Você está na página 1de 5

Solid State Communications 131 (2004) 1519

www.elsevier.com/locate/ssc

NMR chemical shifts in amino acids: effects of environments


in the condensed phase
Young-Gui Yoona,b,*, Bernd G. Pfrommera,b, Steven G. Louiea,b, Andrew Canningc
a
Department of Physics, University of California at Berkeley, Berkeley, CA 94720, USA
Department of Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA
c
National Energy Research Scientific Computing Center, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA
b

Received 18 February 2004; accepted 16 April 2004 by E. Molinari

Abstract
We present calculations of NMR chemical shifts in crystalline phases of some representative amino acids such as glycine,
alanine, and alanyl-alanine. We explore the effects of environment on the chemical shifts in selected glycine geometries ranging
from the crystalline phase to completely isolated molecules. In the crystalline and dilute molecular limits, the calculated distinct
NMR chemical shifts are attributed to intermolecular hydrogen-bonds and dipole electric field effects, respectively.
q 2004 Elsevier Ltd. All rights reserved.
PACS: 82.56.Pp; 76.60.Cq; 71.15.Mb
Keywords: D. NMR chemical shift; A. Amino acid; A. Glycine; D. Hydrogen-bond

1. Introduction
Nuclear magnetic resonance (NMR) chemical shifts
measure the local induced magnetic field on a nucleus,
which is sensitive to the chemical environments. They can
be useful as fingerprints for detailing the structure and
chemical composition of condensed matter systems. The
NMR chemical shifts are becoming increasingly important
for the characterization of biomolecules, especially for
proteins [1,2], which motivated the present work on the
amino acidsglycine, alanine, and a very simple peptide,
alanyl-alanine. A systematic study of the NMR chemical
shifts of amino acids, the building blocks of proteins in
simpler environments, is highly desired, to lay the
foundations for subsequent studies of protein shifts in
complex biological environments. Some features can be
revealed by investigating intermediate configurations
between known crystalline phases and completely isolated
* Corresponding author. Present address: Department of Physics,
Chung-Ang University, 221 Huksuk-dong Dongjak-ku, 156756
Seoul, South Korea. Tel.: 82-282-052-38; fax: 82-282-549-88.
E-mail address: yyoon@cau.ac.kr (Y.G. Yoon).
0038-1098/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ssc.2004.04.023

molecules. Nevertheless, only recently it has become


possible to calculate in an ab initio fashion the NMR
chemical shifts of extended system [5 9].
In this study, we present NMR chemical shift calculations of amino acids in the condensed phase. The
crystalline phase of some representative amino acids are
investigated. We analyze proton chemical shifts in glycine
systematically to address the effects of the environments,
including the effect of the electric field in the molecular
limit and that of amine group rotation.
The calculated of proton shifts in configurations between
that of a crystal and an isolated molecule suggests that there
may be the potential of using NMR chemical shifts for the
analysis of the solvent effects and effects of environments in
the real protein conformations. The proton shifts in the
crystalline limit are analyzed in terms of the effects of
hydrogen-bonded oxygen atoms, while the trend in the shifts
in the dilute molecular limit is attributed to long-range
electric field.
It is well known that amine and methane groups rotate at
ambient temperature in crystalline phases of amino acids at
a rate that is too fast to resolve the three protons in these
groups in NMR chemical shift experiments [3,4]. Indeed,

16

Y.-G. Yoon et al. / Solid State Communications 131 (2004) 1519

there is only one observed line corresponding to the protons


belonging to an amine or a methane group of a-glycine and
L -alanine. Therefore, we have calculated the proton
chemical shifts and averaged them in the following manner.
First, we perform a NMR calculation using the equilibrium
geometries, and the chemical shifts are averaged over the
three equivalent protons belonging to an amine or a methane
group of the a-glycine and L -alanine. Second, we explicitly
rotated the amine group of the a-glycine by the finite angles
of ^30, and 608 around the axis connecting the proton
center of mass and the nitrogen of the amine group, and
performed three additional NMR calculation with those
geometries. Then, the glycine chemical shifts are averaged
over the equilibrium and three additional geometries, which
leads to a total of twelve different proton site considered per
the rotating amine group of the a-glycine. This averaging
procedure improves the agreement with experiment
substantially.

2. Structure
We examine NMR chemical shifts in the crystalline
phases of a-glycine, L -alanine, and L -alanyl-alanine. The
crystal a-glycine [10] with space group 21 =n and L -alanine
[11] with space group P21 21 21 are selected because of their
simple geometry. L -alanyl-L -alanine [3] with space group I4
is included as a dipeptide prototype. Since highly reliable
neutron scattering data are available for the structure of
a-glycine [10] and L -alanine [11] for the heavier elements,
we relax only the position of the hydrogen atoms within the
local density approximation to get an accurate geometry
using the ab initio pseudopotential density functional
method [12]. For L -alanyl-L -alanine, no neutron diffraction
data are available to our knowledge. Therefore, we relax the
whole structure starting from the geometry as given by the
X-ray diffraction experiments [13].

3. Planewave and pseudopotential parameters


In NMR measurements, the chemical shift is one third of
the trace of the chemical shift tensor, sr Trsyr=3;
which connects the induced magnetic field to the
external uniform applied magnetic field through the
relation, Bin r 2syrBext : We compute s following
Refs. [6,8]. The electronic structure is described using
density functional theory in the local density approximation
with pseudopotentials. We expand the wave functions in a
plane wave basis set with an energy cut-off of 100 Ry. Test
of our pseudopotentials for C and N have been reported
previously [6,8].
Following the experimental convention, we will quote
chemical shifts with respect to standard reference systems of
a neat liquid sample with spherical shape at 300 K. The C,H

shifts are given with respect to tetramethylsilane (TMS) by


dTMS sample 2ssample 2 sTMS; where s is the
absolute chemical shift. For N, nitromethane is used as
standard: dCH3 NO2 sample 2ssample 2 sCH3 NO2 :
sTMS and sCH3 NO2 are not directly computed; for the
H dTMS ; we fix sTMS by imposing that the computed
value of dTMS for gas CH4 is equal to the experimental value
of 0.13 ppm [14]. For the C dTMS and the N d3 NO2 ; we use the
procedure in Ref. [8].

4. Results and discussion


Tables 1 and 2 show the computed H, C, and N chemical
shifts for the amino acids studied in their crystalline phase.
The overall agreement with experiment is good. For the
carbon chemical shifts, the calculated shifts are compared
with four available experimental data sets [19,21,23,24]. All
four experiments employed the cross-polarization technique, and two of them [19,23] also employed the magicangle spinning technique. For the case of a-glycine, the
carbon shifts are compared with two different experiments.
Two sets of experimental data are essentially the same. The
degree of agreement with the experiments is dependent on
the kind of carbons. The same dependence is seen in the case
of the L -alanine and L -alanyl-L -alanine carbon chemical
shifts. To analyze the dependence, notice that there are two
kinds of hybridization for carbon atoms. First, C(2) of
a-glycine, C(2), C(3) of L -alanine, and C(1), C(2), C(4),
C(5) of L -alanyl-L -alanine has sp3 hybridization. Secondly,
C(1) of a-glycine, C of L -alanine, and C(3), C(6) of
2
2
L -alanyl-L -alanine has sp hybridization. The sp hybridized
Table 1
Computed NMR chemical shifts dref for crystalline a-glycine.
Reference materials are liquid TMS for C, and H, and liquid
CH3NO2 for N [15]. The chemical shifts calculated with the single
geometry are shown in parentheses in column 2. For the members
not in parentheses, we obtain them by averaging the shifts of glycine
in the crystalline phase over geometries obtained by rotating the
amine group by 0, ^30, and 608 around the axis connecting the
proton center of mass and the nitrogen of the amine group. The atom
labels in parentheses are from Ref. [10]. H(1,2,3) denotes the
averaged proton shift over H(1), H(2) and H(3)

dref

H(1)
H(2)
H(3)
H(1,2,3)
H(4)
H(5)
C(1)
C(2)
N

Theory (ppm)

Experiment (ppm)

(12.8)
(10.1)
(7.0)
9.0(10.0)
6.2(6.0)
4.0(3.9)
151(153)
44(44)
2321(2334)

7.8 [18], 8.5 [19]


4.3 [18], 4.4 [19]
3.5 [18], 3.4 [19]
177 [21], 176 [19]
46 [21], 44 [19]
2348 [19]

Y.-G. Yoon et al. / Solid State Communications 131 (2004) 1519

17

Table 2
Computed NMR chemical shifts dref for L -alanine and L -alanyl-L -alanine. Reference materials are liquid TMS for C, and H, and liquid
CH3NO2 for N [15]. The atom labels are from Ref. [26] for L -alanine and Ref. [13] for L -alanyl-L -alanine. In Ref. [22], data are read from
ultrafast MAS plot. H(1,2,3) and H(5,6,7) denote the averaged proton shifts over H(1), H(2), H(3) and H(5), H(6), H(7), respectively
L -alanine

H(1,2,3)
H(4)
H(5,6,7)
C(1)
C(2)
C(3)
N

dref

L -alanyl-L -alanine

Theory (ppm)

Experiment (ppm)

10.7
4.9
2.2
154
51
22
2320

8.4 [22]
3.6 [22]
1.2 [22]
177 [23]
51 [23]
20 [23]

carbons tend to agree less well with experiment than the sp3
hybridized ones, consistent with previous calculations [6,8].
The nitrogen chemical shifts agree well with experimental
data. The a-glycine nitrogen shift is compared with the
magic-angle spinning cross-polarization experiment [19],
and the L -alanyl-L -alanine nitrogen shifts is compared with
cross-polarization experiment [24]. For the proton chemical
shifts, excellent agreement with experiments in the literature
is obtained. For a-glycine experiments, 1H CRAMPS
(combined rotation and multiple pulse spectroscopy) [19]
and solid-state multiple pulse spectroscopy for deuteron
NMR [18] methods were used in the references. Our
theoretical calculations agree better with more recent and
more accurate 1H CRAMPS results. For L -alanine proton
chemical shifts, the data are from experimental solid-state
ultrafast magic-angle spinning vR =2p 35 kHz measurements [22].
Focusing on the proton chemical shifts of amine group in
glycine, the shifts that are not averaged deviate somewhat
from experiment. Theoretical proton chemical shifts in the
same amine group differ from each other by as much as
5.8 ppm, although there is only one observed peak in
experiment corresponding to these hydrogen atoms. This is
not unexpected because the amine and methane group of
amino acids in the crystalline phase are known to rotate over
a wide range of temperature [3,4]. The experimental rate of
amine group rotation from the spin-lattice relaxation T1
measurement and Arrhenius expression of correlation time
is found to be around per second [4]. This dynamical motion
cannot be resolved in typical NMR chemical shifts
measurements, and the motion averages the shifts over the
rotation. To investigate the dynamical effect, we compare
chemical shifts calculated with the single relaxed equilibrium geometry of the glycine to those averaged over the
geometries corresponding to amine group rotations in Table
1. The chemical shifts calculated with the single geometry
are shown in parentheses in Table 1. For the members not in
parentheses, we obtain them by averaging the shifts of
glycine in the crystalline phase over geometries obtained by

dref
Theory (ppm)

C(2)
C(1)
C(3)
C(4)
C(5)
C(6)
N(1)
N(2)

47
20
150
50
20
156
2327
2257

Experiment (ppm)

170 [24]

2260 [24]

rotating the amine group by 0, ^30, and 608 around the axis
connecting the proton center of mass and the nitrogen of the
amine group. The finite angles are chosen to sample
geometries of continuous angles of the amine group
rotation. We average over protons in the same amine
group, since they are not resolved in the time scale of NMR
experiments. The agreement with experiment is now
significantly improved. The averaged proton chemical shifts
with rotation reduce the error. This calculation also shows
that the effects of rotation of amine group can be reasonably
well described with relatively small sampling of motion. We
note that amine group rotation does not affect the atoms that
do not belong to amine group. (Again, members not in
parentheses are averaged over 4 rotational geometries).
Only the nitrogen shift is affected appreciably.
We perform a simple study of the effects of environments on the chemical shifts in glycine by calculating
configurations between the known crystalline phase and
completely isolated molecules, keeping the internal geometry of individual molecules to be that in the crystal but
with increasing intermolecular distances (Fig. 1). Chemical

Fig. 1. The effects of dilation on the chemical shifts in glycine in


intermediate configurations.

18

Y.-G. Yoon et al. / Solid State Communications 131 (2004) 1519

shift at infinite volume per molecule is extrapolated from a


separate set of supercell calculations to get values that are
accurate and independent of our choice of intermediate
configurations. The proton chemical shifts as a function of
the inverse volume per molecule are fairly linear if the

intermolecular neighbor distance in the crystal is around 4 A


larger. Since the electric field of a dipole moment is
proportional to the inverse of the volume per molecule in the
molecular limit, the linearity at large distance indicates
the influence of long-range electric field effects of the
neighboring molecules. The proton chemical shifts in this
study may shed a light on the trend of the proton chemical
shifts in proteins. Unlike chemical shifts in proteins, which
are expected to have a 10 ppm range from electrostatic
interaction [1], the proton chemical shifts in proteins is
estimated to be less sensitive to electric field with a range
less than 1 ppm. The significant feature of proton chemical
shifts in protein folding will be hydrogen bonding, which is
also pointed out in the Ref. [1]. Most of the change in the
proton chemical shifts between the two limits in Fig. 1 is
attributed to nearest intermolecular neighbor oxygen atoms.
H(1), H(3), and H(4) have two adjacent oxygen atoms that
do not belong to the single glycine molecule unit. H(2) has
only one such oxygen atom, and H(5) has no such oxygen
atom. To see the relation between the oxygen-hydrogen
distance dOH and the chemical shift, we tabulated the
3
quantity obtained by multiplying dOH
and the difference of
the H(2) chemical shift from that the isolated configuration
in Table 3. (We did not tabulated the product in which dOH
because, then, the second nearest dOH is
exceeds 4 A
comparable to the nearest dOH :) Robust linearity is found
between the chemical shifts change and the inverse of the
third power of the intermolecular distance in the hydrogen
H(2) which has only one intermolecular neighbor oxygen.
This is not new in the literature. The same relation between
proton chemical shifts with hydrogen bond length in water
dimer was reported before [20].
In Fig. 2, we present residual chemical shifts after
subtracting the nearest intermolecular neighbor oxygen
contribution by assuming a dsOnearest  138:54 ppm au3 =
3
dOH
relation following Ref. [20]. Indeed, the coefficient
138.54 ppm au3 is quite close to the quantities tabulated in

Fig. 2. Proton chemical shifts in glycine after subtracting hydrogenbonding effects using empirical formula for water, dsOnearest 
3
: H(5) has no hydrogen-bond with oxygen
138:54 ppm au3 =dOH
atoms, and thus, there is no such subtraction for H(5) chemical
shifts.

Table 3, and the discrepancy can be attributed to the fact that


there are other distant oxygen atoms in the surroundings and
that the model does not necessarily take into account all
possible details of quantum mechanical effects. However,
the major contribution of the chemical shift change in the
small intermolecular oxygen hydrogen distance limit is
well described by the empirical formula. The distribution of
the proton chemical shifts is expected to carry information
on the existence of oxygen atoms in unknown environments.
Therefore, proton chemical shifts might be a good oxygen
probe which is insensitive to a long-range electric field.
Another interesting comparison is between theoretical
calculation of amine protons in the crystalline glycine limits
and experimental measurements of various hydrogenbonded glycine amide protons of glycine-containing peptides and polyglycines in the solid state [25]. In Fig. 3, the

Table 3
Relation between oxygen-hydrogen distance dOH and the chemical
shifts of H(2) in a-glycine
O H distance dOH
(au)

dHTMS
(ppm)

3
dTMS dOH 2 dH
TMS 1dOH
(ppm au3)

3.398
3.732
4.163
4.731
5.508
6.638
1

10.08
9.03
8.11
7.41
6.92
6.61
6.07

157
154
147
141
141
156

Fig. 3. Proton chemical shifts in glycine against their hydrogenbond length between amide nitrogen and oxygen atoms RN O :
Experimental data are taken from Ref. [25].

Y.-G. Yoon et al. / Solid State Communications 131 (2004) 1519

proton chemical shifts are plotted against hydrogen-bond


lengths between amide nitrogen and oxygen atoms RN O :
All experimental data are around the shift data of H(2) atom
which has a single hydrogen-bond with an oxygen atom.
The experimental correlation between RN O and the proton
chemical shifts of different peptides and polyglycines is
consistent with the trend of our calculations, which indicates
that analysis of this kind may be useful not only for glycine
crystals and molecules but also for various related materials.

5. Conclusion
In summary, our calculations show that NMR chemical
shift calculation may be helpful in understanding effects of
environments in amino acids and the dynamical effects of
rotating amine groups. Effects of environments on glycine
proton chemical shifts in the crystalline limit is consistent
with the influence of the intermolecular hydrogen-bond with
oxygen atoms as in water, and those in the dilute molecular
limit are attributed to long-range electric field of dipole
moments of neighboring molecules. Proton chemical shifts
averaged over the amine group rotations appear to mimic the
correct number of peaks and their positions in experiments.

Acknowledgements
This work was supported by the NSF under Grant No.
DMR-9520554, and by the Director, Office of Advanced
Scientific Computing Research, Division of Mathematical,
Information and Computational Sciences of the US Department of Energy under contract number DE-AC0376SF00098. Computer time was provided by the NSF at
the National Center for Supercomputing Applications and
by the Office of Science of the DOE at the National Energy
Research Scientific Computing Center.

References
[1] A. de Dios, J.G. Pearson, E. Oldfield, Science 260 (1993)
1491.
[2] A. de Dios, J. Pearson, E. Oldfield, J. Am. Chem. Soc. 115
(1993) 9768.
[3] G.R. Kneller, W. Doster, M. Settles, S. Cusack, J.C. Smith,
J. Chem. Phys. 97 (1992) 8864.

19

[4] Z. Gu, K. Ebisawa, A. McDermott, Sol. Stat. Nucl. Magn. Res.


7 (1996) 161.
[5] F. Mauri, B.G. Pfrommer, S.G. Louie, Phys. Rev. Lett. 77
(1996) 5300.
[6] F. Mauri, B.G. Pfrommer, S.G. Louie, Phys. Rev. Lett. 79
(1997) 2340.
[7] F. Mauri, A. Pasquarello, B.G. Pfrommer, Y.-G. Yoon, S.G.
Louie, Phys. Rev. B 62 (2000) R4786.
[8] Y.-G. Yoon, B.G. Pfrommer, F. Mauri, S.G. Louie, Phys. Rev.
Lett. 80 (1998) 3388.
[9] B.G. Pfrommer, F. Mauri, S.G. Louie, J. Am. Chem. Soc. 122
(2000) 123.
. Kvic, Acta Crystallogr. B 28 (1972) 1827.
[10] P. Jonsson, A
[11] M. Lehmann, T. Koetzle, W. Hamilton, J. Am. Chem. Soc. 94
(1972) 2657.
[12] W. Kohn, L.J. Sham, Phys. Rev. 140 (1965) A1133.
[13] R. Fletterick, C. Tsai, R. Hughes, J. Phys. Chem. 75 (1971)
918.
[14] W. Kutzelnigg, U. Fleischer, M. Schindler, in: P. Diehl (Ed.),
NMR Basic Principles and Progress, vol. 23, Springer, New
York, 1990.
[15] For C, to convert dbenzene of experimental data to dTMS we use
dTMS C; benzene 128:5 ppm [14]. For N, to convert
dNH3 liquid ; and dNH4 Clsaturated of experimental data to
dCH3 NO2 liquid we use dCH3 NO2 liquid N; NH3 liquid 2380
ppm [16] and dCH3 NO2 liquid NH4 Clsaturated 2359:5 ppm
[17].
[16] C.J. Jameson, A.K. Jameson, D. Oppusunggu, S. Wille, P.M.
Burrell, J. Mason, J. Chem. Phys. 74 (1981) 81.
[17] G. Martin, M. Martin, J. Gouesnard, in: P. Diehl (Ed.), NMR
Basic Principles And Progress, vol. 18, Springer, New York,
1981.
[18] C. Muller, W. Schajor, H. Zimmermann, U. Haeberlen,
J. Magn. Res. 56 (1984) 235.
[19] H. Kimura, K. Nakamura, A. Eguchi, H. Sugisawa, K.
Deguchi, K. Ebisawa, E.-I. Suzuki, A. Shoji, J. Mol. Struct.
447 (1998) 247.
[20] R. Ditchfield, J. Chem. Phys. 65 (1976) 3123.
[21] R.A. Haberkorn, R.E. Stark, H. van Willigen, R.G. Griffin,
J. Am. Chem. Soc. 103 (1981) 2534.
[22] I. Schnell, A. Lupulescu, S. Hafner, D.E. Demco, H.W. Spiess,
J. Magn. Res. 133 (1998) 61.
[23] A. Naito, S. Ganapathy, K. Akasaka, C.A. McDowell, J. Chem.
Phys. 74 (1981) 3190.
[24] C. Hartzell, M. Whitfield, T.G. Oas, G.P. Drobny, J. Am.
Chem. Soc. 109 (1987) 5966.
[25] K. Yamauchi, S. Kuroki, K. Fujii, I. Ando, J. Chem. Phys. 324
(2000) 435.
[26] H.J. Simpson Jr., R.E. March, Acta Crystallogr. 20 (1966) 550.

Você também pode gostar