. M., Brown, by magic angle spinning NMR spectroscopy.
L. S., and Ladizhansky, V. (2009) Solid- Proc. Natl. Acad. Sci. USA 101, 711–716.
state NMR study of proteorhodopsin in the 40. Ferguson, N., Becker, J., Tidow, H., Tremmel,
lipid environment: Secondary structure S., Sharpe, T. D., Krause, G., Flinders, J.,
and dynamics. Biochim. Biophys. Acta 1788, Petrovich, M., Berriman, J., Oschkinat, H.,
2563–2574. and Fersht, A. R. (2006) General structural
30. Pfleger, N., Woerner, A. C., Yang, J., Shastri, motifs of amyloid protofilaments. Proc. Natl.
S., Hellmich, U. A., Aslimovska, L., Maier, Acad. Sci. USA 103, 16248–16253.
M. S. M., and Glaubitz, C. (2009) Solid-state 41. Ritter, C., Maddelein, M.-L., Siemer, A. B.,
NMR and functional studies on proteorhodop- Lührs, T., Ernst, M., Meier, B. H., Saupe, S.
sin. Biochim. Biophys. Acta 1787, 697–705. J., and Riek, R. (2005) Correlation of struc-
31. Lange, V., Becker-Baldus, J., Kunert, B., van tural elements and infectivity of the HET-s
Rossum, B.-J., Casagrande, F., Engel, A., prion. Nature 435, 844–848.
Roske, Y., Scheffel, F. M., Schneider, E., and 42. Wasmer, C., Lange, A., Van Melckebeke, H.,
Oschkinat, H. (2010) A MAS NMR Study of Siemer, A. B., Riek, R., and Meier, B. H. (2008)
the Bacterial ABC Transporter ArtMP. Amyloid fibrils of the HET-s(218–289) prion
ChemBioChem. 11, 547–555. form a beta solenoid with a triangular hydro-
32. Li, Y., Berthold, D. A., Frericks, H. L., Gennis, phobic core. Science 319, 1523–1526.
R. B., and Rienstra, C. M. (2007) Partial 43. Luca, S., Yau, W.-M., Leapman, R. D., and
C-13 and N-15 chemical-shift assignments Tycko, R. (2007) Peptide Conformation and
of the disulfide-bond-forming enzyme DsbB Supramolecular Organization in Amylin
by 3D magic-angle spinning NMR spectros- Fibrils: Constraints from Solid-State NMR.
copy. Chembiochem 8, 434–442. Biochemistry 46, 13505–13522.
33. Li, Y., Berthold, D. A., Gennis, R. B., and 44. Madine, J., Jack, E., Stockley, P. G., Radford,
Rienstra, C. M. (2008) Chemical shift assign- S. E., Serpell, L. C., and Middleton, D. A. (2008)
ment of the transmembrane helices of DsbB, Structural Insights into the Polymorphism of
a 20-kDa integral membrane enzyme, by 3D Amyloid-Like Fibrils Formed by Region
magic-angle spinning NMR spectroscopy. 20–29 of Amylin Revealed by Solid-State
Protein Science 17, 199–204. NMR and X-ray Fiber Diffraction. J. Am. Chem.
34. Petkova, A. T., Ishii, Y., Balbach, J. J., Antzutkin, Soc. 130, 14990–15001.
O. N., Leapman, R. D., Delaglio, F., and Tycko, 45. Nielsen, J. T., Bjerring, M., Jeppesen, M. D.,
R. (2002) A structural model for Alzheimer’s Pedersen, R. O., Pedersen, J. M., Hein, K. L.,
β-amyloid fibrils based on experimental con- Vosegaard, T., Skrydstrup, T., Otzen, D. E., and
straints from solid state NMR. Proc. Natl. Nielsen, N. C. (2009) Unique Identification
Acad. Sci. USA 99, 16742–16747. of Supramolecular Structures in Amyloid
35. Petkova, A. T., Leapman, R. D., Guo, Z. H., Fibrils by Solid-State NMR Spectroscopy.
Yau, W. M., Mattson, M. P., and Tycko, R. Angew. Chem. Int. Edt. 48, 2118 –2121.
(2005) Self-propagating, molecular-level 46. Heise, H., Hoyer, W., Becker, S., Andronesi, O. C.,
polymorphism in Alzheimer’s beta-amyloid Riedel, D., and Baldus, M. (2005) Molecular-
fibrils. Science 307, 262–265. level secondary structure, polymorphism, and
16 Deuterated Peptides and Proteins: Structure and Dynamics Studies… 297
dynamics of full-length alpha-synuclein fibrils 59. Gerstein, B. C., Chow, C., Pembleton, R. G.,
studied by solid-state NMR. Proc. Natl. Acad. and Wilson, R. C. (1977) Utility of Pulse
Sci. USA 102, 15871–15876. Nuclear Magnetic Resonance in Studying
47. Kloepper, K. D., Zhou, D. H., Li, Y., Winter, Protons in Coals. J. Phys. Chem. 81, 565–570.
K. A., George, J. M., and Rienstra, C. M. (2007) 60. Burum, D. P. (1990) Combined Rotation
Temperature-dependent sensitivity enhancement and Multiple Pulse Spectroscopy (CRAMPS).
of solid-state NMR spectra of alpha-synuclein Concepts in Magn. Reson. 2, 213–227.
fibrils. J. Biomol. NMR 39, 197–211. 61. Waugh, J. S., Huber, L. M., and Haeberlen,
48. Loquet, A., Luc, Gardiennet, C., Sourigues, Y., U. (1968) Approach to High-Resolution
Wasmer, C., Habenstein, B., Schutz, A., Meier, B. NMR in Solids. Phys. Rev. Lett. 20, 180.
H., Melki, R., and Bockmann, A. (2009) Prion 62. Burum, D. P., and Rhim, W. K. (1979) Analysis
Fibrils of Ure2p Assembled under Physiological of multiple pulse NMR in solids .3. J. Chem.
Conditions Contain Highly Ordered, Natively Phys. 71, 944–956.
Folded Modules. J. Mol. Biol. 394, 108–118. 63. Hohwy, M., Bower, P. V., Jakobsen, H. J.,
49. Sun, S., Siglin, A., Williams, J. C., and and Nielsen, N. C. (1997) A high-order
Polenova, T. (2009) Solid-State and Solution and broadband CRAMPS experiment using
NMR Studies of the CAP-Gly Domain of z-rotational decoupling Chem. Phys. Lett.
Mammalian Dynactin and Its Interaction 273, 297–303.
with Microtubules. J. Am. Chem. Soc. 131, 64. Bielecki, A., Kolbert, A. C., and Levitt, M. H.
10113–10126. (1989) Frequency-switched pulse sequences –
50. Ahmed, S., Sun, S., Siglin, A. E., Polenova, Homonuclear decoupling and dilute spin NMR
T., and Williams, J. C. (2010) Disease- in solids. Chem. Phys. Lett. 155, 341–346.
Associated Mutations in the p150(Glued) 65. Vinogradov, E., Madhu, P. K., and Vega, S.
Subunit Destabilize the CAP-gly Domain. (1999) High-resolution proton solid-state
Biochemistry 49, 5083–5085. NMR spectroscopy by phase-modulated
51. Jehle, S., van Rossum, B.-J., Stout, J. R., Lee-Goldburg experiment. Chem. Phys. Lett.
Noguchi, S. M., Falber, K., Rehbein, K., 314, 443–450.
Oschkinat, H., Klevit, R. E., and Rajagopal, P. 66. Vinogradov, E., Madhu, P. K., and Vega, S.
(2009) alpha B-Crystallin: A Hybrid Solid- (2002) Proton spectroscopy in solid state
State/Solution-State NMR Investigation nuclear magnetic resonance with windowed
Reveals Structural Aspects of the Heterogeneous phase modulated Lee–Goldburg decoupling
Oligomer. J. Mol. Biol. 385, 1481–1497. sequences. Chem. Phys. Lett. 354, 193–202.
52. Baldus, M. (2002) Correlation experiments 67. Lesage, A., Sakellariou, D., Hediger, S.,
for assignment and structure elucidation of Elena, B., Charmont, P., Steuernagel, S., and
immobilized polypeptides under magic angle Emsley, L. (2003) Experimental aspects of
spinning. Prog. NMR Spect. 41, 1–47. proton NMR spectroscopy in solids using
53. Baldus, M. (2006) Molecular interactions phase-modulated homonuclear dipolar decou-
investigated by multi-dimensional solid-state pling. J. Magn. Reson. 163, 105–113.
NMR. Curr. Opin. Struct. Biol 16, 618–623. 68. Madhu, P. K., Zhao, X., and Levitt, M. H.
54. Brown, S. P. (2007) Probing proton–proton (2001) High-resolution H-1 NMR in the
proximities in the solid state. Prog. NMR solid state using symmetry-based pulse
Spect. 50, 199–251 sequences. Chem. Phys. Lett. 346, 142–148.
55. Böckmann, A. (2008) 3D protein structures 69. Samoson, A., Tuherm, T., Past, J., Reinhold,
by solid-state NMR: ready for high resolution. A., Anupold, T., and Heinmaa, N. (2005)
Angew. Chem. Int. Edt. 47, 6110–6113. New horizons for magic-angle spinning
56. Wylie, B. J., and Rienstra, C. M. (2008) NMR. Top. Curr. Chem. 246, 15–31.
Multidimensional solid state NMR of aniso- 70. LeMaster, D. M., and Richards, F. M. (1988)
tropic interactions in peptides and proteins. NMR Sequential Assignment of Escherichia
J. Chem. Phys. 128, 052207. Coli Thioredoxin Utilizing Random Fractional
57. McDermott, A. (2009) Structure and Deuteration. Biochemistry 27, 142–150.
Dynamics of Membrane Proteins by Magic 71. LeMaster, D. M. (1989) Deuteration in
Angle Spinning Solid-State NMR. Ann. Rev. protein proton magnetic resonance. Methods
Biophys. 38, 385–403. Enzymol. 177, 23–43.
58. Schnabel, B., Haubenreisser, U., Scheler, G., 72. Kay, L. E., and Gardner, K. H. (1997)
and Müller, R. (1976) in 19th Congress Solution NMR spectroscopy beyond 25 kDa.
Ampere pp 441, Heidelberg. Curr. Op. Struct. Biol. 7, 722–731.
298 B. Reif
73. McDermott, A. E., Creuzet, F. J., Kolbert, A. domain of alpha-spectrin by MAS solid-
C., and Griffin, R. G. (1992) High-Resolution state NMR. J. Biomol. NMR 31, 295–310.
Magic-Angle-Spinning NMR Spectra of 84. Zhou, D. H., and Rienstra, C. M. (2008)
Protons in Deuterated Solids. J. Magn. Reson. High-Performance Solvent Suppression for
98, 408–413. Proton-Detected Solid-State NMR. J. Magn.
74. Zheng, L., Fishbein, K. W., Griffin, R. G., Reson. 192, 167–172.
and Herzfeld, J. (1993) Two-Dimensional 85. Morcombe, C. R., Paulson, E. K., Gaponenko,
Solid-State 1H NMR and Proton Exchange. V., Byrd, R. A., and Zilm, K. W. (2005) H-1-
J. Am. Chem. Soc. 115, 6254–6261. N-15 correlation spectroscopy of nanocrystal-
75. Zorin, V. E., Brown, S. P., and Hodgkinson, P. line proteins. J. Biomol. NMR 31, 217–230.
(2006) Origins of linewidth in 1H magic- 86. Chevelkov, V., Rehbein, K., Diehl, A., and
angle spinning NMR. J. Chem. Phys. 125, Reif, B. (2006) Ultra-high resolution in pro-
144508. ton solid-state NMR at high levels of deutera-
76. Reif, B., Jaroniec, C. P., Rienstra, C. M., tion. Angew. Chem. Int. Ed. 45, 3878–3881.
Hohwy, M., and Griffin, R. G. (2001) 1H-1H 87. Agarwal, V., and Reif, B. (2008) Residual
MAS Correlation Spectroscopy and Distance Methyl Protonation in Perdeuterated Proteins
Measurements in a Deuterated Peptide. J. Magn. for Multidimensional Correlation Experiments
Reson. 151, 320–327. in MAS solid-state NMR Spectroscopy.
77. Reif, B., and Griffin, R. G. (2003) 1H detected J. Magn. Reson. 194, 16–24.
1
H,15N Correlation Spectroscopy in Rotating 88. Agarwal, V., Diehl, A., Skrynnikov, N., and
Solids. J. Magn. Reson. 160, 78–83. Reif, B. (2006) High Resolution 1H Detected
1
78. Zhou, D. H., Graesser, D. T., Franks, W. T., H,13C Correlation Spectra in MAS Solid-State
and Rienstra, C. M. (2006) Sensitivity and NMR using Deuterated Proteins with Selective
1
resolution in proton solid-state NMR at inter- H,2H Isotopic Labeling of Methyl Groups. J.
mediate deuteration levels: Quantitative lin- Am. Chem. Soc. 128, 12620–12621.
ewidth analysis and applications to correlation 89. Agarwal, V., Xue, Y., Reif, B., and Skrynnikov,
spectroscopy. J. Magn. Reson. 178, 297–307. N. R. (2008) Protein side-chain dynamics as
79. Chevelkov, V., van Rossum, B. J., Castellani, observed by solution- and solid-state NMR: a
F., Rehbein, K., Diehl, A., Hohwy, M., similarity revealed J. Am. Chem. Soc. 130,
Steuernagel, S., Engelke, F., Oschkinat, H., 16611–16621.
and Reif, B. (2003) 1H detection in MAS 90. Goto, N., and Kay, L. E. (2000) New devel-
solid state NMR spectroscopy employing opments in isotope strategies for protein solu-
pulsed field gradients for residual solvent sup- tion NMR spectroscopy. Curr. Opin. Cell
pression. J. Am. Chem. Soc. 125, 7788–7789. Biol. 10, 585–592.
80. Reif, B., van Rossum, B. J., Castellani, F., 91. Zhou, D. H., Shah, G., Cormos, M., Mullen,
Rehbein, K., Diehl, A., and Oschkinat, H. C., Sandoz, D., and Rienstra, C. M. (2007)
(2003) Determination of 1H 1H distances in a Proton-detected solid-state NMR
uniformly 2H,15N labeled SH3 domain by Spectroscopy of fully protonated proteins at
MAS solid state NMR spectroscopy. J. Am. 40 kHz magic-angle spinning. J. Am. Chem.
Chem. Soc. 125, 1488–1489. Soc. 129, 11791–11801.
81. Paulson, E. K., Morcombe, C. R., Gaponenko, 92. Zhou, D. H., Shea, J. J., Nieuwkoop, A. J.,
V., Dancheck, B., Byrd, R. A., and Zilm, K. Franks, W. T., Wylie, B. J., Mullen, C.,
W. (2003) High-Sensitivity Observation of Sandoz, D., and Rienstra, C. M. (2007) Solid-
Dipolar Exchange and NOEs between State Protein-Structure Determination with
Exchangeable Protons in Proteins by 3D Proton-Detected Triple-Resonance 3D
Solid-State NMR Spectroscopy. J. Am. Chem. Magic-Angle-Spinning NMR Spectroscopy.
Soc. 125, 14222–14223. Angew. Chemie Int. Edt. 46, 8380–8383.
82. Paulson, E. K., Morcombe, C. R., Gaponenko, 93. Agarwal, V., Faelber, K., Schmieder, P., and
V., Dancheck, B., Byrd, R. A., and Zilm, K. Reif, B. (2009) High-Resolution Double-
W. (2003) Sensitive High Resolution Inverse Quantum Deuterium Magic Angle Spinning
Detection NMR Spectroscopy of Proteins Solid-State NMR Spectroscopy of Perdeuterated
in the Solid State. J. Am. Chem. Soc. 125, Proteins. J. Am. Chem. Soc. 131, 2–3.
15831–15836. 94. Tosner, Z., Vosegaard, T., Kehlet, C., Khaneja,
83. Chevelkov, V., Faelber, K., Diehl, A., N., Glaser, S. J., and Nielsen, N. C. (2009)
Heinemann, U., Oschkinat, H., and Reif, B. Optimal control in NMR spectroscopy:
(2005) Detection of dynamic water mole- Numerical implementation in SIMPSON.
cules in a microcrystalline sample of the SH3 J. Magn. Reson. 197, 120–134.
16 Deuterated Peptides and Proteins: Structure and Dynamics Studies… 299
95. Agarwal, V., Linser, R., Fink, U., Faelber, K., Backbone Dynamics in a Crystalline Protein
and Reif, B. (2010) Identification of Hydroxyl from Nitrogen-15 Spin-Lattice Relaxation.
Protons, Determination of their Exchange J. Am. Chem. Soc. 127, 18190–18201.
Dynamics, and Characterization of Hydrogen 107. Giraud, N., Blackledge, M., Böckmann, A.,
Bonding by MAS solid-state NMR Spectro- and Emsley, L. (2007) The influence of nitro-
scopy in a Microcrystalline Protein. J. Am. gen-15 proton-driven spin diffusion on the
Chem. Soc. 132, 3187–3195. measurement of nitrogen-15 longitudinal
96. Mamone, S., Dorsch, A., Johannessen, O. G., relaxation times. J. Magn. Reson. 184, 51–61.
Naik, M. V., Madhu, P. K., and Levitt, M. H. 108. Chevelkov, V., Diehl, A., and Reif, B. (2008)
(2008) A Hall effect angle detector for solid- Measurement of 15 N-T1 Relaxation Rates in a
state NMR. J. Magn. Reson. 190, 135–141. Perdeuterated Protein by MAS Solid-State
97. Chevelkov, V., Faelber, K., Schrey, A., NMR Spectroscopy. J. Chem. Phys. 128,
Rehbein, K., Diehl, A., and Reif, B. (2007) 052316.
Differential Line Broadening in MAS solid- 109. Chevelkov, V., Zhuravleva, A. V., Xue, Y., Reif,
state NMR due to Dynamic Interference. B., and Skrynnikov, N. R. (2007) Combined
J. Am. Chem. Soc. 129, 10195–10200. Analysis of 15 N Relaxation Data from Solid-
98. Bak, M., Rasmussen, J. T., and Nielsen, N. C. and Solution-State NMR Spectroscopy. J. Am.
(2000) SIMPSON: A General Simulation Chem. Soc. 129, 12594–12595.
Program for Solid-State NMR Spectroscopy. 110. Chevelkov, V., Xue, Y., Linser, R., Skrynnikov,
J. Magn. Reson. 147, 296–330. N. R., and Reif, B. (2010) Comparison of
99. Wickramasinghe, N. P., Kotecha, M., Solid-State Dipolar Couplings and Solution
Samoson, A., Past, J., and Ishii, Y. (2007) Relaxation Data Provides Insight into Protein
Sensitivity enhancement in C-13 solid-state Backbone Dynamics. J. Am. Chem. Soc. 132,
NMR of protein microcrystals by use of para- 5015–5017.
magnetic metal ions for optimizing H-1 T-1 111. Torchia, D. A., and Szabo, A. (1982) Spin-
relaxation. J. Magn. Reson. 184, 350–356. Lattice Relaxation in Solids. J. Magn. Reson.
100. Linser, R., Chevelkov, V., Diehl, A., and Reif, B. 49, 107–121.
(2007) Sensitivity Enhancement Using 112. Cavanagh, J., Fairbrother, W. J., Palmer, A.
Paramagnetic Relaxation in MAS Solid State G., and Skelton, N. J. (1996) Protein NMR
NMR of Perdeuterated Proteins. J. Magn. Spectroscopy: Principles and Practice,
Reson. 189, 209–216. Academic Press, San Diego.
101. Linser, R., Fink, U., and Reif, B. (2008) 113. Chekmenev, E. Y., Zhang, Q., Waddell, K.
Proton-detected Scalar Coupling based W., Mashuta, M. S., and Wittebort, R. J.
Assignment Strategies in MAS Solid-State (2004) 15 N Chemical Shielding in Glycyl
NMR Spectroscopy applied to Perdeuterated Tripeptides: Measurement by Solid-State
Proteins J. Magn. Reson. 193, 89–93. NMR and Correlation with X-ray Structure.
102. Hologne, M., Faelber, K., Diehl, A., and Reif, J. Am. Chem. Soc. 126, 379–384.
B. (2005) Characterization of Dynamics of 114. Wylie, B. J., Franks, W. T., and Rienstra, C.
Perdeuterated Proteins by MAS Solid-State M. (2006) Determinations of N-15 chemical
NMR. J. Am. Chem. Soc. 127, 11208–11209. shift anisotropy magnitudes in a uniformly
103. Hologne, M., Chen, Z., and Reif, B. (2006) N-15, C-13-labeled microcrystalline protein
Characterization of dynamic processes using by three-dimensional magic-angle spinning
deuterium in uniformly 2H,13C,15N enriched nuclear magnetic resonance spectroscopy.
peptides by MAS solid-state NMR. J. Magn. J. Phys. Chem. B 110, 10926–10936.
Reson. 179, 20–28. 115. Hall, J. B., and Fushman, D. (2006) Variability
104. Cole, H. B. R., and Torchia, D. A. (1991) An of the N-15 chemical shielding tensors in the
NMR-study of the Backbone Dynamics of B3 domain of protein G from N-15 relaxation
Staphylococcal Nuclease in the Crystalline measurements at several fields. Implications
State. Chem. Phys. 158, 271–281. for backbone order parameters. J. Am. Chem.
105. Giraud, N., Böckmann, A., Lesage, A., Penin, Soc. 128, 7855–7870.
F., Blackledge, M., and Emsley, L. (2004) Site- 116. Wylie, B. J., Sperling, L. J., Frericks, H. L.,
Specific Backbone Dynamics from a Crystalline Shah, G. J., Franks, W. T., and Rienstra, C.
Protein by Solid-State NMR Spectroscopy. M. (2007) Chemical-shift anisotropy mea-
J. Am. Chem. Soc. 126, 11422–11423. surements of amide and carbonyl resonances
106. Giraud, N., Blackledge, M., Goldman, M., in a microcrystalline protein with slow magic-
Böckmann, A., Lesage, A., Penin, F., and angle spinning NMR spectroscopy. J. Am.
Emsley, L. (2005) Quantitative Analysis of Chem. Soc. 129, 5318–5319.
300 B. Reif
117. Yao, L., Vögeli, B., Ying, J., and Bax, A. 129. Reif, B., Diener, A., Hennig, M., Maurer, M.,
(2008) NMR Determination of Amide N-H and Griesinger, C. (2000) Cross Correlated
Equilibrium Bond Length from Concerted Relaxation for the Measurement of Angles
Dipolar Coupling Measurements. J. Am. between Tensorial Interactions. J. Magn.
Chem. Soc. 130, 16518–16520. Reson. 143, 45–68.
118. Zilm, K. W., and Grant, D. M. (1981) 130. Chevelkov, V., Diehl, A., and Reif, B. (2007)
Carbon-13 Dipolar Spectroscopy of Small Quantitative Measurement of Differential
Organic Molecules in Argon Matrices. J. Am. 15
N-Hα/β T2 Relaxation Times in a
Chem. Soc. 103, 2913–2922. Perdeuterated Protein by MAS Solid-State
119. Harris, R. K., Packer, K. J., and Thayer, A. M. NMR Spectroscopy. Magn. Reson. Chem. 45,
(1985) Slow Magic-Angle Rotation 13 C S156–S160.
NMR Studies of Solid Phosphonium Iodides. 131. Lorieau, J. L., and McDermott, A. E. (2006)
The Interplay of Dipolar, Shielding and Order parameters based on (CH)-C-13-H-1,
Indirect Coupling Tensors. J. Magn. Reson. (CH2)-C-13-H-1 and (CH3)-C-13-H-1 het-
62, 284–297. eronuclear dipolar powder patterns: a compari-
120. Griffey, D., and Redfield, A. (1987) Proton- son of MAS-based solid-state NMR sequences.
deteceted heteronuclear edited and correlated Magn. Reson. Chem. 44, 334–347.
nuclear-magnetic-resonance and nuclear 132. Lorieau, J. L., and McDermott, A. E. (2006)
Overhauser effect in solution. Quart. Rev. Conformational Flexibility of a Microcrystalline
Biophys. 19, 51–82. Globular Protein: Order Parameters by Solid-
121. Wu, G., Sun, B., Wasylishen, R. E., and State NMR Spectroscopy. J. Am. Chem. Soc.
Griffin, R. G. (1997) Spinning Sidebands in 128, 11505–11512.
Slow-Magic-Angle-Spinning NMR Spectra 133. Lorieau, J. L., Day, L. A., and McDermott, A.
Arising from Tightly J-Coupled Spin Pairs. J. E. (2008) Conformational dynamics of an
Magn. Reson. 124, 366–371 intact virus: Order parameters for the coat
122. Duma, L., Hediger, S., Lesage, A., Sakellariou, protein of Pf1 bacteriophage. Proc. Natl
D., and Emsley, L. (2003) Carbon-13 lineshapes Acad. Sci. USA 105, 10366–10371.
in solid-state NMR of labeled compounds. 134. Chevelkov, V., Fink, U., and Reif, B. (2009)
Effects of coherent CSA-dipolar cross-correla- Accurate Determination of Order Parameters
tion. J. Magn. Reson. 162, 90–101. from 1 H,15N Dipolar Couplings in MAS
123. Igumenova, T. I., and McDermott, A. E. solid-state NMR experiments. J. Am. Chem.
(2003) Improvement of resolution in solid Soc. 131, 14018–14022.
state NMR spectra with J-decoupling: an 135. Wu, X. L., and Zilm, K. W. (1993) Cross-
analysis of lineshape contributions in uni- Polarization with High-Speed Magic-Angle
formly 13 C-enriched amino acids and pro- Spinning. J. Magn. Reson. A 104, 154–165.
teins. J. Magn. Reson. 164, 270–285. 136. Dvinskikh, S. V., Zimmermann, H., Maliniak,
124. Maricq, M. M., and Waugh, J. S. (1979) A., and Sandstrom, D. (2003) Heteronuclear
NMR in rotating solids. J. Chem. Phys. 70, dipolar recoupling in liquid crystals and solids
3300–3316. by PISEMA-type pulse sequences. J. Magn.
125. Skrynnikov, N. R. (2007) Asymmetric doublets Reson. 164, 165–170.
in MAS NMR: coherent and incoherent mech- 137. Dvinskikh, S. V., Zimmermann, H., Maliniak,
anisms. Magn. Reson. Chem. 45, S161–S173. A., and Sandström, D. (2005) Heteronuclear
126. Pervushin, K., Riek, R., Wider, G., and dipolar recoupling in solid-state nuclear mag-
Wüthrich, K. (1997) Attenuated T2 relaxation netic resonance by amplitude-, phase-, and
by mutual cancellation of dipole-dipole cou- frequency-modulated Lee–Goldburg cross-
pling and chemical shift anisotropy indicates polarization. J. Chem. Phys. 122, 044512.
an avenue to NMR structures of very large 138. Lipari, G., and Szabo, A. (1982) Model-Free
biological macromolecules in solution. Proc. Approach to the Interpretation of Nuclear
Natl. Acad. Sci. USA 94, 12366–12371. Magnetic Resonance Relaxation in
127. Tjandra, N., Szabo, A., and Bax, A. (1996) Macromolecules. 1. Theory and Range of
Protein Backbone Dynamics and 15 N Validity. J. Am. Chem. Soc. 104, 4546–4559.
Chemical Shift Anisotropy from Quantitative 139. Clore, G. M., Szabo, A., Bax, A., Kay, L. E.,
Measurement of Relaxation Interference Driscoll, P. C., and Gronenborn, A. M.
Effects. J. Am. Chem. Soc. 118, 6986–6991. (1990) Deviations from the Simple
128. Reif, B., Hennig, M., and Griesinger, C. 2-Parameter Model-Free Approach to the
(1997) Direct Measurement of Angles Interpretation of N-15 Nuclear Magnetic
Between Bond Vectors in High-Resolution Relaxation of Proteins. J. Am. Chem. Soc.
NMR. Science 276, 1230–1233. 112, 4989–4991.
16 Deuterated Peptides and Proteins: Structure and Dynamics Studies… 301
140. Chevelkov, V., Fink, U., and Reif, B. (2009) site on the nicotinic acetylcholine receptor.
Analysis of the Dynamics of Backbone Motion Proc. Natl. Acad. Sci. USA 98, 2346–2351.
in the Solid-State. J. Biomol. NMR 45, 148. Howard, K. P., Liu, W., Crocker, E., Nanda,
197–206. V., Lear, J., Degrado, W. F., and Smith, S. O.
141. Sandström, D., and Zimmermann, H. (2000) (2005) Rotational orientation of monomers
Correlation of deuterium quadrupolar cou- within a designed homo-oligomer transmem-
plings and carbon-13 chemical shifts in brane helical bundle. Protein Sci. 14,
ordered media by multiple-quantum NMR. J. 1019–1024.
Phys. Chem. B 104, 1490–1493. 149. Williams, J. C., and McDermott, A. E. (1995)
142. Spiess, H. (1985) Deuteron NMR – a new Dynamics of the flexible loop of triosephos-
tool for studying chain mobility and orien- phate isomerase – the loop motion is not
tation in polymers. Adv. Polym. Sci. 66, ligand-gated Biochemistry 34, 8309–8319.
23–58. 150. Hologne, M., Chevelkov, V., and Reif, B.
143. Hirschinger, J., Miura, H., Gardner, K. H., (2006) Deuteration of Peptides and Proteins
and English, A. D. (1990) Segmental dynam- in MAS Solid-State NMR. Prog. NMR Spect.
ics in the crystalline phase of Nylon 66 : Solid 48, 211–232.
State 2 H NMR. Macromolecules 23, 151. Hoatson, G. L., and Vold, R. L. (1994) 2 H
2153–2169. NMR Spectroscopy of Solids and Liquid
144. Seelig, J. (1977) Deuterium magnetic reso- Crystals. NMR Basic Principles and Progress
nance: Theory and application to lipid mem- 32, 3–61.
branes. Q. Rev. Biophys. 10, 353–418. 152. Emsley, J. W. (2002) Solid-State NMR
145. Davis, J. H. (1983) The description of mem- Spectroscopy- Principles and Applications,
brane lipid conformation, order and dynamics Duer, M.J. edt., Blackwell Science, Oxford.
by 2 H NMR. Biochim. Biophys. Acta 737, 153. Wittebort, R. J., Olejniczak, E. T., and Griffin,
117–171. R. G. (1987) Analysis of deuterium nuclear
146. Copié, V., McDermott, A. E., Beshah, K., magnetic resonance line shapes in anisotropic
Williams, J. C., Spijker-Assink, M., Gebhard, media. J. Chem. Phys. 86, 5411–5420.
R., Lugtenburg, J., Herzfeld, J., and Griffin, 154. Hologne, M., and Hirschinger, J. (2004)
R. G. (1994) Deuterium Solid-State Nuclear Molecular Dynamics as Studied by Static-
Magnetic Resonance Studies of Methyl Group Powder and MAS 2 H NMR. Solid State NMR
Dynamics in Bacteriorhodopsin and Retinal 26, 1–10.
Model Compounds: Evidence for a 6-s-Trans 155. Reif, B., Xue, Y., Agarwal, V., Pavlova, M. S.,
Chromophore in the Protein. Biochemistry. Hologne, M., Diehl, A., Ryabov, Y. E., and
33, 3280–3286. Skrynnikov, N. R. (2006) Protein Side-Chain
147. Williamson, P. T. F., Watts, J. A., Addona, G. Dynamics Observed by Solution- and Solid-
H., Miller, K. W., and Watts, A. (2001) state NMR: Comparative Analysis of Methyl
Dynamics and orientation of N+ (CD3) 2
H Relaxation Data. J. Am. Chem. Soc. 128,
(3)-bromoacetylcholine bound to its binding 12354–12355.
Chapter 17
Abstract
Protein–protein interactions are vital for many biological processes. These interactions often result in the
formation of protein assemblies that are large in size, insoluble, and difficult to crystallize, and therefore
are challenging to study by structure biology techniques, such as single crystal X-ray diffraction and solu-
tion NMR spectroscopy. Solid-state NMR (SSNMR) spectroscopy is emerging as a promising technique
for studies of such protein assemblies because it is not limited by molecular size, solubility, or lack of long-
range order. In the past several years, we have applied magic angle spinning SSNMR-based methods to
study several protein complexes. In this chapter, we discuss the general SSNMR methodologies employed
for structural and dynamics analyses of protein complexes with specific examples from our work on thiore-
doxin reassemblies, HIV-1 capsid protein assemblies, and microtubule-associated protein assemblies. We
present protocols for sample preparation and characterization, pulse sequences, SSNMR spectra collection,
and data analysis.
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_17, © Springer Science+Business Media, LLC 2012
303
304 S. Sun et al.
Fig. 1. Magic angle spinning: (a) the sample rotor is spun at a 54.7° angle (magic angle) with respect to the static magnetic
field; (b) a Varian 3.2-mm thick wall rotor is loaded into a Varian T3 probe. The stator holds the rotor at the magic angle and
allows the bearing/drive air flow to spin the rotor at desired frequency.
Fig. 2. Pulse sequences for resonance assignments of proteins in MAS solid-state NMR: (a) 2D 13C–13C DARR; (b) 2D
dipolar-based NCA/NCO with SPECIFIC-CP for heteronuclear 15N–13C polarization transfer; (c) 3D dipolar-based NCACX or
NCOCX with SPECIFIC-CP and DARR mixing periods for 15N–13C and 13C–13C polarization transfers, respectively; (d) 3D
dipolar-based NCACB with SPECIFIC-CP and DREAM mixing periods for 15N–13C and 13C–13C polarization transfers, respec-
tively. Filled and open rectangles represent p/2 and p pulses, respectively, unless specified otherwise.
17 Solid-State NMR Spectroscopy of Protein Complexes 307
Fig. 3. Representative solid-state NMR spectra for resonance assignments of protein complexes. (a) 2D spectra of the
1–73(U-13C,15N)/74–108(U-15N) thioredoxin reassembly demonstrating the examples of intraresidue and sequential back-
bone and side chain assignments; (a1) 13C–13C DARR; (a2) NCO and (a3) NCA. All spectra are recorded at 14.1 T with the
MAS frequency of 10 kHz. Reproduced from ref. 22 with permission from John Wiley and Sons. (b) Overlay of 2D DARR
spectra of CAP-Gly/MT (black) and CAP-Gly alone (green). The spectra of free CAP-Gly and of CAP-Gly/MT complex are
acquired at 21.1 T and MAS frequency of 14 kHz. (b2) and (b3) are expansions around selected aliphatic regions (Ca–Cb or
Ca–Cg correlations) to demonstrate chemical shift perturbations of CAP-Gly upon binding to microtubules. Reproduced from
ref. 19 with permission from the American Chemical Society. (c) Sequential backbone connectivity for the sequence stretch
A105-L111 in HIV-1 CA assemblies of conical morphology based on 3D NCOCX, NCACX, and NCACB experiments at 14.1 T
and MAS frequency of 10 kHz. The residue names are shown on top of the spectra at their 15N chemical shift plane.
Negative cross-peaks resulting from two-bond N–Cb correlations in the NCACB spectra are displayed in green. Reproduced
from ref. 18 with permission from the American Chemical Society.
Fig. 4. Pulse sequences for interface studies by solid-state NMR. (a) 15N–13C REDOR–
PAINCP; ( b ) 15N– 15N PDSD–REDOR; ( c ) 1 H– 15N HETCOR–REDOR; ( d ) 1H– 113C
REDOR–HETCOR. Filled and open rectangles represent p and p/2 pulses, respec-
tively, unless specified otherwise. XY-8 phase cycle is used in the rotor-synchronous
REDOR-p pulse train.
Fig. 5. 2D spectra for studies of intermolecular interfaces in 1–73(U-13C,15N)/74–108(U-15 N) thioredoxin reassembly: (a)
REDOR–PAINCP, (b) REDOR–HETCOR, (c) HETCOR–REDOR, and (d) PDSD–REDOR. All spectra are acquired at 14.1 T with a
MAS frequency of 10 kHz. Reproduced from ref. 27 with permission from the American Chemical Society.
sequence may include (in one or all dimensions, as needed): 90° or 60°
shifted sine bell/sine square apodization followed by a Lorentzian-
to-Gaussian transformation (for sensitivity or resolution enhancement,
respectively); forward linear predication in the indirect dimension(s),
zero filling, phase correction, polynomial, or multipoint baseline
correction. Depending on a particular experiment, maximum entropy
reconstruction (57, 58) and/or nonuniform sampling algorithms
(56, 59) may be beneficial.
Numerical simulations of SSNMR spectra are an integral part
of most of the data analysis protocols. Numerical simulations
can be employed for any part of the SSNMR investigation, from
pulse sequence design to interpretation of anisotropic lineshapes to
310 S. Sun et al.
2. Materials
Fig. 6. Morphology of HIV-1CA assemblies, microtubules and CAP-Gly/MT characterized by confocal and TEM microscopy.
(a) Confocal images of HIV-1CA assemblies before and after magic angle spinning of the sample; (b) TEM images of MT
and MT/CAP-Gly assemblies before and after magic angle spinning of the sample.
2.2. Preparation of 1. cDNA encoding gag polyprotein, pr55gag: Obtained from the
HIV-1 CA Assemblies NIH AIDS Research and Reference Reagent Program (88).
2. pET21 vector (EMD).
3. Basal Vitamins Eagle medium.
4. Modified M9 growth medium: Prepared by supplementing the
1 L of standard M9 medium (Subheading 2.1) with 10.0 mL
of Basal Vitamins Eagle medium (65).
5. Growth medium for selective labeling: Prepared by adding a
13
C, 15N isotopically labeled amino acid and the other 19 unlabeled
amino acids to the cultures at 100 mg/L in M9 medium.
6. Anion exchange chromatography buffer: 25 mM sodium
phosphate, pH 7.0, 1 mM DTT, 0.02% NaN3. Adjust pH with
1 M HCl or 1 M NaOH.
17 Solid-State NMR Spectroscopy of Protein Complexes 313
3. Methods
3.1. Preparation For a detailed description of the overexpression system and the
of Thioredoxin purification protocol for E. coli thioredoxin, see refs. 66 and 67.
Reassemblies For a description of proteolytic cleavage of thioredoxin at Arg-73
for Solid-State NMR by trypsin digestion see ref. 68. The salient steps pertaining to the
Studies preparation of the SSNMR samples of differentially enriched thi-
oredoxin reassembly are outlined below.
1. Prepare differentially enriched thioredoxin reassemblies by
overexpressing two batches of thioredoxin in E. coli BL21(DE3)
separately. Use M9 minimal medium containing 15NH4Cl and
U-13C6 glucose for expression of U-13C,15N thioredoxin and
use M9 minimal medium containing 15NH4Cl and natural
abundance glucose for 15N thioredoxin (69).
2. Purify each batch of thioredoxin by loading the crude cell
extract onto a size exclusion (Superdex 75) column. Elute the
protein using size exclusion chromatography buffer.
3. Apply the eluant to an anion exchange (DEAE-cellulose) column.
Elute the protein.
4. Validate the purity of thioredoxin by using SDS-PAGE and
measure its concentration by UV absorbance (extinction
coefficient e 280 = 14,100/M/cm) (66).
5. Cleave each protein batch at the Arg-73 site by trypsin diges-
tion. First, dialyze thioredoxin against citraconylation buffer
and concentrate the protein in an Amicon stirred cell (molecu-
lar weight cut off: 3,000 Da) to 0.3 mM. Second, block the
lysine side chain amine groups by adding 25 mL of citraconic
anhydride at 20 min intervals (total amount of citraconic anhy-
dride is 50 mL for every 1 mmol of thioredoxin) and allow the
reaction to continue for 2 h after the final addition. Add 5 M
NaOH to maintain the pH at 8.5. Third, remove excess reagent
by running the reaction mixture through a desalting column.
Fourth, add trypsin to citraconylated thioredoxin (trypsin to
thioredoxin ratio is 1:100 w/w) and allow the enzyme digestion
to continue for 6 h at 37°C. Finally, lyophilize the mixture and
incubate in 50% acetic acid for 1 h to remove citraconyl groups
on lysine side chains.
6. Separate the two peptide fragments, thioredoxin (1–73) and
thioredoxin (74–108), on a size exclusion chromatography
(Sephadex G-50) using 50% acetic acid as elution buffer.
7. Purify each fragment separately by size exclusion chromatogra-
phy (Sephadex G-25) using citraconylated thioredoxin purifi-
cation elution buffer.
8. Validate the purity of each fragment by SDS-PAGE and mea-
sure the concentration of each fragment by UV absorbance.
316 S. Sun et al.
3.2. Preparation of The cDNA encoding gag polyprotein, pr55gag, was obtained from
HIV-1 CA Assemblies the NIH AIDS Research and Reference Reagent Program (88).
The DNA sequence coding for CA (gag residues 133–363) was
amplified and subcloned into pET21 vector using the NdeI and
XhoI sites (70). The primers used for PCR amplification are 5¢-
GAT ATA CAT ATG CCT ATA GTG CAG AAC ATC CAG
GGG-3¢, and 5¢-GTG GTG CTC GAG TCA TCA CAA AAC TCT
TGC CTT ATG GCC GGG-3¢, respectively. Restriction sites are
underlined.
1. Express U-13C,15N isotopically labeled CA protein in E. coli
Rosetta 2 (DE3) in modified M9 medium using 15NH4Cl and
U-13C6 glucose as the sole nitrogen and carbon sources. Induce
the protein expression with 0.4 mM IPTG at 23°C for 16 h.
2. Express selectively labeled CA protein in E coli Rosetta 2 (DE3)
in M9 growth medium for selective labeling prepared by adding
a 13C, 15N isotopically labeled amino acid and the other 19 unla-
beled amino acids to the cultures at 100 mg/L when 0.4 mM
IPTG is added to induce protein expression (see Note 5).
3. Purify the CA protein by anion exchange chromatography
using anion exchange chromatography buffer. Use a flow rate
of 2 mL/min and collect the flow through (nonbinding part).
4. Purify the CA protein produced in step 3 by cation exchange
chromatography using a gradient formed by cation exchange
17 Solid-State NMR Spectroscopy of Protein Complexes 317
3.3. Transmission The morphologies of the HIV-1 CA assemblies are analyzed using
Electron Microscopy a Zeiss CEM 902 transmission electron microscope operating at
of HIV-1 CA Protein 80 kV. Samples are stained with TEM staining solution, deposited
Assemblies and onto 400 mesh, Formval/carbon-coated copper grids, and dried
CAP-Gly/Microtubule for 40 min. Follow the protocol below to prepare the TEM grids:
Protein Assemblies 1. Transfer 5 mL of CA assemblies slurry and 5 mL of TEM staining
solution to two separate wells of a mini tray.
2. Place the copper grid on the CA assemblies slurry first with the
shiny side down, incubate for 1 min.
3. Use the edge of a filter paper to remove excess solution.
4. Place the grid on top of the staining solution and incubate for
30 s with the same side facing down.
318 S. Sun et al.
3.5. Cryo-SEM The morphologies of the HIV-1 CA assemblies are analyzed using
Microscopy of HIV-1 Cryo-SEM microscopy on a cold stage using a high-pressure
CA Protein Assemblies freezer. The cryo-fixed specimens are cryo-fractured under vacuum
to reveal internal structure. For cryo-SEM imaging of the CA
assemblies, follow the steps below (see Note 11).
1. Place 1 mL of protein assemblies slurry on the gold carrier
plates. Carefully cover the gold carrier plate with copper hat.
2. Transfer the gold plate set to the Leica EM PACT high-
pressure freezer and freeze the sample at 2,000 bar at dT/
dt > 10,000°C/s.
3. Transfer the frozen sample to the precooled sample prepara-
tion chamber (−125°C) with liquid nitrogen.
4. Fracture the copper cover with the knife in the preparation
chamber.
17 Solid-State NMR Spectroscopy of Protein Complexes 319
3.8. Solid-State The typical SSNMR experimental parameters for 2D and 3D reso-
NMR Spectroscopy nance assignment experiments conducted at 14.1 T on a Varian
for Resonance InfinityPlus instrument equipped with 3.2 mm triple tuned T3
Assignments probe are detailed below.
1. For the MAS frequency of 10 kHz, set the radio frequency (rf)
field strengths to 95 kHz (1H), 50 kHz (13C), and 50 kHz
(15N) for hard pulses. For 1H–13C CP or 1H–15N CP, contact
times are 0.85 and 1.1 ms, respectively; 1H radio frequency
field is 50 kHz, and 13C or 15N radio frequency field is ca.
40 kHz in the center of a linear or tangential ramp. Use TPPM
decoupling (72); the decoupling field strengths range between
80 and 100 kHz in different experiments. Recycle delays in all
17 Solid-State NMR Spectroscopy of Protein Complexes 321
3.9. Solid-State NMR The pulse sequences used for backbone dynamics experiments are
Experiments for shown in Fig. 7 and described below.
Probing Protein 15
1. N longitudinal relaxation rates (T1) provide information
Backbone Dynamics about motions on the pico- to nanosecond time scales. Insert
a p/2 − t − p/2 block into the NCA experiment before the
DCP block, by which magnetization is transferred from 15N to
13
Ca. Monitor the decay of 15N magnetization by tracing the
peak intensities in a series of 2D NCA spectra (28). Generate
relaxation curves from the spectra acquired with a series of
delays; for each residue, plot the cross peak intensities as a
function of the delay time, and fit the experimental points to a
single-exponential function I = I0exp(−R1t) to extract the residue-
specific relaxation rates R1 (see Note 18).
2. For qualitative detection of submillisecond motions in backbone
amide protons, insert a t/2 − p − t/2 echo before the 1H–15N CP
block. Proton magnetization is rapidly dephased during the
two t/2 delays if it is in the rigid environment of the solid
protein. Only the amide protons with high mobility on the
submillisecond time scale can survive the relatively long delay
(e.g., 400 ms), and the backbone nitrogen atoms bonded to
these protons detected (see Note 18).
322 S. Sun et al.
Fig. 7. Pulse sequences for dynamics studies of proteins and protein assemblies by MAS solid-state NMR. (a) 3D 1H T2¢
filtered NCA experiment; (b) 3D NCA-based 15N T1 relaxation experiment; (c) 3D DIPSHIFT-NCA experiment with R1817 block
employed for dipolar recoupling; (d) 3D ROCSA-NCA experiment. Filled and open rectangles represent p/2 and p pulses,
respectively, unless specified otherwise.
15
3. N chemical shift anisotropy (CSA) is also sensitive to protein
dynamics. The CSA is reduced in the presence of motions
occurring at frequencies faster than the magnitude of the CSA
interaction. Therefore, the ratio of the anisotropy in the pres-
ence of dynamics to the static-limit anisotropy as well as the
asymmetry parameter of the dynamically averaged CSA tensor
are a probe of the amplitude and geometry of the motions. 15N
CSA at 14.1 T is ca. 10 kHz, and therefore, it is sensitive to
the motions occurring on the time scales faster than 100 ms.
Measure the 15N CSA tensors site-specifically by introducing a
ROCSA CSA recoupling block (73) before the 15N chemical
shift evolution period in the NCA experiment in a 3D ROCSA-
NCA (74) experiment (see Note 19). Representative 15N CSA
lineshapes are illustrated for reassembled thioredoxin in Fig. 8.
4. 1H–15N dipolar couplings are also sensitive to motions on the
time scales of less than 100 ms. Record residue-specific 1H–15N
dipolar lineshapes in a 3D DIPSHIFT-NCA experiment by
introducing a DIPSHIFT period (75) in the basic NCA
sequence. A number of dipolar recoupling sequences can be
employed (the original DIPSHIFT (75), TMREV (76), LGCP
(77–79)). In our work, we employ an RN-type recoupling
block, R1817 (43), for the recoupling of the 1H–15N dipolar
coupling and at the same time the suppression of the 1H–1H
homonuclear dipolar interactions. During the R1817 recoupling
period, the 15N chemical shift is refocused by a spin echo (see
Note 20). Representative 1H–15N dipolar lineshapes are illus-
trated for reassembled thioredoxin in Fig. 8.
5. Extract the CSA and dipolar tensor parameters from the 3D
ROCSA-NCA and 3D DIPSHIFT-NCA experiments, by
numerical simulations of the experimental lineshapes in
17 Solid-State NMR Spectroscopy of Protein Complexes 323
Fig. 8. Dynamics information extracted from 3D-ROCSA and 3D-DIPSHIFT experiments in 1–73(U-13C,15N)/74–108(U-15N)
reassembled thioredoxin. (a) Representative 15N CSA lineshapes; the fit values are: G21, ds = 34 ± 2 ppm, h = 1.0 ± 0.25;
R73, ds = 75 ± 5 ppm, h = 0.16 ± 0.11; V25, ds = 90 ± 4 ppm, h = 0.20 ± 0.10; T8, ds = 97 ± 4 ppm, h = 0.20 ± 0.08.
(b) Chemical shift anisotropy ds plotted as a function of the residue number; (c) Representative 15N–1H dipolar lineshapes;
(d) Dipolar order parameters <S> plotted as a function of the residue number. Reproduced from ref. 28 with permission
from the American Chemical Society.
3.10. Solid-State 1. The REDOR block (80) is incorporated into a family of NMR
NMR Experiments pulse sequences to differentiate between 15N nuclei in two
for Structural Analysis distinct environments. REDOR reintroduces the dipolar
of Protein Interfaces coupling between 15N and 13C (or between 13C and 1H) nuclei,
which would otherwise be suppressed by MAS. Therefore, in
the differentially enriched 1–73(U-13C,15N)/74–108(U-15N)
thioredoxin reassembly the 15N or 1H magnetization in the
U-13C,15N-enriched fragment can be selectively dephased during
the REDOR period by the reintroduced 15N–13C or 1H–13C
heteronuclear dipolar coupling, respectively, while the 15N or 1H
magnetization in the U-15N fragment is retained allowing
for subsequent polarization transfer through the interface or within
the 15N fragment resulting in either intermolecular or intramo-
lecular isotopically edited correlations, depending on the desired
information (see Note 21).
324 S. Sun et al.
4. Notes
Acknowledgments
References
1. Yool A. J. (2007) Aquaporins: Multiple roles from solid-state NMR Spectroscopy. J. Phys.
in the central nervous system. Neuroscientist Chem. B 111, 10340–10351.
13, 470–485. 8. Lange A., et al. (2006) Toxin-induced confor-
2. Vale R. D. (2003) The molecular motor mational changes in a potassium channel
toolbox for intracellular transport. Cell 112, revealed by solid-state NMR. Nature 440,
467–480. 959–962.
3. Grunewald K. & Cyrklaff M. (2006) Structure 9. Porcelli F., Buck-Koehntop B. A., Thennarasu
of complex viruses and virus-infected cells by S., Ramamoorthy A., & Veglia G. (2006)
electron cryo tomography. Curr. Opin. Microbiol. Structures of the dimeric and monomeric vari-
9, 437–442. ants of magainin antimicrobial peptides
4. Klein K. C., Reed J. C., & Lingappa J. R. (2007) (MSI-78 and MSI-594) in micelles and bilayers,
Intracellular destinies: Degradation, targeting, determined by NMR spectroscopy. Biochemistry
assembly, and endocytosis of HIV gag. AIDS 45, 5793–5799.
Rev. 9, 150–161. 10. Zheng Z., Yang R., Bodner M. L., & Weliky
5. Uysal H., et al. (2010) Antibodies to citrulli- D. P. (2006) Conformational flexibility and strand
nated proteins: molecular interactions and arrangements of the membrane-associated
arthritogenicity. Immunol. Rev. 233, 9–33. HIV fusion peptide trimer probed by solid-
6. Goldbourt A., Gross B. J., Day L. A., & state NMR spectroscopy. Biochemistry 45,
McDermott A. E. (2007) Filamentous phage 12960–12975.
studied by magic-angle spinning NMR: 11. Lewandowski J. R., De Paepe G., & Griffin R.
Resonance assignment and secondary struc- G. (2007) Proton assisted insensitive nuclei
ture of the coat protein in Pf1. J. Am. Chem. cross polarization. J. Am. Chem. Soc. 129,
Soc. 129, 2338–2344. 728–729.
7. Hong M. (2007) Structure, topology, and 12. Chimon S. & Ishii Y. (2005) Capturing interme-
dynamics of membrane peptides and proteins diate structures of Alzheimer’s beta-amyloid,
17 Solid-State NMR Spectroscopy of Protein Complexes 329
A beta(1–40), by solid-state NMR spectroscopy. 26. Marulanda D., Tasayco M. L., Cataldi M.,
J. Am. Chem. Soc. 127, 13472–13473. Arriaran V., & Polenova T. (2005) Resonance
13. Jaroniec C. P., et al. (2004) High-resolution assignments and secondary structure analysis
molecular structure of a peptide in an amyloid of E. coli thioredoxin by magic angle spinning
fibril determined by magic angle spinning solid-state NMR spectroscopy. J. Phys. Chem. B
NMR spectroscopy. Proc. Natl. Acad. Sci. USA 109, 18135–18145.
101, 711–716. 27. Yang J., Tasayco M. L., & Polenova T. (2008)
14. Petkova A. T., et al. (2004) Solid state NMR Magic angle spinning NMR experiments for
reveals a pH-dependent antiparallel beta-sheet structural studies of differentially enriched
registry in fibrils formed by a beta-amyloid protein interfaces and protein assemblies. J.
peptide. J. Mol. Biol. 335, 247–260. Am. Chem. Soc. 130, 5798–5807.
15. Shewmaker F., Wickner R. B., & Tycko R. (2006) 28. Yang J., Tasayco M. L., & Polenova T. (2009)
Amyloid of the prion domain of Sup35p has an Dynamics of Reassembled Thioredoxin Studied
in-register parallel beta-sheet structure. Proc. by Magic Angle Spinning NMR: Snapshots
Natl. Acad. Sci. USA 103, 19754–19759. from Different Time Scales. J. Am. Chem. Soc.
16. Siemer A. B., et al. (2006) Observation of highly 131, 13690–13702.
flexible residues in amyloid fibrils of the HET-s 29. Franks W., Kloepper K., Wylie B., & Rienstra
prion. J. Am. Chem. Soc. 128, 13224–13228. C. (2007) Four-dimensional heteronuclear
17. Tycko R. (2006) Molecular structure of correlation experiments for chemical shift
amyloid fibrils: insights from solid-state NMR. assignment of solid proteins. J. Biomol. NMR
Q. Rev. Biophys. 39, 1–55. 39, 107–131.
18. Han Y., et al. (2010) Solid-State NMR Studies 30. Andrew E. R., Bradbury A., & Eades R. G.
of HIV-1 Capsid Protein Assemblies. J. Am. (1958) Nuclear Magnetic Resonance Spectra
Chem. Soc. 132, 1976–1987. from a Crystal Rotated at High Speed. Nature
19. Sun S. J., Siglin A., Williams J. C., & Polenova 182, 1659–1659.
T. (2009) Solid-State and Solution NMR 31. Schaefer J., McKay R. A., & Stejskal E. O.
Studies of the CAP-Gly Domain of Mammalian (1979) Double-cross-polarization NMR of
Dynactin and Its Interaction with Microtubules. solids. J. Magn. Reson. 34, 443–447.
J. Am. Chem. Soc. 131, 10113–10126. 32. Baldus M., Petkova A. T., Herzfeld J., & Griffin
20. Etzkorn M., Bockmann A., Lange A., & Baldus R. G. (1998) Cross polarization in the tilted
M. (2004) Probing molecular interfaces using frame: assignment and spectral simplification
2D magic-angle-spinning NMR on protein in heteronuclear spin systems. Mol. Phys. 95,
mixtures with different uniform labeling. J. Am. 1197–1207.
Chem. Soc. 126, 14746–14751. 33. Szeverenyi N. M., Sullivan M. J., & Maciel G.
21. Marulanda D., et al. (2004) Magic angle spinning E. (1982) Observation of spin exchange by
solid-state NMR spectroscopy for structural two-dimensional fourier transform 13C cross
studies of protein interfaces. Resonance assign- polarization-magic-angle spinning. J. Magn.
ments of differentially enriched Escherichia Reson. 47, 462–475.
coli thioredoxin reassembled by fragment 34. Takegoshi K., Nakamura S., & Terao T. (2001)
complementation. J. Am. Chem. Soc. 126, 13C-1H dipolar-assisted rotational resonance
16608–16620. in magic-angle spinning NMR. Chem. Phys.
22. Yang J., et al. (2007) Magic angle spinning NMR Lett. 344, 631–637.
spectroscopy of thioredoxin reassemblies. Magn. 35. Verel R., Baldus M., Ernst M., & Meier B. H.
Reson. Chem. 45, S73-S83. (1998) A homonuclear spin-pair filter for
23. Castellani F., et al. (2002) Structure of a protein solid-state NMR based on adiabatic-passage
determined by solid-state magic-angle-spinning techniques. Chem. Phys. Lett. 287, 421–428.
NMR spectroscopy. Nature 420, 98–102. 36. Bennett A. E., et al. (1998) Homonuclear
24. Hong M. & Jakes K. (1999) Selective and radio frequency-driven recoupling in rotating
extensive 13 C labeling of a membrane protein solids. J. Chem. Phys. 108, 9463–9479.
for solid-state NMR investigations. J. Biomol. 37. Hohwy M., Rienstra C. M., Jaroniec C. P., &
NMR 14, 71–74. Griffin R. G. (1999) Fivefold symmetric homo-
25. Muchmore D. C., McIntosh L. P., Russell C. nuclear dipolar recoupling in rotating solids:
B., Anderson D. E., & Dahlquist F. W. (1989) Application to double quantum spectroscopy.
Expression and nitrogen-15 labeling of proteins J. Chem. Phys. 110, 7983–7992.
for proton and nitrogen-15 nuclear magnetic 38. Ernst M., Detken A., Bockmann A., & Meier B. H.
resonance. Methods Enzymol. 177, 44–73. (2003) NMR spectra of a microcrystalline
330 S. Sun et al.
protein at 30 kHz MAS. J. Am. Chem. Soc. 125, analysis of NMR data. J. Biomol. NMR 4,
15807–15810. 603–614.
39. Chen L. L., et al. (2007) J-based 2D homo- 53. Goddard T. D. & Kneller D. G. Sparky 3
nuclear and heteronuclear correlation in (University of California, San Francisco).
solid-state proteins. Magn. Reson. Chem. 45, 54. Vranken W. F., et al. (2005) The CCPN data
S84-S92. model for NMR spectroscopy: development of
40. Chen L., et al. (2006) Constant-Time a software pipeline. Proteins 59, 687–696.
Through-Bond 13C Correlation Spectroscopy 55. Kraulis P. J. (1989) ANSIG: A program for the
for Assigning Protein Resonances with Solid- assignment of protein 1H 2D NMR spectra by
State NMR Spectroscopy. J. Am. Chem. Soc. interactive computer graphics. J. Magn. Reson.
128, 9992–9993. 84, 627–633.
41. Chen L., et al. (2007) Backbone assignments 56. Matsuki Y., Eddy M. T., & Herzfeld J. (2009)
in solid-state proteins using J-based 3D Spectroscopy by Integration of Frequency and
Heteronuclear correlation spectroscopy. J. Am. Time Domain Information for Fast Acquisition
Chem. Soc. 129, 10650–10651. of High-Resolution Dark Spectra. J. Am.
42. Griffin R. G. (1998) Dipolar recoupling in Chem. Soc. 131, 4648–4656.
MAS spectra of biological solids. Nat. Struct. 57. Laue E. D., Skilling J., Staunton J., Sibisi S., &
Biol. 5 Suppl, 508–512. Brereton R. G. (1985) Maximum Entropy Method
43. Zhao X., Edén M., & Levitt M. H. (2001) in Nuclear Magnetic Resonance Spectroscopy.
Recoupling of heteronuclear dipolar interactions J. Magn. Reson. 62, 437–452.
in solid-state NMR using symmetry-based pulse 58. Hoch J. C., Stern A. S., Donoho D. L., &
sequences. Chem. Phys. Lett. 342, 353–361. Johnstone I. M. (1990) Maximum-Entropy
44. Ladizhansky V. (2009) Homonuclear dipolar Reconstruction of Complex (Phase-Sensitive)
recoupling techniques for structure determi- Spectra. J. Magn. Reson. 86, 236–246.
nation in uniformly 13C-labeled proteins. 59. Barna J. C. J., Laue E. D., Mayger M. R.,
Solid State Nucl. Magn. Reson. 36, 119–128. Skilling J., & Worrall S. J. P. (1987) Exponential
45. Balayssac S. p., Bertini I., Lelli M., Luchinat C., Sampling, an Alternative Method for Sampling
& Maletta M. (2007) Paramagnetic Ions Provide in Two-Dimensional NMR Experiments.
Structural Restraints in Solid-State NMR of J. Magn. Reson. 73, 69–77.
Proteins. J. Am. Chem. Soc. 129, 2218–2219. 60. de Bouregas F. S. & Waugh J. S. (1992)
46. Nadaud P. S., Helmus J. J., Kall S. L., & Jaroniec ANTIOPE, a program for computer experiments
C. P. (2009) Paramagnetic Ions Enable Tuning on spin dynamics. J. Magn. Reson. 96, 280–289.
of Nuclear Relaxation Rates and Provide Long- 61. Smith S. A., Levante T. O., Meier B. H., &
Range Structural Restraints in Solid-State Ernst R. R. (1994) Computer Simulations in
NMR of Proteins. J. Am. Chem. Soc. 131, Magnetic Resonance. An Object-Oriented
8108–8120. Programming Approach. J. Magn. Reson., Ser
47. Xu X., et al. (2009) Intermolecular dynamics A 106, 75–105.
studied by paramagnetic tagging. J. Biomol. 62. Blanton W. B. (2003) BlochLib: a fast NMR
NMR 43, 247–254. C++ tool kit. J. Magn. Reson. 162, 269–283.
48. Lian L.-Y. & Middleton D. A. (2001) Labelling 63. Bak M., Rasmussen J. T., & Nielsen N. C.
approaches for protein structural studies by (2000) SIMPSON: A General Simulation
solution-state and solid-state NMR. Prog. Program for Solid-State NMR Spectroscopy. J.
Nucl. Magn. Reson. Spectrosc. 39, 171–190. Magn. Reson. 147, 296–330.
49. Schubert M., Manolikas T., Rogowski M., & 64. Veshtort M. & Griffin R. G. (2006)
Meier B. H. (2006) Solid-state NMR spectros- SPINEVOLUTION: A powerful tool for the
copy of 10% 13C labeled ubiquitin: spectral simpli- simulation of solid and liquid state NMR
fication and stereospecific assignment of isopropyl experiments. J. Magn. Reson. 178, 248–282.
groups. J. Biomol. NMR 35, 167–173. 65. Erickson-Viitanen S., et al. (1989) Cleavage of
50. Delaglio F., et al. (1995) NMRPipe: a multidi- HIV-1 gag polyprotein synthesized in vitro:
mensional spectral processing system based on sequential cleavage by the viral protease. AIDS
UNIX pipes. J. Biomol. NMR 6, 277–293. Res. Hum. Retroviruses 5, 577–591.
51. Mobli M., Maciejewski M. W., Gryk M. R., & 66. Langsetmo K., Fuchs J., & Woodward C.
Hoch J. C. (2007) Automatic maximum entropy (1989) Escherichia coli Thioredoxin Folds into
spectral reconstruction in NMR. J. Biomol. NMR 2 Compact Forms of Different Stability to Urea
39, 133–139. Denaturation. Biochemistry 28, 3211–3220.
52. Johnson B. A. & Blevins R. A. (1994) NMR View: 67. Tasayco M. L. & Chao K. (1995) NMR study
A computer program for the visualization and of the reconstitution of the beta-sheet of
17 Solid-State NMR Spectroscopy of Protein Complexes 331
thioredoxin by fragment complementation. 78. van Rossum B. J., de Groot C. P., Ladizhansky
Proteins 22, 41–44. V., Vega S., & de Groot H. J. M. (2000) A
68. Slaby I. & Holmgren A. (1975) Reconstitution method for measuring heteronuclear (1H-13C)
of Escherichia coli thioredoxin from comple- distances in high speed MAS NMR. J. Am.
menting peptide fragments obtained by cleav- Chem. Soc. 122, 3465–3472.
age at methionine-37 or arginine-73. J. Biol. 79. Lorieau J. L. & McDermott A. E. (2006)
Chem. 250, 1340–1347. Conformational flexibility of a microcrystalline
69. Sambrook J., Fritsch E. F., & Sambrook J. globular protein: order parameters by solid-
(1989) Molecular cloning: a laboratory manual state NMR spectroscopy. J. Am. Chem. Soc.
(Cold Spring Harbor Laboratory, Cold Spring 128, 11505–11512.
Harbor, N.Y.) 2nd Ed. 80. Gullion T. & Schaefer J. (1989) Rotational-
70. Byeon I. J., et al. (2009) Structural conver- echo double-resonance NMR. J. Magn. Reson.
gence between Cryo-EM and NMR reveals 81, 196–200.
intersubunit interactions critical for HIV-1 81. Bielecki A., Kolbert A. C., & Levitt M. H.
capsid function. Cell 139, 780–790. (1989) Frequency-switched pulse sequences:
71. Mossessova E. & Lima C. D. (2000) Ulp1- Homonuclear decoupling and dilute spin NMR
SUMO Crystal Structure and Genetic Analysis in solids. Chem. Phys. Lett. 155, 341–346.
Reveal Conserved Interactions and a 82. Wickramasinghe N. P., Kotecha M., Samoson
Regulatory Element Essential for Cell Growth A., Past J., & Ishii Y. (2007) Sensitivity
in Yeast. Mol. Cell 5, 865–876. enhancement in (13)C solid-state NMR of
72. Bennett A. E., Rienstra C. M., Auger M., protein microcrystals by use of paramagnetic
Lakshmi K. V., & Griffin R. G. (1995) metal ions for optimizing (1)H T(1) relax-
Heteronuclear Decoupling in Rotating Solids. ation. J. Magn. Reson. 184, 350–356.
J. Chem. Phys. 103, 6951–6958. 83. Wickramasinghe N. P., et al. (2009) Nanomole-
73. Chan J. C. C. & Tycko R. (2003) Recoupling scale protein solid-state NMR by breaking
of chemical shift anisotropies in solid-state intrinsic 1HT1 boundaries. Nat. Methods 6,
NMR under high-speed magic-angle spinning 215–218.
and in uniformly 13C-labeled systems. J. Chem. 84. Carravetta M., Edén M., Zhao X., Brinkmann
Phys. 118, 8378–8389. A., & Levitt M. H. (2000) Symmetry principles
74. Wylie B. J., Franks W. T., & Rienstra C. M. for the design of radiofrequency pulse sequences
(2006) Determinations of 15N Chemical Shift in the nuclear magnetic resonance of rotating
Anisotropy Magnitudes in a Uniformly solids. Chem. Phys. Lett. 321, 205–215.
15
N,13C-Labeled Microcrystalline Protein by 85. Holl S. M., McKay R. A., Gullion T., &
Three-Dimensional Magic-Angle Spinning Schaefer J. (1990) Rotational-echo triple-reso-
Nuclear Magnetic Resonance Spectroscopy. J. nance NMR. J. Magn. Reson. 89, 620–626.
Phys. Chem. B 110, 10926–10936. 86. Gullion T., Baker D. B., & Conradi M. S. (1990)
75. Munowitz M., Aue W. P., & Griffin R. G. New, compensated Carr-Purcell sequences.
(1982) Two-dimensional separation of dipolar J. Magn. Reson. 89, 479–484.
and scaled isotropic chemical shift interactions 87. Vinogradov E., Madhu P. K., & Vega S. (1999)
in magic angle NMR spectra. J. Chem. Phys. High-resolution proton solid-state NMR spec-
77, 1686–1689. troscopy by phase-modulated Lee-Goldburg
76. Hohwy M., Jaroniec C. P., Reif B., Rienstra C. experiment. Chem. Phys. Lett. 314, 443–450.
M., & Griffin R. G. (2000) Local structure 88. Erickson-Viitanen S., Manfredi J., Viitanen P.,
and relaxation in solid-state NMR: Accurate Tribe D. E., Tritch R., Hutchison C. A., 3rd,
measurement of amide N-H bond lengths and Loeb D. D., & Swanstrom R. (1989) Cleavage
H-N-H bond angles. J. Am. Chem. Soc. 122, of HIV-1 gag polyprotein synthesized in vitro:
3218–3219. sequential cleavage by the viral protease. AIDS
77. Hong M., Yao X., Jakes K., & Huster D. (2002) Res. Hum. Retroviruses 5, 577–591.
Investigation of Molecular Motions by Lee- 89. Marley J., Lu M., & Bracken C. (2001) A
Goldburg Cross-Polarization NMR Spectroscopy. method for efficient isotopic labeling of recom-
J. Phys. Chem. B 106, 7355–7364. binant proteins. J. Biomol. NMR 20, 71–75.
Chapter 18
Abstract
Membrane proteins function as receptors, channels, transporters, and enzymes. These proteins are gener-
ally difficult to express and purify in a functional form due to the hydrophobic nature of their membrane
spanning sequences. Studies on membrane proteins with a single membrane spanning helix have been
particularly challenging. Single-pass membrane proteins will often form dimers or higher order oligomers
in cell membranes as a result of sequence motifs that mediate specific transmembrane helix interactions.
Understanding the structural basis for helix association provides insights into how these proteins function.
Nevertheless, nonspecific association or aggregation of hydrophobic membrane spanning sequences can
occur when isolated transmembrane domains are reconstituted into membrane bilayers or solubilized into
detergent micelles for structural studies by solid-state or solution NMR spectroscopy. Here, we outline the
methods used to synthesize, purify, and characterize single transmembrane segments for structural studies.
Two synthetic strategies are discussed. The first strategy is to express hydrophobic peptides as protein
chimera attached to the maltose binding protein. The second strategy is by direct chemical synthesis.
Purification is carried out by several complementary chromatography methods. The peptides are solubi-
lized in detergent for solution NMR studies or reconstituted into model membranes for solid-state NMR
studies. We describe the methods used to characterize the reconstitution of these systems prior to NMR
structural studies to establish if there is nonspecific aggregation.
Key words: Membrane protein, NMR spectroscopy, Gp55-P, Epo receptor, Transmembrane
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_18, © Springer Science+Business Media, LLC 2012
333
334 M. Itaya et al.
1.1. Synthesis of One of the limiting factors of membrane protein structure deter-
Peptides and Proteins mination is the increased difficulty of membrane protein produc-
with Single TM Helices tion and purification. Many different production and purification
schemes have been described for structural studies (22, 33–37).
For short membrane protein sequences, two particularly useful
methods are solid-phase peptide synthesis (38–40) and expression
in E. coli (41–43).
1.1.2. Protein Many recombinant protein expression methods have been devel-
Overexpression oped, but the basic differences are in the properties of the vector
that govern the purification method (affinity tag) and the mecha-
nism of separation of the affinity tag from the protein of interest
(49). Common affinity tags are the hexahistidine (His), glutathi-
one-S-transferase (GST), and Flag tags. Incubating the cellular
lysate with a modified resin that interacts with the tag allows puri-
fication of the protein of interest. These tags, while useful for
expression and purification of soluble proteins, have had mixed
success when used with hydrophobic proteins. More recent
approaches focus on combining a traditional affinity tag with
another tag that promotes solubility of the otherwise hydrophobic
fusion protein. For example, maltose binding protein (MBP) is
336 M. Itaya et al.
1.2.2. Protein Purification Once the fusion protein has been expressed and the TM domain-
containing construct has been separated from the His-MBP tag, the
protein of interest must be separated from this mix. Two different
methods can be employed to accomplish this: organic extraction
(Subheading 3.2.3) and aqueous (Subheading 3.2.4) purification.
Constructs that have a high percentage of hydrophobic residues can
be solubilized from a dried mixture after TEV protease cleavage
using organic solvents. This method depends on two factors: a high
hydrophobic content of the protein of interest and low solubility in
organic solvents of the other contaminants (His-TEV, His-MBP,
His-MBP-fusion protein). A drawback of this method is that the
denaturation of the protein of interest is likely. In contrast, the
aqueous method leaves the protein of interest in its natively folded
state and uses the His tag present on all contaminants to remove
them from solution. To follow the purification and cleavage reac-
tions, we use SDS-PAGE (Fig. 1) and mass spectrometry (Fig. 2).
Sample conditions require that detergent which is present in
excess when purifying membrane proteins; therefore, most of the
preceding steps are done at relatively high detergent concentra-
tion. Because the signals from the detergent alkyl groups will be
overwhelming in some NMR experiments, it would be preferable
to have the sample in a deuterated detergent. It is impractical to
conduct the purification in deuterated detergents because of the
expense, but it is possible to exchange the purified sample into a
deuterated detergent. Detergent exchange can lead to longer pro-
tein relaxation times. However, this is generally not appreciable.
Fig. 1. Analysis of protein expression, cleavage, and purification using 15% SDS-PAGE. The ~48-kDa fusion protein (His-
MBP EpoR218–268) is the major protein in the lysate. The eluted protein is nearly pure and, when exposed to TEV protease for
22 h, almost 100% cleavage occurs, resulting in a 5.9-kDa band corresponding to EpoR218–268 (see boxed area). NEB MWM
NEB broad range molecular weight markers.
338 M. Itaya et al.
a b
1000
Lane 1
M 1
TEV
66 kDa His-MBP EpoR 800
Intensity (AU)
His-MBP His-MBP
600
27 kDa His-TEV
20 kDa 400
14 kDa
200
0
20 40 60 80 100 120 (103)
m/z
c M 1 2
d
300 Lane 2
66 kDa
Intensity (AU)
200
27 kDa EpoR
20 kDa EpoR(218-368)
+2
100
14 kDa
0
10 20 30 40 50 (103)
m/z
Fig. 2. Characterization of a typical aqueous purification by 15% SDS-PAGE and mass spectrometry. (a) SDS-PAGE gel from
a typical purification. Lane 1 contains the TEV cleavage mixture ~24 h after the addition of the TEV protease. The compo-
nents in the sample are His-TEV protease (27 kDa), His-MBP (43 kDa), and EpoR218–368 (17.4 kDa). (b) MALDI-TOF mass
spectrum (large MW window) of the TEV cleavage mixture showing peaks for the His-TEV protease and His-MBP. (c) SDS-
PAGE gel of the same sample mixture from (a) after the aqueous purification described in the text. Lanes 1 and 2 contain
1 and 5 μL of sample, respectively. (d) MALDI-TOF mass spectrum (small MW window) of the EpoR218–368 peptide after
aqueous purification. The 17.4-kDa band is clearly visible without contaminants from His-MBP or His-TEV protease. The
mass chromatograms shown here were obtained using a Bruker AutoFlex II MALDI-TOF–TOF mass spectrometer.
1.3. Characterization Mass spectrometry (MS) (Subheading 3.3) is used to verify the
of NMR Samples molecular weight and purity of the final purified sample. Figure 2
presents a combination of SDS-PAGE gels and mass spectra to
1.3.1. Mass Spectrometry
show the usefulness of mass spectrometry for assaying the purity of
expressed hydrophobic TM peptides after the removal of His-MBP
and His-TEV by aqueous extraction. Mass spectrometry is also
routinely used for assessing the purity of peptides produced by
solid phase synthesis and organic extraction.
18 Synthesis, Purification, and Characterization of Single Helix Membrane¼ 339
1.3.2. Size Exclusion One of the pitfalls in the purification of membrane proteins is
Chromatography aggregation. While centrifugation can remove large aggregates
from solution, microaggregrates and oligomers can remain and are
undesirable for solution NMR spectroscopy. Changing the type or
the amount of detergent used for solubilization can disrupt micro-
aggregates. Size exclusion chromatography (SEC) can optimize
the peptide solubilization by providing a picture of the aggregation
state. Figure 3 shows that adding more detergent disperses micro-
aggregates of the EpoR218–268 peptide.
a b
80
15 EpoR(218-268)
60 dimer
A 220 (mAU)
A 220 (mAU)
EpoR(218-268)
10 40
aggregates
5 20
0 0
8 10 12 14 16 18 20 22 8 10 12 14 16 18 20 22
Elution Volume (mL) Elution Volume (mL)
Fig. 3. Fast protein liquid chromatography. (a) FPLC analysis of the EpoR218–268 peptide solubilized in DPC at 20× CMC
shows that the peptide elutes in the first fractions from the gel filtration column, indicating the presence of molecular
aggregates. (b) FPLC analysis of the same sample as in (a) after the addition of DPC to 100× CMC and sonication for 1 min
shows that the peptide elutes at the molecular weight corresponding to a dimer. The FPLC separations were performed
with a Superdex 200 10/300 GL column calibrated with known standards. The buffer for the samples and standards con-
tained 100 mM sodium phosphate and 150 mM NaCl at pH 7.0, and the injected sample volume was 0.1 mL. SDS-PAGE
confirmed the identity of the elution fractions (data not shown).
340 M. Itaya et al.
0.2
0.15
Absorbance
0.1
0.05
0
1800 1760 1720 1680 1640 1600
Wavenumber (cm–1)
Fig. 4. Polarized ATR-FTIR spectroscopy. Polarized ATR-FTIR spectra were obtained of the gp55-P peptide reconstituted
into DMPC bilayers by detergent dialysis using IR light polarized parallel (solid line) and perpendicular (dotted line) rela-
tive to the bilayer normal. Only the region between 1,600cm-1 and 1,800cm-1 is shown. The amide I vibration is observed
at 1,654cm-1, indicating that the reconstituted peptide has an α-helical conformation. The intense vibration at 1,735cm-1
is due to the C=O stretching vibration of the lipid acyl chains. The dichroic ratio (A║/A┴) of the amide I band of 3.3
corresponds to a helix orientation of ~20°. The FTIR spectrum shown here was obtained using a Bruker IFS 66 V/S
spectrometer.
ratios of less than ~3. Figure 4 presents the amide I region from
the polarized FTIR spectrum of gp55-P reconstituted into
dimyristoylphosphatidylcholine (DMPC) bilayers. There is a
single amide I band observed at a frequency of 1,654cm-1, char-
acteristic of helical secondary structure. The dichroic ratio of the
1,654cm-1 vibration is 3.3, which corresponds to a tilt of the helix
axis of ~20° relative to the membrane normal. Together, the
frequency and dichroic ratio of the amide I vibration indicate that
the gp55-P peptide is properly reconstituted in a homogeneous
α-helical conformation.
1.3.4. Circular Dichroism Circular dichroism (CD) spectroscopy is a widely used technique
Spectroscopy to deduce protein secondary structure (59). CD can inform us
about the helical content of the sample, confirming that the sample
is folded or refolded properly. FTIR and CD spectroscopy provide
complementary information on secondary structure. CD can easily
distinguish random coil from α-helical and β-sheet secondary
structure. Light scattering from membrane vesicles often leads to a
red shift and damping of the CD bands making it more difficult to
distinguish the signature bands for α-helix and β-sheet. In con-
trast, α-helix and β-sheet are easily distinguished by FTIR spec-
troscopy, but they are more difficult to distinguish from random
coil. Figure 5 presents the CD spectrum of the pure EpoR218–268
peptide in dodecylphosphocholine (DPC) at a detergent concen-
tration corresponding to 1.4 times its CMC. It exhibits negative
CD absorption bands at approximately 208 and 222 nm character-
istic of α-helical secondary structure. The peptide sequence includes
18 Synthesis, Purification, and Characterization of Single Helix Membrane¼ 341
30
20
Ellipticity (millideg)
10
0
200 220 240 260 280
–10
Wavelength (nm)
1.4. NMR Deuterium NMR spectroscopy is well suited for probing dynamic
Spectroscopy processes in membrane proteins (60–62). Deuterium NMR has
been widely used to look at lipid dynamics, but much less so to
1.4.1. Solid-State 2H NMR
investigate side chain dynamics in membrane proteins because of
Spectroscopy
sensitivity and selectivity issues (63, 64). However, by combining
magic angle spinning (MAS) with specifically deuterated mem-
brane proteins and peptides, the sensitivity issue can be resolved
(61). We use deuterium MAS NMR to assess homo-oligomerization
of the reconstituted membrane peptides prior to undertaking
structural studies with more advanced NMR methods. There are
several advantages of deuterium MAS NMR. The first is that it is a
simple experiment, using a single pulse at relatively low MAS fre-
quencies. The spinning side bands in slow MAS spectra reveal the
envelope of the static 2H lineshape. Second, the measurements are
carried out in the liquid crystalline phase (above the lipid phase
transition temperature). Third, the experiments are comparative in
nature. Figure 6 presents deuterium MAS NMR spectra of gp55-P
selectively labeled at four consecutive leucines (Leu396–399) in
the middle of the TM region. Comparison of the deuterium MAS
side band patterns shows that Leu399 has the most restricted
motion and consequently is likely to be oriented toward the TM
dimer interface. Leu397 and Leu398 exhibit the narrowest deute-
rium side band patterns. The narrow lineshapes correspond to
342 M. Itaya et al.
Leu397 Leu396
396 398
L
397 L 399 L
L
399
Leu398 L Leu399
397
398 L
L 396
L
gp55-p
Restricted motion of
Mobile side chain
side chain
RRPPWFTTLISTIMGSLIILLLLLILLIWTLYS
Fig. 6. Deuterium MAS NMR spectroscopy. Gp55-P has five consecutive leucines in the middle of the TM region that allow
us to map out the dimer interface by using deuterium MAS NMR (9). Four different gp55-P TM peptides were chemically
synthesized, each with one of four sequential leucines methyl deuterium labeled. By examining the deuterium spectrum of
each peptide, we can infer which leucines are in the helix dimer interface and which ones face the lipid side chains.
Comparison of lineshapes of deuterated leucines clearly show that Leu396 and Leu399 are most likely in the dimer inter-
face and Leu397 and Leu398 oriented away from the dimer interface.
1.4.2. Solution-State NMR After production of a pure sample, preliminary scouting experi-
Spectroscopy ments must be performed to determine sample conditions that
promote both long-term sample stability and sample homogeneity.
Several factors that must be considered are buffer type and concen-
tration, detergent type and concentration, pH, and temperature
for data collection. These variables are covered well elsewhere
(2, 65). Solubility/stability assessments may be carried out with unla-
beled protein. However, the proton-nitrogen heteronuclear single
quantum correlation (15N-HSQC) experiment should be used to
assess the resolution and the sensitivity of particular sample condi-
tions; this requires 15N-labeled protein. These experiments can be
conducted at a relatively low protein concentration (~0.1 mM),
18 Synthesis, Purification, and Characterization of Single Helix Membrane¼ 343
ppm
112
113
114
115
116
117
15
N 118
119
120
121
122
123
124
125
8.8 8.6 8.4 8.2 8.0 7.8 7.6 7.4 7.2 ppm
1
H
Fig. 7. Solution NMR spectroscopy. The 1H-15N HSQC spectrum of the EpoR220–248 peptide is
shown. The peptide was prepared by the aqueous extraction method and solubilized in
d38-DPC (100× CMC). The spectrum was collected with 32 scans at a temperature of
313 K. About 30 N-H peaks are visible, which are expected from the 31 amino acid
construct, along with the pair of N-H resonances from the single Asn residue side chain.
2. Materials
2.3. Mass 1. Saturated sinapinic acid matrix solution: Dissolve ~20 mg/mL
Spectroscopy in 50% acetonitrile and 0.1% TFA. Spin down excess matrix by
centrifugation at 15,000 × g for 10 s at room temperature and
use only the supernatant (saturated sinapinic acid solution).
2. Washing buffer: 10 mM ammonium phosphate, monobasic
dissolved in 0.1% TFA.
3. Recrystallization buffer: 60% ethanol, 30% acetone, and 10%
water with 0.1% TFA.
3. Methods
3.2.2. Protein Purification 1. Lyse the cells using one of several methods (e.g., French press,
cell homogenizer).
2. Clarify the lysate by centrifuging at 25,000 × g for 25 min at
4°C.
3. Transfer the supernatant to a tube and add β-OG to ~2× its
CMC (CMC = 0.25 mM), nutate ~5 min at room temperature
to dissolve.
4. Apply the supernatant to a Ni+/NTA column (12.5 mL bed
volume) previously equilibrated with binding buffer, nutate (in
column) for 2–4 h at 4°C to bind the fusion protein, and then
allow the lysate to flow through.
5. Wash with 14 column volumes of wash buffer, collecting all
fractions for SDS-PAGE analysis. Figure 1 shows that most of
the contaminating protein is removed in the flow-through and
the first wash.
6. Elute the column with 3 column volumes of elution buffer,
collecting 1 mL fractions. Measure the optical absorbance at
280 nm (A280), combine the fractions that have protein, mea-
sure the A280 of that mixture (for a rough determination of the
protein concentration). Add DDM to 10× its CMC, and dis-
solve by sonication or nutation.
7. Add an appropriate amount of His-TEV protease (see Note 7),
nutate for 16–24 h at 23°C to cleave. Monitor the cleavage by
the loss of the band corresponding to the His-MBP-fusion
protein and the appearance of two lower molecular weight
bands corresponding to His-MBP and the TM peptide on
SDS-PAGE (see Fig. 1, lane 10).
At this point, further purification proceeds using either the
organic extraction or the aqueous extraction method.
348 M. Itaya et al.
3.2.3. Organic Extraction 1. Precipitate the protein in the TEV protease mixture
(Subheading 3.2.2) by adding trichloroacetic acid to a final
concentration of 6% (w/v).
2. Centrifuge at 9,000 × g for 20 min at 4°C to collect the precipi-
tate. Decant the supernatant, wash with milliQ water, centri-
fuge again, decant the supernatant; repeat once. Lyophilize to
remove water.
3. Nutate the dried pellet for 2 h at room temperature with
10 mL of 90% methanol/10% chloroform.
4. Filter the supernatant through a 0.2-μm syringe filter (PTFE)
to remove particulate matter. Continue to Subheading 3.4 for
reconstitution of organic solvent purified samples.
5. Assess the purity of the protein by SDS-PAGE analysis and
mass spectrometry as described in Fig. 2.
3.2.4. Aqueous Extraction 1. Dialyze the sample overnight against binding buffer contain-
ing DDM at 1× its CMC to remove the imidazole and glycerol
from the Ni+/NTA purification and TEV cleavage, respectively
(1 kDa MW cutoff is appropriate).
2. Nutate the dialyzed sample in a column with a 20-mL bed
volume of Ni+/NTA (pre-equilibrated with binding buffer plus
1× CMC of DDM) for 2–4 h at 4°C.
3. Collect the flow-through and wash the beads with 2 col-
umn volumes of binding buffer containing DDM at 1× its
CMC. Combine the flow-through with the first column vol-
ume in the wash step. Elute the column with 6 column volumes
of elution buffer.
4. Re-equilibrate the Ni+/NTA with binding buffer. Perform a
second incubation using the combined flow-through and
washes from above. Again, combine the flow-through and first
wash fractions.
5. Dialyze overnight against the final NMR buffer, or if detergent
exchange will occur, against the detergent exchange dialysis
buffer of choice (see Subheading 3.2.5, step 1).
6. If NMR will be performed directly on this sample, concentrate
the result of step 5 using an Amicon spin column (3 kDa
MWCO), and verify the purity and molecular weight by SDS-
PAGE and mass spectrometry (see Fig. 2).
3.2.5. Detergent Exchange 1. Dialyze the aqueous purification sample against detergent
exchange dialysis buffer without deuterated detergent (see
Note 8).
2. Concentrate the sample to 1–3 mL using an Amicon Ultra spin
concentrator.
3. Centrifuge at 20,000 × g for 15 min at 4°C to remove particulate
matter or aggregated protein, save the supernatant.
18 Synthesis, Purification, and Characterization of Single Helix Membrane¼ 349
3.3. Characterization Mass spectrometry is used to verify the molecular weight and purity
of Purified Protein of the final purified sample.
1. Mix the sample and saturated sinapinic acid matrix solution in
a volume ratio of 1:5 to 1:10 (see Note 9). Apply 1–2 μL of
solution onto the target plate for MALDI-TOF MS and let it
dry at room temperature (67).
2. Wash the target with 5–10 μL of washing buffer.
3. Remove the washing buffer after a few seconds with a pipet
and let the liquid evaporate.
4. Apply 0.5–1 μL of recrystallization buffer on the washed spot
and let it dry.
3.4. NMR Sample 1. Co-solubilize β-OG and DMPC in TFE. The amount of deter-
Preparation gent is determined by the final detergent concentration (5%
w/v) when the sample is rehydrated in step 5. In general, a
3.4.1. Reconstitution into
1:50 peptide-to-lipid molar ratio is used (see Note 10). Freeze
Membrane Lipids
the solution with liquid nitrogen and remove under vacuum; a
small amount of water can be added if a powder cannot be
obtained with TFE alone.
2. Dissolve the dry mixture of β-OG and DMPC in a minimum
amount of water (~0.5 mL).
3. Dissolve the appropriate amount of peptide in 1 mL of TFE
and incubate for 6 h at 37°C (for solid-state NMR experi-
ments, we typically use ~2 μmol peptide).
4. Add β-OG /DMPC drop-wise to the peptide solution while
stirring. Then add water drop-wise while stirring the sample
until bubbles form. Bubbling can be achieved when the ratio
of organic solution to water is between 1:2 and 1:4 (19). After
the titration is complete, freeze and lyophilize the sample (see
Note 11).
5. Rehydrate the sample with 4 mL of phosphate dialysis buffer
for 6 h while stirring slowly at 37°C. If the peptide sequence
contains cysteine, use MES rehydration buffer to rehydrate the
sample.
6. Dialyze the rehydrated sample against 2 L of phosphate dialysis
buffer for 48 h at 37°C. If the peptide sequence contains
350 M. Itaya et al.
4. Notes
Acknowledgments
References
1. Opella, S. J., and Marassi, F. M. (2004) dipolar coupling spectra obtained with polar-
Structure determination of membrane proteins ization inversion spin exchange at the magic
by NMR spectroscopy. Chem. Rev. 104, angle and magic-angle sample spinning
3587–3606. (PISEMAMAS). Solid State Nucl. Magn. Reson.
2. Sanders, C. R., and Sönnichsen, F. (2006) 4, 387–392.
Solution NMR of membrane proteins: Practice 13. Wang, J. F., Kim, S., Kovacs, F., and Cross, T.
and challenges. Magn. Reson. Chem. 44, A. (2001) Structure of the transmembrane
S24–S40. region of the M2 protein H+ channel. Protein
3. McDermott, A. (2009) Structure and dynam- Sci. 10, 2241–2250.
ics of membrane proteins by magic angle spin- 14. Kim, M. J., Park, S. H., Opella, S. J., Marsilje,
ning solid-state NMR. Ann. Rev. Biophys. 38, T. H., Michellys, P. Y., Seidel, H. M., and Tian,
385–403. S. S. (2007) NMR structural studies of interac-
4. Ketchem, R. R., Hu, W., and Cross, T. A. tions of a small, nonpeptidyl Tpo mimic with
(1993) High-resolution conformation of gram- the thrombopoietin receptor extracellular jux-
icidin A in a lipid bilayer by solid-state NMR. tamembrane and transmembrane domains. J.
Science 261, 1457–1460. Biol. Chem. 282, 14253–14261.
5. Andronesi, O. C., Becker, S., Seidel, K., Heise, 15. Bocharov, E. V., Pustovalova, Y. E., Pavlov, K.
H., Young, H. S., and Baldus, M. (2005) V., Volynsky, P. E., Goncharuk, M. V.,
Determination of membrane protein structure Ermolyuk, Y. S., Karpunin, D. V., Schulga, A.
and dynamics by magic-angle-spinning solid- A., Kirpichnikov, M. P., Efremov, R. G., et al.
state NMR spectroscopy. J. Am. Chem. Soc. (2007) Unique dimeric structure of BNip3
127, 12965–12974. transmembrane domain suggests membrane
6. Cady, S. D., Schmidt-Rohr, K., Wang, J., Soto, permeabilization as a cell death trigger. J. Biol.
C. S., DeGrado, W. F., and Hong, M. (2010) Chem. 282, 16256–16266.
Structure of the amantadine binding site of 16. Bocharov, E. V., Mineev, K. S., Volynsky, P. E.,
influenza M2 proton channels in lipid bilayers. Ermolyuk, Y. S., Tkach, E. N., Sobol, A. G.,
Nature 463, 689–692. Chupin, V. V., Kirpichnikov, M. P., Efremov, R.
7. Chu, S. D., Coey, A. T., and Lorigan, G. A. G., and Arseniev, A. S. (2008) Spatial structure
(2010) Solid-state 2H and 15N NMR studies of of the dimeric transmembrane domain of the
side-chain and backbone dynamics of phospho- growth factor receptor ErbB2 presumably cor-
lamban in lipid bilayers: Investigation of the responding to the receptor active state. J. Biol.
N27A mutation. Biochim. Biophys. Acta- Chem. 283, 6950–6956.
Biomembr. 1798, 210–215. 17. Roosild, T. P., Greenwald, J., Vega, M.,
8. Etzkorn, M., Martell, S., Andronesi, O. C., Castronovo, S., Riek, R., and Choe, S. (2005)
Seidel, K., Engelhard, M., and Baldus, M. NMR structure of Mistic, a membrane-inte-
(2007) Secondary structure, dynamics, and grating protein for membrane protein expres-
topology of a seven-helix receptor in native sion. Science 307, 1317–1321.
membranes, studied by solid-state NMR spec- 18. MacKenzie, K. R., Prestegard, J. H., and
troscopy. Angew. Chem. Int. Ed. Engl. 46, Engelman, D. M. (1997) A transmembrane
459–462. helix dimer: Structure and implications. Science
9. Liu, W., Crocker, E., Constantinescu, S. N., 276, 131–133.
and Smith, S. O. (2005) Helix packing and ori- 19. Sulistijo, E. S., and MacKenzie, K. R. (2009)
entation in the transmembrane dimer of gp55- Structural basis for dimerization of the BNIP3
P of the spleen focus forming virus. Biophys. J. transmembrane domain. Biochemistry 48,
89, 1194–1202. 5106–5120.
10. Smith, S. O., Song, D., Shekar, S., Groesbeek, 20. Oxenoid, K., and Chou, J. J. (2005) The struc-
M., Ziliox, M., and Aimoto, S. (2001) Structure ture of phospholamban pentamer reveals a
of the transmembrane dimer interface of glyco- channel-like architecture in membranes. Proc.
phorin A in membrane bilayers. Biochemistry Natl. Acad. Sci. USA 102, 10870–10875.
40, 6553–6558. 21. Hu, J., Qin, H., Li, C., Sharma, M., Cross, T.
11. Smith, S. O., Smith, C. S., and Bormann, B. J. A., and Gao, F. P. (2007) Structural biology of
(1996) Strong hydrogen bonding interactions transmembrane domains: Efficient production
involving a buried glutamic acid in the trans- and characterization of transmembrane pep-
membrane sequence of the neu/erbB-2 recep- tides by NMR. Protein Sci. 16, 2153–2165.
tor. Nat. Struct. Biol. 3, 252–258. 22. Tamm, L. K., and Liang, B. Y. (2006) NMR of
12. Ramamoorthy, A., and Opella, S. J. (1995) membrane proteins in solution. Prog. Nucl.
Two-dimensional chemical shift/heteronuclear Magn. Reson. Spectrosc. 48, 201–210.
18 Synthesis, Purification, and Characterization of Single Helix Membrane¼ 355
23. Mineev, K. S., Bocharov, E. V., Pustovalova, Y. E., 34. Loll, P. J. (2003) Membrane protein structural
Bocharova, O. V., Chupin, V. V., and Arseniev, biology: the high throughput challenge. J.
A. S. (2010) Spatial structure of the transmem- Struct. Biol. 142, 144–153.
brane domain heterodimer of ErbB1 and 35. Wang, D. N., Safferling, M., Lemieux, M. J.,
ErbB2 receptor tyrosine kinases. J. Mol. Biol. Griffith, H., Chen, Y., and Li, X. D. (2003)
400, 231–243. Practical aspects of overexpressing bacterial
24. White, S. H. (2009) Biophysical dissection of secondary membrane transporters for struc-
membrane proteins. Nature 459, 344–346. tural studies. Biochim. Biophys. Acta-Biomembr.
25. Ruscetti, S. K., Janesch, N. J., Chakraborti, A., 1610, 23–36.
Sawyer, S. T., and Hankins, W. D. (1990) 36. Klammt, C., Löhr, F., Schäfer, B., Haase, W.,
Friend spleen focus-forming virus induces fac- Dötsch, V., Rüterjans, H., Glaubitz, C., and
tor independence in an erythropoietin-depen- Bernhard, F. (2004) High level cell-free expres-
dent erythroleukemia cell line. J. Virol. 64, sion and specific labeling of integral membrane
1057–1062. proteins. Eur. J. Biochem. 271, 568–580.
26. Hoatlin, M. E., Kozak, S. L., Lilly, F., 37. Laage, R., and Langosch, D. (2001) Strategies
Chakraborti, A., Kozak, C. A., and Kabat, D. for prokaryotic expression of eukaryotic mem-
(1990) Activation of erythropoietin receptors brane proteins. Traffic 2, 99–104.
by Friend viral gp55 and by erythropoietin and 38. Bormann, B. J., Knowles, W. J., and Marchesi,
down-modulation by the murine Fv-2r resis- V. T. (1989) Synthetic peptides mimic the
tance gene. Proc. Natl. Acad. Sci. USA 87, assembly of transmembrane glycoproteins.
9985–9989. J. Biol. Chem. 264, 4033–4037.
27. Li, J. P., D’Andrea, A. D., Lodish, H. F., and 39. Kochendoerfer, G. G., Salom, D., Lear, J. D.,
Baltimore, D. (1990) Activation of cell growth Wilk-Orescan, R., Kent, S. B. H., and DeGrado,
by binding of Friend spleen focus-forming virus W. F. (1999) Total chemical synthesis of the
gp55 glycoprotein to the erythropoietin recep- integral membrane protein influenza A virus
tor. Nature 343, 762–764. M2: Role of its C-terminal domain in tetramer.
28. Constantinescu, S. N., Keren, T., Socolovsky, Biochemistry 38, 11905–11913.
M., Nam, H. S., Henis, Y. I., and Lodish, H. F. 40. Fisher, L. E., and Engelman, D. M. (2001)
(2001) Ligand-independent oligomerization High-yield synthesis and purification of an
of cell-surface erythropoietin receptor is medi- α-helical transmembrane domain. Anal.
ated by the transmembrane domain. Proc. Natl. Biochem. 293, 102–108.
Acad. Sci. USA 98, 4379–4384. 41. Tian, C. L., Karra, M. D., Ellis, C. D., Jacob,
29. Gurezka, R., Laage, R., Brosig, B., and J., Oxenoid, K., Sonnichsen, F., and Sanders,
Langosch, D. (1999) A heptad motif of leucine C. R. (2005) Membrane protein preparation
residues found in membrane proteins can drive for TROSY NMR screening. Meth. Enzymol.
self-assembly of artificial transmembrane seg- 394, 321–334.
ments. J. Biol. Chem. 274, 9265–9270. 42. Page, R. C., Moore, J. D., Nguyen, H. B.,
30. Constantinescu, S. N., Keren, T., Russ, W. P., Sharma, M., Chase, R., Gao, F. P., Mobley, C.
Ubarretxena-Belandia, I., Malka, Y., Kubatzky, K., Sanders, C. R., Ma, L., Sonnichsen, F. D.,
K. F., Engelman, D. M., Lodish, H. F., and et al. (2006) Comprehensive evaluation of
Henis, Y. I. (2003) The erythropoietin recep- solution nuclear magnetic resonance spectros-
tor transmembrane domain mediates complex copy sample preparation for helical integral
formation with viral anemic and polycythemic membrane proteins. J. Struct. Funct. Genomics
gp55 proteins. J. Biol. Chem. 278, 7, 51–64.
43755–43763. 43. Qin, H. J., Hu, J., Hua, Y. Z., Challa, S. V.,
31. Constantinescu, S. N., Huang, L. J. S., Nam, Cross, T. A., and Gao, F. P. (2008) Construction
H. S., and Lodish, H. F. (2001) The erythro- of a series of vectors for high throughput clon-
poietin receptor cytosolic juxtamembrane ing and expression screening of membrane pro-
domain contains an essential, precisely oriented, teins from Mycobacterium tuberculosis. BMC
hydrophobic motif. Mol. Cell 7, 377–385. Biotechnol. 8, 51–59.
32. Witthuhn, B. A., Quelle, F. W., Silvennoinen, 44. Kent, S. B. H. (1988) Chemical synthesis of
O., Yi, T. L., Tang, B., Miura, O., and Ihle, J. peptides and proteins. Annu. Rev. Biochem. 57,
N. (1993) Jak2 associates with the erythropoi- 957–989.
etin receptor and is tyrosine-phosphorylated 45. Carpino, L. A. (1993) 1-Hydroxy-7-
and activated following stimulation with eryth- Azabenzotriazole – an efficient peptide coupling
ropoietin. Cell 74, 227–236. additive. J. Am. Chem. Soc. 115, 4397–4398.
33. Dawson, P. E., and Kent, S. B. H. (2000) 46. Glover, K. J., Martini, P. M., Vold, R. R., and
Synthesis of native proteins by chemical liga- Komives, E. A. (1999) Preparation of insoluble
tion. Annu. Rev. Biochem. 69, 923–960. transmembrane peptides: Glycophorin-A, prion
356 M. Itaya et al.
(110–137), and FGFR (368–397). Anal. A transmembrane helix dimer. Biophys. J. 82,
Biochem. 272, 270–274. 2476–2486.
47. Heukeshoven, J., and Dernick, R. (1982) 58. Tamm, L. K., and Tatulian, S. A. (1997)
Reversed-phase high-performance liquid-chro- Infrared spectroscopy of proteins and peptides
matography of virus proteins and other large in lipid bilayers. Q. Rev. Biophys. 30, 365–429.
hydrophobic proteins in formic-acid contain- 59. Johnson, W. C. (1999) Analyzing protein cir-
ing solvents. J. Chromatogr. 252, 241–254. cular dichroism spectra for accurate secondary
48. Mutter, M., Nefzi, A., Sato, T., Sun, X., Wahl, structures. Proteins 35, 307–312.
F., and Wohr, T. (1995) Pseudo-prolines (Psi- 60. Siminovitch, D. J. (1998) Solid-state NMR
Pro) for accessing inaccessible peptides. Pept. studies of proteins: the view from static 2H
Res. 8, 145–153. NMR experiments. Biochem. Cell Biol. 76,
49. Arnau, J., Lauritzen, C., Petersen, G. E., and 411–422.
Pedersen, J. (2006) Current strategies for the 61. Ying, W. W., Irvine, S. E., Beekman, R. A.,
use of affinity tags and tag removal for the puri- Siminovitch, D. J., and Smith, S. O. (2000)
fication of recombinant proteins. Protein Expr. Deuterium NMR reveals helix packing interac-
Purif. 48, 1–13. tions in phospholamban. J. Am. Chem. Soc.
50. Kapust, R. B., and Waugh, D. S. (1999) 122, 11125–11128.
Escherichia coli maltose-binding protein is 62. Sharpe, S., Barber, K. R., Grant, C. W. M.,
uncommonly effective at promoting the solu- Goodyear, D., and Morrow, M. R. (2002)
bility of polypeptides to which it is fused. Organization of model helical peptides in lipid
Protein Sci. 8, 1668–1674. bilayers: Insight into the behavior of single-
51. Nallamsetty, S., and Waugh, D. S. (2006) span protein transmembrane domains. Biophys.
Solubility-enhancing proteins MBP and NusA J. 83, 345–358.
play a passive role in the folding of their fusion 63. Siminovitch, D. J., Ruocco, M. J., Olejniczak,
partners. Protein Expr. Purif. 45, 175–182. E. T., Das Gupta, S. K., and Griffin, R. G.
52. Korepanova, A., Gao, F. P., Hua, Y. Z., Qin, H. (1988) Anisotropic 2H-nuclear magnetic reso-
J., Nakamoto, R. K., and Cross, T. A. (2005) nance spin-lattice relaxation in cerebroside- and
Cloning and expression of multiple integral phospholipid-cholesterol bilayer membranes.
membrane proteins from Mycobacterium tuber- Biophys. J. 54, 373–381.
culosis in Escherichia coli. Protein Sci. 14, 64. Bloom, M., and Smith, I. C. P. (1985)
148–158. Manifestations of lipid-protein interactions in
53. Amor-Mahjoub, M., Suppini, J. P., Gomez- deuterium NMR, in Progress in Protein-Lipid
Vrielyunck, N., and Ladjimi, M. (2006) The Interactions (Watts, A. & De Pont, J. J. H. H.
effect of the hexahistidine-tag in the oligomer- M., Eds.) pp 61–88, Elsevier, Amsterdam.
ization of HSC70 constructs. J. Chromatogr. B 65. Krueger-Koplin, R. D., Sorgen, P. L., Krueger-
Analyt. Technol. Biomed. Life Sci. 844, Koplin, S. T., Rivera-Torres, A. O., Cahill, S.
328–334. M., Hicks, D. B., Grinius, L., Krulwich, T. A.,
54. Stols, L., Gu, M. Y., Dieckman, L., Raffen, R., and Girvin, M. E. (2004) An evaluation of
Collart, F. R., and Donnelly, M. I. (2002) A detergents for NMR structural studies of mem-
new vector for high-throughput, ligation-inde- brane proteins. J. Biomol. NMR 28, 43–57.
pendent cloning encoding a tobacco etch virus 66. Nallamsetty, S., and Waugh, D. S. (2007) A
protease cleavage site. Protein Expr. Purif. 25, generic protocol for the expression and purifi-
8–15. cation of recombinant proteins in Escherichia
55. Lew, S., and London, E. (1997) Simple proce- coli using a combinatorial His6-maltose bind-
dure for reversed-phase high-performance liq- ing protein fusion tag. Nat. Protoc. 2,
uid chromatographic purification of long 383–391.
hydrophobic peptides that form transmem- 67. Karas, M., Bachmann, D., Bahr, U., and
brane helices. Anal. Biochem. 251, 113–116. Hillenkamp, F. (1987) Matrix-assisted ultravio-
56. Fleming, K. G., Ackerman, A. L., and let laser desorption of non-volatile compounds.
Engelman, D. M. (1997) The effect of point Int. J. Mass Spectrom. Ion Process. 78, 53–68.
mutations on the free energy of transmembrane 68. Bollhagen, R., Schmiedberger, M., and Grell,
alpha-helix dimerization. J. Mol. Biol. 272, E. (1995) High-Performance Liquid-
266–275. Chromatographic Purification of Extremely
57. Smith, S. O., Eilers, M., Song, D., Crocker, E., Hydrophobic Peptides – Transmembrane
Ying, W. W., Groesbeek, M., Metz, G., Ziliox, Segments. J. Chromatogr. A 711, 181–186.
M., and Aimoto, S. (2002) Implications of 69. Sato, T., Kawakami, T., Akaji, K., Konishi, H.,
threonine hydrogen bonding in the glycophorin Mochizuki, K., Fujiwara, T., Akutsu, H., and
18 Synthesis, Purification, and Characterization of Single Helix Membrane¼ 357
Aimoto, S. (2002) Synthesis of a membrane 71. Goto, N. K., and Kay, L. E. (2000) New
protein with two transmembrane regions. J. developments in isotope labeling strategies for
Pept. Sci. 8, 172–180. protein solution NMR spectroscopy. Curr.
70. Gardner, K. H., and Kay, L. E. (1998) The use Opin. Struct. Biol. 10, 585–592.
of 2H, 13C, 15N multidimensional NMR to study 72. Gennis, R. B. (1989) Membrane Dynamics and
the structure and dynamics of proteins. Annu. Protein-Lipid Interactions, in Biomembranes pp
Rev. Biophys. Biomol. Struct. 27, 357–406. 166–198, Springer-Verlag, New York.
Chapter 19
Abstract
A first step toward the analysis of the structure, dynamics, and interactions of proteins by NMR is obtain-
ing an acceptable level of resonance assignments. This process is nontrivial in most eukaryotic kinases given
their size and suboptimal behavior in solution. Using inactive ERK2 as a representative example, we
describe the procedures we utilized to achieve a significant degree of completeness of backbone resonance
assignment.
Key words: MAP kinase, ERK2, TROSY, Backbone resonance assignment, Selective labeling, Spin-
labeled ATP
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_19, © Springer Science+Business Media, LLC 2012
359
360 A. Piserchio et al.
2. Assignment
of Backbone
Resonances
for Full-Length The size of ERK2 (42 kDa), its tendency for nonspecific aggrega-
Inactive ERK2 tion at concentrations above approximately 200 mM, and dynamics
on the slow to intermediate timescale leading to line-broadening
effects make the assignment of backbone resonances quite difficult.
Given these issues, the resonance assignment procedure for full-
length inactive ERK2 is described below in some detail.
2.1. Standard For inactive ERK2 (referred to hereon forward as ERK2), TROSY-
Triple-Resonance based experiments (10, 16, 17) consistently displayed significantly
Experiments narrower line widths, even at fields as low as 600 MHz, and were
therefore preferred over their non-TROSY counterparts. However,
for samples prepared in D2O-based media, the slow back exchange
of several well-protected amide groups resulted in reduced sensi-
tivity in backbone-directed NMR experiments, complicating their
analysis. This problem is evident when an 15N-, 1H-TROSY spec-
trum of perdeuterated ERK2 (prepared in a D2O-based medium)
was compared to that of a sample prepared from cells grown in a
H2O-based medium supplemented with uniformly 2H-, 15N-labeled
amino acids. The latter spectrum displayed additional sets of reso-
nances not visible in the former. Incomplete back exchange com-
plicates resonance assignment both by restricting the number of
detectable resonances, and limiting the degree of correlations
obtained (complicating the so-called backbone walk) for unambigu-
ous correspondences between the resonances that are observed.
This complication adds to the problem of the overall quality of the
triple-resonance experiments being poor presumably because of the
aforementioned aggregation phenomenon and dynamics. A TROSY-
HNCO experiment collected at 800 MHz exhibits around 65–70%
of the expected peaks. Clearly, this represents the upper limit for
19 Assignment of Backbone Resonances in a Eukaryotic Protein Kinase – ERK2... 361
2.2. Use of Predicted Several crystal structures of ERK2 can be found in the PDB, in
Chemical Shifts both the inactive (19) and dual-phosphorylated (on T183 and
Y185) active (20) forms. A simple way to take advantage of the
available structural data is to utilize them to predict the protein
NMR chemical shifts. This can be done with reasonable accuracy
for 13C resonances, as long as the crystal structure reflects the struc-
ture in solution. Toward this purpose, we used the software Sparta
(21), freely available from the Bax group. For well-ordered regions,
like sheets and helices, these predictions are expected to be more
accurate than for loops and regions of noncanonical secondary
structure. Discrepancies may also be introduced in highly struc-
tured areas by the presence (or by the lack of) ligands or by inter-
molecular interactions either in solution (aggregation) or in crystallo
(crystal-packing forces). The latter scenario is especially true in flex-
ible and highly dynamic molecules (22, 23) as the protein kinases
are known to be. It should be noted that in most of the crystal
structures of ERK2 available in the PDB it is in complex with a
variety of ligands. In addition, the measured chemical shift values
362 A. Piserchio et al.
2.3. Use of Structural Clearly, the assignments obtained using predicted chemical shifts
Information should not be considered reliable until confirmed using more con-
ventional experimental spectroscopy-based approaches. An obvi-
ous way to accomplish is to take advantage of the known
three-dimensional structure of ERK2, and use the internuclear dis-
tances available from them for comparison with cross peaks between
amide protons that appear in a three-dimensional 15N-edited
NOESY-TROSY experiment. The verification of the existence (or
the absence) of specific cross peaks predicted from the crystal struc-
ture is an effective way to validate assignments, especially for
b-strands and loops. In the case of b-sheets, the expected (and
observed) NOEs are mainly long range, allowing confirmation of
sequentially nonproximal stretches of residues that comprise indi-
vidual strands of a b-sheet that have been independently assigned.
The reproduction of proper patterns of internuclear distances from
incorrectly assigned resonances is highly unlikely. In case of loops,
generally only few specific residues would be expected to be well-
structured and generate amide–amide NOEs, so again the observed
NOE pattern can be used to confirm a tentative assignment. For
helices, however, the expected NOEs are mostly short range and
only involve amino acids in the particular helical segment, so they
19 Assignment of Backbone Resonances in a Eukaryotic Protein Kinase – ERK2... 363
2.4. Use of Selective Selective amino acid labeling represents another route to aid in the
Labeling Strategies linking of neighboring spin systems and to fill gaps in assignments
when the information content of the triple-resonance experiments,
especially in the Cb region, is poor. Usually, selective labeling (25)
is performed by supplementing the M9 medium with unlabeled
(14 N) ammonium chloride, 1H-12C glucose, or similar nutrients
(sometimes, LB is used directly (26)), a particular 15N-labeled (15N,
12
C, 1H) amino acid, and sometimes an unlabeled (14N, 12C, 1H)
pool of the remaining amino acids. Then, a simple 15N, 1H HSQC
experiment should highlight the amide resonances belonging to
the residue type selected. This method can also be more rigorously
applied using E. coli strains that are auxotrophic for the specific
amino acid to be labeled (27). Unfortunately, this labeling approach
did not perform well when applied to ERK2. Independently of the
specific amino acid tested, the resulting HSQC spectrum lacked
discernable peaks. We attributed this problem to extensive line
broadening resulting from efficient 1H–1H relaxation in the absence
of deuteration. Reducing the overall 1H density by growing the
bacteria in D2O did not significantly improve the quality of the
spectra suggesting that this was the result of the contribution of
the local dipolar interactions between the amide and alpha protons
of the selectively labeled amino acids. This phenomenon leads to
an increase in the contribution of the 1H homonuclear R1 to the
relaxation rate of the antiphase term between the amide 15N and
1
H nuclei, and results in a broadening of the resonances in an 15N,
1
H-HSQC experiment. We then decided to alter our selective
labeling approach and use amino acids selectively 13C-labeled only
at the carbonyl position (14N, 12C, 13CO,1H) in a uniformly
15
N-labeled, deuterated background. As shown by Takeuchi and
364 A. Piserchio et al.
Fig. 1. (a) Structure of ERK2 with the N- and C-terminal lobes colored light and dark grey, respectively. The MAP kinase
insert and the C-terminal extension are colored black. Side chains for the regulatory T183 and Y185 residues are shown
and labeled. Side chains for the tight turn encompassing residues T92-M96 are shown on the structure and expanded on
the right panel. (b) Strips taken from an 15N-edited NOESY-TROSY spectrum collected with a 400-ms mixing time at
800 MHz on a uniformly 2H-, 13C-, 15N-labeled inactive ERK2 sample in a buffer containing 150 mM NaCl, 2 mM DTT, 10 mM
MgCl2, 2 mM ADP, 50 mM phosphate, pH 6.8, 10% 2H2O. Shown here is the effect of spin diffusion generating long-range
connections among the amides of the segment comprising residues T92-M96. The lines highlight the total correlation-like
(as in a TOCSY experiment, where transfer occurs through scalar rather than dipolar couplings) effect of the magnetization
transfer. The source (first label) and target (last label) amide 1HN nuclei for the cross peaks are labeled. Only a single label
is used for the diagonal peaks.
Fig. 2. 13CO, 1H planes for TROSY-based HNCO spectra for representative examples (Leu, Ala) of residue-selective
13
CO-labeled samples of ERK2 in a uniformly 15N-labeled, perdeuterated background. Also shown in the extreme left panel
is the corresponding plane from uniformly 13C-, 2H-, 15N-labeled ERK2. The labels correspond to the residue that contributes
the 13CO nucleus (i.e., the i − 1 residue).
2.5. Use of Spin- Like all protein and indeed nonprotein kinases, ERK2 binds ATP
Labeled ATP Analogs and ADP. However, the chemical shift perturbations induced by
binding of these molecules (or corresponding slowly hydrolyzed
analogs) are not limited to the ATP-binding pocket; therefore, the
shifts of unknown resonances can be difficult to correlate to a spe-
cific portion of the structure simply by monitoring chemical shift
perturbations. It has already been shown for PKA that spin-labeled
ATP (sl-ATP) molecules can be successfully employed to highlight
those residues within a certain distance from the nucleotide binding
pocket (18). The sl-ATP we employed, sl-N3-ATP (a kind gift from
Dr. Pia Vogel, SMU), carries a stable nitroxide spin label as part of a
2,2,5,5 tetramethyl 3-pyrroline scaffold attached to the 3¢ (70–80%)
or 2¢ (20–30%) positions of the ribose moiety (29). A crystal struc-
ture of ERK2 bound to this specific ligand does not exist; therefore,
366 A. Piserchio et al.
3. Conclusions
Acknowledgments
This research has been supported by the following grants from the
National Institutes of Health: GM084278 (to RG), GM059802
(to KND), and 5G12 RR03060 (toward partial support of the
NMR facilities at The City College of New York). RG is a member
of the New York Structural Biology Center, NYSTAR facility.
KND is a recipient of a grant from the Welch Foundation (F-1390).
The authors thank Dr. Pia Vogel (SMU) for the kind gift of spin-
labeled ATP.
References
1. Murphy, L. O., and Blenis, J. (2006) MAPK pathway: application as anticancer drugs. Prog.
signal specificity: the right place at the right Cell Cyc. Res. 5, 219–224.
time. Trends Biochem. Sci. 31, 268–275. 7. Roux, P. P., and Blenis, J. (2004) ERK and p38
2. Chen, Z., Gibson, T. B., Robinson, F., Silvestro, MAPK-activated protein kinases: a family of
L., Pearson, G., Xu, B., Wright, A., Vanderbilt, protein kinases with diverse biological func-
C., and Cobb, M. H. (2001) MAP kinases. tions. Microbiol. Mol. Biol. Rev. 68, 320–344.
Chem. Rev. 101, 2449–2476. 8. Piserchio, A., Dalby, K. N., and Ghose, R.
3. Pearson, G., Robinson, F., Beers Gibson, T., (2012) Expression and Purification of Src-
Xu, B. E., Karandikar, M., Berman, K., and family Kinases for Solution NMR Studies.
Cobb, M. H. (2001) Mitogen-activated pro- Meth. Mol. Biol. 831, 111–132.
tein (MAP) kinase pathways: regulation and 9. Sattler, M., Schleucher, J., and Griesinger, C.
physiological functions. Endocrine Rev. 22, (1999) Heteronuclear multidimensional NMR
153–183. experiments for the structure determination of
4. Fang, J. Y., and Richardson, B. C. (2005) The proteins in solution employing pulsed field gra-
MAPK signalling pathways and colorectal can- dients. Prog. NMR Spectr. 34, 93–158.
cer. The Lancet Oncol. 6, 322–327. 10. Salzmann, M., Pervushin, K., Wider, G., Senn,
5. Kohno, M., and Pouyssegur, J. (2006) H., and Wuthrich, K. (1998) TROSY in tri-
Targeting the ERK signaling pathway in cancer ple-resonance experiments: new perspectives
therapy. Annal. Med. 38, 200–211. for sequential NMR assignment of large pro-
6. Kohno, M., and Pouyssegur, J. (2003) teins. Proc. Natl. Acad. Sci. USA 95,
Pharmacological inhibitors of the ERK signaling 13585–13590.
368 A. Piserchio et al.
11. Masterson, L. R., Mascioni, A., Traaseth, N. J., 21. Shen, Y., and Bax, A. (2007) Protein backbone
Taylor, S. S., and Veglia, G. (2008) Allosteric chemical shifts predicted from searching a data-
cooperativity in protein kinase A. Proc. Natl. base for torsion angle and sequence homology.
Acad. Sci. USA 105, 506–511. J. Biomol. NMR 38, 289–302.
12. Masterson, L. R., Cheng, C., Yu, T., Tonelli, 22. Fushman, D., Xu, R., and Cowburn, D. (1999)
M., Kornev, A., Taylor, S. S., and Veglia, G. Direct determination of changes of interdo-
(2010) Dynamics connect substrate recogni- main orientation on ligation: use of the orien-
tion to catalysis in protein kinase A. Nature tational dependence of 15 N NMR relaxation in
Chem. Biol. 6, 821–828. Abl SH(32). Biochemistry 38, 10225–10230.
13. Wiesner, S., Wybenga-Groot, L. E., Warner, N., 23. Piserchio, A., Nair, P. A., Shuman, S., and
Lin, H., Pawson, T., Forman-Kay, J. D., and Ghose, R. (2010) Solution NMR studies of
Sicheri, F. (2006) A change in conformational Chlorella virus DNA ligase-adenylate. J. Mol.
dynamics underlies the activation of Eph recep- Biol. 395, 291–308.
tor tyrosine kinases. EMBO J. 25, 4686–4696. 24. Wüthrich, K. (1986) NMR of proteins and
14. Vajpai, N., Strauss, A., Fendrich, G., Cowan- nucleic acids, John Wiley and Sons, New York.
Jacob, S. W., Manley, P. W., Grzesiek, S., and 25. Muchmore, D. C., McIntosh, L. P., Russell, C.
Jahnke, W. (2008) Solution conformations and B., Anderson, D. E., and Dahlquist, F. W. (1989)
dynamics of ABL kinase-inhibitor complexes Expression and nitrogen-15 labeling of proteins
determined by NMR substantiate the different for proton and nitrogen-15 nuclear magnetic
binding modes of imatinib/nilotinib and dasa- resonance. Meth. Ezymnol. 177, 44–73.
tinib. J. Biol. Chem. 283, 18292–18302.
26. Englander, J., Cohen, L., Arshava, B., Estephan,
15. Vogtherr, M., Saxena, K., Hoelder, S., Grimme, R., Becker, J. M., and Naider, F. (2006)
S., Betz, M., Schieborr, U., Pescatore, B., Selective labeling of a membrane peptide with
Robin, M., Delarbre, L., Langer, T., Wendt, K. 15
N-amino acids using cells grown in rich
U., and Schwalbe, H. (2006) NMR character- medium. Biopolymers 84, 508–518.
ization of kinase p38 dynamics in free and
ligand-bound forms. Angew. Chem. Intl. Ed. 27. LeMaster, D. M., and Cronan, J. E., Jr. (1982)
Engl. 45, 993–997. Biosynthetic production of 13 C-labeled amino
acids with site-specific enrichment. J. Biol.
16. Riek, R., Pervushin, K., and Wuthrich, K. Chem. 257, 1224–1230.
(2000) TROSY and CRINEPT: NMR with
large molecular and supramolecular structures 28. Takeuchi, K., Ng, E., Malia, T. J., and Wagner,
in solution. Trends Biochem. Sci. 25, 462–468. G. (2007) 1-13 C amino acid selective labeling
in a 2H15N background for NMR studies of
17. Pervushin, K. (2000) Impact of transverse
large proteins. J. Biomol. NMR 38, 89–98.
relaxation optimized spectroscopy (TROSY)
on NMR as a technique in structural biology. 29. Vogel-Claude, P., Schafer, G., and Trommer,
Q. Rev. Biophys. 33, 161–197. W. E. (1988) Synthesis of a photoaffinity-spin-
labeled derivative of ATP and its first applica-
18. Langer, T., Vogtherr, M., Elshorst, B., Betz,
tion to F1-ATPase. FEBS Lett. 227, 107–109.
M., Schieborr, U., Saxena, K., and Schwalbe,
H. (2004) NMR backbone assignment of a 30. Prowse, C. N., and Lew, J. (2001) Mechanism
protein kinase catalytic domain by a combina- of activation of ERK2 by dual phosphoryla-
tion of several approaches: application to the tion. J. Biol. Chem. 276, 99–103.
catalytic subunit of cAMP-dependent protein 31. Li, Y., and Palmer, A. G., III. (2010) Narrowing
kinase. ChemBioChem 5, 1508–1516. of protein NMR spectral lines broadened by
19. Zhang, F., Strand, A., Robbins, D., Cobb, M. chemical exchange. J. Am. Chem. Soc. 132,
H., and Goldsmith, E. J. (1994) Atomic 8856–8857.
structure of the MAP kinase ERK2 at 2.3 Å 32. Piserchio, A., Warthaka, M., Devkota, A. K.,
resolution. Nature 367, 704–711. Kaoud, T. S., Lee, S., Abramczyk, O., Ren, P.,
20. Canagarajah, B. J., Khokhlatchev, A., Cobb, Dalby, K. N., and Ghose R. (2011) Solution
M. H., and Goldsmith, E. J. (1997) Activation NMR insights into docking interactions
mechanism of the MAP kinase ERK2 by dual involving inactive ERK2. Biochemistry, 50,
phosphorylation. Cell 90, 859–869. 3660–3672.
Chapter 20
Abstract
Electrostatic interactions at the protein–aqueous interface modulate the reactivity of solvent-exposed backbone
amides by a factor of at least a billion fold. The brief (~10 ps) lifetime of the peptide anion formed during
the hydroxide-catalyzed exchange reaction helps enable the experimental rates to be robustly predictable
by continuum dielectric methods. Since this ability to predict the structural dependence of exchange reac-
tivity also applies to the protein amide hydrogens that are only rarely exposed to the bulk solvent phase,
electrostatic analysis of the experimental exchange rates provides an effective assessment of whether a given
model ensemble is consistent with the properly weighted Boltzmann conformational distribution of the
protein native state.
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_20, © Springer Science+Business Media, LLC 2012
369
370 G. Hernández et al.
2. Steric
Interpretation
of Protein
Hydrogen Before the first protein X-ray structure was reported, Linderstrøm-
Exchange Lang and colleagues (1) described the so-called EX2 analysis of
hydrogen exchange from structurally buried backbone amides, as
2.1. Hydrogen summarized in the following kinetic scheme:
Exchange as a kop kch
Measure of Solvent closed open → exchanged
kcl
Accessibility
If the rate of the closing reaction is rapid compared to the open
state chemical exchange step (i.e., kcl > > kch), a preequilibrium of
the open and closed conformational states is established and the
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 371
2.0 E53
S25
1.0
log kex (s–1)
K2
0.0 K3 D14
D35 K46
D36
I12
K29
S47
D21 K51
–1.0
V38*
Fig. 1. Magnetization transfer-based hydrogen exchange rate measurements on the solvent-exposed amides of P. furiosus
A2K rubredoxin. CLEANEX-PM [71, 72] measurements were carried out at 25°C. Dashed lines with slope of 1.0 were
drawn for the pH dependent data of each solvent-exposed amide, indicating a simple hydroxide ion dependence on the
exchange rates over most of the pH range. The exchange rate value for Val 38, marked with an asterisk, is derived by
extrapolation from measurements at 52°C. Reprinted from ref. 22 with permission from the American Chemical Society.
β1 β2 α β3 β4 β5
0
–1
Fig. 2. The fraction of conformations in which the backbone amide hydrogen is predicted
to be exposed to solvent for each residue of ubiquitin. Estimations based on protection
factor analysis [3, 23] of hydrogen exchange measurements [20], normalized to model
peptide values, are indicated (filled circle). Illustrated as well is the fraction of conforma-
tions in the 2NR2 (filled triangle) and 2K39 (filled inverted triangle) NMR-restrained
ensembles for which the solvent accessibility of the amide hydrogen is greater than
0.5 Å2. The position of the secondary structure elements of ubiquitin are indicated along
the top of the figure. Reprinted from ref. 33 with permission from Elsevier Limited.
3. Kinetics
and Electrostatics
of Hydrogen
Exchange Hydroxide-catalyzed amide hydrogen exchange is a straightforward
acid-base reaction. A number of researchers (36–41) have pointed
3.1. Implications out that electrostatic interactions modulate the kinetics of amide
of Amides Being Weak
hydrogen exchange. Nevertheless, many of the earlier reported
electrostatic effects have appeared to be relatively modest when
Normal Eigen Acids
compared with the 107 to 108-fold decrease in exchange rates,
which is commonly observed for the most slowly exchanging
amides of moderately stable proteins. Consistent with such an
assessment, it has been argued that titrating the formal charges of
the side chains modulates the observed hydrogen exchange rates
much more strongly via an indirect effect on protein stability than
they do via a direct electrostatic interaction (42).
As first predicted by Eigen (43), amides have been experimen-
tally demonstrated (44, 45) to act as normal Eigen acids such that
the reaction rate with hydroxide ions is attenuated from the diffu-
sion limit by the fraction of forward-reacting encounters Ki/
(Ki + 1), where Ki is the equilibrium constant for the transfer of a
proton from the amide to an hydroxide ion. Therefore, the ther-
modynamic acidity of an amide directly predicts its kinetic acidity
as monitored by the hydrogen exchange reaction. Most all protein
backbone amides have appreciably lower thermodynamic acidities
than that of water. As a result, nearly every collision with a neutral
water molecule will quench the peptide anion charge state. This
low acidity implies that near neutral pH, most backbone amides
will be in the peptide anion state at a fractional population of less
than one part in 1010.
The key advantage in predicting the ionization behavior of the
backbone amide, as compared to predicting ionization of protein
side chains, stems from the short lifetime of the peptide anion (22,
46, 47). In contrast to the μs–ms lifetimes for the charge states of
the ionizable side chains near neutral pH, the range of protein
conformational responses to the peptide anion charge state is
strongly limited by its brief lifetime. Although a direct measure-
ment of how rapidly the peptide anion is quenched by a neutral
water molecule has not been reported, NMR relaxation studies
indicate that the residence lifetime of an hydroxide ion in water
is ~ 5 ps (48), and lifetimes near 10 ps have often been observed
for photoactivated strong acids and bases (49, 50). Given that the
dominant phase of the Debye dielectric relaxation profile for water
has a time constant of 8 ps at 25°C (51), it has been argued that
the dynamics of water reorientation are limiting in these fast pro-
ton transfer reactions (49, 50). By analogy, the lifetime of the
peptide anion intermediate is likewise anticipated to be ~10 ps
(22, 46, 47).
376 G. Hernández et al.
3.2. Electronic As long discussed in electron transfer theory (52), dielectric shielding
Polarizability is frequency dependent. The lifetime of a transient charge state
in the Dielectric determines the range of conformational motions that can give rise to
Shielding effective dielectric shielding since conformational transitions that are
of the Peptide Anion slower than the charge state lifetime cannot adjust rapidly enough to
stabilize that state. As a result, the dielectric shielding of the hydro-
gen exchange reaction that arises from the protein molecule is
expected to be dominated by electronic polarizability (47). Owing
to the highly transient peptide anion charge state, the kinetics of
amide hydrogen exchange provides a “snapshot” of the Boltzmann
conformational distribution that is nearly independent from the
dynamics of interchange between protein conformations.
The electrostatic free energy of a generalized-Born ion of
charge Q and radius R is given by the formula (53):
O R O R H O R H
OH– OH–
N N N
N N N
H O R H O R O R
DV
Fig. 3. Electrostatic free energy of peptide ionization. The reaction of hydroxide ion forms
a peptide anion at one or another site along the protein backbone. The differential electro-
static free energy for these two species is given by the product of the charge and the
difference in electrostatic potential for the two sites ΔeV which, in turn, is proportional to
the ΔpK for these two amide nitrogens. Reprinted from ref. 22 with permission from the
American Chemical Society.
4. Hydrogen
Exchange
Techniques
Rapid hydrogen exchange can be monitored by magnetization
4.1. Magnetization transfer techniques in which the water resonance is selectively
Transfer Methods excited. The NMR experiment then monitors the transfer of this
magnetization to the amide resonances. A particularly robust
implementation of magnetization transfer-based hydrogen
exchange monitoring is that of CLEANEX-PM (70, 71), in which
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 379
4.2. Solvent Exchange In the case of ubiquitin, exchange kinetics for the amides that do
by 1H Exchange-In not exhibit exchange in the CLEANEX-PM experiments have been
Protocol analyzed using an 1H exchange-in protocol (20). By preexchange
of the amide hydrogen positions with deuterium and then dissolu-
tion of the protein sample in a 1H2O-containing buffer, one can
circumvent the complications from the isotope dependence of sol-
vent, buffer and protein side chain ionizations that plague quanti-
tative interpretation of exchange rates measured using the
conventional 2H exchange-in protocol. Furthermore, no correc-
tion is needed for the significant differences in protein stability that
can result from comparing measurements in normal and heavy
water buffer solutions (75).
When compared to the magnetization transfer-based hydro-
gen exchange measurements (47), the 1H exchange-in protocol
suffers mainly from the differential effect in breakage of an N–D or
an N–H amide bond. Measurements on poly d,l-alanine indicate a
0.08 shift in the log rate constant for this isotope effect in the
hydroxide-catalyzed exchange reaction (76). An additional benefit
380 G. Hernández et al.
5. Continuum
Dielectric Analysis
for Exchange of
Static Solvent Hydroxide-catalyzed exchange rate constants were determined for
Accessible Protein those amides of rubredoxin, FK506-binding protein (FKBP12),
Amides ubiquitin and chymotrypsin inhibitor 2 (CI2) that are solvent-
accessible in the high-resolution X-ray structures (22, 47). The
5.1. Electrostatic
acidity of these amides were calculated using the Poisson–Boltzmann
finite difference algorithm DelPhi (78) as a function of the nonpo-
Parameter Set
larizable electrostatic parameter set, the internal dielectric value
Dependence of Peptide
and the charge distribution of the peptide anion. As illustrated in
Acidity Predictions
Fig. 4, the best performance was obtained using the CHARMM22
electrostatic atomic partial charge and radius parameters (79)
(these parameters are preserved in the current CHARMM27 force
field), an ab initio-derived peptide anion charge distribution (47),
and an internal dielectric value of 3. These parameters yielded an
rmsd value of 7 for the 56 amide exchange rate constants ranging
from 100.67 to 109.0 M−1 s−1. The optimal internal dielectric value
was obtained via its (1/eint) scaling effect on the differences in elec-
trostatic potential for the various peptide anions predicted by the
Poisson–Boltzmann calculations and linear correlation against the
experimental hydrogen exchange rates.
The OPLS-AA electrostatic parameter set (80) yielded compa-
rably robust predictions, as might be expected from its strong simi-
larity to the CHARMM atomic charge and radii set. By contrast,
the nonpolarizable AMBER parm99 (81) and AMBER ff03 (82)
parameter sets performed more poorly. As illustrated in Fig. 5, the
parm99 electrostatic parameters from the AMBER force field do
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 381
Fig. 4. Dependence of amide acidity predictions on the atomic charge distribution of the
peptide anion. Protein amide pK values predicted using CHARMM22 atomic charge and
radius parameters [79] and an internal dielectric constant of three at 25°C, with the
excess anion charge density distributed throughout the peptide unit as predicted from
B3LYP DFT calculations [47]. Reprinted from ref. 47 with permission from the American
Chemical Society.
sets that do not incorporate large shifts in the atomic charges of the
backbone atoms, indicates that the magnitude of charge migration
within the individual amino acids that is modeled into the AMBER
parm99 electrostatic parameter set appears to be well beyond what
might be needed to rationalize local sequence-dependent varia-
tions. These considerations apply even more markedly to calcula-
tions using the AMBER ff03 electrostatic parameters (47).
5.2. Atomic Charge Our initial rubredoxin hydrogen exchange predictions assumed
Distribution that the excess negative charge of the peptide anion resides exclu-
in the Peptide Anion sively on the nitrogen, following the earlier results from continuum
dielectric modeling of hydrogen exchange in simple peptides by
McCammmon and colleagues (41). That assumption conflicts with
the long-standing tradition of representing the product formed by
deprotonation of an amide as an imidate anion. However, in con-
trast to predictions from early valence bond theory studies, there
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 383
Fig. 6. Dependence of amide acidity predictions on the atomic charge distribution of the
peptide anion. Protein amide pK values predicted using CHARMM22 electrostatic param-
eters and an internal dielectric constant of 3 at 25°C, with the excess anion charge den-
sity localized to the carbonyl oxygen. Reprinted from ref. 47 with permission from the
American Chemical Society.
5.3. Dominant Acidic Although these continuum dielectric calculations were based on
Conformer Analysis high resolution X-ray structures, a discrete set of adjustments was
applied to several side chain types during the calculation of the
intraresidue peptide acidity. The most significant case involved
aspartate side chains in which the χ1 side chain torsion angle is
gauche to the backbone nitrogen. This orientation places the neg-
atively charged carboxylate near the intraresidue amide, thus
strongly suppressing its predicted ionization. The sterically unhin-
dered rotation of an Asp carboxylate to a trans rotamer can
enhance the acidity of the intraresidue amide by 5 pH units or
more (22, 35, 47).
A second type of systematic modulation in the predicted elec-
trostatic free energy of the peptide anions as a function of the resi-
due side chain conformation was applied when the χ1 side chain
torsion angle is near +60°. In this rotamer, the Cγ is gauche to both
the main chain nitrogen and carbonyl carbon and is often in van
der Waals contact with the amide hydrogen. Unhindered rotation
to another c1 rotamer tends to increase the solvation of the amide
anion with a resultant increase the peptide acidity of that residue.
Regarding the physical validity of these ad hoc side chain rota-
tions for identifying energetically favorable conformations near the
X-ray coordinates that have enhanced peptide acidities, it should
be noted that the model conformation (molecule 92) within the
independently generated 2NR2 ubiquitin ensemble (further dis-
cussed below) that most accurately predicts the experimental
hydrogen exchange has undergone the Asp and gauche+ side chain
rotamer transitions that were identified by this earlier published
side chain reorientation protocol (33).
The assumption of limited protein conformational reorganiza-
tion during the lifetime of the peptide anion surely can not apply
generally to the side chain hydroxyl hydrogens, since the analo-
gous reorientation of the hydrogens on water molecules gives rise
to the dominant dielectric shielding of that phase. Particularly for
side chain hydroxyl hydrogens that are not involved in an intramo-
lecular hydrogen bond, continuum dielectric calculations based on
a fixed orientation are potentially misleading. This is most notably
the case when amide acidity is estimated with an intraresidue serine
or threonine hydroxyl in either a gauche+ or gauche− c1 rotamer.
Given that the exchange rates are similar for serine- and threonine-
containing model peptides, as compared to the alanine reference
(23), the side chain hydroxyl does not generally serve as a catalyst
for peptide hydrogen exchange. Consistent with that observation,
the peptide acidity analyses for serine and threonine residues with
a gauche c1 rotamer assume that the dielectric shielding of the side
chain hydroxyl is equal to that of the equivalent volume of water.
In such cases, the serine side chain is computationally truncated to
alanine, and threonine is tranformed into α-aminobutyrate.
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 385
6. Ensemble
Averaging in
Prediction of
Ubiquitin Hydrogen All model ensembles that are justified by their collective ability to
Exchange predict experimental measurements necessarily invoke the assump-
tion that they represent an accurate Boltzmann sampling of con-
6.1. Population formational space. Yet not all experimental approaches offer
comparable sensitivity to variations in the model distribution. In
Averaging of
contrast to experimental techniques that are equally sensitive to
Conformer Acidities
every protein conformation and thus are generally dominated by
the most populated states, hydrogen exchange reactivity is highly
sensitive to conformation. The fact that structurally buried amides
are effectively unreactive to hydrogen exchange forms the basis for
the widespread application of this experimental technique for
monitoring rare conformational states. On the one hand, the highly
exposed amides exhibit exchange rates that are quite sensitive to
the well-populated protein conformations. On the other hand, this
sensitivity to conformation applies to the rarely exposed amides as
well so that hydrogen exchange measurements for these sites pro-
vide a powerful experimental monitor of both the population and
conformation of the transient exchange-competent state.
Although population averaging of the conformer acidities
(ΣKi) is more formally correct (85), population averaging of the
conformer pKi values (ΣpKi or, equivalently, averaging of the con-
former electrostatic potential values) has occasionally been used to
estimate the effect of conformer sampling in the prediction of pro-
tein side chain ionization. Karplus and colleagues (86) have con-
cluded that, in assessment of the ionization midpoint for each
titrating residue, averaging over the Ki values or averaging over the
log Ki values usually has little effect on the predicted pK values.
The issue of Ki vs. pKi averaging is markedly different in the
estimation of hydroxide-catalyzed hydrogen exchange rates near
neutral pH where, in most cases, less than 1 out of every 1010 mol-
ecules will have a given amide in the ionized state. As a result,
whenever there is a substantial range in conformer acidities, the
most acidic conformers can make the dominant contribution to
the observed hydrogen exchange rate, even if they constitute only
a modest fraction of the overall conformer population.
Molecular simulation techniques have been increasingly
employed to predict the Boltzmann-weighted conformational dis-
tribution of the protein native state. In principle, under the assump-
tion of ergodicity, an unconstrained constant temperature molecular
dynamics simulation can provide the Boltzmann conformational
distribution. In practice, given the roughness of protein energy
landscapes, even simulations extending for hundreds of nanosec-
onds will generally suffer from incomplete conformational sam-
pling. Furthermore, force field parameterizations are only
approximate. As a result, the predicted conformational distribution
can drift away from the physical values.
386 G. Hernández et al.
6.2. Consistency Concerns arising from approximate force fields and incomplete
of Ubiquitin Model sampling have been approached by incorporating experimentally
Ensembles derived restraints into molecular dynamics simulations, as applied
with the Native State to ubiquitin. The MUMO algorithm of Vendruscolo and col-
Conformational leagues (31) introduced NOE-derived distance bound restraints,
Distribution averaged over subsets of protein conformations, as a mechanism
for maintaining the predicted molecular dynamics ensemble distri-
bution to within the neighborhood of the experimentally deter-
mined structure. In parallel, order parameters S2, derived from
backbone 15N and side chain 13C methyl NMR relaxation measure-
ments, were incorporated into the restrained molecular simulation
where they enforce enhanced conformational sampling. The resul-
tant set of 144 protein conformations (PDB code 2NR2 (31))
serves as a model for the random sampling of the native state
Boltzmann distribution of ubiquitin.
In generating an alternate model ensemble (PDB code 2K39),
de Groot and colleagues (32) applied the CONCOORD algorithm
(87) using the same 2,727 NOE constraints from the 1D3Z solu-
tion structure analysis (88) to generate 1,000 model conforma-
tions of ubiquitin. In the EROS (ensemble refinement with
orientational restraints) protocol a subset of 400 conformations
were initially selected as most consistent with the residual dipolar
coupling (RDC) data. An iterative process of simulated annealing
followed by reselection against the RDC data was then applied
until the initial set of 1,000 conformations was winnowed down to
a final set of 116 conformations.
When ensemble averaging of hydrogen exchange reactivity was
applied to the NOE, S2-restrained 2NR2 ubiquitin ensemble (31),
the hydroxide-catalyzed exchange rates for nearly all of the highly
exposed amide hydrogens (solvent-accessible in >50% of confor-
mations) were quite accurately predicted (black circles in Fig. 7)
(20, 33). For 16 of these highly exposed amides (Gly 47 and Asp
52 discussed below), the 2NR2 ensemble predicted the 105-fold
range in experimental rates, yielding an rmsd of 0.51 and a correla-
tion coefficient r = 0.94 for the log kOH− values. This correlation is
markedly better than that obtained using a single crystallographi-
cally derived ubiquitin structure (47).
Most strikingly, for the backbone amides that are exposed to
solvent above 0.5 Å2 in more than one but less than half of the
models in the NMR relaxation-restrained ensemble, with the
exception of Lys 48, the amide pKa predictions are nearly as accu-
rate (rmsd for log kOH− of 0.69) as those for the more highly
exposed sites (Fig. 7). Despite being structurally buried by most
conventional criteria, the exchange rate constants for these 12 resi-
dues, spanning nearly a million-fold range, are predictable to within
a factor of 5.
The underestimation of the hydrogen exchange rates for residues
in which only one model conformation has an amide hydrogen
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 387
10
G47
K48
8
–2
–2 0 2 4 6 8 10
log kOH– (M–1 s–1)
14
16
18
20
pKPB
22
24
26
28
0 10 20 30 40 50 60 70
residue
Fig. 8. The range of conformer acidities for amide hydrogens exposed to solvent by at least 0.5 Å2 in the NMR relaxation-
restrained 2NR2 ensemble of ubiquitin. Residues for which the amide hydrogen is solvent-exposed in less than 50% of
the ensemble models are indicated in gray, while those that are solvent-accessible in more than 50% of the models
are marked in black. The peptide acidities are placed on an absolute scale based on their normal Eigen acid behavior, the
diffusion-limited rate for hydroxide-catalyzed exchange of 2 × 1010 M−1 s−1, and the pK of 15.7 for water at 25°C. These
properties imply an exchange rate constant of 1.0 M−1 s−1 for an amide with a pK value of 26.0 [47]. Reprinted from ref. 20
with permission from the American Chemical Society.
6.3. Comparison Of particular significance is the pattern of exchange rates seen for
Against the Set the proteasome targeting interaction site around residue Lys 48.
of Known Ubiquitin- In explicit contrast to the 2NR2 ubiquitin ensemble of Vendruscolo
Protein Complexes and colleagues (31), de Groot and colleagues (32) contended that
their 2K39 ensemble spans a conformational space that includes
the ubiquitin structures found in all of the available X-ray studies
of ubiquitin-protein complexes (41 complexed-ubiquitin mole-
cules +5 X-ray structures of uncomplexed ubiquitin). These authors
further claimed that conformations of ubiquitin found in these
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 389
10
G47
K48
8
F45
4 I44 D52
0
0 2 4 6 8 10
log kOH– (M–1 s–1)
Fig. 9. Hydroxide-catalyzed rate constants predicted from the NMR residual dipolar coupling-
restrained 2K39 ensemble of ubiquitin. For residues in which the amide hydrogen is
exposed to solvent by more than 0.5 Å2 in at least one ensemble model, conformer acidi-
ties were predicted for all solvent-exposed amides. Each residue is distinguished accord-
ing to whether the amide hydrogen is exposed to solvent by more than 0.5 Å2 in at least
50% of the models (filled circle) or exposed in only a single model (filled square). The other
transiently exposed amides are denoted as (filled diamond ). Asp 52 and the residues
involved in the primary interaction site for proteasome targeting are individually identified.
Reprinted from ref. 20 with permission from the American Chemical Society.
6.4. Hydrogen Ideally, these two ubiquitin ensembles represent ~102 random sam-
Exchange Analysis plings of the Boltzmann conformational distribution so that amides,
as a Monitor which become exposed to solvent at less than a 1% frequency, will
for Completeness generally be unrepresented in these peptide acidity predictions. As
of Ensemble Sampling indicated in Fig. 2, for nearly every case in which an amide hydro-
gen is exposed to solvent in at least one conformation from either
the 2NR2 or 2K39 ensembles, the experimental exchange rate is
less than what would be predicted for a model peptide having the
same fraction of solvent-exposed conformations (only for Thr 9 is
the apparent solvent accessibility estimated from peptide normaliza-
tion significantly above that from both of the ensembles).
The log exchange rate constants for most model peptides are
>8 (23). Hence, one may anticipate that for a proper 1% Boltzmann
sampling of the conformational distribution nearly all backbone
amides having log kOH− values >6 should have solvent-accessible
conformations within that 1% sampling. Indeed, each of the 23
ubiquitin amides that have experimental log kOH− values >6 are
exposed to solvent in at least one conformation in both the 2NR2
and 2K39 ensembles (20). On the contrary, there are some amides
of ubiquitin, which are exposed to solvent in these two ensembles,
that have predicted and observed exchange log rate constants that
are significantly less than 6, reflecting the fact that their exchange-
competent conformations have strongly depressed exchange reac-
tivities. Nevertheless, a number of backbone amides that are solvent
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 391
7. Continuum
Dielectric Analysis
of Hydrogen
Exchange in Model Predicting the experimental side-chain-dependent differential
Peptides exchange rates for conformationally unstructured peptides pres-
ents a rather stringent challenge. Although early studies argued
7.1. Backbone that the differences in exchange rates among various simple model
Conformation
peptides arise from chemical induction effects (44, 93), more
recent electrostatic calculations have indicated that amide hydro-
Dependence of Side
gen exchange rates are strongly dependent upon the relative orien-
Chain Correction
tation of the adjacent peptide groups (41, 94).
Factors for Hydrogen
Experimentally measured hydroxide-catalyzed amide exchange
Exchange
rates for conformationally unstructured alanine peptides are essen-
tially unaffected by the intraresidue substitution of a methionine
side chain, while substitution of a phenylalanine, tyrosine, or tryp-
tophan side chain decreases the exchange rate by approximately
twofold (23). A significantly larger (~fivefold) attenuation of the
exchange rates results from substituting any of the branched side
chains from valine, leucine, or isoleucine. A similar pattern is
obtained when these side chain substitutions are introduced into
the residue preceding the site of amide exchange, although the
magnitude of the variations in exchange rates is approximately four-
fold smaller. As a result of these fairly small effects, a compelling
correlation between predicted and observed side-chain-dependent
hydrogen exchange rate differences for conformationally unstruc-
tured peptides requires substantially more accurate predictions than
have yet been demonstrated in protein studies.
392 G. Hernández et al.
7.2. Dependence Poisson–Boltzmann analysis was carried out utilizing the Protein
of Peptide Acidity Coil Library of Rose and colleagues (101) as a model for the
on Backbone Boltzmann-weighted distribution of the unstructured state. In this
Conformation structural library, protein segments lying outside of regular second-
ary structures were identified from high resolution X-ray analysis.
Generally implicit in the application of these coil libraries is the
assumption that the other forms of long range interactions that are
present in the protein crystal structure do not systematically shift the
average conformational distribution of the individual residue types
away from that of conformationally disordered polypeptides.
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 393
100
N-Acetyl-Ala-Ala-N-Methylamide
*
80
40
20
0
–3.0 –2.0 –1.0 0 1.0 2.0
ΔpK (NMA)
Fig. 10. Peptide acidities of Ala–Ala conformers. The electrostatic potential was calculated
for the central peptide anion (asterisks) formed from blocked peptides derived from the
679 Ala–Ala segments found in the Protein Coil Library [101], utilizing the CHARMM22
atomic charge and radius parameters [79] and an internal dielectric value of 3. The
N-methylacetamide anion was used as reference for the electrostatic potential calcula-
tions. Reprinted from ref. 34 with permission from Elsevier Limited.
180
150
120
90
60
30
psi
0
–30
–60
–90
–120
–150
–180
–180 –150 –120 –90 –60 –30 0
phi
Fig. 11. The backbone conformational distribution of the most acidic and least acidic
N-acetyl-Ala-Ala-N-methylamide conformers in the Protein Coil Library. The (f,ψ) torsion
angle values for the 50 most acidic peptides are plotted in gray, while the values for the
30 least acidic peptides are plotted in black. The N-terminal residues are denoted by cir-
cles and the C-terminal residues by triangles. Dotted lines are used to correlate the N- and
C-terminal residue backbone torsion angles for peptides that do not bridge between the
extended and α conformational regions. None of the most acidic peptides have positive f
torsion angles. Reprinted from ref. 34 with permission from Elsevier Limited.
15
16
17
18
pKPB (Ala-Val)
19
20
21
22
23
23 2
22 21 20 19 18 17 16 15
pKPB (Ala-Ala)
Fig. 12. Relative contributions of side chain and main chain interactions determining the
peptide acidities of N-acetyl-[Ala-Val]-N-methylamide conformers. For each of the
N-acetyl-[Ala-Val]-N-methylamide conformers, the methyl groups were truncated to form
an [Ala-Ala] conformer, and the electrostatic free energy was calculated. The correspond-
ing pairs of amide pK values are denoted by their χ1 side chain rotamer (g− as diamonds,
g+ as squares, and t as circles). Reprinted from ref. 35 with permission from Elsevier
Limited.
7.3. Hydrogen The degree to which the backbone geometry contributes to the
Exchange for Nonpolar predicted conformer acidities can be assessed by analysis of
Side Chains and the Protein Coil Library [Ala-Val] peptide conformations with the
Nonadditivity of valine side chain truncated to alanine. For each of the three c1
Correction Factors rotamers of the valine side chain, the acidity of the central amide is
closely correlated with that of the [Ala-Ala] peptide in the same
backbone geometry (Fig. 12). However, conformers with the
gauche− side chain rotamer, in which both methyl groups are ori-
ented gauche to the backbone nitrogen, are predicted to have
appreciably lower amide acidities (on average ~0.7 pH units).
Throughout the range of amide acidities, differences in the side
chain c1 torsion angle give rise to variations in pK values spanning
~1 pH unit.
For each of the nonpolar N-acetyl-[X-Ala]-N-methylamides
and N-acetyl-[Ala-Y]-N-methylamides, the predicted acidities of
the central amide in the individual conformers of each peptide span
nearly a million-fold range. Nevertheless, population averaging of
the conformer reactivities predicts the standard side-chain-depen-
dent hydrogen exchange correction factors for model peptides to
within a factor of 30% (100.11) with a correlation coefficient r = 0.91
(Fig. 13).
396 G. Hernández et al.
0.2
YA
FA
0 WA MA
LA
IA AA
–0.6
AV
AI
–0.8
–0.8 –0.6 –0.4 –0.2 0 0.2
D log kOH– (exp)
Fig. 13. Predicted and observed nonpolar side-chain-dependent differences in the hydrox-
ide-catalyzed log rate constants for model peptides. Poisson–Boltzmann electrostatic free
energies were calculated for the N-acetyl-[X-Ala]-N-methylamide conformers and
N-acetyl-[Ala-Y]-N-methylamide conformers derived from the Protein Coil Library [101].
Hydrogen exchange rate constants were predicted from the ensemble averaging of the
conformer exchange reactivities and were then compared to the standard experimental
side-chain-dependent hydrogen exchange correction factors [23]. Reprinted from ref. 35
with permission from Elsevier Limited.
7.4. Evidence It should be noted that for the nonpolar side chains illustrated in
for Dielectric Shielding Fig. 13, an assumed internal dielectric value of three provides an
from Conformational optimal correlation between the experimental and predicted pep-
Reorganization in tide acidities. As expected, the slope of this correlation exhibits the
Carboxamide Side anticipated inverse dependence on that dielectric value. Hence,
Chains even in the case of the highly mobile nonpolar peptides, the con-
formational contribution to dielectric shielding of amide ioniza-
tion appears to be severely limited. However, evidence for a modest
contribution from conformational reorganization was observed for
the Asn and Gln side chains. Poisson–Boltzmann calculations on
the coordinates of the Asn and Gln residues in Protein Coil Library
distribution yielded amide reactivity predictions that were appre-
ciably less than the experimentally determined values (open sym-
bols in Fig. 14). In contrast to unhindered rotamer transitions
around the sp3–sp3 bonds that generally occur in the timeframe of
hundreds of picoseconds to nanoseconds (107, 108), the sp3–sp2
hybridization of the carboxamide side chain results in more rapid
dihedral angle transitions due to the lower intrinsic torsional poten-
tial barrier. Quantum mechanical analysis indicates a barrier of only
0.15 kcal/mol for acetamide (109). As a result, within each c1
rotamer state of Asn, extensive sampling of the c2 torsion angle can
potentially occur during the peptide anion lifetime. Given the large
dipole of the side chain carboxamide group, such a bond rotation
can substantially alter the degree of stabilization provided for the
peptide anion.
Calculations were conducted to estimate the magnitude of the
shielding effect from conformational reorganization of the Asn
side chain by assuming rapid averaging around the c2 torsion angle.
Upon deprotonation of the peptide unit, the statistical weighting
398 G. Hernández et al.
0.6
0.4
AN
0.2 NA
–0.2 AQ
–0.4
–0.6
–0.8
–0.8 –0.6 –0.4 –0.2 0 0.2 0.4 0.6
D log kOH– (exp)
Fig. 14. The effect of conformational reorganization within the individual side chain
rotamer states for Asn (χ1) and Gln (χ1 and χ2). Electrostatic free energies were calculated
for the [Asn-Ala], [Ala-Asn], [Gln-Ala], and [Ala-Gln] methylamide conformers derived from
the Protein Coil Library [101]. Hydrogen exchange rate constants were predicted from the
ensemble averaging of these conformer exchange reactivities with (filled circle) and with-
out (open circle) allowance for conformational reorganization of the peptide anions within
each side chain rotamer state. The other data are displayed as given in Fig. 13. Reprinted
from ref. 35 with permission from Elsevier Limited.
7.5. Deviations The Ser, Thr, Cys and His+ intraresidue side chains all accelerate
in Dielectric hydroxide-catalyzed peptide hydrogen exchange (23), consistent
Continuum Modeling with an electron-withdrawing effect from the substituent. However,
of Hydrogen Exchange the challenges facing an adequate modeling of the electrostatic
Arising from Chemical potential for these side chains complicate the deconvolution of an
Induction and additional contribution to peptide acidity arising from chemical
Aspartate Side Chain induction (35).
Interactions With the exception of the Asp residue, the other charged side
chains yield correction factors for model peptide hydrogen
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 399
16
18
pKPB (Ala-Asp)
20
22
24
24 22 20 18 16
pKPB (Ala-Ala)
Fig. 15. The relative contributions of side chain and main chain interactions in determining
the peptide acidities of N-acetyl-[Ala-Asp]-N-methylamide conformers. For each N-acetyl-
[Ala-Asp]-N- methylamide, the carboxylate group was truncated to form an [Ala-Ala] con-
former, and the electrostatic free energy was then calculated. The corresponding pairs of
amide pK values are denoted by their c1 side chain rotamer (g− as diamonds, g+ as
squares, and t as circles). Reprinted from ref. 35 with permission from Elsevier Limited.
8. Future
Directions
Future studies will provide insight into the degree to which the
residual inaccuracies in predictions of hydrogen exchange for either
proteins or model peptides reflect errors in the modeling of the
Boltzmann conformational distribution or rather reflect inadequa-
cies in the electrostatic modeling used to analyze those conforma-
tions. Both of these avenues for improved predictive capability will
require pursuit. However, as recently observed by Senn and Thiel
(110), despite many years of intense research effort, there are as yet
no generally established polarizable biomolecular force fields.
Fortunately, the present studies further demonstrate that, in the
context of dielectric shielding without substantial conformational
reorganization, the classic paradigm of uniform volume polariz-
ability is strikingly robust. As such, continued insights into model-
ing of the Boltzmann conformational distribution from hydrogen
exchange analysis can be anticipated on the basis of the continuum
dielectric representation.
References
1. Berger, A., and Linderstrøm-Lang, K. (1957) exchange protection factors. J. Mol. Biol. 262,
Deuterium exchange of poly-DL-alanine in 756–772.
aqueous solution. Arch. Biochem. Biophys. 69, 7. Wallqvist, A., Smythers, G.W., and Covell,
106–118. D.G. (1997) Identification of cooperative
2. Hvidt, A., and Nielsen, S.O. (1966) Hydrogen folding units in a set of native proteins. Prot.
exchange in proteins. Advances in Protein Sci. 6, 1627–1642.
Chem. 21, 287–386. 8. Bahar, I., Wallqvist, A., Covell, D.G., and
3. Bai, Y.W., Milne, J.S., Mayne, L., and Jernigan, R.L. (1998) Correlation between
Englander, S.W. (1994) Protein stability native-state hydrogen exchange and coopera-
parameters measured by hydrogen exchange. tive residue fluctuations from a simple model.
Proteins: Struct., Funct., Genet. 20, 4–14. Biochemistry 37, 1067–1075.
4. Huyghues-Despointes, B.M.P., Scholtz, J.M., 9. Sheinerman, F.B., and Brooks, C.L. (1998)
and Pace, C.N. (1999) Protein conformational Molecular picture of folding of a small a/ß
stabilities can be determined from hydrogen protein. Proc. Natl. Acad. Sci USA 95,
exchange rates. Nat. Struct. Biol. 6, 910–912. 1562–1567.
5. Li, R., and Woodward, C. (1999) The hydro- 10. Garcia, A.E., and Hummer, G. (1999)
gen exchange core and protein folding. Prot. Conformational dynamics of cytochrome c:
Sci. 8, 1571–1591. Correlation to hydrogen exchange. Prot.
6. Hilser, V.J., and Freire, E. (1996) Structure- Struct. Funct. Genet. 36, 175–191.
based calculations of the equilibrium folding 11. Dixon, R.D.S., Chen, Y., Ding, F., Khare,
pathway of proteins. Correlation with hydrogen S.D., Prutzman, K.C., Schaller, M.D.,
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 401
Campbell, S.L., and Dokholyan, N.V. (2004) 24. Hernández, G., Anderson, J.S., and LeMaster,
New insights into FAK signaling and localiza- D.M. (2008) Electrostatic stabilization and
tion based on detection of a FAT domain fold- general base catalysis in the active site of the
ing intermediate. Structure 12, 2161–2171. human protein disulfide isomerase a domain
12. Livesay, D.R., Dallakyan, S., Wood, G.G., and monitored by hydrogen exchange.
Jacobs, D.J. (2004) A flexible approach for ChemBioChem 9, 768–778.
understanding protein stability. FEBS Lett. 25. Radford, S.E., Buck, M., Topping, K.D.,
576, 468–476. Dobson, C.M., and Evans, P.A. (1992) Hydrogen
13. Freire, E. (1999) The propagation of binding exchange in native and denatured states of hen
interactions to remote sites in proteins: egg-white lysozyme. Proteins 14, 237–248.
Analysis of the binding of the monoclonal 26. Hollien, J., and Marqusee, S. (2002)
antibody D1.3 to lysozyme. Proc. Natl. Acad. Comparison of the folding processes of T.
Sci. USA 96, 10118–10122. thermophilus and E. coli Ribonucleases H. J.
14. Pan, H., Lee, J.C., and Hilser, V.J. (2000) Mol. Biol. 316, 327–340.
Binding sites in Escherichia coli dihydrofolate 27. Bau, R., Rees, D.C., Kurtz-Jr., D.M., Scott,
reductase communicate by modulating the R.A., Huang, H.S., Adams, M.W.W., and
conformational ensemble. Proc. Natl. Acad. Eidsness, M.K. (1998) Crystal-structure of
Sci USA 97, 12020–12025. rubredoxin from Pyrococcus furiosus at 0.95
15. Wrabl, J.O., Larson, S.A., and Hilser, V.J. Angstrom resolution, and the structures of
(2001) Thermodynamic propensities of amino N-terminal methionine and formylmethionine
acids in the native state ensemble: Implications variants of Pf Rd. Contributions of N-terminal
for fold recognition. Prot. Sci. 10, 1032–1045. interactions to thermostability. J. Biol. Inorg.
Chem. 3, 484–493.
16. Wrabl, J.O., Larson, S.A., and Hilser, V.J.
(2002) Thermodynamic environments in pro- 28. LeMaster, D.M., Tang, J., Paredes, D.I., and
teins: Fundamental determinants of fold spec- Hernández, G. (2005) Enhanced thermal sta-
ificity. Prot. Sci. 11, 1945–1957. bility achieved without increased conforma-
tional rigidity at physiological temperatures:
17. Babu, C.R., Hilser, V.J., and Wand, A.J. Spatial propagation of differential flexibility in
(2004) Direct access to the cooperative sub- rubredoxin hybrids. Proteins 61, 608–616.
structure of proteins and the protein ensemble
via cold denaturation. Nat. Struct. Mol. Biol. 29. Lee, B., and Richards, F.M. (1971) The
11, 352–357. Interpretation of Protein Structures: Estimation
of Static Accessibility. J. Mol. Biol. 55, 379–400.
18. Wang, S.W., Gu, J., Larson, S.A., Whitten,
30. Hiller, R., Zhou, Z.H., Adams, M.W.W., and
S.T., and Hilser, V.J. (2008) Denatured-state
Englander, S.W. (1997) Stability and dynam-
energy landscapes of a protein structural data-
ics in a hyperthermophilic protein with melt-
base reveal the energetic determinants of a
ing temperature close to 200 degrees C. Proc.
framework model for folding. J. Mol. Biol.
Natl. Acad. Sci. USA 94, 11329–11332.
381, 1184–1201.
31. Richter, B., Gsponer, J., Varnai, P., Salvatella,
19. Cremades, N., Sancho, J., and Freire, E. X., and Vendruscolo, M. (2007) The MUMO
(2006) The native-state ensemble of proteins (minimal under-restraining minimal over-
provides clues for folding, misfolding and restraining) method for the determination of
function. Trends Biochem. Sci. 31, 494–496. native state ensembles of proteins. J. Biomol.
20. LeMaster, D.M., Anderson, J.S., and Hernández, NMR 37, 117–135.
G. (2009) Peptide conformer acidity analysis of 32. Lange, O.F., Lakomek, N.A., Fares, C.,
protein flexibility monitored by hydrogen Schroder, G.F., Walter, K.F.A., Becker, S.,
exchange. Biochemistry 48, 9256–9265. Meiler, J., Grubmuller, H., Griesinger, C.,
21. Anderson, J.S., Hernandez, G., and LeMaster, and deGroot, B.L. (2008) Recognition
D.M. (2010) Conformational Electrostatics dynamics up to microseconds revealed from
in the Stabilization of the Peptide Anion. an RDC-derived ubiquitin ensemble in solu-
Curr. Org. Chem. 14, 162–180. tion. Science 320, 1471–1475.
22. Anderson, J.S., Hernández, G., and LeMaster, 33. Hernández, G., Anderson, J.S., and LeMaster,
D.M. (2008) A billion-fold range in acidity D.M. (2010) Assessing the native state con-
for the solvent-exposed amides of Pyrococcus formational distribution of ubiquitin by pep-
furiosus rubredoxin. Biochemistry 47, tide acidity. Biophys. Chem., doi: 10.1016/j.
6178–6188. bpc.2010.10.006.
23. Bai, Y.W., Milne, J.S., Mayne, L., and 34. Anderson, J.S., Hernández, G., and LeMaster,
Englander, S.W. (1993) Primary structure D.M. (2009) Backbone conformational
effects on peptide group hydrogen-exchange. dependence of peptide acidity. Biophys. Chem.
Proteins: Struct., Funct., Genet. 17, 75–86. 141, 124–130.
402 G. Hernández et al.
35. Anderson, J.S., Hernandez, G., and LeMaster, 48. Luz, Z., and Meiboom, S. (1964) The activa-
D.M. (2010) Sidechain conformational tion energies of proton transfer reactions in
dependence of hydrogen exchange in model water. J. Am. Chem. Soc. 86, 4768–4769.
peptides. Biophys. Chem. 151, 61–70. 49. Tolbert, L.M., and Solntsev, K.M. (2002)
36. Kim, P.S., and Baldwin, R.L. (1982) Influence Excited-state proton transfer: From con-
of charge on the rate of amide proton strained systems to “Super” photoacids to
exchange. Biochemistry 21, 1–5. superfast proton transfer. Acc. Chem. Res. 35,
37. Tüchsen, E., and Woodward, C. (1985) 19–27.
Hydrogen kinetics of peptide amide protons 50. Leiderman, P., Genosar, L., and Huppert, D.
at the bovine pancreatic trypsin inhibitor pro- (2005) Excited-state proton transfer:
tein-solvent interface. J. Mol. Biol. 185, Indication of three steps in the dissociation
405–419. and recombination process. J. Phys. Chem A
38. Delepierre, M., Dobson, C.M., Karplus, M., 109, 5965–5977.
Poulsen, F.M., States, D.J., and Wedin, R.E. 51. Ellison, W.J., Lamkaouchi, K., and Moreau,
(1987) Electrostatic effects and hydrogen J.M. (1996) Water: A dielectric reference.
exchange behavior in proteins. The pH depen- J. Molec. Liquids 68, 171–279.
dence of exchange rates in lysozyme. J. Mol. 52. Marcus, R.A. (1964) Chemical and electro-
Biol. 197, 111–122. chemical electron-transfer theory. Annu. Rev.
39. Dempsey, C.E. (1995) Hydrogen bond sta- Phys. Chem. 15, 155–196.
bilities in the isolated alamethicin helix: pH- 53. Schaefer, M., and Karplus, M. (1996) A com-
dependent amide exchange measurements prehensive analytical treatment of continuum
in methanol. J. Am. Chem. Soc. 117, electrostatics. J. Phys. Chem. 100,
7526–7534. 1578–1599.
40. Forsyth, W.R., and Robertson, A.D. (1996) 54. Richards, F.M. (1974) The interpretation of
Intramolecular electrostatic interactions accel- protein structures: Total volume, group vol-
erate hydrogen exchange in diketopiperazine ume distributions and packing density. J. Mol.
relative to 2-piperidone. J. Am. Chem. Soc. Biol. 82, 1–14.
118, 2694–2698. 55. Tsai, J., Taylor, R., Chothia, C., and Gerstein,
41. Fogolari, F., Esposito, G., Viglino, P., Briggs, M. (1999) The packing density in proteins:
J.M., and McCammon, J.A. (1998) pKa shift Standard radii and volumes. J. Mol. Biol. 290,
effects on backbone amide base-catalyzed 253–266.
hydrogen exchange rates in peptides. J. Am. 56. Mertz, E.L., and Krishtalik, L.I. (2000) Low
Chem. Soc. 120, 3735–3738. dielectric response in enzyme active site. Proc.
42. Matthew, J.B., and Richards, F.M. (1983) Natl. Acad. Sci. USA 97, 2081–2086.
The pH dependence of hydrogen exchange in 57. Hawranek, J.P., Wrzeszcz, W., Muszyñski,
proteins. J. Biol. Chem. 258, 3039–3044. A.S., and Pajdowska, M. (2002) Infrared dis-
43. Eigen, M. (1964) Proton transfer, acid-base persion of liquid triethylamine. J. Non-Crystal.
catalysis, and enzymatic hydrolysis. (I) Solids 305, 62–70.
Elementary processes. Angew. Chem. Int. Ed. 58. LeMaster, D.M., Anderson, J.S., and
3, 1–19. Hernández, G. (2006) Role of native-state
44. Molday, R.S., and Kallen, R.G. (1972) structure in rubredoxin native-state hydrogen
Substituent effects on amide hydrogen exchange. Biochemistry 45, 9956–9963.
exchange rates in aqueous solution. J. Am. 59. Antosiewicz, J., McCammon, J.A., and Gilson,
Chem. Soc. 94, 6739–6745. M.K. (1994) Prediction of pH dependent
45. Wang, W.H., and Cheng, C.C. (1994) General properties of proteins. J. Mol. Biol. 238,
base catalyzed proton exchange in amides. 415–436.
Bull. Chem. Soc. Jpn. 67, 1054–1057. 60. Antosiewicz, J., McCammon, J.A., and Gilson,
46. LeMaster, D.M., Anderson, J.S., and M.K. (1996) The determinants of pKas in
Hernández, G. (2007) Spatial distribution of proteins. Biochemistry 35, 7819–7833.
dielectric shielding in the interior of Pyrococcus 61. Demchuk, E., and Wade, R.C. (1996)
furiosus rubredoxin as sampled in the sub- Improving the continuum dielectric
nanosecond timeframe by hydrogen exchange. approach to calculating pKa’s of ionizable
Biophys. Chem. 129, 43–48. groups in proteins. J. Phys. Chem. 100,
47. Hernández, G., Anderson, J.S., and LeMaster, 17373–17387.
D.M. (2009) Polarization and polarizability 62. Georgescu, R.E., Alexov, E.G., and Gunner,
assessed by protein amide acidity. Biochemistry M.R. (2002) Combining conformational
48, 6482–6494. flexibility and continuum electrostatics for
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 403
calculating pKas in proteins. Biophys. J. 83, 74. Chevelkov, V., Xue, Y., Rao, D.K., Forman-
1731–1748. Kay, J.D., and Skrynnikov, N.R. (2010)
63. Wisz, M.S., and Hellinga, H.W. (2003) An N-15(H/D)-SOLEXSY experiment for accu-
empirical model for electrostatic interactions rate measurement of amide solvent exchange
in proteins incorporating multiple geometry- rates: application to denatured drkN SH3.
dependent dielectric constants. Proteins 51, J. Biomolec. NMR 46, 227–244.
360–377. 75. Makhatadze, G.I., Clore, G.M., and
64. Song, Y., Mao, J., and Gunner, M.R. (2009) Gronenborn, A.M. (1995) Solvent isotope
MCCE2: Improving protein pKa calculations effect and protein stability. Nat. Struct. Biol.
with extensive side chain rotamer sampling. J. 2, 852–855.
Comput. Chem. 30, 2231–2247. 76. Connelly, G.P., Bai, Y.W., Jeng, M.F., and
65. Simonson, T., and Perahia, D. (1995) Internal Englander, S.W. (1993) Isotope effects in
and interfacial dielectric properties of cyto- peptide group hydrogen-exchange. Proteins:
chrome c from molecular dynamics in aque- Struct., Funct., Genet. 17, 87–92.
ous solution. Proc. Natl. Acad. Sci. USA 92, 77. Sivaraman, T., Arrington, C.B., and Robertson,
1082–1086. A.D. (2001) Kinetics of unfolding and folding
66. Harms, M.J., Schlessman, J.L., Chimenti, from amide hydrogen exchange in native
M.S., Sue, G.R., Damjanovic, A., and García- ubiquitin. Nat. Struct. Biol. 8, 331–333.
Moreno, B.E. (2008) A buried lysine that 78. Rocchia, W., Sridharan, S., Nicholls, A., Alexov,
titrates with a normal pKa: Role of conforma- E., Chiabrera, A., and Honig, B. (2002) Rapid
tional flexibility at the protein-water interface grid-based construction of the molecular sur-
as a determinant of pKa values, Prot. Sci. 17, face and the use of induced surface charge to
833–845. calculate reaction field energies: Applications to
67. Bordwell, F.G., W. J. Boyle, J., and Yee, K.C. the molecular systems and geometric objects.
(1970) Equilibrium and kinetic acidities of J. Comput. Chem. 23, 128–137.
nitroalkanes and their relationship to transi- 79. MacKerell.Jr., A.D., Bashford, D., Bellott,
tion state structures. J. Am. Chem. Soc. 92, M., Dunbrack.Jr., R.L., Evanseck, J.D., Field,
5926–5932. M.J., Fischer, S., Gao, J., Guo, H., Ha, S.,
Joseph-McCarthy, D., Kuchnir, L., Kuczera,
68. Bernasconi, C.F. (1987) Intrinsic barriers of
K., Lau, F.T.K., Mattos, C., Michnick, S.,
reactions and the principle of nonperfect syn-
Ngo, T., Nguyen, D.T., Prodhom, B., Reiher-
chronization. Acc. Chem. Res. 20, 301–308.
III, W.E., Roux, B., Schlenkrich, M., Smith,
69. Costentin, C., and Saveant, J.M. (2004) Why J.C., Stote, R., Straub, J., Watanabe, M.,
are proton transfers at carbon slow? Self- Wiorkiewicz-Kuczera, J., Yin, D., and Karplus,
exchange reactions. J. Am. Chem. Soc. 126, M. (1998) All-atom empirical potential for
14787–14795. molecular modeling and dynamics studies of
70. Hwang, T.L., Mori, S., Shaka, A.J., and vanZijl, proteins. J. Phys. Chem. B 102, 3586–3616.
P.C.M. (1997) Application of Phase-Modulated 80. Jorgensen, W.L., Maxwell, D.S., and Tirado-
CLEAN Chemical EXchange Spectroscopy Rives, J. (1996) Development and testing of
(CLEANEX-PM) to detect water-protein pro- the OPLS all-atom force field on conforma-
ton exchange and intermolecular NOEs. J. Am. tional energetics and properties of organic liq-
Chem. Soc. 119, 6203–6204. uids. J. Am. Chem. Soc. 118, 11225–11236.
71. Hwang, T.L., vanZijl, P.C.M., and Mori, S. 81. Cheatham, T.E., Cieplak, P., and Kollman,
(1998) Accurate quantitation of water-amide P.A. (1999) A modified version of the Cornell
proton exchange rates using the phase-modu- et al. force field with improved sugar pucker
lated CLEAN chemical EXchange phases and helical repeat. J. Biomolec. Struct.
(CLEANEX-PM) approach with a fast-HSQC Dynamics 16, 845–862.
(FHSQC) detection scheme. J. Biomol. NMR 82. Duan, Y., Wu, C., Chowdhury, S., Lee, M.C.,
11, 221–226. Xiong, G.M., Zhang, W., Yang, R., Cieplak,
72. Hernández, G., and LeMaster, D.M. (2003) P., Lou, R., Lee, T., Caldwell, J., Wang, J.M.,
Relaxation compensation in chemical exchange and Kollman, P. (2003) A point-charge force
measurements for the quantitation of amide field for molecular mechanics simulations of
hydrogen exchange in larger proteins. Magn. proteins based on condensed-phase quantum
Reson. Chem. 41, 699–702. mechanical calculations. J. Comp. Chem. 24,
73. Griesinger, C., and Ernst, R.R. (1987) 1999–2012.
Frequency offset effects and their elimination 83. Bayly, C.I., Cieplak, P., Cornell, W.D., and
in NMR rotating-frame cross-relaxation spec- Kollman, P.A. (1993) A well behaved electro-
troscopy. J. Magn. Reson. 75, 261–271. static potential-based method using charge
404 G. Hernández et al.
restraints for deriving atomic charges: The RESP 96. Chen, K., Liu, Z., Zhou, C., Bracken, W.C.,
model. J. Phys. Chem. 97, 10269–10280. and Kallenbach, N.R. (2007) Spin relaxation
84. Becke, A.D. (1993) Density-functional ther- enhancement confirms dominance of extended
mochemistry III. The role of exact exchange. conformations in short alanine peptides.
J. Chem. Phys. 98, 5648–5652. Angew. Chem. Int. Ed. 46, 9036–9039.
85. You, T.J., and Bashford, D. (1995) 97. Graf, J., Nguyen, P.H., Stock, G., and
Conformation and hydrogen ion titration of Schwalbe, H. (2007) Structure and dynamics
proteins: A continuum electrostatic model of the homologous series of alanine peptides:
with conformational flexibility. Biophys. J. 69, A joint molecular dynamics/NMR study. J.
1721–1733. Am. Chem. Soc. 129, 1179–1189.
86. vanVlijmen, H.W.T., Schaefer, M., and 98. Wickstrom, L., Okur, A., and Simmerling, C.
Karplus, M. (1998) Improving the accuracy of (2009) Evaluating the Performance of the
protein pKa calculations: Conformational ff99SB Force Field Based on NMR Scalar
averaging versus the average structure. Proteins Coupling Data. Biophys. J. 97, 853–856.
33, 145–158. 99. Pizzanelli, S., Forte, C., Monti, S.,
87. deGroot, B.L., vanAalten, D.M.F., Scheek, Zandomeneghi, G., Hagarman, A., Measey,
R.M., Amadei, A., Vriend, G., and Berendsen, T.J., and Schweitzer-Stenner, R. (2010)
H.J.C. (1997) Prediction of protein confor- Conformations of Phenylalanine in the
mational freedom from distance constraints. Tripeptides AFA and GFG Probed by
Proteins 29, 240–251. Combining MD Simulations with NMR,
88. Cornilescu, G., Marquardt, J.L., Ottiger, M., FTIR, Polarized Raman, and VCD
and Bax, A. (1998) Validation of Protein Spectroscopy. J. Phys. Chem B 114,
Structure from Anisotropic Carbonyl Chemical 3965–3978.
Shifts in a Dilute Liquid Crystalline Phase. 100. Tsai, M., Xu, Y.J., and Dannenberg, J.J.
J. Am. Chem. Soc. 120, 6836–6837. (2009) Ramachandran Revisited. DFT Energy
89. Boehr, D.D., Nussinov, R., and Wright, P.E. Surfaces of Diastereomeric Trialanine Peptides
(2009) The role of dynamic conformational in the Gas Phase and Aqueous Solution.
ensembles in biomolecular recognition. J. Phys. Chem B 113, 309–318.
Nature Chem. Biol. 5, 789–796. 101. Fitzkee, N.C., Fleming, P.J., and Rose, G.D.
90. Mittermaier, A.K., and Kay, L.E. (2009) (2005) The Protein Coil Library: A structural
Observing biological dynamics at atomic reso- database of nonhelix, nonstrand fragments
lution using NMR. Trends Biochem. Sci. 34, derived from the PDB. Prot. Struct. Funct.
601–611. Bioinform. 58, 852–854.
91. Dikic, I., Wakatsuki, S., and Walters, K.J. 102. Avbelj, F., and Baldwin, R.L. (2006) Limited
(2009) Ubiquitin-binding domains - from validity of group additivity for the folding
structures to functions. Nature Rev. Molec. energetics of the peptide group. Prot. Struct.
Cell Biol. 10, 659–671. Funct. Bioinform. 63, 283–289.
92. Tjandra, N., Feller, S.E., Pastor, R.W., and 103. Flory, P.J. Statistical mechanics of chain mole-
Bax, A. (1995) Rotational diffusion anisot- cules, (Wiley Interscience, New York, 1969).
ropy of human ubiquitin from 15 N NMR 104. Penkett, C.J., Redfield, C., Dodd, I., Hubbard,
relaxation. J. Am. Chem. Soc. 117, J., McBay, D.L., Mossakowska, D.E., Smith,
12562–12566. R.A.G., Dobson, C.M., and Smith, L.J.
93. Sheinblatt, M. (1970) Determination of an (1997) NMR analysis of main-chain confor-
acidity scale for peptide hydrogens from mational preferences in an unfolded fibronec-
nuclear magnetic resonance kinetic studies. J. tin-binding protein. J. Mol. Biol. 274,
Am. Chem. Soc. 92, 2505–2509. 152–159.
94. Avbelj, F., and Baldwin, R.L. (2004) Origin 105. Keskin, O., Yuret, D., Gursoy, A., Turkay, M.,
of the neighboring residue effect on peptide and Erman, B. (2004) Relationships between
backbone conformation. Proc. Natl. Acad. amino acid sequences and backbone torsion
Sci. USA 101, 10967–10972. angle preferences. Prot. Struct. Funct.
95. Makowska, J., Rodziewicz-Motowid³o, S., Bioinform. 55, 992–998.
Bagiñska, K., Vila, J.A., Liwo, A., Chmurzyñski, 106. Jha, A.K., Colubri, A., Zaman, M.H., Koide,
L., and Scheraga, H.A. (2006) Polyproline II S., Sosnick, T.R., and Freed, K.F. (2005)
conformation is one of many local conforma- Helix, sheet, and polyproline II frequencies
tional states and is not an overall conforma- and strong nearest neighbor effects in a
tion of unfolded peptides and proteins. Proc. restricted coil library. Biochemistry 44,
Natl. Acad. Sci. USA 103, 1744–1749. 9691–9702.
20 Electrostatics of Hydrogen Exchange for Analyzing Protein Flexibility 405
107. LeMaster, D.M. (1999) NMR relaxation time-scale side-chain motions. J. Am. Chem.
order parameter analysis of the dynamics of Soc. 124, 6449–6460.
protein sidechains. J. Am. Chem. Soc. 121, 109. Darley, M.G., and Popelier, P.L.A. (2008) Role
1726–1742. of short-range electrostatics in torsional poten-
108. Skrynnikov, N.R., Millet, O., and Kay, L.E. tials. J. Phys. Chem A 112, 12954–12965.
(2002) Deuterium spin probes of side-chain 110. Senn, H.M., and Thiel, W. (2009) QM/MM
dynamics in proteins. 2. Spectral density methods for biomolecular systems. Angew.
mapping and identification of nanosecond Chem. Int. Ed. 48, 1198–1229.
Chapter 21
Abstract
Probing protein structure, dynamics, and interaction surfaces by NMR requires initial backbone resonance
assignment. The protocol for this step has been progressively developed in the last 15 years to provide
robust assignments. However, even in the case of favorable conditions (high field magnets and cryogenically
cooled probes, small globular proteins, high sample concentration), the assignment step generally takes
several days of data collection and analysis, thus precluding studies of unstable proteins and limiting high-
throughput applications. Recently, we have introduced the BATCH strategy for fast protein backbone
resonance assignment. BATCH benefits from the combination of several tools (BEST/ASCOM/Targeted-
Sampling/COBRA/HADAMAC) for time-optimized and highly automated NMR data acquisition, processing,
and analysis. In this chapter, we discuss the individual steps of the BATCH method and describe its practical
implementation to obtain the backbone resonance assignment of small globular proteins in a few hours of time.
Key words: Protein, Fast NMR, Resonance assignment, Chemical shift, Amino acid type discrimination,
Algorithm
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_21, © Springer Science+Business Media, LLC 2012
407
408 B. Brutscher and E. Lescop
2. Description
of the Various
Tools Implemented
in BATCH In this section, we briefly describe the set of methods implemented
in BATCH allowing for significantly faster data collection and analysis.
We emphasize the main characteristics of the introduced techniques
and describe their benefits in the context of the BATCH strategy.
21 Fast Protein Backbone NMR Resonance Assignment Using the BATCH Strategy 409
2.1. The BEST Principle Sequential resonance assignment is based on a set of 3D triple reso-
for Fast Pulsing nance (H–N–C) experiments. One crucial parameter governing the
Multidimensional NMR overall experimental time is the interscan (or recycling) delay that is
required for magnetization recovery between successive repetitions
of the pulse sequence. In order to maximize the signal-to-noise ratio
per unit of time (sensitivity), this delay is usually set to ~1–1.5 s,
accounting for the average 1H T1 in proteins. It has been demon-
strated that in experiments that excite and detect amide protons,
amide 1H T1 can be significantly reduced by leaving all nonamide
protons unperturbed throughout the pulse sequence (9–11). This
so-called longitudinal relaxation enhancement effect is mainly due
to 1H–1H dipolar interaction-mediated polarization (energy) trans-
fer from the excited amide 1H to other nearby 1H in thermal equi-
librium and chemical exchange between labile amide protons and
water protons. This observation has led to the development of
BEST-type HNC correlation experiments used for backbone assign-
ment (6, 7). BEST pulse sequences differ from conventional pulse
sequences by the extensive use of shaped pulses for amide 1H band
selective excitation, inversion, and refocusing (PC9 (12), EBURP2
(13) and REBURP (13) shapes), as well as pairs of broadband inver-
sion pulses (BIP (14)) that achieve minimal perturbation of aliphatic
and water 1H magnetizations. Composite broadband 1H decoupling
sequences, such as WALTZ or DIPSI are incompatible with the
BEST effect. Therefore another difference in BEST pulse sequences
is the use of a simple 1H inversion (BIP) pulse for refocusing 1H–15N
or 1H–13C coupling evolution. This slightly reduces the overall sen-
sitivity of BEST triple-resonance experiments for fast relaxing sys-
tems such as high molecular weight proteins. In BEST experiments,
maximal sensitivity is obtained for interscan delays of ~200–400 ms.
This yields a reduction in the overall experimental time by a factor of
~3 without compromising spectral resolution and sensitivity.
2.4. Targeted Sampling In contrast to FT processing, the COBRA processing tool is com-
Approach for Rapid patible with regular but incomplete sampling of the indirect 13C
Sampling of the 13C dimensions (5). Typically, the 13C time domain is sampled over two
Dimension time regions (targeted sampling): t1 = 0 to 3 ms (standard STD
region) and t1 = 25 to 28 ms (time-shifted TS region). This second time
window is chosen such that the signal attenuation due to evolution
under 13Ca–13Cb scalar coupling is negligible: (cos (pJ CCt 1 ) ≅ −1).
A COBRA map is calculated separately for each time window and
subsequently combined. The additional sampling of long t1 values
in the TS time window improves frequency discrimination in the
final COBRA map compared with the STD window-only based
COBRA map. The targeted sampling approach combined with
COBRA analysis provides a high level of frequency discrimination
while requiring only a small number of t1 increments in the 13C
dimension. In practice, the targeted sampling scheme is applied to
the most sensitive pair of experiments (H–N–CA).
2.5. The HADAMAC Amino acid type information for individual residues (15N HSQC
Experiment for Amino cross peaks) is required to assign fragments of sequentially con-
Acid Typing nected 1H–15N frequency pairs to a particular location on the pro-
tein sequence. In the BATCH strategy, we use the recently
introduced HADAMAC experiment (4) for amino acid-type dis-
crimination. In this experiment, the 15N HSQC cross peaks are
edited along an additional “amino acid-type” dimension according
412 B. Brutscher and E. Lescop
2.6. 13C Chemical To extract the 13C chemical shift information present in the avail-
Shift Extraction able triple resonance experiments, an efficient algorithm has been
developed (3). Briefly, 1D 13C time domain data are extracted at
the 1H/15N position of each 15N HSQC cross peak from a given 3D
experiment. These traces are then subjected to Fourier transforma-
tion, resulting in complex valued S(w) traces. Analogous to COBRA4
⎛ f (ω) ⎞
−⎜
f ⎟
processing, the traces are weighted as F (ω) = ℜ(S (ω))* e ⎝ cut ⎠
where f(w) corresponds to the angular phase of S(w). The phase
weighting transformation greatly improves frequency resolution.
Using the same fcut as for COBRA processing is recommended. In
the case of delayed acquisition (TS region defined as t1 = t10 + k*Dt1),
zero and first order phase corrections are required and are calcu-
lated as f0 = t10/Dt1*p and f1 = t10/Dt1*2p. An additional signal
inversion (180° zero-order phase correction) is applied to account
for sign inversion due to 13C–13C scalar coupling evolution. When
the STD and TS regions are collected for the H–N–CA pair, the
traces obtained from the two regions are multiplied point-by-point,
resulting in a single trace. In principle, a given 13C nucleus gives
rise to peaks in sequential and intraresidue experiments at the
1
H/15N positions of sequential residues. Prior to chemical shift
extraction, the two corresponding traces are also multiplied. One
13
C chemical shift is extracted for every trace (residue) as the fre-
quency of maximum (absolute) amplitude.
21 Fast Protein Backbone NMR Resonance Assignment Using the BATCH Strategy 413
3. Experimental
Setup
3.1. Protein Sample Protein backbone resonance assignment using the BATCH strategy
requires a 15N/13C labeled protein sample, typically in the concen-
tration range of 100 mM to a few mM. Since high-level deuteration
of aliphatic protons is incompatible with the BATCH strategy, this
approach is best suited for proteins below ~20 kDa. Aliphatic pro-
tons are required as a relaxation source for amide protons in BEST-
type triple-resonance experiments, and as starting magnetization
for the amino acid-type edited HADAMAC experiment. A file
containing the sequence of the protein in NMRView format is also
needed.
3.2. NMR Hardware The time efficiency of the assignment relies on the spectrometer
and Pulse Sequences sensitivity. Therefore high magnetic fields equipped with cryogenic
probes are preferred. A 600 MHz spectrometer may represent a
good trade-off to limit the loss of sensitivity due to the CSA-
induced increase in 13CO transverse relaxation at high field
strengths. In the following sections, we assume that the spectrom-
eter (Bruker or Agilent) is equipped with Topspin 2.1 (or newer)
or VNMRJ (installed with the last version of BioPack) software.
We also assume that pulses on the 15N and 13C channels have been
correctly calibrated and that 1H amplifiers have been linearized.
This is particularly important for automated calibration of shaped
pulses in BEST-type sequences.
For the BATCH strategy, several NMR pulse sequences should
be available on the spectrometer. They include the BEST versions
for 15N HSQC and the following triple resonance experiments:
HN(CO)CA, iHNCA, HN(CO)CB, iHNCB, HNCO, and
iHNCO as well as HADAMAC. The iHNCA (iHNCB, and
iHNCO) only retains the intraresidual H–N–C coherence transfers
as described elsewhere (15). The HN(CO)CB experiment is iden-
tical to HN(CO)CACB with the exception of the longer Ca→Cb
transfer delay for performing a full CA→CB transfer. Therefore, in
HN(CO)CB (and iHNCB) only one cross peak is present for each
1
H/15N frequency pair. In principle, the sensitivity enhanced ver-
sions of these pulse sequences should be used. However, for fast
relaxing systems, shorter INEPT-based versions can be used to
obtain higher signal to noise ratios. The HADAMAC experiment
should be collected first on a protein sample with known resonance
assignments for the assignment of each plane to the corresponding
amino acid group. For VNMRJ users, all pulse sequences are pro-
vided within the latest updated Biopack pulse sequence library. For
Bruker users, most experiments are directly available for Topspin
2.1 and later versions. Pulse sequences are also available upon
request from the authors of this chapter.
414 B. Brutscher and E. Lescop
3.3. Software The BATCH strategy requires the ASCOM tool for 15N spectral
optimization, which consists of a simple Perl script that can be run
on the spectrometer. The software is available at the Web site http://
www.icsn.cnrs-gif.fr/download/nmr. For Agilent spectrometers
equipped with the most recent versions of BioPack, the macro
BestSW allows for automated peak picking of 2D 15N HSQC spectra
and provides the ASCOM optimized 15N spectral width. The
BATCH software platform is written in Tcl language and is embed-
ded in the NMRView software (http://www.onemoonscientific.
com/nmrview) (16) as a new functionality to make use of the variety
of NMRView native functions. BATCH was validated for the Aqua
(for MaxOS) and C (Linux system) versions of the NMRView soft-
ware and is also available for the platform independent Java version.
The BATCH software package can be downloaded from the Web site
http://www.icsn.cnrs-gif.fr/download/nmr. The downloaded pack-
age includes instructions for installation, a manual, as well as a tuto-
rial. The NMRPipe processing software (http://www.nmrscience.
com/nmrpipe.html) (17) is also required to execute scripts gener-
ated by BATCH. The final assignment can be carried out using the
native algorithm embedded in the BATCH software. Alternatively,
the BATCH software provides a convenient interface for the Mars
( http://www.mpibpc.mpg.de/groups/zweckstetter/_links/
software_mars.htm) (18) or SmartNotebook (http://www.bionmr.
ualberta.ca/bds/software/snb/) (19) software packages that may
also be installed. For the following description, we assume that the
user has a minimal proficiency with NMRView and NMRPipe
software.
4. NMR Data
Collection
In the following sections, specific commands for Topspin and
VNMRJ are given in small caps and italics, respectively, for exam-
ple: PULSECAL and pulsecal.
4.1. Experimental After sample injection and temperature regulation, the probe is
Setup and Pulse tuned and shimmed. The 90° 1H pulse is calibrated, typically by
Calibration determining the 360° pulse duration measured on the on-resonance
water frequency or by using the Bruker PULSECAL tool. A water sup-
pressed 1H spectrum is collected as a first evaluation of spectral
quality, and for 1H chemical shift calibration in case of the presence
of an internal reference (e.g., DSS, TSP).
4.2. The 15N HSQC In a new folder, load the 15N HSQC experiment (RPAR BHSQC,
Experiment best_Nhsqc). Amide 1HN band selective pulses are adjusted to cover
the 1HN spectral width (typically 4 ppm centered at 8.5 ppm). Care
is taken, however, to avoid water saturation for an efficient BEST
effect and solvent suppression. Correct pulse calibration for 1H and
21 Fast Protein Backbone NMR Resonance Assignment Using the BATCH Strategy 415
15
N channels are set (GETPROSOL 1H 9 1.5; pw = 9). For Bruker, the
getprosol command automatically generates the power level for all
shaped pulses present in the loaded pulse sequence. For VNMRJ,
the power levels are automatically adjusted “on the fly”using the
Pbox tool. The 15N HSQC experiment is run using the following
parameters by default: recycling delay (d1) of 250 ms, 15N spectral
width of 35 ppm centered at 118 ppm, 2 scans, 16 dummy scans,
and 80–100 increments in the 15N dimension. For cryogenic probe
safety, it is advised to reduce the acquisition time (AQ, at) to less
than 100 ms and the power level of the 15N decoupling sequence
applied during acquisition. The 15N HSQC is acquired within
~1–2 min. After double FT processing, the spectrum is visually
inspected in terms of averaged signal-to-noise ratio, intensity het-
erogeneity, and spectral dispersion to assess the amenability of
sequential resonance assignment through the BATCH approach.
Possible aliasing of backbone 1H/15N cross peaks is detected and
the experiment is possibly run again with increased 15N spectral
width. This step is important to extract exact 15N frequencies before
ASCOM optimization. Care should be taken for side chain H–N
moieties (arginine, asparagine, and glutamine residues). The auto-
mated optimal 15N spectral width is obtained by ASCOM either
“on the fly” (BestSW) or after peak picking and analysis (Bruker),
and should be memorized for the following experiments. The
extent of spectral compression by ASCOM software should be
carefully set. In the course of COBRA processing and for increas-
ing the signal-to-noise ratio, the residue specific 13C time domain is
obtained by the integration of 3D matrices along the 1H and 15N
dimensions (for the typical 1H/15N line widths) around the 1H/15N
chemical shifts position. To avoid cross talk between 13C time
domains corresponding to different residues, the cross peaks should
be as separate as possible in the 1H/15N planes. Therefore, the cut-
off parameters (RH and RN), defining whether two 1H/15N cross
peaks overlap or not in ASCOM software, are set to the averaged
1
H and 15N line widths to limit partial peak overlapping.
4.3. The HADAMAC In a new folder the HADAMAC pulse sequence (RPAR HADAMAC,
Experiment hadamac) is loaded, preferably the HADAMAC-2 version for
increased sensitivity. Set the correct pulse power values (GETPROSOL
1H 9 1.5; pw = 9), the number of repetitions to 2, the relaxation
delay d1 to 0.8–1 s, the 15N spectral width (ASCOM-optimized or
not) and the maximum number of 15N increments as allowed by
the constant-time period inserted in the pulse sequence for
increased resolution. For VNMRJ, set had_flg to (1, 2, …, 8),
phase = 1,2 and array=“(had_flg, phase).” Depending on the cho-
sen 15N spectral width, the HADAMAC experiment lasts from
~30 min to 1 h. This time can be used for the initial setup of
BATCH, inspection of the 15N HSQC spectrum (Subheading 5.2),
and the setup of triple resonance experiments.
416 B. Brutscher and E. Lescop
4.4. The Triple In a new folder, the sequential BEST HN(CO)CA pulse sequence
Resonance is loaded (RPAR BHNCOCA, best_hncocaP). Set the correct pulse
Experiments power values (vide supra), the number of repetitions to 2, the
relaxation delay (d1) to 250 ms, the ASCOM-optimized 15N spec-
tral width with the maximum number of increments as allowed by
the constant-time period inserted in the pulse sequence, and a typi-
cal 13C spectral window (20 ppm) centered at 56 ppm. The initial
evolution delay under 13C chemical shifts is set to 0 (d0 = 0, d1 = 0),
and 10 complex points are recorded to sample the 0…3 ms time
region (for a 600 MHz 1H field strength). Repeat the procedure
for the BEST iHNCA pulse sequence. This pair of experiments
represents the standard (STD) H–N–CA pair. For correct COBRA
processing, the 13C time domain of the sequential and intraresidual
experiments should be collected in an exactly identical manner,
including the pulse sequence element for 13C chemical shift evolu-
tion, sampling points, and the 13C carrier frequency. The phase
weighted correlation coefficient embedded in COBRA is able to
discriminate 13C signals that are 180° phase shifted (corresponding
to peaks of opposite sign in frequency domain). Such sign inver-
sion occurs, for example, when the first t1 time value is set to half
dwell in the 15N dimension in the case of aliased cross peaks. For
correct COBRA processing, the first t1 time value should be set to
0 ms. Using the same procedure, the time shifted (TS) H–N–CA
pair of experiments is prepared in two additional folders. This pair
is exactly identical to the previous STD H–N–CA pair except that
the initial evolution delay under 13C chemical shifts is set to 25 ms
(d0 = 25 m, d1 = 25 m) instead of 0 ms. Alternatively, the TS time
region can be stored in the same files as the STD H–N–CA and can
also be analyzed by the BATCH software. The same procedure is
then applied to prepare the H–N–CB pair, consisting of the BEST
HN(CO)CB and iHNCB pulse sequences and the H–N–CO pair,
consisting of the BEST HNCO and iHNCO pulse sequences. For
the 13Cb dimension, set the spectral window to 60 ppm centered at
46 ppm and collect 30 complex points to sample the t = 0…3 ms
time region (for a 600 MHz 1H frequency). For the carbonyl 13C
dimension, set the spectral window to 10 ppm centered at 176 ppm
and collect 10 complex increments. Owing to the intrinsic low
sensitivity of the iHNCB and the iHNCO experiments, the num-
ber of transients may be set to 4. Launch all experiments in a row.
The time required for acquisition of the triple resonance data can
be used to process and analyze the HADAMAC experiment
(Subheading 5.2).
After completion of the first pair of triple-resonance experi-
ments (STD H–N–CA), they can be processed (Subheading 5.2),
analyzed using the COBRA algorithm (Subheading 6.2), and a
first attempt at resonance assignment can be done (Subheading 7).
Every newly collected pair can be immediately processed and
21 Fast Protein Backbone NMR Resonance Assignment Using the BATCH Strategy 417
5.1. General A few paths have to be set in the PATH Window (Fig. 1b): the
Parameters location of the directory containing BATCH scripts (see installa-
tion procedure), the base directory for NMR data containing one
folder per experiment and the sequence file. An additional path can
be set for already existing 2D peak lists in the xpk (NMRView)
format. Define the names of the folders containing the collected
experiments. Additional parameter definitions will be automatically
set in the course of preprocessing. The last section of this window
allows the setting of a signal-free region of the 15N HSQC and of
the integration box. The signal-free region is defined by a 1H/15N
cross peak position and is used to estimate the noise in H–N–C spec-
tra for the automated phase cutoff setting algorithm during COBRA
processing. Owing to the extensive folding in 15N dimension in H–N–C
experiments, it is advised to locate the cross peak in a signal-free region
418 B. Brutscher and E. Lescop
Fig. 1. (a) Main graphical interface of the BATCH software. (b) Window for paths definition and peak integration. (c) Window
for rapid preprocessing of NMR data.
5.2. Preprocessing The 15N HSQC, the HADAMAC and the triple resonance experi-
of the Raw Data ments require different preprocessing steps: the 15N HSQC spec-
trum is processed using conventional double FT, the HADAMAC
experiment is subjected to double FT followed by Hadamard
decoding, and triple resonance experiments are Fourier trans-
formed along the 1H and 15N dimensions, while leaving the 13C
dimension in the time domain for COBRA processing (files stored
in the cobra folder). An additional Fourier transformation is also
applied along the 13C dimension of triple resonance experiments
(files stored in ft folder and converted to the NMRview format) for
visualization (Load NvFile). Preprocessing is performed with
NMRPipe in two steps: data conversion to NMRPipe format (fid.
com script file) and data manipulation (nmrproc.com script file).
Since for a given experimental setup (spectrometer and pulse
sequences), the same NMRPipe processing script files can be used
with minor modifications, the BATCH scripts directory contains
one nmrproc.com-type file for each experiment. These files are
modified and validated for the specific experimental setup only
once and can be trustfully used for following protein studies. We
put emphasis on the adequate processing of the HADAMAC
experiment for correct subspectra assignment.
The preprocessing step, facilitated by the “Process Window”
(Fig. 1c), consists of going through each collected and previously
defined experiment. Basically, for each recorded dataset, an
nmrDraw window is opened (fid.com) for the preparation of the
conversion script. The label of the 1H and 15N dimensions should
be consistent over all experiments and should be set to “HN” and
“N15” respectively. After adequate parameter adjustment, the
conversion is executed. In the next step, the processing file is
opened (nmrproc.com). This file may be chosen from different
locations: default script library, folder of the current experiment
or from the STD H–N–CA experiment (in the case of TS
H–N–CA). The nmrproc.com script is then adjusted for the spe-
cific protein (mainly 1H phase correction), saved in the correct
subdirectory and executed by NMRPipe (Save & Execute). The
individual spectra can be inspected in NMRDraw (Check in
NMRDraw) for optimal processing. This initial setting is carried
out without linear prediction (LP) in the 15N dimensions to save
time. An additional button (Process All with LP) allows repro-
cessing of all datasets with additional LP. During the preprocess-
ing step, the paths to all newly generated files are automatically set
(see Path window).
420 B. Brutscher and E. Lescop
6. 2D Peak-
Picking, HADAMAC
Analysis and
COBRA Calculation One unique (2D) peak list is required in BATCH, containing the
protein’s 1H/15N chemical shifts. Although such a peak list can be
6.1. 2D Peak List obtained directly from the 15N HSQC through the 2D peak-pick-
Generation and ing procedure embedded in NMRView, the better spectral disper-
Extraction of Amino
sion in the pseudo-3D HADAMAC spectra makes this experiment
well suited to discriminate partially overlapping peaks. The
Acid Type
HADAMAC subspectra are loaded (Show HADAMAC spectra)
and inspected for spectral quality. A dedicated algorithm for the
HADAMAC-based 2D 1H/15N peak list generation is called
(AutoPeakPick), and the resulting peak list is inspected and manu-
ally adjusted (Fig. 2b). The peak list (defined as the Current 2D
peak list in Fig. 1a) is then visualized on the 15N HSQC spectrum
(LoadHSQC). Possible folded peaks in the 15N dimensions are
identified and unfolded. More details about this procedure (spe-
cific to NMRView) are provided in the BATCH manual. The 2D
peak list should contain exact (nonfolded) 15N frequencies for the
subsequent extraction of 13C time domain at the correct 1H/15N
chemical shifts position in triple resonance experiments. The indi-
vidual box sizes, used for the extraction of amino acid type infor-
mation, should be manually adjusted for partially overlapping cross
peaks. The cross-peak-dependent amino acid type information
is stored into the Comment section of the NMRView peak list
window (Transfer HADAMAC information to Current Peaklist).
Fig. 2. (a) 15N HSQC spectrum of Hyl1. (b) Overlay of the HADAMAC subspectra of Hyl1. The amino acid groups are colored
as follows: AVI (black), Gly (red), Thr (yellow), Ser (Blue), Asx (Magenta), Cys-Arom (Green), and Rest (Cyan). Of note, the
cross peaks corresponding to the Cys-Arom and Rest groups are present in the same plane but with opposite signs. The
result from the automated peak picking is shown as boxes with (random) numbering.
21 Fast Protein Backbone NMR Resonance Assignment Using the BATCH Strategy 421
6.2. COBRA Analysis The COBRA maps are calculated from the main BATCH window
(Fig. 1a). Each pair of H–N–C experiments (i.e., one COBRA
map) corresponds to one line allowing the setting of three param-
eters: the first point (usually 1), the number of points in the 13C
time domain (−1 to use all available data) to be used for COBRA
calculation and the phase cutoff parameter fcut. In case the STD
and TS sampled regions are stored in the same file for the H–N–CA
pair, an additional list of 13C evolution delays can be defined on the
same dataset. Only selected pairs of experiments are used for
COBRA calculation (Calculate COBRA maps) and the individual
and final maps are displayed (Show COBRA maps). At the same
time the contour level for the visualization is automatically set to
have on average ~1–2 visible COBRA elements per column and
row (Fig. 3a), and can be further adjusted using classical NMRView
tools. The contour level will be used later for binning. Spectral
noise present in the individual triple resonance experiments propa-
gates through the COBRA calculation. Noise in experimental 13C
traces translates into overall reduction of COBRA elements values.
In addition, noise leads to the deviation of the angular phase f
value of the correlation coefficient corr when compared with noise-
free 13C traces. This deviation is larger for low “averaged” signal-
to-noise ratio and has to be taken into account for setting the phase
cutoff parameter fcut. Default initial cutoff phase values have to be
set to low values (e.g., 15°).
In absence of assignment, the cross peak numbering is random
and real (physical) connectivities are not apparent as a (shifted)
diagonal. In addition, for nonoptimized phase settings, several
nonphysical connectivities (reconstruction artifacts) may appear as
additional nonzero values in the COBRA map and/or correct con-
nectivites may be absent. For these reasons, the informational con-
tent of a given map is difficult to assess visually. To facilitate the
COBRA processing, a new peak list can be generated (Order
Peaklist) that differs from the current one by the cross peak num-
bering (ordering). The new order of the peaks is guided by the
unambiguous connectivities extracted from the current COBRA
map to create unambiguously defined fragments of different
lengths. An unambiguous connectivity between cross peaks i and j
is obtained if the M(i,j) element is larger than the cutoff value
(defined in the main BATCH window) and if no other values
higher than the cutoff value exist along the ith column and the jth
row. This intensity cutoff value is automatically set during processing
422 B. Brutscher and E. Lescop
Fig. 3. COBRA maps calculated from the H–N–CA experiments collected on Hyl1 with different phase cutoff values for the
STD (fSTD) and TS (fTS) region and different peak ordering. (a) (fSTD = 15°; fTS = 15°) and original (random) peak ordering.
(b) (fSTD = 15°; fTS = 15°) and peak ordering obtained after unambiguous fragment identification. (c) (fSTD = 15°; fTS = 45°)
and same peak order as in (b). (d) (fSTD = 15°; fTS = 45°) and peaks ordered according the final assignment.
and is also used for contour level definition. Optionally, it can also
be adjusted manually. The same cutoff value is used for binning in
the assignment step (vide infra). After peak ordering and calculation
of the new COBRA maps (Calculate COBRA maps), the identified
fragments are ordered from the longest one to the shortest one as
illustrated in Fig. 3b-c. This ordering step is recommended once a
reasonably good COBRA map is obtained (see rule 1 below).
21 Fast Protein Backbone NMR Resonance Assignment Using the BATCH Strategy 423
Table 1
Strategy for setting cutoff phases in COBRA algorithm
(c) If the right part of the COBRA product map (after ordering)
contains many columns with more than 2 possible connectivi-
ties, this suggests either a lack of 13C frequency discrimination
due to close or overlapping 13C chemical shifts for different
residues in the protein or to partially overlapping 1H/15N. In
the former case, the HADAMAC information may be suffi-
cient to alleviate the ambiguities in the course of the assign-
ment. The latter case can be identified from the inspection of
the 15N HSQC spectrum and the integration box around
1
H/15N frequencies can be reduced (see Path window).
(d) The HNCO/iHNCO pair of experiments should be used
carefully during the course of assignment using COBRA for
glycine residues. The iHNCO experiment contains one Ca
selective pulse designed to optimize the Ca -> CO transfer
while limiting Ca -> Cb transfer. As a consequence, signals
corresponding to glycine residues may be severely attenuated
in the iHNCO experiment. This translates into missing COBRA
connectivities for glycine residues in the final COBRA maps
(after combination of all available COBRA maps).
(e) Overall, less than ~5 COBRA calculations should in principle be
sufficient to obtain a COBRA map that reflects the information
content of the triple resonance experiments. Otherwise, the
suitability of the BATCH method for the particular protein
under investigation should be questioned.
7. Backbone
Assignment
Based on the previous steps, a 2D 1H–15N peak list containing the
cross-peak-dependent HADAMAC amino acid group and the cor-
responding COBRA sequential connectivity map are available.
Three alternative options with different degrees of automation
have been developed to facilitate sequential resonance assignment.
They include the embedded BATCH algorithm (3), the interface
to MARS (18) and SmartNotebook (19) software packages. The
COBRA product map (Subheading 6.2) contains element values
between 0 and 1. Before application of any of the three methods,
this map is first binned to 0 and 1 values according to the intensity
cutoff parameter defined in the main BATCH window. The cutoff
parameter automatically proposed after COBRA map calculation
often represents an excellent starting value. However, it can be
manually adjusted from the (visual) inspection of the COBRA
maps. Based on the newly calculated matrix, fragments of unam-
biguously connected cross peaks are built and fed to the assign-
ment algorithm together with the additional (ambiguous)
sequential connectivities and the HADAMAC information.
21 Fast Protein Backbone NMR Resonance Assignment Using the BATCH Strategy 425
Fig. 4. Pseudo 2D spectrum processed as described in Subheading 2.6. 13C frequency traces from the intraresidual (left
half ) and sequential (right half ) H–N–CA experiments are shown for each 1H–15N cross peak (ordered according to the final
assignment). The two halves of the spectrum show similar patterns due to the presence of traces containing signals at the
same 13C frequency in the two experiments. For the sequential experiment, cross peak numbers are incremented by n (total
number of cross peaks).
and the sequential (right half) experiments for every 15N HSQC
cross peak (Fig. 4). This calculation is based on the same phase
cutoff parameters as defined for COBRA map calculation. Residue-
specific 13C chemical shifts are automatically extracted for the
assigned cross peaks from the pseudo-spectra (Get Shifts) and are
eventually unfolded to account for possible aliasing in the 13C
dimension (Unfold?). This operation is based on the expected
chemical shifts for known amino acids. 1H, 15N, and available 13C
chemical shifts data are stored in the NMRView assignment table
and can be directly analyzed in terms of secondary structure
prediction by CSI (21) within NMR view.
9. Conclusions
Acknowledgments
References
1. Jaravine, V. A., Zhuravleva, A. V., Permi, P., time-optimized protein resonance assignment.
Ibraghimov, I., and Orekhov, V. Y. (2008) J. Magn. Reson. 187, 163–169.
Hyperdimensional NMR spectroscopy with 7. Schanda, P., Van Melckebeke, H., and
nonlinear sampling. J. Am. Chem. Soc. 130, Brutscher, B. (2006) Speeding up three-dimen-
3927–3936. sional protein NMR experiments to a few min-
2. Hiller, S., Fiorito, F., Wuthrich, K., and Wider, utes. J. Am. Chem. Soc. 128, 9042–9043.
G. (2005) Automated projection spectroscopy 8. Lescop, E., Schanda, P., Rasia, R., and
(APSY). Proc. Natl. Acad. Sci. USA 102, Brutscher, B. (2007) Automated spectral com-
10876–10881. pression for fast multidimensional NMR and
3. Lescop, E., and Brutscher, B. (2009) Highly increased time resolution in real-time NMR
automated protein backbone resonance assign- spectroscopy. J. Am. Chem. Soc. 129,
ment within a few hours: the “BATCH” strat- 2756–2757.
egy and software package. J. Biomol. NMR 44, 9. Deschamps, M., and Campbell, I. D. (2006)
43–57. Cooling overall spin temperature: protein
4. Lescop, E., Rasia, R., and Brutscher, B. (2008) NMR experiments optimized for longitudinal
Hadamard amino-acid-type edited NMR relaxation effects. J. Magn. Reson. 178,
experiment for fast protein resonance assign- 206–211.
ment. J. Am. Chem. Soc. 130, 5014–5015. 10. Pervushin, K., Vogeli, B., and Eletsky, A.
5. Lescop, E., and Brutscher, B. (2007) (2002) Longitudinal (1)H relaxation optimi-
Hyperdimensional protein NMR spectroscopy zation in TROSY NMR spectroscopy. J. Am.
in peptide-sequence space. J. Am. Chem. Soc. Chem. Soc. 124, 12898–12902.
129, 11916–11917. 11. Diercks, T., Daniels, M., and Kaptein, R. (2005)
6. Lescop, E., Schanda, P., and Brutscher, B. (2007) Extended flip-back schemes for sensitivity
A set of BEST triple-resonance experiments for enhancement in multidimensional HSQC-type
428 B. Brutscher and E. Lescop
Abstract
This chapter gives an overview of automated protein structure determination by nuclear magnetic resonance
(NMR) with the UNIO protocol that enables high to full automation of all NMR data analysis steps
involved. Four established algorithms, namely, the MATCH algorithm for sequence-specific resonance
assignment, the ASCAN algorithm for side-chain resonance assignment, the CANDID algorithm for NOE
assignment, and the ATNOS algorithm for signal identification in NMR spectra, are assembled into three
principal UNIO NMR data analysis components (MATCH, ATNOS/ASCAN, and ATNOS/CANDID)
that are accessed thanks to a particularly intuitive and flexible, yet powerful graphical user interface (GUI).
UNIO is designed to work independently or in association with other NMR software. The principal data
analysis components for sequence-specific backbone, side-chain and NOE assignment may be run separately
or out of sequence. User-intervention at individual stages is encouraged and facilitated by graphical tools
included for the preparation, analysis, validation, and subsequent presentation of the NMR structure.
Key words: Protein structure, NMR structure determination, Resonance assignment, NOE assign-
ment, Automated NMR structure determination, MATCH, ASCAN, ATNOS, CANDID, UNIO
protocol
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_22, © Springer Science+Business Media, LLC 2012
429
430 P. Guerry and T. Herrmann
Fig. 1. Scheme of the stepwise standard protocol for protein structure determination by NMR.
2. Specificities
of the UNIO
Program
The simplicity and ease of use of the UNIO suite ensure that a
structure calculation is up and running within only a few man-
hours of installation, making UNIO attractive to expert and casual
users alike. The following list provides an overview of computer
requirements, compatible input file formats, and molecular dynam-
ics programs that can be used in combination with UNIO.
1. Display resolution of 1,024 × 768 pixel or higher. True color
display (16-bit or 32-bit depth). Computer with either Linux
kernel 2.4 or above, or Mac OSX operating system 10.5 or
higher with Intel processors. A minimum of 100 megabytes of
disk space is required.
2. UNIO software application suite for automated protein NMR
structure determination. UNIO is free-of-charge for academic
use at http://www.unio-nmr.eu.
3. CNS (56), XPLOR-NIH (57), or CYANA(58) software pack-
age for NMR structure calculation by simulated annealing.
4. Input file for the amino acid sequence in any of the following for-
mats: BioMagResBank (59), XEASY (11), FASTA, ANSIG (9),
NMRVIEW (10), SPARKY (14), CYANA, CNS, XPLOR-NIH.
22 Comprehensive Automation for NMR Structure Determination of Proteins 433
3. Standard
UNIO Protocol
for Protein NMR
Structure The UNIO protocol for substantially automated protein NMR
Determination structure determination comprises sequence-specific backbone and
side-chain assignment followed by NOE assignment and NMR
structure calculation (Fig. 2).
1. Sequence-specific backbone resonance assignment with MATCH.
The standard experimental input data for MATCH (47) con-
sists of a set of three APSY datasets: 4D APSY-HACANH, 5D
APSY-HACACONH, and 5D APSY-CBCACONH (60, 61).
The MATCH algorithm yields polypeptide backbone reso-
nance assignments for the 1HN, 1Ha, 15N, 13Ca, 13Cb, and 13C¢
atoms.
2. Side-chain resonance assignment with ATNOS/ASCAN. The
standard experimental input data for ATNOS/ASCAN (48,
54) comprises the previously obtained sequence-specific back-
bone resonance assignments and a set of three NOESY spectra:
3D [1H,1H]-NOESY-15N-HSQC and two 3D [1H,1H]-
NOESY-13C-HSQC with the 13C carrier frequency in the ali-
phatic and the aromatic spectral regions. The ATNOS/ASCAN
approach yields meaningful side-chain resonance assignments,
i.e., resonance frequencies of atoms involved in many NOEs.
3. Automated NOESY assignment and NMR structure calcula-
tion. The standard experimental input data for ATNOS/
CANDID (19, 54) consists of the previously determined back-
bone and side-chain resonance assignments, and the three
aforementioned 3D NOESY spectra. ATNOS/CANDID in
combination with a simulated annealing program yields listings
of assigned NOESY peaks and the 3D protein structure.
In the following sections, the individual UNIO data analysis
components are presented in detail. Automated NMR signal identifi-
cation is described in Subheading 4, automated backbone assignment
in Subheading 5, automated side-chain assignment in Subheading 6,
and automated NOESY assignment in Subheading 7. The entire
UNIO protocol presented here has been successfully applied to more
434 P. Guerry and T. Herrmann
Fig. 2. Schematic outline of the UNIO protocol for highly automatic NMR protein structure determination. The three principal
UNIO modules are in bold font inside solid boxes along with the standard input data and output, in dashed and shaded
rectangles, respectively. The input data common to all three modules is shown in a thick dashed box at the top. Cyclic
symbols denote reevaluation of the experimental input data at the start of each iteration guided by the output of the previ-
ous cycle (new resonance assignments for ATNOS/ASCAN, intermediate protein structures for ATNOS/CANDID). In case, the
UNIO validation criteria are not met, the required interactive refinement is facilitated by UNIO reports.
4. Automated
NMR Signal
Identification
In the UNIO protocol, automated NOESY peak picking and
NOE signal identification in 2D homonuclear and heteronuclear-
resolved 3D [1H, 1H]-NOESY spectra is performed with the
ATNOS algorithm (54) in association with either ASCAN (48)
automated side-chain assignment (see Subheading 6) or CANDID
(19) automated NOE assignment and NMR structure calculation
(see Subheading 7).
4.1. Overview The main elements of ATNOS for NOESY spectral analysis are
of the ATNOS local baseline correction and evaluation of local noise level ampli-
Algorithm tudes, automated determination of spectrum-specific threshold
parameters, the use of symmetry relations, and the inclusion of
chemical shift information and the intermediate protein structures
to distinguish between NOE cross peaks and artifacts.
1. Input data for ATNOS. The input data consists of the amino
acid sequence of the protein, resonance frequencies of the
assigned atoms, and 2D or 3D NOESY spectra.
2. Determination of local baseline and local noise level. These tech-
niques are based on those previously introduced by the FLATT
(62) and AUTOPSY (53) algorithms.
3. Generation of a comprehensive set of NMR signals. Highly per-
missive criteria are applied that only require an initial minimal
signal-to-noise ratio and a local minimum.
436 P. Guerry and T. Herrmann
5. Automated
Backbone
Assignment
In the UNIO protocol, automated sequence-specific polypeptide
backbone NMR assignment is performed with the MATCH algo-
rithm (47). MATCH employs local optimization for tracing partial
sequence-specific assignments within a global, population-based
search environment, where the simultaneous application of local
and global optimization heuristics guarantees high efficiency and
robustness. MATCH thus makes combined use of the two pre-
dominant concepts in use for automated assignment of proteins
(see Subheading 5.2).
5.1. Overview The MATCH algorithm is founded on two main building blocks:
of the MATCH initialization and optimization. Novel concepts in MATCH are
Algorithm dynamic transition and inherent mutation that enable automatic
adaptation to the variable quality of experimental input data. The
concept of dynamic transition is incorporated in all major building
blocks of the MATCH algorithm, where it enables switching
between local and global optimization heuristics at any time dur-
ing the assignment process. Inherent mutation restricts the intrin-
sically required randomness of the evolutionary algorithm to those
regions of the conformation space that are compatible with the
experimental input data.
1. Input data for MATCH. The input consists of the amino acid
sequence of the protein, a statistical analysis of chemical shift
values of proteins contained in the BioMagResBank, and the
experimental NMR data in form of the frequency coordinates
of the NMR correlation signals.
22 Comprehensive Automation for NMR Structure Determination of Proteins 439
5.2. Memetic In general, algorithms for solving the resonance assignment prob-
Algorithm lem employ either local or global optimization. Local optimization
algorithms refine a preliminary solution by screening the adjacent
configuration space in search of information on the best candidate
solution. They can work in a highly deterministic fashion, follow-
ing a concrete optimization strategy. The benefit of local optimiza-
tion is high efficiency resulting from the assumption that the
underlying data do not contain information that is incompatible
with the rationale used by the algorithm. This efficiency is predict-
ably gained at the expense of robustness. Global optimization
algorithms, on the contrary, solve combinatorial problems by opti-
mizing all problem parameters independently and at once. They
are usually implemented in a population-based fashion such that
multiple candidate solutions located in different regions of the
configuration space are optimized simultaneously. A certain degree
of randomness may be involved, analogous to mutation in biologi-
cal evolution, e.g., genetic algorithms, the deteriorating influence
of misleading experimental input data is thus muted, and the risk
of getting trapped in local minima is greatly reduced. Overall, a
population-based global optimization approach has high robust-
ness but low efficiency due to the fact that numerous candidate
solutions have to be managed concurrently.
A memetic algorithm is the logical attempt to merge both
approaches, since it contains a local optimization routine embed-
ded in an evolutionary, global optimization algorithm. The evolu-
tionary algorithm is meant to explore the overall problem space,
while the local search heuristic refines discrete areas of this space.
With MATCH, the advantages of both approaches are exploited to
the fullest extent as local optimization efficiently traces partial solu-
tions inside a population-based (genetic) environment that pre-
serves robustness.
6. Automated
Side-Chain
Assignment
In the UNIO protocol, automated sequence-specific NMR assign-
ment of amino acid side-chain atoms is performed according to the
ATNOS/ASCAN approach (48, 54). ATNOS/ASCAN operates
on the 3D heteronuclear-resolved [1H,1H]-NOESY datasets that
are subsequently used to collect the input of NOE-distance con-
straints for the structure calculation. ATNOS/ASCAN makes use
of the chemical shift lists for the previously assigned backbone
atoms, and the knowledge of the covalent polypeptide structure. To
make inevitable imperfections of experimental input NMR data
tractable, the chemical shifts of the previously assigned backbone
and Cb atoms are used to guide both the peak-picking of the NOESY
spectra and the search for new side-chain resonance assignments.
6.2. Resonance As is the case for interactive side-chain assignment, the ATNOS/
Assignments Obtained ASCAN procedure performs better on interior, buried residues
by the ATNOS/ASCAN than on extensively solvent-exposed residues. In general, for both
Approach approaches, the completeness of the side chain assignments corre-
lates inversely with the degree of solvent-accessibility. This is read-
ily rationalized if one considers that a much larger number of NOEs
is generally observed for interior atoms than for atoms at or near
the protein surface. ATNOS/ASCAN thus primarily assigns side-
chain atoms that are involved in numerous inter-residue NOEs.
Note that it is a special advantage of the [1H,1H]-NOESY-
based ATNOS/ASCAN approach that the same datasets are used
for the amino acid side chain assignments and for the collection of
NOE upper distance constraints. Since adjustments of polypeptide
backbone chemical shifts have already been made when preparing
the input for ATNOS/ASCAN, this eliminates the need for further
chemical shift adjustments between datasets recorded with differ-
ent experimental conditions, which is an intrinsically laborious
procedure that may introduce unnecessary ambiguity into the fol-
lowing step of NOESY assignment and structure calculation.
7. Automated
NOESY Assignment
and NMR Structure
Calculation In the UNIO protocol, automated NOESY spectral analysis fol-
lows the ATNOS/CANDID approach (19, 54) that proceeds, as
all commonly used NOE assignment algorithms, in iterative cycles,
each consisting of exhaustive NOE signal identification and, in
part, ambiguous NOE assignments followed by a structure calcula-
tion. But in contrast to many other NOE assignment approaches
that operate on listings of peak positions and chemical shifts invari-
ant in all NOE assignment cycles, the combined use of ATNOS
NOE signal identification and CANDID NOE assignment waives
the common requirement for multiple rounds of manual peak list
preparation and refinement, and leads to a dramatic increase in the
efficiency and reproducibility of the NOESY spectral analysis.
7.1. Overview Each cycle of the iteratively performed NOESY spectral analysis
of the ATNOS/CANDID consists of automated NOESY peak picking with ATNOS, use of
Approach the resulting lists of peak positions and peak intensities as input for
CANDID automated NOE assignment, and use of a set of NOE
distance restraints from CANDID as input for the structure calcu-
lation. Between subsequent ATNOS/CANDID cycles, information
is transferred exclusively through the intermediate 3D structures,
in that the protein molecular structure obtained in a given cycle is
444 P. Guerry and T. Herrmann
7.2. Ambiguous The high NOE assignment ambiguity at the outset of a protein
Distance Restraints structure determination can be resolved by temporarily ignoring
cross peaks with too many (typically, more than two) assignment
possibilities and instead generating distance restraints for all assign-
ment possibilities of the remaining cross peaks. However, such a
procedure requires highly accurate chemical shift values and NOE
cross peak positions to be present in the input data and is hardly
achievable under realistic, experimental conditions. A more elegant
way for handling the initial chemical shift-based assignment ambi-
guity is given by the concept of ambiguous distance restraints
(63, 67). When using ambiguous distance restraints, each individ-
ual NOE cross peak is treated as the superposition of n degenerate
signals arising from each of its multiple initial chemical shift-based
assignments, using relative weights proportional to the inverse
sixth power of the corresponding interatomic distance. A NOE
cross peak uniquely assigned to a pair of hydrogen atoms, α and β ,
gives rise to an upper distance limit b for the corresponding dis-
tance dαβ ≤ b . A NOESY cross peak with two or more assignment
possibilities (n ≥ 2 ) is then interpreted as an ambiguous distance
restraint with an effective, d eff , or r −6 -summed distance
1
−
⎛ n ⎞ 6
d eff
= ⎜ ∑ di−6 ⎟
⎝ i =1 ⎠
7.3. Network- The concept of ambiguous distance restraint is quite efficient for
Anchored NOE improving and completing the NOESY assignment once a correct
Assignment preliminary polypeptide 3D fold is available, e.g., based on a lim-
ited set of interactively assigned NOESY cross peaks. However,
obtaining a correct initial protein fold at the outset of a de novo
structure determination often proves to be difficult, because struc-
ture-based filters used for the detection and elimination of errone-
ous cross peaks in the input data and for the discrimination between
multiple initial chemical shift-based cross peak assignments are not
yet operational.
To achieve reliable and robust automated NOE assignment for
de novo protein NMR structure determination, the NOE assign-
ment process cannot solely rely on chemical shift agreement between
resonance frequencies of assigned atoms and frequency coordinates
of the NMR signals, and the subsequent use of ambiguous distance
restraints. Indeed, techniques to remove artifacts prior to any
knowledge of a structure model must also be included.
One powerful concept for robust automated NOE assignment
is network-anchored assignment (19). Network-anchoring imitates
the modus operandi of an experienced spectroscopist who typically
decides on the assignment of an individual NOE cross peak on the
basis of the set of already assigned NOE cross peaks. Network-
anchored assignment exploits the observation that the correctly
assigned restraints form a self-consistent subset in any network of
distance restraints that is sufficiently dense for the determination of
a protein 3D structure. Network-anchoring thereby evaluates the
self-consistency of the NOE assignments independently of any
knowledge about the 3D protein structure, and in this way com-
pensates for the absence of 3D structural information at the outset
of a de novo structure determination. The requirement that each
NOE assignment must be embedded in the network of all other
assignments makes network-anchoring a sensitive approach for
detecting erroneous restraints that might artificially constrain
unstructured parts of the protein. Such restraints might not lead to
systematic constraint violation during the structure calculation,
and therefore might also escape 3D structure-based filtering meth-
ods. The concept of network-anchored assignment has proved effi-
cient and reliable in searching for the correct fold especially in the
initial phase of de novo NMR structure determinations.
448 P. Guerry and T. Herrmann
References
1. Billeter, M., Wagner, G., and Wüthrich, K. 10. Johnson, B. A., and Blevins, R. A. (1994)
(2008) Solution NMR structure determination Nmr View - a Computer-Program for the
of proteins revisited. J. Biomol. NMR 42, Visualization and Analysis of Nmr Data.
155–158. J. Biomol. NMR 4, 603–614.
2. Williamson, M. P., and Craven, C. J. (2009) 11. Bartels, C., Xia, T. H., Billeter, M., Güntert, P.,
Automated protein structure calculation from and Wüthrich, K. (1995) The Program Xeasy
NMR data. J. Biomol. NMR 43, 131–143. for Computer-Supported Nmr Spectral-Analysis
3. Wüthrich, K. (1986) NMR of Proteins and of Biological Macromolecules. J. Biomol. NMR
Nucleic Acids. Wiley, New York. 6, 1–10.
4. Altieri, A. S., and Byrd, R. A. (2004) Automation 12. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu,
of NMR structure determination of proteins. G., Pfeifer, J., and Bax, A. (1995) Nmrpipe - a
Curr. Opin. Struct. Biol. 14, 547–553. Multidimensional Spectral Processing System
5. Baran, M. C., Huang, Y. J., Moseley, H. N. B., Based on Unix Pipes. J. Biomol. NMR 6,
and Montelione, G. T. (2004) Automated anal- 277–293.
ysis of protein NMR assignments and struc- 13. Neidig, K. P., Geyer, M., Gorler, A., Antz, C.,
tures. Chem. Rev. 104, 3541–3555. Saffrich, R., Beneicke, W., et al. (1995) Aurelia,
6. Huang, Y. P. J., Moseley, H. N. B., Baran, M. a Program for Computer-Aided Analysis of
C., Arrowsmith, C., Powers, R., Tejero, R., Multidimensional Nmr-Spectra. J. Biomol.
et al. (2005) An integrated platform for auto- NMR 6, 255–270.
mated analysis of protein NMR structures. 14. Goddard, T. D., and Kneller, D. G. (2001)
Methods Enzymol. 394, 111–141. SPARKY 3. University of Californai, San
7. Gronwald, W., and Kalbitzer, H. R. (2004) Francisco.
Automated structure determination of proteins 15. Keller, R. L. J. (2004) Optimizing the process
by NMR spectroscopy. Prog. Nucl. Magn. of nuclear magnetic resonance spectrum analy-
Reson. Spectrosc. 44, 33–96. sis and computer aided resonance assignment.
8. Güntert, P. (2009) Automated structure deter- Ph.D. thesis. Diss. ETH Nr. 15947. ETH Zurich,
mination from NMR spectra. Eur. Biophys. J. Zurich, Switzerland.
38, 129–143. 16. Kobayashi, N., Iwahara, J., Koshiba, S.,
9. Kraulis, P. J. (1989) Ansig - a Program for the Tomizawa, T., Tochio, N., Güntert, P., et al.
Assignment of Protein H-1 2d-Nmr Spectra by (2007) KUJIRA, a package of integrated
Interactive Computer-Graphics. J. Magn. modules for systematic and interactive analysis
Reson. 84, 627–633. of NMR data directed to high-throughput
22 Comprehensive Automation for NMR Structure Determination of Proteins 449
NMR structure studies. J. Biomol. NMR 39, multidimensional nuclear magnetic resonance
31–52. spectra. J. Comput. Chem. 18, 139–149.
17. Mumenthaler, C., Güntert, P., Braun, W., and 28. Choy, W. Y., Sanctuary, B. C., and Zhu, G.
Wüthrich, K. (1997) Automated combined (1997) Using neural network predicted sec-
assignment of NOESY spectra and three- ondary structure information in automatic pro-
dimensional protein structure determination. J. tein NMR assignment. J. Chem. Inf. Comput.
Biomol. NMR 10, 351–362. Sci. 37, 1086–1094.
18. Gronwald, W., Moussa, S., Elsner, R., Jung, A., 29. Buchler, N. E. G., Zuiderweg, E. R. P., Wang,
Ganslmeier, B., Trenner, J., et al. (2002) H., and Goldstein, R. A. (1997) Protein NMR
Automated assignment of NOESY NMR spec- assignments using mean-field simulated anneal-
tra using a knowledge based method ing. Biophys. J. 72, Wp447–Wp447.
(KNOWNOE). J. Biomol. NMR 23, 271–287. 30. Croft, D., Kemmink, J., Neidig, K. P., and
19. Herrmann, T., Güntert, P., and Wüthrich, K. Oschkinat, H. (1997) Tools for the automated
(2002) Protein NMR structure determination assignment of high-resolution three-dimen-
with automated NOE assignment using the sional protein NMR spectra based on pattern
new software CANDID and the torsion angle recognition techniques. J. Biomol. NMR 10,
dynamics algorithm DYANA. J. Mol. Biol. 319, 207–219.
209–227. 31. Zimmerman, D. E., Kulikowski, C. A., Huang,
20. Linge, J. P., Habeck, M., Rieping, W., and Y. P., Feng, W. Q., Tashiro, M., Shimotakahara,
Nilges, M. (2003) ARIA: automated NOE S., et al. (1997) Automated analysis of protein
assignment and NMR structure calculation. NMR assignments using methods from artifi-
Bioinformatics 19, 315–316. cial intelligence. J. Mol. Biol. 269, 592–610.
21. Kuszewski, J., Schwieters, C. D., Garrett, D. S., 32. Gronwald, W., Willard, L., Jellard, T., Boyko,
Byrd, R. A., Tjandra, N., and Clore, G. M. R. E., Rajarathnam, K., Wishart, D. S., et al.
(2004) Completely automated, highly error- (1998) CAMRA: Chemical shift based com-
tolerant macromolecular structure determina- puter aided protein NMR assignments.
tion from multidimensional nuclear overhauser J. Biomol. NMR 12, 395–405.
enhancement spectra and chemical shift assign- 33. Leutner, M., Gschwind, R. M., Liermann, J.,
ments. J. Am. Chem. Soc. 126, 6258–6273. Schwarz, C., Gemmecker, G., and Kessler, H.
22. Huang, Y. J., Tejero, R., Powers, R., and (1998) Automated backbone assignment of
Montelione, G. T. (2006) A topology-con- labeled proteins using the threshold accepting
strained distance network algorithm for protein algorithm. J. Biomol. NMR 11, 31–43.
structure determination from NOESY data. 34. Moseley, H. N. B., Monleon, D., and
Methods Enzymol 62, 587–603. Montelione, G. T. (2001) Automatic determi-
23. Güntert, P. (2003) Automated NMR protein nation of protein backbone resonance assign-
structure calculation. Prog. Nucl. Magn. Reson. ments from triple resonance nuclear magnetic
Spectrosc. 43, 105–125. resonance data. Nuc. Magn. Reson. Biol.
24. Bernstein, R., Cieslar, C., Ross, A., Oschkinat, H., Macromol. 339, 91–108.
Freund, J., and Holak, T. A. (1993) Computer- 35. Coggins, B. E., and Zhou, P. (2003) PACES:
Assisted Assignment of Multidimensional Nmr- Protein sequential assignment by computer-
Spectra of Proteins - Application to 3d assisted exhaustive search. J. Biomol. NMR 26,
Noesy-Hmqc and Tocsy-Hmqc Spectra. 93–111.
J. Biomol. NMR 3, 245–251. 36. Malmodin, D., Papavoine, C. H. M., and
25. Olson, J. B., and Markley, J. L. (1994) Evaluation Billeter, M. (2003) Fully automated sequence-
of an Algorithm for the Automated Sequential specific resonance assignments of heteronuclear
Assignment of Protein Backbone Resonances - a protein spectra. J. Biomol. NMR 27, 69–79.
Demonstration of the Connectivity Tracing 37. Hitchens, T. K., Lukin, J. A., Zhan, Y. P.,
Assignment Tools (Contrast) Software Package. McCallum, S. A., and Rule, G. S. (2003)
J. Biomol. NMR 4, 385–410. MONTE: An automated Monte Carlo based
26. Lukin, J. A., Gove, A. P., Talukdar, S. N., and approach to nuclear magnetic resonance assign-
Ho, C. (1997) Automated probabilistic method ment of proteins. J. Biomol. NMR 25, 1–9.
for assigning backbone resonances of (C-13,N- 38. Moseley, H. N. B., Riaz, N., Aramini, J. M.,
15)-labeled proteins. J. Biomol. NMR 9, Szyperski, T., and Montelione, G. T. (2004) A
151–166. generalized approach to automated NMR peak
27. Bartels, C., Güntert, P., Billeter, M., and list editing: application to reduced dimension-
Wüthrich, K. (1997) GARANT - A general ality triple resonance spectra. J. Magn. Reson.
algorithm for resonance assignment of 170, 263–277.
450 P. Guerry and T. Herrmann
39. Eghbalnia, H. R., Bahrami, A., Wang, L. Y., Using Automatic Computer-Analysis of Contour
Assadi, A., and Markley, J. L. (2005) Diagrams. J. Magn. Reson. 95, 214–220.
Probabilistic identification of spin systems and 51. Carrara, E. A., Pagliari, F., and Nicolini, C.
their assignments including coil-helix inference (1993) Neural Networks for the Peak-Picking
as output (PISTACHIO). J. Biomol. NMR 32, of Nuclear-Magnetic-Resonance Spectra.
219–233. Neural Networks 6, 1023–1032.
40. Lin, H. N., Wu, K. P., Chang, J. M., Sung, T. 52. Antz, C., Neidig, K. P., and Kalbitzer, H. R.
Y., and Hsu, W. L. (2005) GANA - a genetic (1995) A General Bayesian Method for an
algorithm for NMR backbone resonance assign- Automated Signal Class Recognition in 2d
ment. Nucleic Acids Res. 33, 4593–4601. Nmr-Spectra Combined with a Multivariate
41. Masse, J. E., Keller, R., and Pervushin, K. Discriminant-Analysis. J. Biomol. NMR 5,
(2006) SideLink: Automated side-chain assign- 287–296.
ment of biopolymers from NMR data by rela- 53. Koradi, R., Billeter, M., Engeli, M., Güntert,
tive-hypothesis-prioritization-based simulated P., and Wüthrich, K. (1998) Automated peak
logic. J. Magn. Reson. 181, 45–67. picking and peak integration in macromolecu-
42. Masse, J. E., and Keller, R. (2005) AutoLink: lar NMR spectra using AUTOPSY. J. Magn.
Automated sequential resonance assignment of Reson. 135, 288–297.
biopolymers from NMR data by relative- 54. Herrmann, T., Güntert, P., and Wüthrich, K.
hypothesis-prioritization-based simulated logic. (2002) Protein NMR structure determination
J. Magn. Reson. 174, 133–151. with automated NOE-identification in the
43. Wang, J. Y., Wang, T. Z., Zuiderweg, E. R. P., NOESY spectra using the new software
and Crippen, G. M. (2005) CASA: An efficient ATNOS. J. Biomol. NMR 24, 171–189.
automated assignment of protein mainchain 55. Dancea, F., and Gunther, U. (2005) Automated
NMR data using an ordered tree search algo- protein NMR structure determination using
rithm. J. Biomol. NMR 33, 261–279. wavelet de-noised NOESY spectra. J. Biomol.
44. Kamisetty, H., Bailey-Kellogg, C., and NMR 33, 139–152.
Pandurangan, G. (2006) An efficient random- 56. Brünger, A. T., Adams, P. D., Clore, G. M.,
ized algorithm for contact-based NMR back- DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W.,
bone resonance assignment. Bioinformatics 22, et al. (1998) Crystallography & NMR system: A
172–180. new software suite for macromolecular structure
45. Vitek, O., Bailey-Kellogg, C., Craig, B., and determination. Acta Cryst. D 54, 905–921.
Vitek, J. (2006) Inferential backbone assign- 57. Schwieters, C. D., Kuszewski, J. J., Tjandra,
ment for sparse data. J. Biomol. NMR 35, N., and Clore, G. M. (2003) The Xplor-NIH
187–208. NMR molecular structure determination pack-
46. Wu, K. P., Chang, J. M., Chen, J. B., Chang, C. age. J. Magn. Reson. 160, 65–73.
F., Wu, W. J., Huang, T. H., et al. (2006) 58. Güntert, P., Mumenthaler, C., and Wüthrich,
RIBRA - An error-tolerant algorithm for K. (1997) Torsion angle dynamics for NMR
the NMR backbone assignment problem. structure calculation with the new program
J. Comput. Biol. 13, 229–244. DYANA. J. Mol. Biol. 273, 283–298.
47. Volk, J., Herrmann, T., and Wüthrich, K. 59. Doreleijers, J. F., Mading, S., Maziuk, D.,
(2008) Automated sequence-specific protein Sojourner, K., Yin, L., Zhu, J., et al. (2003)
NMR assignment using the memetic algorithm BioMagResBank database with sets of experi-
MATCH. J. Biomol. NMR 41, 127–138. mental NMR constraints corresponding to the
48. Fiorito, F., Herrmann, T., Damberger, F. F., structures of over 1400 biomolecules deposited
and Wüthrich, K. (2008) Automated amino in the Protein Data Bank. J. Biomol. NMR 26,
acid side-chain NMR assignment of proteins 139–146.
using C-13- and N-15-resolved 3D [H-1,H- 60. Hiller, S., Wider, G., and Wüthrich, K. (2008)
1]-NOESY. J. Biomol. NMR 42, 23–33. APSY-NMR with proteins: practical aspects and
49. Neidig, K. P., Saffrich, R., Lorenz, M., and backbone assignment. J. Biomol. NMR 42,
Kalbitzer, H. R. (1990) Cluster-Analysis and 179–195.
Multiplet Pattern-Recognition in 2-Dimensional 61. Hiller, S., Fiorito, F., Wüthrich, K., and Wider,
Nmr-Spectra. J. Magn. Reson. 89, 543–552. G. (2005) Automated projection spectroscopy
50. Garrett, D. S., Powers, R., Gronenborn, A. M., (APSY). Proc. Natl. Acad. Sci. USA 102,
and Clore, G. M. (1991) A Common-Sense 10876–10881.
Approach to Peak Picking in 2-Dimensional, 62. Güntert, P., and Wüthrich, K. (1992) Flatt - a
3-Dimensional, and 4-Dimensional Spectra New Procedure for High-Quality Base-Line
22 Comprehensive Automation for NMR Structure Determination of Proteins 451
Abstract
In solution or solid-state, determining the three-dimensional structure of biomolecules by Nuclear
Magnetic Resonance (NMR) normally requires the collection of distance information. The interpretation
of the spectra containing this distance information is a critical step in an NMR structure determination. In
this chapter, we present the Ambiguous Restraints for Iterative Assignment (ARIA) program for auto-
mated cross-peak assignment and determination of macromolecular structure from solution and solid-state
NMR experiments. While the program was initially designed for the assignment of nuclear Overhauser
effect (NOE) resonances, it has been extended to the interpretation of magic-angle spinning (MAS) solid-
state NMR data. This chapter first details the concepts and procedures carried out by the program. Then,
we describe both the general strategy for structure determination with ARIA 2.3 and practical aspects of
the technique. ARIA 2.3 includes all recent developments. such as an extended integration of the
Collaborative Computing Project for the NMR community (CCPN), the incorporation of the log-har-
monic distance restraint potential and an automated treatment of symmetric oligomers.
Key words: Ambiguous distance restraint, Structure calculation, Automated assignment, MAS,
Solid-state NMR, CCPN, NOE, ARIA, PDSD, CHHC
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_23, © Springer Science+Business Media, LLC 2012
453
454 B. Bardiaux et al.
2. Materials
2.1. ARIA Software The following software packages are required to use ARIA.
Package
1. ARIA software package. ARIA (6) is written in the program-
ming language Python (15). The current version is 2.3.
Instructions on how to install ARIA can be found in the ARIA
installation archive, which should be downloaded from http://
aria.pasteur.fr. ARIA can be installed on computers operating
under Linux, Windows or Mac OS X.
2. CNS software. To enable specific features used by ARIA, it is
necessary to compile the CNS program (16) with libraries pro-
vided within the ARIA package.
3. Optional: CCPNmr Analysis software package (version 2 or
later) (17). ARIA uses the CCPN data model to read input
data and to store all results in a general format.
4. Optional: Access to a computer cluster for distributed
calculation.
2.2. Input Data The minimal set of data required by ARIA consists of (see Note 1):
1. Definition of the molecular system.
2. List(s) of chemical shift assignments of 1H (for 2D-NOESY)
and 13C/15N if necessary for 3D-NOESY or for MAS solid-
state NMR spectra (see Note 2).
3. One or more lists of cross-peaks with chemical shift positions
in each dimension and peak volumes/intensities. Individual
peaks can be either fully assigned, partially assigned or com-
pletely unassigned. A list of cross-peaks generally corresponds
to the peaks picked in a particular spectrum. It is recommended
that similar experiments performed with different mixing times
are entered as separated lists.
ARIA also integrates various data types for additional experi-
mental information. All restraints must be in CNS “tbl” format
(see Note 1).
1. Hydrogen bonds: The distance between hydrogen donor and
acceptor as well as the distance between acceptor and hydrogen.
2. Dihedral angles: Dihedral angle restraints incorporated using a
flat-bottom harmonic-wall potential.
3. J-couplings: Calculated J-couplings are directly refined against
observed J-couplings.
4. Residual dipolar couplings: Residual dipolar coupling (RDC)
data as restraints.
5. Distance restraints: Preformatted distance restraints, e.g., from
manual assignments.
456 B. Bardiaux et al.
2.3. Software Additional software required to analyze the quality of the final
for Structural Quality structure ensembles:
Checks
1. PROCHECK (18).
2. WHAT IF (or WHAT_CHECK) (19).
3. ProSa II (or ProSa 2003) (20).
4. MolProbity suite (21).
3. Methods
3.1. Preparation Phase Before cross-peak assignment and structure calculation, the fol-
lowing steps are automatically performed by ARIA. First, the data
are checked and filtered for errors and inconsistencies. The pro-
gram then creates the molecular topology of the system.
3.1.1. Data Filtering When checking the chemical shift assignments for consistency,
ARIA considers three possible situations:
1. A unique assignment consisting of a single atom and a single
chemical shift.
2. A degenerate chemical shift assignment, where one group of
equivalent atoms is assigned to exactly one chemical shift.
3. An assignment of the two substituents of a prochiral group,
which can have one or two chemical shifts.
In the latter case, floating chirality assignment (22) is used in
the resulting restraints (cf. Subheading 3.3.5). Peaks that lack fre-
quency information or with incorrect/missing peak sizes are
removed (see Note 3).
23 ARIA for Solution and Solid-State NMR 457
a b
ltering
Initial cross-peak assignment
Molecular topology creation
ARIA
Iterative protocol
c
Chemical
Cross-peak lists Calibration
shits assignments
d
Violation Analysis
Molecular Structure
ntion ensemble e
Structure ensemble
Noise peaks removal
f
Partial Assignment
Additional restraints g
(dihedrals, RDC,
Restraints Merging
restraints
Distance
distance)
Structure Calculation
ARIA nition h
(GUI)
Floating chirality assignment
j i
Generation of report
nement in explicit solvent
Quality analysis
Structure quality
statistics
Fig. 1. Description of the ARIA protocol workflow. Rounded rectangles indicate steps performed by ARIA, folded rectangles
correspond to user provided input-data and trapezoids represent results.
3.1.2. Molecular Topology From the definition of the molecular system provided as input data,
Creation ARIA creates a molecular topology file (MTF) with the program
CNS (16). Name, chemical type, charge and mass of each atom as
well as the covalent connectivity are defined in the MTF. An
extended conformation of the molecule is then generated by CNS
and the coordinates are stored in a PDB file (cf. Subheading 3.8).
The molecular topology is created automatically for standard bio-
polymers. If applicable, topological features can be easily defined
by the user through the graphical interface (cf. Subheading 3.7).
3.2. Initial Cross-Peak For every cross-peak, ARIA uses the chemical shift lists from the
Assignment sequential resonance assignment to derive possible assignments. As
illustrated in Fig. 2, the peak position is defined by its frequency
3.2.1. Chemical-Shift
coordinates (c1, c2) in each dimension of the spectrum. To account
Based Assignment
for the limited precision in chemical shift measurements, for the
uncertainty of the cross-peak coordinates and for systematic exper-
imental errors, chemical shift tolerances (d1, d2) are applied around
the peak position. The tolerances should be chosen to be sufficiently
458 B. Bardiaux et al.
dimension 2
c1−δ1 c1 c1+δ1
c2+δ2
pz
py
c2
px
c2−δ2
pa pb pc pd dimension 1
Fig. 2. Illustration of the assignment of a cross-peak. c1,c2 denote the peak coordinates in
frequency space. The assignment frequency window is indicated by the solid black square,
defined from the chemical shifts tolerances d1 and d2. The coordinates of the (hypotheti-
cal) correct assignment are represented by the gray dashed lines (pb, py). Multiple reso-
nances within the tolerance window (pa, pb, pc, pd in dimension 1 and px, py, pz in the other
dimension) give rise to 12 assignment possibilities.
3.2.2. Structural Rules ARIA can use information about the secondary structure organiza-
for Symmetric Oligomers tion of the system under investigation to remove unlikely assign-
ments. ARIA uses simple rules (23) to assign some cross-peaks as
intermonomer before the structure calculation, using the predicted
23 ARIA for Solution and Solid-State NMR 459
3.2.3. Network Anchoring ARIA implements a network anchoring approach (8) to reduce the
number of possibilities of cross-peak assignments prior to structure
calculation. The approach is based on the ranking of each assign-
ment, calculated using the information about the assignments of
neighboring nuclei in 3D space, and is efficient because true assign-
ments form a self-consistent subset of the network of all possible
assignments (see ref. 8, 13 for details). The behavior of network
anchoring is controlled by a set of user-defined parameters:
1. “High network-anchoring (NA) score per residue threshold”
high
(N res ).
min
2. “Minimal NA score per residue threshold” (N res ).
min
3. “Minimal NA score per residues threshold” (N atom ).
A peak is conserved if one of the following rules is verified:
S res ≥ N res
high
(1)
S res ≥ N res
min
and S atom ≥ N atom
min
(2)
where Sres and Satom are respectively the residue-wise and atom-wise
network anchoring score. Even though the network anchoring
approach does not directly rely on 3D structure information, it is
still possible to use it after the first ARIA iteration.
3.3. Iterative Structure The most important idea that underlies the ARIA methodology is
Calculation the concept of Ambiguous Distance Restraints (ADR) (2). In the
framework of the ADR, each NOESY cross-peak is treated as the
3.3.1. Ambiguous Distance
superposition of the signals from each of its multiple assignments
Restraints
possibilities: the NOE intensity depends on the sum of the inverse
sixth power of all the individual proton–proton distances that con-
tribute to the signal. An effective distance D is thus derived as:
1
−
⎛ Nc ⎞ 6
D = ⎜ ∑ dc−6 ⎟ (3)
⎝ c =1 ⎠
3.3.2. Distance Calibration The simplest model to derive distances from NOE signal intensity
is the Isolated Spin Pair Approximation (ISPA), which considers
only the observed spin pair, neglecting spin diffusion through third
nuclei. For short mixing times, ISPA provides a good approxima-
tion to relate an NOE volume (Vij) to the distance dij of two inter-
acting spins i and j:
Vij = Cdij−6 (4)
∑V exp
C= i
(5)
∑ dˆ
i
i
−6
where dˆi is the average effective distance for NOE i in the con-
former ensemble. In the case of multiple assignment possibilities,
dˆi is calculated according to equation of ADR Eq. 3. Finally, the
calibrated distance is obtained by:
1
−
d = (C −1V exp ) 6
(6)
In the case of NOE between two groups of magnetically equiv-
alent spins (e.g., methyl groups and aromatic rings), averaging
effects are taken into account by expanding Eq. 4 (see Note 9).
Magnetization can also be transferred from one spin to another
not only directly but also by spin diffusion, i.e., indirectly via other
spins in the vicinity. For longer mixing times, the spin-diffusion
phenomenon must be considered in the estimation of the distance.
When applying ISPA the resulting interproton distances are there-
fore mostly underestimated. ARIA employs relaxation matrix the-
ory to account for indirect magnetization transfer. In this formalism,
cross-peak volumes at mixing time tm can be calculated given the
volumes at tm = 0 and the matrix of auto- and cross-relaxation rates,
R (24):
Vij (t m ) = CVij (0)(exp(−Rt m ))ij (7)
where d̂ is the average effective distance, and V exp and V th are the
experimental and theoretical NOE volumes, respectively. When
using spin-diffusion corrected distances, the distance bounds cal-
culated from the theoretical volume may also be of use for the
structure calculation (25). In ARIA 2.3, the spin-diffusion correc-
tion is performed by the python core of ARIA and not by CNS
routines. It is also important to note that every spectrum is inde-
pendently calibrated. Still, these models are approximate and it is
common practice to restrain the distance to an interval to account
for uncertainties in the distances (see Note 10). This interval is
thus defined by lower and upper distance bounds, L and U:
L = d − Δ,U = d + Δ where Δ = 0.125d 2 (9)
3.3.3. Violation Analysis To identify incorrect assignments and noise peaks, the calibrated
and Noise Peak Removal restraints are treated with a violation analysis, following the struc-
tural consistency hypothesis (3, 26): incorrectly assigned peaks or
noise peaks are not consistent with the 3D structure determined
with all experimental data. To assess whether a particular restraint
follows the “general trends” imposed on the structures by the entire
data set, the obtained distance bounds are compared to the corre-
sponding distances found in the conformer ensemble. A restraint is
considered as violated if the distance found in the structure lies
outside the bounds by more than a user-defined violation tolerance,
t. To identify systematically violated restraints, each conformer in
the ensemble is analyzed. The fraction, f i , of conformers violating
restraint i is calculated according to:
1 S
fi = ∑ max(Θ(Li − t − di(k) ), Θ(di(k) −U i − t ))
S k =1
(10)
where Li and Ui denote the lower and upper bounds of the i-th
restraint, di(k) designates the distance found in the k-th conformer;
Q is the Heaviside step function and S is the total number of con-
formers analyzed. A restraint is classified as violated if f i exceeds a
user-defined violation threshold (50% by default). The correspond-
ing cross-peak is thus removed from the list of active peaks for the
next iteration. During the course of the protocol, the violation
tolerance, is reduced from iteration to iteration to ensure that most
of the inconsistent peaks are removed.
Nc
∑w
c =1
c =1 (12)
∑w
1
c ≥p (13)
3.3.5. Calculation On the basis of the merged restraints list, a new structure ensemble
of Structure Ensemble is calculated with the program CNS (16) through a molecular
dynamics simulated annealing (MDSA) protocol. ARIA provides
two forms of molecular dynamics : in Cartesian or torsion angle
space. Torsion angle molecular dynamics (TAD) (27) reduces the
calculation time and allows for higher MDSA temperatures, while
generally increasing the convergence radius. The molecular struc-
tures obtained with TAD also provide better local geometries. The
MDSA protocol used in ARIA is divided into two phases : an initial
high temperature search phase, and a cooling phase where the tem-
perature slowly decreases. The second part of the cooling stage is
performed in Cartesian coordinates. The length of the cooling
stages determines the slope of the bath temperature cooling func-
tion. It has been shown that this parameter plays an important role
in the convergence properties of the ARIA calculation for highly
ambiguous data (28). The MDSA protocols implemented in ARIA
(3) are optimized for the application of ambiguous distance
restraints and for the violation analysis method. The minimization
protocols are based primarily on separate scaling of different energy
terms with relatively low force constants. Any other structural
23 ARIA for Solution and Solid-State NMR 463
Table 1
Important protocol parameters, their location in the GUI, and
defaults values (if applicable)
3.3.6. Restraint Energy The aim of the MDSA protocol is to find a global energy minimum
Function of an objective function that incorporates experimental data and
physical energy. The latter is quantified by using a molecular
dynamics force field. Experimental data are integrated in the form
of conformational restraints entering the objective function via an
energy potential. For distance restraints, ARIA employs an flat-
bottom harmonic-wall potential with zero-energy between the dis-
tance bounds and linear asymptotes (3). This potential allows for
large distance violations as may occur in an automated assignment
procedure. Nevertheless, it is still difficult to correctly evaluate the
bounds and the relative weight to apply to the data. Recently, we
have introduced an new error-tolerant potential where lower and
upper bounds are replaced by a bounds-free log-harmonic potential
(14). This potential derives from a Bayesian analysis showing that
NOEs and the derived distances ideally follow the log-normal dis-
tribution (29, 30). In ARIA, we also retain another important fea-
ture of this Bayesian approach: automatic determination of the
optimal weight for the experimental data (31). The log-harmonic
potential is applied during the second cooling stage of the MDSA
and during water refinement. The weight for the distance restraints,
wdata , is iteratively evaluated as:
n
wdata = (14)
χ (X )
2
where, for each restraint i, dˆi is the effective distance Eq. 3 calcu-
lated from the current structure, and di is the target distance of
the restraint. This approach was shown to generally improve the
accuracy as well as the quality of the structures calculated from
assigned restraints (14). Our initial experience in using ARIA with
real (noisy and ambiguous) data indicates that the log-harmonic
restraint potential is preferable.
3.3.7. Symmetric The symmetry of the system is maintained during the calculation
Oligomers by adding a symmetry target function to the objective energy func-
tion (32). This target function contains terms that ensure the
symmetry relation between the monomers and keep them in the
vicinity of each other (Packing, see Note 11).
3.3.8. Floating Chirality The treatment of unassigned prochiral groups is realized with a
Assignment floating chirality assignment approach (22). The two substituents
of a prochiral center (methylene protons or methyl protons of iso-
propyl groups) are often difficult to assign stereo-specifically, in
terms of chemical shifts. In each proton dimension, a resonance
23 ARIA for Solution and Solid-State NMR 465
3.4. Solvent The simplified force field parameters for nonbonded contacts
Refinement applied to structure calculations in vacuo often produce structures
that contain artifacts (unrealistic side-chain packing and unsatisfied
hydrogen bond donors or acceptors). Therefore, the final struc-
tures of the last ARIA iteration are automatically refined in a shell
of explicit solvent (water or DMSO molecules). This refinement
consists in a short MD with a complete force field, which includes
coulombic and Lennard-Jones potentials. The covalent parameters
used in the refinement (33) are consistent with the force field used
for structure calculation and validation, thus avoiding systematic
differences that could influence validation results. It has been
shown that the refinement in solution significantly improves the
quality of the structure (33–35).
3.5. Results Export
At the end of the ARIA protocol, assigned peak lists, restraint lists,
and Generation
along with violations, and final structure ensembles (last iteration
of Output Files
and solvent refined) are automatically exported into a CCPN proj-
3.5.1. Export to CCPN ect (see Fig. 3). Data exchange, further analysis of results, and
management of ARIA runs are then facilitated through the use of
the CCPN program suite (cf. Subheading 3.11).
ARIA
Cross-peak assignments
CCPN
project CCPN Analysis
Fig. 3. Communication interface between ARIA and CCPN for import of input data and export of results.
466 B. Bardiaux et al.
3.5.2. Report Files For every iteration, ARIA creates the following report files:
1. report summarizes analyses of the restraint lists and the
structure ensemble (number of restraints applied, violations,
ensemble precision).
2. noe_restraints.unambig, noe_restraints.ambig
tabulates information about unambiguous and ambiguous
restraints, respectively. For each restraint, the reference cross-
peak, restraint bounds and the average distance found in the
ensemble are provided. The result of violation analysis is also
given here (see Note 12).
3. noe_restraints.violations lists all violated restraints.
4. noe_restraints.assignments lists the tentative assign-
ments corresponding to every restraint. The nature of the
assignment(s) is also given (fully, partially or unassigned cross-
peaks).
5. noe_restraints.xml, noe_restraints.pickle stores
the complete list of cross-peak based distance restraints in
XML format and Python binary format. The latter is required
for further assignment analysis in the ARIA GUI (cf.
Subheading 3.10.2).
3.5.3. Quality Checks To evaluate the structural quality of both the final set of structures
and the solvent-refined ensemble, ARIA makes use of the programs
WHAT IF (19), PROCHECK (18), ProSa (20) and MolProbity
(21). Separate report files are generated for every program, named
quality_checks.*, and are stored in the directories of the
respective ensembles (last iteration and solvent-refined). Overall
quality scores are tabulated in the file quality_checks, whereas
WHAT-IF score profiles along the molecular sequence are gener-
ated in both textual and graphical forms (cf. Subheading 3.10.3).
3.5.4. CNS Analyses CNS scripts calculate restraint energies, ensemble RMSDs, an opti-
mal superposition of the final structure ensemble (with automated
determination of flexible and rigid regions), and an unminimized
average structure. Analyses of restraints from complementary
experimental data are also given. Results are stored in the directory
analysis/.
In the following sections, we detail the typical procedure to be
followed by a user to perform an ARIA calculation. In a structure
determination project, the general procedure consists of repeated
ARIA runs using revised results from a previous calculation as input
data (Fig. 4).
3.6. Conversion Since most NMR software packages use proprietary formats for
of Input Data data storage, the interconversion step required to transfer data with
other applications such as ARIA can lead to a loss of information.
23 ARIA for Solution and Solid-State NMR 467
Fig. 4. A series of ARIA runs in a typical structure determination project, with several cycles of structure calculations and
cross-peak assignments punctuated by manual inspection and correction of experimental input data.
3.7. Specification 1. Project creation. All program parameters and locations of the
of ARIA Project input data are stored in single project file (in XML format). To
Parameters conveniently change or review the project settings, ARIA pro-
vides a Graphical User Interface (GUI) (Fig. 5). Entering the
following command will start the GUI and load the project
definition from project.xml (see Note 15)
aria2 --gui project.xml
Fig. 5. Graphical User Interface of ARIA 2.3 for project management, where data and protocol settings can be modified
graphically.
23 ARIA for Solution and Solid-State NMR 469
angle restraints (see Note 8), RDC (see Note 16) and J-couplings
(see Note 17) should be defined.
6. Symmetry. ARIA can treat oligomers with C2, C3, C5 or D2
symmetry (see Note 11).
7. Specifying topology patches. By default, ARIA supports the fol-
lowing cases: Disulfide bridges (unambiguous or ambiguous)
(2), Histidine protonation states, cis-proline and tetrahedral
coordination of Zinc ions. In the case of nonstandard residues
or other chemical compounds, manual intervention of the user
is required (see Note 18).
8. Iteration parameters. The mode of restraint calibration has to
be specified : ratio of average (default), spin-diffusion correction
or fixed bounds (see Note 10). For every iteration, default val-
ues are provided for protocol parameters (Table 1) and the
network-anchoring thresholds (see Note 19).
9. Job Manager. Distributing structure calculations to multiple
processors speeds up the ARIA protocol. ARIA provides sup-
port for several job submission modes (see Note 20). The
appropriate command should be entered and the correct path
to the remote CNS program executable should be specified.
10. Structure calculation parameters. The remaining parameters
are related to the molecular dynamics simulated annealing, and
in particular the number of steps, restraint force constants and
potential shape (flat-bottom-harmonic-wall and log-harmonic).
3.8. Project Setup At this point, the project must be set up with the following
command.
aria2 --setup project.xml
The project is then validated and ARIA creates the directory
tree for the project (directory run1). As shown in Fig. 6, the results
of the successive iterations are stored in structures/, each iteration
having its own subdirectory, e.g., structures/it0/. Experimental
data files are copied into their respective directory in data/ (see
Note 21). Report files for the cross-peak filtering procedure are
stored in data/spectra/. All data, protocols, parameters, and topol-
ogy files used by CNS reside in the cns/ subdirectory.
3.9. Starting an ARIA It is now possible to launch the ARIA calculation, using the follow-
Run ing command:
aria2 project.xml
ARIA will then automatically perform all the steps listed in
Subheadings 3.1–3.5. The main ARIA job will be executed on the
local machine where it has been started. According to the job man-
ager settings of the project, the structure calculations will be
23 ARIA for Solution and Solid-State NMR 471
begin le
it1
analysis Various analysis results (performed by CNS)
run1 structures ...
graphics les (PostScript)
ARIA run directory les for each it8
iteration are stored here molmol le to visualize restraints
Last iteration
cns les
Fig. 6. Illustration of the directory tree of an ARIA project and details about the content. Final results can be found in the
directories marked in gray.
3.10. Checking In the next paragraphs, we list the points of interests when inspect-
the Results ing the calculation results, along with some guidance on how to
correct input data and adapt the protocol parameters.
3.10.2. Automated 1. The report files listed in Subheading 3.5 provide analyses on all
Assignments restraints and particularly which restraints have been classified
as violated. Restraints showing consistent violation greater
than 0.1 Å should be inspected manually. Restraints with large
upper-bound violations (³5 Å) in the majority of the conform-
ers (³85%) usually result from incorrect assignments. Restraints
detected as such should not be used in a later ARIA run and
the corresponding cross-peak removed from its respective
spectrum. Other assignments should be considered as “reli-
able” in a subsequent run.
2. Analyzing text files for violations and assignments can be a
tedious task. ARIA also provides ways to investigate this in a
graphical manner (37). Postscript files describing the restraints,
based on the RMS of violations are generated automatically
during a run. These values are displayed at the residue level, in
the form of a profile along the protein sequence, or as a contact
map for the RMS of violations per residue pair (Fig. 7a). The
contact map displays the sum of the RMS of violations per resi-
due pair. In the profile, the sum of the RMS of violations per
residues is plotted along the protein sequence. In addition, the
program provides an interactive tool to browse assignments at
the residue level (Peak map). A peak-map can be viewed for all
iterations in the ARIA GUI (Fig. 8). Clicking on a contact
Fig. 7. Per-residue quality plots. (a) Contact map displaying the sums of RMS deviations and a profile of the RMS deviations.
(b) WHATIF score profiles along the sequence. The RMS deviations are plotted on a color scale (figure adapted from ref. 25).
23 ARIA for Solution and Solid-State NMR 473
Fig. 8. Interactive peak map. Right panel of the ARIA 2.3 GUI showing the interactive peak map at iteration 8 of an ARIA run.
Each pixel of the map located between residues i and j is clickable and opens an assignment report, which contains the
list of peaks that exist between residues i and j, along with their contributions (figure adapted from ref. 25).
Fig. 9. Screenshot of CCPNmr Analysis windows showing the result of an ARIA run.
3.11. Preparing To use the result of an ARIA run to further improve the structure,
a New Run it may be necessary to correct the input data. At this stage, we rec-
ommend preparing a new ARIA project for better bookkeeping.
CCPNmr Analysis also offers a utility to manage the input and
output of successive ARIA runs (Fig. 9). The same CCPN project
can be used in multiple ARIA runs.
3.11.2. Adjusting In the new project file, protocol parameters may also be changed
Parameters according to the result of a previous calculation. We list here the
most important parameters that ought to be adapted.
1. The number of dynamic steps required for convergence is
determined by the system size and the level of ambiguity or
incompleteness of the input data. Default values work well for
systems up to about 100 residues studied with NOESY.
However, for larger systems (e.g., symmetric oligomers) or
when MAS solid-state NMR data are used, it might become
necessary to increase the number of steps in the cooling stage
476 B. Bardiaux et al.
4. Notes
15. A user can also choose the “New” item in the GUI menu
“Project” to create a new project. As an alternative, the follow-
ing command
aria2 --project_template project.xml
will create a new project file.
16. Residual dipolar coupling data can be incorporated as restraints
following two alternative approaches: direct (SANI) or indirect
(VEAN). For SANI, the user has to specify the rhombicity and
magnitude of the alignment tensor (65). Several methods exist
to predict these parameters, from the distribution of the RDC
values (66) or from the shape of the molecule (67). VEAN
uses intervector projection angle restraints which must be gen-
erated with a separate program (68).
17. The correlation between a three-bond measured J-coupling
and the corresponding dihedral angle is modeled by the Karplus
curve. Default values for the parameters of the Karplus curve
are given for 3J(HNHa).
18. An MTF can be specified in the project file. Changes must be also
made to the CNS topology, linkage, and parameter files. Definitions
of the additional residues or compounds must be added to the
ARIA dictionary (files atomnames.xml and iupac.xml).
A detailed explanation is given on the ARIA Web site.
19. We recommend the use of the network-anchoring only for the
first 3 iterations. Too stringent thresholds or an application of
network-anchoring during more ARIA iterations may bias the
assignment process toward an incorrect structure (13).
20. Jobs can be submitted via ssh commands or with the follow-
ing batch queuing systems: PBS (69), SGE (70) or Condor
(71). Alternatively, CCPN users can submit their ARIA calcu-
lation to the CCPNGrid portal server at http://www.webapps.
ccpn.ac.uk/ccpngrid/.
21. Only local copies of data files are used for structure calculation.
Changes in the original files will thus become active only in the
next project setup.
22. For systems of about 100 residues, well converged ensembles
show average energies of the order of 1,000 kcal/mol. Normal
energy variation is about 10%, the total average energy scaling
is approximately linear with the system size.
23. Others methods are available to estimate the credibility of the
structures, notably by scrutinizing the information content of
the data (72). For instance, the completeness (73) of a restraint
set provides insight into the local reliability of each structure.
The completeness is the ratio between the number of observed
restraints and the number of expected restraints. We recom-
mend the method AQUA (73) to perform such analysis.
480 B. Bardiaux et al.
Acknowledgments
References
1. Wuthrich, K. (1986) NMR of Proteins and 6. Rieping, W., Habeck, M., Bardiaux, B.,
Nucleic Acids, Wiley-Interscience New York. Bernard, A., Malliavin, T., and Nilges, M.
2. Nilges, M. (1995) Calculation of protein struc- (2007) ARIA2: automated NOE assignment
tures with ambiguous distance restraints. and data integration in NMR structure calcula-
Automated assignment of ambiguous NOE tion. Bioinformatics 23, 381–382.
crosspeaks and disulphide connectivities. J. Mol. 7. Castellani, F., van Rossum, B., Diehl, A.,
Biol. 245, 645–660. Schubert, M., Rehbein, K., and Oschkinat, H.
3. Nilges, M. and O’Donoghue, S. I. (1998) (2002) Structure of a protein determined by
Ambiguous NOEs and automated NOESY solid-state magic-angle-spinning NMR spec-
assignment. Prog. NMR Spec. 32, 107–139. troscopy. Nature 420, 98–102.
4. Linge, J. P., O’Donoghue, S. I., and Nilges, M. 8. Herrmann, T., Güntert, P., and Wüthrich, K.
(2001) Automated assignment of ambiguous (2002) Protein NMR structure determination
nuclear overhauser effects with ARIA. Methods with automated NOE assignment using the
Enzymol. 339, 71–90. new software CANDID and the torsion angle
5. Linge, J. P., Habeck, M., Rieping, W., and dynamics algorithm DYANA. J. Mol. Biol. 319,
Nilges, M. (2003) ARIA: automated NOE 209–227.
assignment and NMR structure calculation. 9. Fossi, M., Castellani, F., Nilges, M., Oschkinat,
Bioinformatics 19, 315–316. H., and van Rossum, B. (2005) SOLARIA: a
23 ARIA for Solution and Solid-State NMR 481
protocol for automated cross-peak assignment Arendall, W. B., Snoeyink, J., Richardson, J. S.,
and structure calculation for solid-state magic- and Richardson, D. C. (2007) MolProbity: all-
angle spinning NMR spectroscopy. Angew. atom contacts and structure validation for pro-
Chem. Int. Ed. Engl. 44, 6151–6154. teins and nucleic acids. Nucleic Acids Res. 35,
10. Loquet, A., Bardiaux, B., Gardiennet, C., W375–383.
Blanchet, C., Baldus, M., Nilges, M., Malliavin, 22. Folmer, R. H., Hilbers, C. W., Konings, R. N.,
T., and Böckmann, A. (2008) 3D Structure and Nilges, M. (1997) Floating stereospecific
Determination of the Crh Protein from Highly assignment revisited: application to an 18 kDa
Ambiguous Solid-State NMR Restraints. protein and comparison with J-coupling data.
J. Am. Chem. Soc. 130, 3579–3589. J. Biomol. NMR 9, 245–258.
11. Manolikas, T., Herrmann, T., and Meier, B. 23. Duggan, B., Legge, G., Dyson, H., and Wright,
(2008) Protein structure determination from P. (2001) SANE (Structure Assisted NOE
(13)C spin-diffusion solid-state NMR spectros- Evaluation): an automated model-based
copy. J. Am. Chem. Soc. 130, 3959–3966. approach for NOE assignment. J. Biomol. NMR
12. Wasmer, C., Lange, A., Melckebeke, H. V., 19, 321–329.
Siemer, A., Riek, R., and Meier, B. (2008) 24. Görler, A. and Kalbitzer, H. R. (1997) Relax, a
Amyloid fibrils of the HET-s(218–289) prion flexible program for the back calculation of
form a beta solenoid with a triangular hydro- NOESY spectra based on complete relaxation
phobic core. Science 319, 1523–1526. matrix formalism. J. Magn. Reson. 124,
13. Bardiaux, B., Bernard, A., Rieping, W., Habeck, 177–188.
M., Malliavin, T. E., and Nilges, M. (2009) 25. Linge, J., Habeck, M., Rieping, W., and Nilges,
Influence of different assignment conditions on M. (2004) Correction of spin diffusion during
the determination of symmetric homodimeric iterative automated NOE assignment. J. Magn.
structures with ARIA. Proteins 75, 569–585. Reson. 167, 334–342.
14. Nilges, M., Bernard, A., Bardiaux, B., Malliavin, 26. Mumenthaler, C. and Braun, W. (1995)
T., Habeck, M., and Rieping, W. (2008) Automated assignment of simulated and exper-
Accurate NMR structures through minimisa- imental NOESY spectra of proteins by feedback
tion of an extended hybrid energy. Structure filtering and self-correcting distance geometry.
16, 1305–1312. J. Mol. Biol. 254, 465–480.
15. van Rossum, G., http://www.python.org/. 27. Stein, E. G., Rice, L. M., and Brünger, A. T.
16. Brünger, A. T., Adams, P. D., Clore, G. M., (1997) Torsion-angle molecular dynamics as a
DeLano, W. L., Gros, P., Grosse-Kunstleve, new efficient tool for NMR structure calcula-
R. W., Jiang, J.-S., Kuszewski, J., Nilges, M., tion. J. Magn. Reson. 124, 154–164.
Pannu, N. S., Read, R. J., Rice, L. M., 28. Fossi, M., Oschkinat, F., Nilges, M., and Ball,
Simonson, T., and Warren, G. L. (1998) L. (2005) Quantitative study of the effects of
Crystallography and NMR system (CNS): A chemical shift tolerances and rates of SA cool-
new software suite for macromolecular struc- ing on structure calculation from automatically
ture determination. Acta Cryst. sect. D 54, assigned NOE data. J. Magn. Reson. 175,
905–921. 92–102.
17. Vranken, W. F., Boucher, W., Stevens, T. J., 29. Rieping, W., Habeck, M., and Nilges, M. (2005)
Fogh, R. H., Pajon, A., Llinas, M., Ulrich, E. Modeling errors in NOE data with a log-normal
L., Markley, J. L., Ionides, J., and Laue, E. D. distribution improves the quality of NMR struc-
(2005) The CCPN data model for NMR spec- tures. J. Am. Chem. Soc. 127, 16026–16027.
troscopy: development of a software pipeline. 30. Rieping, W., Habeck, M., and Nilges, M.
Proteins 59, 687–696. (2005) Inferential Structure Determination.
18. Laskowski, R. A., MacArthur, M. W., Moss, D. Science 309, 303–306.
S., and Thornton, J. M. (1993) PROCHECK: 31. Habeck, M., Rieping, W., and Nilges, M.
a program to check the stereochemical quality (2006) Weighting of experimental evidence in
of protein structures. J. Appl. Cryst. 26, macromolecular structure determination. Proc.
283–291. Natl. Acad. Sci. USA 103, 1756–1761.
19. Vriend, G. (1990) WHAT IF: a molecular 32. Nilges, M. (1993) A calculation strategy for the
modeling and drug design program. J. Mol. structure determination of symmetric dimers
Graph. 8, 52–56. by 1 H NMR. Proteins 17, 297–309.
20. Sippl, M. J. (1993) Recognition of errors in 33. Linge, J. P., Williams, M. A., Spronk, C. A.,
three-dimensional structures of proteins. Bonvin, A. M., and Nilges, M. (2003)
Proteins Struct. Funct. Genet. 17, 355–362. Refinement of protein structures in explicit sol-
21. Davis, I. W., Leaver-Fay, A., Chen, V. B., Block, vent. Proteins Struct. Funct. Genet. 20,
J. N., Kapral, G. J., Wang, X., Murray, L. W., 496–506.
482 B. Bardiaux et al.
34. Linge, J. P. and Nilges, M. (1999) Influence of contour diagrams., J. Magn. Reson. 95,
non-bonded parameters on the quality of NMR 214–220.
structures: a new force-field for NMR structure 46. Kjær, M., Andersen, K. V., and Poulsen, F. M.
calculation. J. Biomol. NMR 13, 51–59. (1994) Automated and semiautomated analysis
35. Nederveen, A., Doreleijers, J., Vranken, W., of homo- and heteronuclear multidimensional
Miller, Z., Spronk, C., Nabuurs, S., Guntert, nuclear magnetic resonance spectra of proteins:
P., Livny, M., Markley, J., Nilges, M., Ulrich, the program PRONTO. Methods Enzymol.
E., Kaptein, R., and Bonvin, A. M. (2005) 239, 288–308.
RECOORD: a REcalculated COORdinates 47. Bartels, C., Xia, T.-H., Billeter, M., Güntert,
Database of 500+ proteins from the PDB using P., and Wüthrich, K. (1995) The program
restraints from the BioMagResBank. Proteins XEASY for computer-supported NMR spectral
59, 662–672. analysis of biological macromolecules. J. Biomol.
36. The World Wide Web Consortium (2008), NMR 5, 1–10.
Extensible Markup Language (XML) 1.0 (Fifth 48. Güntert, P., Braun, W., and Wüthrich, K.
Edition), http://www.w3.org/TR/xml/. (1991) Efficient computation of three-dimen-
37. Bardiaux, B., Bernard, A., Rieping, W., Habeck, sional protein structures in solution from
M., Malliavin, T., and Nilges, M. (2008) nuclear magnetic resonance data using the pro-
Graphical analysis of NMR structural quality gram DIANA and the supporting programs
and interactive contact map of NOE assign- CALIBA, HABAS and GLOMSA. J. Mol. Biol.
ments in ARIA. BMC Struct. Biol. 8, 30–34. 217, 517–530.
38. Spronk, C. A. E. M., Nabuurs, S. B., Krieger, 49. Hall, S. R. and Cook, A. P. F. (1995) STAR
E., Vriend, G., and Vuister, G.W. (2004) dictionary definition language: Initial specifica-
Validation of protein structures derived by tion. J. Chem. Inf. Comput. Sci. 35, 819–825.
NMR spectroscopy. Progress in Nuclear 50. Markley, J. L., Bax, A., Arata, Y., Hilbers, C. W.,
Magnetic Resonance Spectroscopy 45, 315–337. Kaptein, R., Sykes, B. D., Wright, P. E., and
39. Saccenti, E. and Rosato, A. (2008) The war of Wüthrich, K. (1998) Recommendations for the
tools: how can NMR spectroscopists detect presentation of NMR structures of proteins
errors in their structures? J. Biomol. NMR 40, and nucleic acids. J. Mol. Biol. 280, 933–952.
251–261. 51. Güntert, P., Mumenthaler, C., and Wütrich, K.
40. Nabuurs, S., Krieger, E., Spronk, C., Nederveen, (1997) Torsion Angle Dynamics for NMR
A., Vriend, G., and Vuister, G. (2005) Strucutre Calculation with the New Program
Definition of a new information-based per-res- DYANA. J. Mol. Biol. 273, 283–298.
idue quality parameter. J. Biomol. NMR 33, 52. Wüthrich, K., Billeter, M., and Braun, W.
123–134. (1983) Pseudo-structures for the 20 common
41. Nabuurs, S., Spronk, C., Vuister, G., and Vriend, G. amino acids for use in studies of protein con-
(2006) Traditional biomolecular structure deter- formations by measurements of intramolecular
mination by NMR spectroscopy allows for major proton-proton distance constraints with nuclear
errors. PLoS Comput. Biol. 2, e9. magnetic resonance. J Mol Biol 169, 949–961.
42. Kraulis, P., Domaille, P. J., Campbell-Burk, S. 53. Lange, A., Luca, S., and Baldus, M. (2002)
L., van Aken, T., and Laue, E. D. (1994) Structural constraints from proton-mediated
Solution structure and dynamics of ras p21. rare-spin correlation spectroscopy in rotating
GDP determined by heteronuclear three- and solids. J. Am. Chem. Soc. 124, 9704–9705.
four-dimensional NMR spectroscopy. 54. Szeverenyi, N., Sullivan, M., and Maciel, G.
Biochemistry 33, 3515–3531. (1982) Observation of spin exchange by two-
43. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, dimensional fourier transform 13 C cross polar-
G., Pfeifer, J., and Bax, A. (1995) NMRPipe: a ization-magic-angle spinning. J. Magn. Reson.
multidimensional spectral processing system 47, 462–475.
based on UNIX pipes. J. Biomol. NMR 6, 55. Castellani, F., van Rossum, B., Diehl, A.,
277–293. Rehbein, K., and Oschkinat, H. (2003)
44. Johnson, B. A. and Blevins, R. A. (1994) Determination of solid-state NMR structures
NMRView: A computer program for the visu- of proteins by means of three-dimensional
alization and analysis of NMR data. J. Biomol. 15 N-13 C-13 C dipolar correlation spectros-
NMR 4, 603–614. copy and chemical shift analysis. Biochemistry
45. Garrett, D., Powers, R., Gronenborn, A., and 42, 11476–11483.
Clore, G. (1991) A common sense approach to 56. Takegoshi, K., Nakamura, S., and Terao, T.
peak picking two-, three- and four-dimensional (2003) 13 C-1 H dipolar-driven 13 C-13 C recou-
spectra using automatic computer analysis of pling without 13 C rf irradiation in nuclear
23 ARIA for Solution and Solid-State NMR 483
magnetic resonance of rotating solids. J. Chem. 67. Zweckstetter, M. and Bax, A. (2000) Prediction
Phys. 118, 2325–2341. of sterically induced alignment in a dilute liquid
57. Lewandowski, J. R., Paëpe, G. D., Eddy, M. T., crystalline phase: Aid to protein structure
and Griffin, R. G. (2009) (15)N-(15)N proton determination by NMR. J. Am. Chem. Soc.
assisted recoupling in magic angle spinning 122, 3791–3792.
NMR. J. Am. Chem. Soc. 131, 5769–5776. 68. Meiler, J., Blomberg, N., Nilges, M., and
58. Fossi, M., Linge, J., Labudde, D., Leitner, D., Griesinger, C. (2000) A new approach for
Nilges, M., and Oschkinat, H. (2005) Influence applying residual dipolar couplings as restraints
of chemical shift tolerances on NMR structure in structure calculations. J. Biomol. NMR 16,
calculations using ARIA protocols for assigning 245–252.
NOE data. J. Biomol. NMR 31, 21–34. 69. Jones, J. P. (2002) PBS: portable batch system,
59. Wishart, D. S. and Sykes, B. D. (1994) The 13 C Beowulf cluster computing with Linux, MIT
chemical-shift index: a simple method for the Press, Cambridge, MA, USA, 369–390.
identification of protein secondary structure 70. Gentzsch, W. (2001) Sun Grid Engine: Towards
using 13 C chemical-shift data. J. Biomol. NMR creating a compute power grid, CCGRID ’01:
4, 171–180. Proceedings of the 1st International Symposium
60. Cornilescu, G., Delaglio, F., and Bax, A. (1999) on Cluster Computing and the Grid,
Protein backbone angle restraints from search- IEEE Computer Society, Washington, DC,
ing a database for chemical shift and sequence USA, 35.
homology. J. Biomol. NMR 13, 289–302. 71. Thain, D., Tannenbaum, T., and Livny, M.
61. Cheung, M.-S., Maguire, M. L., Stevens, T. J., (2005) Distributed computing in practice: the
and Broadhurst, R. W. (2010) DANGLE: A Condor experience. Concurr. Comput.: Pract.
Bayesian inferential method for predicting pro- Exper. 17, 323–356.
tein backbone dihedral angles and secondary 72. Nabuurs, S., Spronk, C., Krieger, E.,
structure. J. Magn. Reson. 202, 223–33. Maassen, H., Vriend, G., and Vuister, G.
62. Loquet, A., Gardiennet, C., and Böckmann, A. (2003) Quantitative evaluation of experimental
(2010) Protein 3D structure determination by NMR restraints. J. Am. Chem. Soc. 125,
high-resolution solid-state NMR. Comptes. 12026–12034.
Rendus - Chimie 13, 423–430. 73. Doreleijers, J. F., Raves, M. L., Rullmann, T.,
63. Gardiennet, C., Loquet, A., Etzkorn, M., and Kaptein, R. (1999) Completeness of NOEs
Heise, H., Baldus, M., and Böckmann, A. in protein structure: a statistical analysis of
(2008) Structural constraints for the Crh pro- NMR data. J. Biomol. NMR 14, 123–132.
tein from solid-state NMR experiments. J. 74. Bhattacharya, A., Tejero, R., and Montelione,
Biomol. NMR. 40, 239–250. G. T. (2007) Evaluating protein structures
64. LeMaster, D. M. and Kushlan, D. M. (1996) determined by structural genomics consortia.
Dynamical mapping of E. coli thioredoxin via Proteins 66, 778–795.
13 C NMR relaxation analysis. J. Am. Chem. 75. Doreleijers, J. F., Vranken, W. F., Schulte, C.,
Soc. 118, 9255–9264. Lin, J., Wedell, J. R., Penkett, C. J., Vuister, G.
65. Tjandra, N., Garrett, D. S., Gronenborn, A. W., Vriend, G., Markley, J. L., and Ulrich, E. L.
M., Bax, A., and Clore, G. M. (1997) Defining (2009) The NMR restraints grid at BMRB for
long range order in NMR structure determina- 5,266 protein and nucleic acid PDB entries.
tion from the dependence of heteronuclear J. Biomol. NMR 45, 389–396.
relaxation times on rotational diffusion anisot- 76. Jehle, S., Rajagopal, P., Bardiaux, B., Markovic,
ropy. Nature Struct. Biol. 4, 443–449. S., Kühne, R., Stout, J. R., Higman, V. A.,
66. Clore, G., Gronenborn, A., and Bax, A. (1998) A Klevit, R. E., van Rossum, B.-J., and Oschkinat,
robust method for determining the magnitude of H. (2010) Solid-state NMR and SAXS studies
the fully asymmetric alignment tensor of oriented provide a structural basis for the activation of
macromolecules in the absence of structural alphaB-crystallin oligomers. Nat. Struct. Mol.
information. J. Magn. Reson. 133, 216–221. Biol. 17, 1037–1042.
Chapter 24
Abstract
Motions are essential for protein function, and knowledge of protein dynamics is a key to our understanding
the mechanisms underlying protein folding and stability, ligand recognition, allostery, and catalysis. In the
last two decades, NMR relaxation measurements have become a powerful tool for characterizing backbone
and side chain dynamics in complex biological macromolecules such as proteins and nucleic acids. Accurate
analysis of the experimental data in terms of motional parameters is an essential prerequisite for developing
physical models of motions to paint an adequate picture of protein dynamics. Here, I describe in detail how
to use the software package DYNAMICS that was developed for accurate characterization of the overall
tumbling and local dynamics in a protein from nuclear spin-relaxation rates measured by NMR. Step-by-
step instructions are provided and illustrated through an analysis of 15N relaxation data for protein G.
Key words: Relaxation, Protein dynamics, Order parameter, Spectral density, Dipolar coupling,
Chemical shift anisotropy, CSA, Overall tumbling, Rotational diffusion tensor, Monomer–dimer
equilibrium
1. Introduction
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3_24, © Springer Science+Business Media, LLC 2012
485
486 D. Fushman
1.1. The Underlying The experimental spin relaxation parameters (longitudinal and
Equations transverse relaxation rates, R1 and R2, and the steady-state hetero-
nuclear NOE) are directly related to power spectral densities, J(w),
which are Fourier transforms of the corresponding correlation
functions describing reorientations of the internuclear vector of
interest. In the case of the backbone amide 15N nucleus, the major
24 Determining Protein Dynamics from 15N Relaxation Data… 487
2. Description
of DYNAMICS
Software
DYNAMICS is a computer program for model-free analysis of spin
2.1. Highlights relaxation data. The current version of the program (version 3.0)
of the Program includes the following features:
DYNAMICS ● Overall tumbling. All possible models of the overall rotational
diffusion are allowed: isotropic, axially symmetric, and fully
anisotropic. The overall rotational diffusion tensor can be an
input variable, but can also be determined simultaneously with
the model-free analysis of the relaxation data.
● Multiple-field data. Simultaneous or separate analysis of experi-
mental data from measurements at multiple magnetic fields.
Data sets at various fields do not have to be complete.
● Chemical shift anisotropy. The CSA can be treated as uniform
(fixed) or site-specific, the program also allows determining
site-specific CSA values simultaneously with data analysis.
490 D. Fushman
2.2. The Overall The overall organization of DYNAMICS is depicted in the flow-
Organization chart in Fig. 1. The program is written in Matlab (The MathWorks,
of the Program Inc); the current version of the program is compatible with Matlab
versions 6.5 and newer. It is assumed here that the user is familiar
with very basic Matlab commands that allow loading and saving
data and navigation to the desired folder/directory.
4
j(ω,τc), τc=10 ns
j(ω,τc), τc=5 ns
3 j(ω,τe), τc=10 ns, τloc=100 ps
j(ω,τ), a.u.
0.2
2
0.1
1 0.0
100 900
0
1 10 100 900
ω/2π, MHz
Fig. 2. Relative contributions to the power spectral density J (w) from the overall tumbling
and local motion. Shown as a function of frequency w are j ( w,tc) (see Eq. 7) for tc = 5 ns
(green) and 10 ns (blue), and j (w,te) for tloc = 100 ps and tc =10 ns (red ). The factors S 2
and (1−S 2) are not included.
2.3. Treating As shown in Fig. 2, the contribution to the spectral density func-
the Overall Rotational tion from the overall tumbling is quite substantial (if not domi-
Diffusion nant) and often overshadows that from local motions. As our main
of a Molecule goal here is to characterize internal motions, accurate treatment of
the overall tumbling is absolutely critical for accurate analysis of the
local dynamics in a protein (20). Thus, the first and foremost step
in relaxation data analysis is to determine, and “subtract,” the con-
tribution from the overall tumbling. Significant attention in the
past was paid to developing tools for accurate analysis of the overall
rotational diffusion (16, 21–28).
In principle, the overall rotational diffusion can also be charac-
terized simultaneously with the analysis of local dynamics, and in
fact, DYNAMICS includes a mechanism for doing this (see
Subheadings 3.2 and 3.3). However, beyond the simplest case of
isotropic tumbling, this determination becomes less straightfor-
ward and can require significant effort, as multiple parameters need
to be optimized manually,. Therefore, if protein atom coordinates
are available, the most straightforward and reliable way to charac-
terize the overall rotational diffusion is directly from relaxation
data and separately from (and prior to) the analysis of local motions.
The underlying reason for this is based on the fact that, for well-
defined structural regions in a protein, the “reduced” relaxation
rates R1¢ and R2¢ are both proportional (to a good approximation)
to the squared order parameter. (The “reduction” is achieved by
subtracting from Eqs. 1–2 the contributions from the high-fre-
quency components, J(wH) and J(wH ± wN), of the spectral density
function, e.g., see ref. 27). Thus the R2¢/R1¢ ratio is S2-independent,
and the determination of the overall motion can be de-convoluted
492 D. Fushman
2.4. Selection At the heart of the DYNAMICS program is the model selection
of the Appropriate algorithm, which, based on how a particular model of local motion
Model for Local fits experimental data, selects the most appropriate model. It is
Motion similar to the approach described in (31) and is based on the
Occam’s razor principle, in that the simplest model that fits the
data is considered sufficient. All models of local motion used in
DYNAMICS are listed in Table 1. The model selection process
starts with the simplest model, LS_00, and first determines if it is
acceptable, i.e., the following two criteria are satisfied: (1) the
model yields physically reasonable values of the microdynamic
parameters (in this case, 0 £ S2 £1, but more generally for all mod-
els: 0 £ S2, Sfast2 £ 1; tloc, tfast >5 ps; 100 ps < tslow < tc, and Rex ³ sR2)
and (2) it provides a reasonable fit to the experimental data, i.e.,
passes the goodness-of-fit test (32). If this model is acceptable, the
program proceeds to the next-level-complexity models (in this
case, LS_tl and LS_ex) and applies the same acceptance rules as
above. If any of these models are acceptable and yield lower residu-
als of fit (c2) than the lower-complexity model (in this case, LS_00),
the program uses the F-statistics test to determine if this improve-
ment in the fit is genuine and reflects a better-fit model or merely
24 Determining Protein Dynamics from 15N Relaxation Data… 493
Table 1
Microdynamic parameters for the various models of local motion used
in DYNAMICS
Modela S 2 or Sslow2b tloc or tslowb Sfast2 tfast Rex #expc Npard Indexe
LS_00 Vf 0 1 N/A 0 1 1 0
LS_tl V V 1 N/A 0 1 2 1
LS_ex V 0 1 N/A V 1 2 2
LS_tx V V 1 N/A V 1 3 3
CL_00 V 0 V V 0 2 3 4
CL_tl V V V V 0 2 4 5
CL_ex V 0 V V V 2 4 6
CL_tx V V V V V 2 5 7
Matlab name S2 g TAUloc S2f TAUf Rex
or TAUsl b
a
The name of the corresponding model of local motion as used in DYNAMICS
b
Naming convention used in DYNAMICS: the corresponding motional parameters in the monoexponen-
tial model are S2 and tloc, whereas in the double-exponential (“extended”) model these parameters are
called Sslow2 and tslow
c
The number of exponentials in the corresponding correlation function of local motion see Eqs. 4 and 5
d
The total number of fitting parameters in a given model
e
A numerical index of the model in DYNAMICS, plotted in the output graphs (see Figs. 3 and 4)
f
“V ” indicates that the corresponding parameter is present in a given model and is fitted (not fixed);
N/A = not applied
g
The name of the corresponding Matlab variable in DYNAMICS output. Note that in the case of
“extended” models the reported S2 value is in fact S2 = Sfast2 × Sslow2
2.5. Running All the scripts and functions of DYNAMICS package come in a
DYNAMICS single compressed file. When you uncompress it (using one of the
standard programs), it will by default put all the content of the
2.5.1. Getting Started
package in a folder called dynamics.
I recommend that you run all the analysis from the directory
containing your relaxation data, which is separate from the dynamics
directory: this will prevent you from “littering” the latter with out-
put files that DYNAMICS creates automatically (see Subheading 3.5).
For this, you will need to add the dynamics directory to your Matlab
path, for example by using the following command:
> > path(path,’c:/MyMatlab/dynamics’)
(here I assumed that all DYNAMICS scripts are located in the
folder c:/MyMatlab/dynamics on your computer).
2.5.2. Before You Run Navigate to your data directory and load all required input param-
the Program eters (Table 2) into the Matlab workspace (use Matlab function
load for this). Make sure that all of the parameters are in the
proper format and units as specified in Table 2. The auxiliary pro-
gram pdb2nh (see Subheading 3.6.1) will help you retrieve
NH-vector coordinates from the protein coordinate file.
2.5.3. Run-Time Dialog To start the program type the following command in the Matlab
Command window:
> > dynamics
If you added the dynamics directory to the Matlab path, you
can type this command directly from your data directory (recom-
mended). At the start, the program performs preliminary analysis
of the input data and outputs on the screen various estimates of the
overall rotational correlation time and the statistics of the distribu-
tion of the R2/R1 ratios. Here is an example of such output for 15N
relaxation data at 14.1 Tesla (1H frequency = 600.13 MHz) for
GB3 (19). We use these data throughout this chapter. If relaxation
data at more than one field are included, the analysis and the out-
put will be done for each field separately.
Table 2
Input parameters for DYNAMICS
- - - - - - - 600.13 MHz - - - - - - - - -
TOTAL: MEAN = 2.2131 SD(MEAN) = 0.12582 TAU = 3.3506
L&S: MEAN = 2.1806 SD(MEAN) = 0.062933 TAU(MEAN)
= 3.2995
MEAN(TAU) = 3.2981 SD_TAU = 0.098668
TAUmc = 3.3001 SD_TAUmc = 0.11508
resid. with the R2/R1 within MEAN +/− SD: 37
resid. with the R2/R1 above MEAN + SD: 9
resid. with the R2/R1 below MEAN - SD: 5
The purpose of this analysis is to estimate the overall rotational
1 6R2
correlation time (as t c = − 7 , (33)) and to count, in
2wN R1
the spirit of (34), how many residues have the R2/R1 ratio within
one standard deviation (SD) from the mean R2/R1 value. These
residues are expected to fit into the “standard” Lipari & Szabo
model (10, 11). Residues with the R2/R1 ratio more than one
standard deviation below the mean value could require the Rex
term, see Eq. 2, while those residues that have R2/R1 more than
one standard deviation above the mean value might need the
“extended” model-free model (12).
At the start the program plots the experimental data (R2, R1,
NOE, and residue-specific CSAs, if applicable) as a function of resi-
due number (see Figs. 3 and 4). This output can be suppressed by
setting kplot to 0 (or any number other than 1).
The run-time dialog that follows is shown step-by-step below.
Note that many questions that appear on the screen have a default
answer (indicated in the square brackets): this answer will be assumed
if you press ENTER, and if the question was about a parameter
involved in computations, the program will output a message con-
firming that the corresponding value was assumed.
Input a CSA value [−160] ==>
This line appears if kcsa is set to 0 (default), i.e., a fixed uni-
form CSA value will be used. Input the desired value (only numeric
input) or simply press ENTER: in this case CSA = −160 ppm will be
assumed. Note that if kcsa was set to −1, a list of fixed (site-spe-
cific) CSA values must exist in the workspace; otherwise, the pro-
gram will output an error message and exit.
If you did not define TAUc value(s), the program will ask you
the following:
Input TAUc value(s) (in ns) ==>
Here you can input a single value (e.g., 3.3) or a list of values,
e.g., [3.28 3.3 3.32].
If you selected the isotropic rotational diffusion model (i.e.,
kovrl was set to 0 on undefined), the program will proceed to actual
model-free analysis and model selection on a residue by residue
498 D. Fushman
Fig. 3. Output of DYNAMICS analysis of backbone motions in GB3 from 15 N relaxation data at 600 MHz. (a) Input data; (b)
the results of analysis assuming isotropic overall tumbling with TAUc = 3.33 ns; and (c) the results of analysis assuming
anisotropic (axially symmetric) overall tumbling with TAUc = 3.33 ns and other diffusion tensor characteristics presented in
Subheading 2.3. A uniform 15N CSA value of −174.2 ppm was assumed throughout the protein. The circles on the “model”
plot in b indicate residues that fall into the NOMOD category.
Fig. 4. Output of DYNAMICS analysis of backbone motions in GB3 from 15N relaxation data at five magnetic fields. (a) Input
data. (b–c) the results of analysis assuming (axially symmetric) anisotropic overall tumbling (b) with a uniform (fixed)
CSA = −174.2 ppm (as in Fig. 3) and (c) site-specific CSAs obtained simultaneously with the microdynamic parameters
from fitting these relaxation data. The circles on the “model” plots indicate residues that fall into the NOMOD category.
values (if there is more than one value for each of these parame-
ters), or stops and waits for user’s input. If the isotropic tumbling
model was selected (kovrl = 0 or 2), the message on the screen will
read as follows:
Input TAUc (TAUc < = [0] - break)==>
Entering a positive number will trigger another round of cal-
culations with this TAUc value, whereas zero or a negative number
will be interpreted as the signal to proceed to exit or error analysis.
Note that the latest positive TAUc value will be taken as the final/
accepted value and used for error analysis. If the TAUc value that
you want to accept is not the latest one, you need to reenter the
desired value, let the program run through all residues again (this
is quite fast anyway), and only after that enter 0 or a negative TAUc
to exit or proceed to error analysis.
In the case of anisotropic tumbling (kovrl = −1 or 1), the mes-
sage on the screen reads as follows:
Satisfied? (1-yes(calc.err), [0]-cont.(beta-
range), 2-man.input, -1-stop/exit)==>
Enter 1 here to proceed to error analysis, 0 to continue
computations with other β values (if more than one b value was
entered above), 2 if you want to return to manual input of the dif-
fusion tensor parameters (see above), and −1 to exit the program.
If you choose to exit the program, it will automatically remove
unnecessary (run-time) variables from the workspace and finish.
If you choose to proceed to error analysis, the program will ask
you to select the method of error estimation:
Choose MC simulation of exper.data(0) or fitted
params(1 or 2(vis = on)) ==>
Selecting option 0 will generate synthetic experimental data
(assuming normally distributed noise with the standard deviation
sR1, sR2, or sNOE), and for each set of generated data will per-
form the fit using the same model of local motion as selected for
the real data. By default, 500 runs will be performed for each resi-
due, and the standard deviation will be displayed and included in
the ERR array and in the final report RESERR. If you select option
1, the program will determine experimental errors using the con-
stant c2-boundaries method (32), which assumes that the residuals
of fit are distributed according to a c2 distribution, and therefore a
deviation of the fitted parameters from the optimal value by one
standard deviation would result in a specific increase in c2 that
depends on the number of fitting parameters (e.g., Dc2 =1, 2.3 or
3.53 for Npar = 1, 2 or 3, respectively). Thus the program deter-
mines the confidence boundaries for the fitted parameters by gen-
erating their values randomly and keeping only those values that
502 D. Fushman
2.5.4. Understanding DYNAMICS outputs the results of the analysis on the screen, both
the Output in numerical format and as plots, and stores them in several output
parameters/arrays, summarized in Table 3.
As discussed above, during each run the program outputs the
results of analysis for each residue. In addition to the obvious
issues, such as the model and the actual values of the microdynamic
parameters, the user should also pay attention to the residuals of fit
(chi2 = c2), which are represented by the last number in each row
(or the one before last if CSA is also a fitting parameter). The chi2
information is important, because it tells you how well the data fit
the model. Ideally, a good fit would give chi2 values of about 1 per
degree of freedom. Thus, chi2 numbers in the range of single dig-
its (» df) or lower indicate a reasonable fit, whereas much higher
chi2 values indicate a potential problem with data analysis for a
particular residue: either the “best”-fit model is not ideal for that
residue or perhaps the experimental errors are underestimated
(hence elevated chi2).
24 Determining Protein Dynamics from 15N Relaxation Data… 503
Table 3
Output parameters created by program DYNAMICS
NOMOD List of residues in the NOMOD category (see Subheading 2.4), i.e., residues for which
none of the models of local motion passed the goodness-of-fit test (c 2 too high),
although at least one model provides a physically meaningful set of microdynamic
parameters
EXCL List of residues that have been excluded by the program because none of the tested
models of local motion are able to provide a physically meaningful set of micrody-
namic parameters (see Subheading 2.4)
RES The results of fit in the following format (array Nres x 10): [Residue# tc S2 tloc Rex Sfast2
tfast model-Index c 2 CSA]
ERR The results of error analysis in the following format (array Nres x 7): [Residue# dtc dS 2
d tloc dRex dSfast2 dtfast]
RESERR Combined results of fit (RES) and error analysis (ERR) in the following format (array
Nres x 16): [Residue# tc dtc S2 d S2 tloc d tloc Rex d Rex Sfast2 dSfast2 tfast dtfast model-Index
c 2 CSA]
TAUCHI Record of all evaluations performed during the current DYNAMICS run in the isotropic
tumbling mode or monomer–dimer equilibrium (empty if anisotropic tumbling).
Each line is a summary statistics for all residues, in the following format: tc, c 2(mod),
Nnomod, c 2(nomod), runs, df, c 2(total), c 2(total)/df, Nexcl
ANISO Record of all evaluations performed during the current DYNAMICS run in the aniso-
tropic mode (empty matrix in the isotropic or monomer-dimer equilibrium). Each line
is a summary of statistics for all residues, in the following format (for axially symmetric
model): tc, Dz/Dx,b,a, c 2(mod),Nnomod,c 2(nomod), runs, df, c 2(total), c 2(total)/df,
NexclIn the case of fully anisotropic tumbling model, the format is
tc, Dz/Dx, Dy/Dx b,a,g, c 2(mod),Nnomod,c 2(nomod), runs, df, c 2(total), c2(total)/df, Nexcl
runs This parameter counts how many times the selected model switches between model-free
and “extended” model-free models in adjacent residues along the protein sequence
df Total number of degrees of freedom, df = Ndat – Npar
chi2 Residuals of fit, c 2
⎡⎛ R exp − R calc ⎞ 2 ⎛ R exp − R calc ⎞ 2 ⎛ NOEexp − NOEcalc ⎞ 2 ⎤
χ2 = ∑ ⎢⎜ 1
freq ⎢ ⎝ sR
1
⎟⎠ + ⎜⎝
2
sR
2
⎟⎠ + ⎜⎝ sNOE
⎟⎠ ⎥
⎥⎦
⎣ 1 2
3. Miscellaneous
Issues
3.1. General Notes 1. Make sure the input list of frequencies freq contains all the
on Using DYNAMICS pertinent 1H frequencies (magnetic fields) for the data that you
want to analyze. It is critical that the order in which the fre-
quencies are listed in the freq-list is coordinated with the sec-
ond “index” in the relaxation parameters names. For example,
if freq = [500, 600], then r11, r21, and r31 should be R1, R2,
and NOE data at 500 MHz, respectively, while r12, r22, and
r32 should be R1, R2, and NOE data at 600 MHz.
2. Small Rex values (<1 s−1) could be artifacts of model selection
(for example, due to an inadequate overall tumbling model, as
illustrated in Fig. 3) or a result of elevated 15N CSA, rather
than real exchange contributions. Relaxation data at more than
one magnetic field are required (analyzed separately or
together) to verify the consistency of the model selection as
well as that the resulting Rex term has the expected field depen-
dence (∝ Bo2).
3. To fit CSA values, relaxation data at more than one magnetic
field are required.
4. When conformational exchange (Rex) is present, the CSA val-
ues obtained from the analysis could be biased, since both the
Rex and the CSA contributions to relaxation rates have a similar
field dependence (∝ Bo2).
5. Errors in the derived CSA values, as well as the associated
errors in microdynamic parameters from such analysis
(kcsa = 1), can currently be calculated only using the “experi-
mental data simulation” option.
6. In the output RES matrix, the CSA values are included as col-
umn #10, following the c2 values. In the output RESERR
matrix the CSA values and errors in CSA are included as the
last two columns (#16, 17).
7. Even if you input a single TAUc value for the isotropic or
monomer–dimer equilibrium model, the program will run cal-
culations for this TAUc twice. This is done to compare the
results of at least two runs and, because DYNAMICS is fast,
does not take too much time.
3.2. Strategy What can one do if protein coordinates are not available? One can
for Determining always use the isotropic overall tumbling model in DYNAMICS,
the Overall Rotational which does not require knowledge of protein structure. However,
Correlation Time When one should be aware of the fact that the microdynamic parameters
the Structure derived from such analysis could be biased by the possible oversim-
is Unknown or RotDif plification of the tumbling model used, particularly if the protein
Results are Unreliable shape turns out to be far from spherical.
506 D. Fushman
b 20 Rex-models
CL-model
10
a
Number of
residues
0
10
Number of
Rex-models 4
residues
5 CL-model NOMOD
2
0 0
4.5 45 4.5 40
4.0 4.0
3.5 40 3.5
35
3.0 3.0
χ2/df
χ2/df
df
df
2.5 35 2.5
2.0 2.0
30
1.5 30 1.5
1.0 1.0
3.25 3.30 3.35 3.40 3.45 3.15 3.20 3.25 3.30 3.35 3.40 3.45
TAUc, ns TAUc, ns
Fig. 5. Illustration of the use of DYNAMICS to optimize tc simultaneously with model-free analysis of 15N relaxation data for
GB3. (a) The analysis assumed an anisotropic tumbling model with the diffusion tensor’s anisotropy and orientation as
specified in Subheading 2.3. (b) The analysis assumed isotropic tumbling model. The 15N CSA was set to −174.2 ppm
(fixed) in both cases. Only secondary structure residues were included; in addition, residues D36 and Y45 were excluded
from this analysis because of high c 2 values. Solid line in both panels shows c 2/df, while the dashed line depicts the total
number of degrees of freedom, df. The vertical bars in the top plots show the number of residues with Rex (blue) (LS_ex
and LS_tx models) or with extended model (CL_00) (red) selected for each value of TAUc, as well as the number of
NOMOD residues (black). There were no NOMOD residues in the analysis in a. The RotDif analysis gives the tc value of
3.34 ± 0.14 ns (19) (or 3.37 ± 0.20 ns for these selected residues); a very similar value (3.36 ns) calculated from the mean
R2/R1 value was reported at the beginning of the DYNAMICS run.
TAUc value (or range) in the “middle” region, where the total
chi2 is at or close to its minimum and at the same time the
total df is at or close to its maximum. There might not be a
clear single TAUc value, but one should remember there is
always an uncertainty in tc even from RotDif analysis. Examples
of such analyses are shown in Fig. 5. In addition, to help deal
with an artificial selection of the CL_00 model, DYNAMICS
also uses another parameter, called runs (Table 3). The ratio-
nale behind it is to avoid an artificial situation when a single
residue shows large-amplitude motions (this is what the
extended model is usually needed for) while its neighbors do
not. To avoid this, runs counts how many times the selected
model switches between “LS” and “CL” models in adjacent
residues. Naturally, one wants to minimize the number of such
switches, because intuitively one would expect that large-
amplitude fluctuations involve several adjacent residues in the
polypeptide chain, and not just a single one.
508 D. Fushman
One should also bear in mind that the total chi2/df (as well
as chi2) is generally not a smooth function because changes in
the selected model of local motion for individual residues
would result in abrupt changes in the residuals of fit and df.
5. After the TAUc value has been optimized, use this value to run
DYNAMICS again, this time for all residues in the protein.
You might need to do several iterations of such an analysis.
3.3. Strategy The strategy here is similar to that described in the previous sec-
for Determining tion, except that (1) protein atom coordinates are required and (2)
the Overall Rotational you will need to vary several parameters as the same time. For
Diffusion Tensor example, for the axially symmetric model, you might want to vary
when RotDif Results Dz/Dx and the angles a and b. While doable, this optimization is
are Unreliable not straightforward, can require significant effort, and can result in
a local rather than a global minimizer. Thus, I would recommend
using it only when no other option is available.
3.4. Cleaning DYNAMICS generates and keeps most of the necessary intermedi-
the Workspace ate variables in the current Matlab workspace (computer memory).
They are automatically removed from the workspace when the
program terminates successfully. If the program run was termi-
nated prematurely either by the user (e.g., via CTRL/C) or in case
of a run-time error, these variables will remain in the workspace.
This might cause an interruption in the normal program execution
when you start it next time during the same Matlab session. To
ensure uninterrupted program flow, it is recommended to remove
the remaining intermediate variables from the computer memory
before restarting DYNAMICS. Removal of only intermediate vari-
ables can be achieved by issuing the following command:
> > dynclean
3.5. Automatic Saving To prevent accidental loss of the computed data, the results (RES,
of the Results NOMOD, EXCL, TAUCHI, ANISO) are automatically saved to a
Matlab file after completion (and acceptance) of the model-free
analysis and again after error analysis (the same parameters as above
plus ERR and RESERR). To reduce the chance of overwriting this
file when you run DYNAMICS again, the name of the file contains
the current date followed by a random number from 0 to 99, e.g.,
dyn16jan2011_92.mat.
3.6.2. Reldata, Reldatae Given all pertinent parameters of the overall and local dynamics, as
well as the orientation of the NH vector (if necessary), the reldata
program computes 15N relaxation rates: R1, R2, and NOE. The input
options also include the ability to add random noise to the data.
The program reldatae performs the same task as reldata, but in
addition also computes the longitudinal (hz) and transverse (hxy)
cross-correlation rates between the 1H–15N dipolar interaction and
15
N CSA, e.g., see ref. 19, 36.
3.6.3. conv2temp This program allows conversion between tc values at different tem-
peratures, by taking into account the temperature dependence of
water viscosity, see e.g., ref. 37.
3.6.4. Demo Scripts The package includes several demo scripts, designed to help the
user learn how to run DYNAMICS:
Acknowledgments
References
1. Palmer, A. G., 3 rd. (2004) NMR characteriza- determination of residue-specific 15 N chemical
tion of the dynamics of biomacromolecules. shift tensors in proteins in solution: protein
Chem. Rev. 104, 3623–3640. dynamics, structure, and applications of trans-
2. Sheppard, D., Sprangers, R., and Tugarinov, V. verse relaxation optimized spectroscopy, in
(2010) Experimental approaches for NMR Methods in Enzymology (James, T., Schmitz, U.,
studies of side-chain dynamics in high-molecu- and Doetsch, V., Eds.), 339, 109–126.
lar-weight proteins. Prog. Nucl. Magn. Reson. 10. Lipari, G., and Szabo, A. (1982) Model-free
Spectrosc. 56, 1–45. approach to the interpretation of nuclear mag-
3. Godoy-Ruiz, R., Guo, C., and Tugarinov, V. netic resonance relaxation in macromolecules.
(2010) Alanine methyl groups as NMR probes 2. J. Am. Chem. Soc. 104, 4559–4570.
of molecular structure and dynamics in high- 11. Lipari, G., and Szabo, A. (1982) Model-free
molecular-weight proteins. J. Am. Chem. Soc. approach to the interpretation of nuclear mag-
132, 18340–18350. netic resonance relaxation in macromolecules.
4. Cavanagh, J., Fairbrother, W. J., III, A. J. P., 1. Theory and range of validity, J. Am. Chem.
and Skelton, N. J. (1996) Protein NMR Soc. 104, 4546–4559.
Spectroscopy, Academic Press, San Diego. 12. Clore, G. M., Szabo, A., Bax, A., Kay, L. E.,
5. Fushman, D., and Cowburn, D. (1999) The Driscoll, P. C., and Gronenborn, A. M. (1990)
effect of noncollinearity of 15 N-1 H dipolar and Deviations from the simple two-parameter
15
N CSA tensors and rotational anisotropy on model-free approach to the interpretation of
15
N relaxation rates, CSA/DD cross correla- nitrogen-15 nuclear magnetic relaxation of
tion, and TROSY. J. Biomol. NMR 13, proteins. J. Am. Chem. Soc 112, 4989–4936.
139–147. 13. Woessner, D. (1962) Nuclear spin relaxaion in
6. Fushman, D., Ohlenschlager, O., and Rüterjans, ellipsoids undergoing rotational brownian
H. (1994) Determination of the backbone motion. J.Chem.Phys. 37, 647–654.
mobility of ribonuclease T1 and its 2’GMP 14. Favro, D. L. (1960) Theory of the Rotational
complex using molecular dynamics simulations Brownian Motion of a Free Rigid Body. Phys.
and NMR relaxation data. J. Biomol. Struct. Rev. 119, 53–62.
Dyn. 11, 1377–1402. 15. Ryabov, Y. E., and Fushman, D. (2007) A Model
7. Pfeiffer, S., Fushman, D., and Cowburn, D. of Interdomain Mobility in a Multidomain
(2001) Simulated and NMR derived backbone Protein. J. Am. Chem. Soc. 129, 3315–3327.
dynamics of a protein with significant flexibil- 16. Tjandra, N., Feller, S. E., Pastor, R. W., and
ity: A comparison of spectral densities for Bax, A. (1995) Rotational diffusion anisotropy
the < beta > ARK PH domain. J. Am. Chem. Soc. of human ubiquitin from 15 N NMR relaxation.
123, 3021–3036. J. Am. Chem. Soc. 117, 12562–12566.
8. Maragakis, P., Lindorff-Larsen, K., Eastwood, 17. Palmer, A. G., 3 rd, Grey, M. J., and Wang, C.
M. P., Dror, R. O., Klepeis, J. L., Arkin, I. T., (2005) Solution NMR spin relaxation methods
Jensen, M. O., Xu, H., Trbovic, N., Friesner, for characterizing chemical exchange in high-
R. A., Palmer, A. G., and Shaw, D. E. (2008) molecular-weight systems. Methods Enzymol.
Microsecond molecular dynamics simulation 394, 430–465.
shows effect of slow loop dynamics on back- 18. Fushman, D., Cahill, S., and Cowburn, D.
bone amide order parameters of proteins. (1997) The main chain dynamics of the dynamin
J. Phys. Chem. B 112, 6155–6158. pleckstrin homology (PH) domain in solution:
9. Fushman, D., and Cowburn, D. (2001) Analysis of 15 N relaxation with monomer/dimer
Nuclear magnetic resonance relaxation in equilibration. J. Mol. Biol. 266, 173–194.
24 Determining Protein Dynamics from 15N Relaxation Data… 511
19. Hall, J. B., and Fushman, D. (2006) Variability diffusion of a protein from 15 N relaxation
of the 15 N Chemical Shielding Tensors in the measurements and hydrodynamic calculations,
B3 Domain of Protein G from 15 N Relaxation in Protein NMR techniques (Methods in
Measurements at Several Fields. Implications Molecular Biology) (A.K.Downing, Ed.),
for Backbone Order Parameters. J.Am.Chem. pp 139–160, Humana Press Inc.
Soc. 128, 7855–7870. 29. Fushman, D., and Cowburn, D. (2002)
20. Fushman, D., and Cowburn, D. (1998) Characterization of Inter-Domain Orientations
Studying protein dynamics with NMR relax- in Solution Using the NMR Relaxation
ation, in Structure, Motion, Interaction and Approach, in Protein NMR for the Millenium
Expression of Biological Macromolecules (Sarma, (Biological Magnetic Resonance Vol 20) (N. R.
R., and Sarma, M., Eds.), pp 63–77, Adenine Krishna, L. B., Ed.), pp 53–78, Kluwer.
Press, Albany, NY. 30. Fushman, D. (2002) Determination of protein
21. Blackledge, M., Cordier, F., Dosset, P., and dynamics using 15 N relaxation measurements,
Marion, D. (1998) Precision and uncertainty in BioNMR in drug research (O.Zerbe, Ed.),
in the characterization of anisotropic rotational pp 283–308, Wiley-VCH.
diffusion by 15 N relaxation. J.Am.Chem.Soc. 31. Mandel, A. M., Akke, M., and Palmer, A. G. I.
120, 4538–4539. (1995) Backbone dynamics of E. coli
22. Fushman, D., Xu, R., and Cowburn, D. (1999) Ribonuclease HI: correlations with structure
Direct determination of changes of interdo- and function in an active enzyme. J. Mol. Biol.
main orientation on ligation: use of the orien- 246, 144–163.
tational dependence of 15 N NMR relaxation in 32. Press, W. H., Teukolsky, S. A., Vetterling, W.
Abl SH(32). Biochemistry 38, 10225–10230. T., and Flannery, B. P. (1992) Numerical
23. Fushman, D., Ghose, R., and Cowburn, D. Recipes in C, Cambridge University Press, NY.
(2000) The effect of finite sampling on the 33. Fushman, D., Weisemann, R., Thüring, H.,
determination of orientational properties: A and Rüterjans, H. (1994) Backbone dynamics
theoretical treatment with application to inter- of ribonuclease T1 and its complex with 2’GMP
atomic vectors in proteins. J. Am. Chem. Soc. studied by two-dimensional heteronuclear
122, 10640–10649. NMR spectroscopy. J. Biomol. NMR 4, 61–78.
24. Dosset, P., Hus, J. C., Blackledge, M., and 34. Kay, L. E., Torchia, D. A., and Bax, A. (1989)
Marion, D. (2000) Efficient analysis of macro- Backbone dynamics of proteins as studies by
molecular rotational diffusion from heteronuclear N15 inverse detected heteronuclear NMR
relaxation data. J. Biomol. NMR 16, 23–28. spectroscopy: application to staphylococcal
25. Ghose, R., Fushman, D., and Cowburn, D. nuclease. Biochemistry 28, 8972–8979.
(2001) Determination of the Rotational 35. Hall, J. B., and Fushman, D. (2003)
Diffusion Tensor of Macromolecules in Characterization of the overall and local dynam-
Solution from NMR Relaxation Data with a ics of a protein with intermediate rotational
Combination of Exact and Approximate anisotropy: Differentiating between conforma-
Methods - Application to the Determination of tional exchange and anisotropic diffusion in the
Interdomain Orientation in Multidomain B3 domain of protein G. J. Biomol. NMR 27,
Proteins. J. Magn. Reson. 149, 214–217. 261–275.
26. Walker, O., Varadan, R., and Fushman, D. (2004) 36. Hall, J. B., and Fushman, D. (2003) Direct
Efficient and accurate determination of the over- measurement of the transverse and longitudinal
all rotational diffusion tensor of a molecule from 15
N chemical shift anisotropy-dipolar cross-
15
N relaxation data using computer program correlation rate constants using 1 H-coupled
ROTDIF. J. Magn. Reson. 168, 336–345. HSQC spectra. Mag. Res. in Chemistry 41,
27. Fushman, D., Varadan, R., Assfalg, M., and 837–842.
Walker, O. (2004) Determining domain orienta- 37. Ryabov, Y. E., Geraghty, C., Varshney, A., and
tion in macromolecules by using spin-relaxation Fushman, D. (2006) An efficient computa-
and residual dipolar coupling measurements. tional method for predicting rotational diffu-
Prog. NMR Spectros. 44, 189–214. sion tensors of globular proteins using an
28. Hall, J. B., Walker, O., and Fushman, D. (2004) ellipsoid representation. J. Am. Chem. Soc. 128,
Characterization of the overall rotational 15432–15444.
INDEX
Alexander Shekhtman and David S. Burz (eds.), Protein NMR Techniques, Methods in Molecular Biology, vol. 831,
DOI 10.1007/978-1-61779-480-3, © Springer Science+Business Media, LLC 2012
513
PROTEIN NMR TECHNIQUES
514 Index
Chromatography n-dodecyl-β-D-maltopyranoside
affinity, Ni-NTA........................................................ 189 (DDM) .......................................... 345, 347, 348
desalting..................................................... 185, 190–191 dodecylphosphocholine (DPC) ..........339–341, 343, 345
heparin....................................................... 185, 189–191 n-octyl-β-D-glucopyranoside (β-OG) ...................... 345
ion exchange (IEC) ............................116, 118, 120, 124 Deuterated target proteins ................................................. 28
reverse phase ............................... 168, 172, 229, 336, 352 Dielectric shielding...........................376–378, 384, 397–398
size-exclusion (SEC) .................................118, 120, 121, 4,4-Dimethyl-4-silapentane-1-sulfonic acid
125, 189, 198, 210, 310, 313, 315, 317, 339 (DSS) ........................ 92, 100–101, 346, 351, 414
Circular dichroism (CD) spectroscopy .................... 340–341 Dimyristoylphosphatidylcholine (DMPC)
CLEAN chemical EXchange-phase modulator bilayer ............................................ 340, 345, 349
(CLEANEX-PM) ......................... 372, 378–380 Dipolar coupling
1
CMC. See Critical micelle concentration H, 1H dipolar coupling ..................................... 281, 487
Collaborative Computing Project for the NMR N–H dipolar coupling, d .................................... 288–289
(CCPN) .................................454–456, 465, 467, Dipole-dipole (DD) coupling.......................................... 142
469, 473–476, 478, 479 Discoidin domain of DDR2 .............................................. 28
Combined rotation and multiple pulse spectroscopy Dissociation constant ....................... 225, 239, 243, 496, 500
(CRAMPS) technique ................................... 281 DNA processing ...................................................... 181, 182
Confocal microscopy DNA template...................... 75, 83, 181, 198, 199, 201–203
HIV-1 CA assemblies ........ 305, 307, 310–313, 316–318 DOPC liposome preparation ................................ 91, 97, 98
Conformation distribution ..............................369, 370, 373, Double colony selection ................................2, 3, 5, 7–10, 12
374, 376, 385–388, 390, 392, 394, 396, 400 Double cross polarization (DCP) block .......................... 321
Conformer acidities .................. 385, 387–389, 393, 395, 399 1D proton experiment ..................................................... 144
Constraint combination........................................... 444, 448 DREAM sequence .......................................... 321, 326–327
COREX algorithm .......................................................... 371 3D ROCSA-NCA experiment ............................... 322–323
Correlation spectroscopy ..........................281–286, 304, 321 DYNAMICS protocol ............................................ 485–509
Co-transfection of insect cells ......................... 43, 44, 46–47
CP. See Cross polarization E
1
H–15N CP block ..................................................... 320, 321 Eigen acid... ..................................................... 375, 378, 388
Critical micelle concentration (CMC) ................... 339–341, Electronic polarizability........................................... 376–378
343, 345, 347–348 Electrophoretic mobility shift assay (EMSA).................. 224
Cross polarization (CP) ................................... 281, 304, 327 Episomal vectors.......................................................... 56–57
Cryo-SEM Eukaryotic protein kinase-2 (ERK2)....................... 359–367
HIV-1 CA assemblies ....................................... 318–319 Expression system
Crystallography and NMR system (CNS) bacteria, E. coli
software.................................................. 138, 455 BL21(DE3) ........................... 7, 74, 77, 79, 135, 168,
CSA. See Chemical shift anisotropy 171, 268, 273, 312, 314, 315, 319, 344, 346, 351
CSA/DD cross-correlated cross-relaxation rates ..... 153–157 BL21(DE3) pLysS .......................168, 171, 183, 184
15
N CSA shielding tensor ................................................ 290 Rosetta 2 (DE3) .......................................... 313, 316
CSP. See Chemical shift perturbation baculovirus
CYANA............................................432–433, 437, 444, 454 Autographa californica ............................................. 40
multicapsid nucleopolyhedrovirus
D
(AcMNPV) ............................................... 40, 42
DARR sequence ............................... 304, 306, 307, 321, 326 insect cells
DD coupling. See Dipole-dipole coupling Spodoptera frugiperda Sf9/Sf21 ......................... 43–49
DelPhi....... ...................................................................... 380 Trichoplusia ni BTI 5B1-4 (High Five™)
Detergent micelles cells .................................................................. 43
reconstitution of M2 into .................................. 169, 172 mammalian cells
Detergents (membrane proteins) human embryonic kidney 293
d38-DPC ........................................................... 343, 345 (HEK293) cells .........................39, 57–59, 61, 63
dihexanoyl-sn-glycerol-3-phosphocholine optimization .............................................................. 125
(DHPC) .........................167, 169, 172–174, 177 yeast, Kluyveromyces lactis (K. lactis)
dimyristoylphosphatidylcholine acetamidase gene (amdS) ....................................... 21
(DMPC) ........................................ 340, 345, 349 Lac4 promoter........................................................ 21
1,2-dioleoyl-sn-glycero-3-phosphocholine yeast, Pichia pastoris (P. pastoris)
(DOPC) ...................................60, 90, 91, 97, 98 AOX1 promoter/gene............................................. 20
PROTEIN NMR TECHNIQUES
Index
515
T T7 RNA polymerase
preparation of .................................................. 74, 79–81
TEM. See Transmission electron microscopy TROSY. See Transverse relaxation optimized spectroscopy
TOCSY..... ............................... 208, 226, 363, 364, 379, 435
TPPM decoupling ................................................... 320, 327 U
Transfection .................................... 46, 50, 51, 55–57, 59, 63
Transmembrane (TM) helix protein................................ 334 Ubiquitin... ..............................................373, 374, 379, 380,
Transmission electron microscopy (TEM) 384–391
CAP-Gly/microtubule assemblies ............. 317–318, 320 UNIO protocol
HIV-1 CA assemblies ....................................... 317–318 ASCAN algorithm ............................ 433–436, 441–443
Transverse cross-correlated cross-relaxation rates .... 143, 155 ATNOS algorithm ............................................ 435–438
Transverse relaxation optimized spectroscopy (TROSY ) CANDID algorithm ......................... 433–437, 443–446
methyl-TROSY ......................................... 134, 137, 139 MATCH algorithm............................433, 435, 438–441
15
N-edited NOESY-TROSY ............................. 362–364
15
V
N relaxation dispersion CPMG TROSY
experiment ..................................................... 176 Violation analysis .....................................456, 461, 462, 466
TROSY-HNCO ....................................................... 360 Violation tolerance .................................................. 461, 463
Transverse relaxation rate (T2)
Carr-Purcell-Meiboom-Gill (CPMG) W
echo train ................ 143, 147, 150–151, 157, 158
WATERGATE ............................................................... 247
single echo .......................... 143, 147, 149–150, 157–159
Water-Ligand Observed via Gradient Spectroscopy
Triple resonance NMR experiments
(WaterLOGSY ) .................................... 240, 244
HA(CA)NH.............................................................. 267
Water suppression ...................................159, 209, 242, 247,
HNCA .......................................172–173, 267, 286, 361
251, 256, 281
HNCACB ..........................................172–173, 286, 361
WHAT IF. ....................................... 456, 466, 472, 473, 475
HNCACO ................................................................ 286
HNCO .............................. 145, 172–173, 267, 360–361, X
365, 366, 413, 416, 424
HNCOCA ................................................................ 286 Xplor-NIH ........................138, 169, 175, 432, 433, 437, 444
iHNCA ..................................................... 411, 413, 416
Z
iHNCB.............................................................. 413, 416
iHNCO ..................................................... 413, 416, 424 Zeocin.............................................................. 20–22, 27, 29