Você está na página 1de 9

Analytical Biochemistry 425 (2012) 104–112

Contents lists available at SciVerse ScienceDirect

Analytical Biochemistry
journal homepage: www.elsevier.com/locate/yabio

Separation and detection of individual Ab aggregates by capillary

electrophoresis with laser-induced fluorescence detection
Ryan A. Picou a, David P. Schrum b, Ginger Ku c, Rebekah A. Cerqua a, Indu Kheterpal c,
S. Douglass Gilman a,⇑
Department of Chemistry, Louisiana State University, Baton Rouge, LA 70803, USA
Department of Chemistry, University of Redlands, Redlands, CA 92373, USA
Pennington Biomedical Research Center, LSU System, Baton Rouge, LA 70808, USA

a r t i c l e i n f o a b s t r a c t

Article history: The separation and detection of individual amyloid beta (Ab) aggregates by capillary electrophoresis with
Received 31 December 2011 laser-induced fluorescence detection (CE–LIF) was demonstrated. Samples were prepared with either Ab
Received in revised form 8 March 2012 (1–40) or Ab (1–42) peptides and were characterized by CE with ultraviolet (UV) absorbance detection
Accepted 14 March 2012
and transmission electron microscopy (TEM). Using thioflavin T (ThT) in the electrophoresis buffer, elec-
Available online 21 March 2012
trophoresis of aggregate-containing samples (5.0-s injection) produced up to several hundred narrow
(<20 ms FWHM [full width at half maximum]) fluorescence peaks. Injection of Ab (1–40) monomer sam-
ples resulted in no additional peaks compared with controls. The CE–LIF results were validated by bulk
Capillary electrophoresis
ThT fluorescence measurements for the same samples. The potential of laser-induced fluorescence
Amyloid anisotropy (LIFA) with CE to characterize individual Ab aggregates also was investigated.
Fluorescence anisotropy Ó 2012 Elsevier Inc. All rights reserved.
Protein aggregation

Senile plaques composed of neurofibrillary tangles and poly- of time for the entire reaction landscape from monomer to fully
morphic amyloid beta (Ab)1 fibrils are hypothesized to be responsi- formed fibrils, where all forms of Ab would be distinguishable.
ble for the neurological degeneration associated with Alzheimer’s Current analytical techniques are far from this ideal. A variety
disease (AD) [1,2]. Aggressive research efforts have focused on devel- of techniques have been employed to study static Ab structures,
oping methods to dissociate or prevent the formation of amyloid mainly monomers or fully formed fibrils. These include imaging
fibrils [2,3]. Effective development of such treatments requires a techniques such as transmission electron microscopy (TEM),
thorough understanding of the formation and composition of Ab pla- atomic force microscopy, and total internal reflection fluorescence
ques. Ab (1–40) and Ab (1–42) peptides readily form amyloid fibrils microscopy [4,5]. In addition, spectroscopic and spectrometric
in vitro. The aggregation reaction from soluble monomeric peptide techniques such as circular dichroism, thioflavin T (ThT) and Con-
to insoluble mature fibrils is a complex multistep process that in- go red fluorescence, fluorescence correlation spectroscopy, light
cludes heterogeneous populations of dynamic polymorphic interme- scattering, surface plasmon resonance, nuclear magnetic reso-
diates [2]. These intermediates have been implicated to be cytotoxic; nance, and mass spectrometry have been applied to study Ab
however, the identity and structure of the cytotoxic species remain structures [6–12].
unknown. These multiple potential therapeutic targets make devel- Detecting and distinguishing monomeric Ab peptides, various
oping a treatment for AD extremely challenging. intermediate oligomeric aggregates, and fully formed fibrils within
An ideal technique to analyze Ab peptide aggregation would a sample mixture using one technique is an important challenge
provide concentration and structural information as a function that has not yet been met. Most of the aforementioned techniques
can detect larger peptide aggregates or fibrils, but they are incapa-
ble of simultaneously detecting smaller aggregated peptide struc-
⇑ Corresponding author. Fax: +1 225 578 3465.
tures. Imaging techniques, for example, have a size threshold
E-mail address: sdgilman@lsu.edu (S.D. Gilman).
Abbreviations used: Ab, amyloid beta; AD, Alzheimer’s disease; TEM, transmission
below which the peptide is not detectable. This size threshold also
electron microscopy; ThT, thioflavin T; SEC, size exclusion chromatography; CE, exists for light scattering detection. ThT fluorescence measure-
capillary electrophoresis; LIF, laser-induced fluorescence; LIFA, laser-induced fluo- ments allow the detection of peptide aggregates that have folded
rescence anisotropy; Tris, tris(hydroxymethyl) aminomethane; MO, mesityl oxide; into a b-sheet conformation that is more pronounced as a function
TFA, trifluoroacetic acid; HFIP, hexafluoroisopropanol; PBS, phosphate-buffered
of increasing aggregate size. The signal measured by bulk fluores-
saline; HPLC, high-performance liquid chromatography; UV, ultraviolet; PMT,
photomultiplier tube; FWHM, full width at half maximum. cence is also dominated by larger Ab aggregates that mask the

0003-2697/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
Analysis of Ab aggregates by CE–LIF / R.A. Picou et al. / Anal. Biochem. 425 (2012) 104–112 105

signal produced from smaller aggregates. Larger aggregates also 334 were prepared at 50 nM in Tris buffer. ThT was purchased
dominate bulk light scattering measurements. Although there is from Sigma (St. Louis, MO, USA). A stock solution of 289 lM ThT
no one technique to monitor all forms of Ab, all of the current tech- solution was prepared in Tris buffer, and working solutions of
niques are very informative and important within their limitations. ThT were prepared at 15.0 lM in Tris buffer.
Characterizing the structural pathway from monomeric pep-
tides to fibrillar aggregates involves measuring and characterizing Ab peptide sample preparation and characterization
the intermediate dynamic oligomeric structures for which little
information is available. Separations of Ab aggregates based on size Ab (1–40) peptide was purchased from the W.M. Keck Founda-
have been performed using several techniques, including ultracen- tion Biotechnology Research Laboratory at Yale University (New
trifugation, sedimentation velocity analysis, gel electrophoresis, Haven, CT, USA), and Ab (1–42) peptide was purchased from rPep-
size exclusion chromatography (SEC), and capillary electrophoresis tide (Bogart, GA, USA). Four types of Ab sample were studied in this
(CE) [13–23]. Although CE was first applied to the analysis of Ab work: Ab (1–40) monomer, Ab (1–40) fibril, Ab (1–42) monomer,
monomer in 1993 by Sweeney and coworkers [20], only a few re- and Ab (1–42) fibril. All Ab samples were prepared as described
ports regarding the application of CE to Ab analysis have been pub- previously [27,35]. Ab peptides were treated with trifluoroacetic
lished [13,21–27]. CE offers fast separations with higher peak acid (TFA) and hexafluoroisopropanol (HFIP) to remove any preex-
capacities and lower mass detection limits relative to other separa- isting aggregates. For the Ab monomer sample, the TFA and HFIP
tion methods applied to study Ab aggregation. Rapid separations were evaporated off and the peptide was dissolved in 10.00 mM
are especially important to minimize the disruption of various Tris at pH 7.79. For the Ab (1–40) fibril sample, the TFA and HFIP
aggregate structures when they are physically separated from the were evaporated off and the peptides were dissolved stepwise in
aggregate mixture. Another advantage of CE relative to gel electro- equal volumes of 2.0 mM NaOH and 2 phosphate-buffered saline
phoresis and SEC is that it does not require a gel or stationary (PBS) containing 22.8 mM phosphate, 274 mM NaCl, 5.4 mM KCl,
phase that can disrupt or otherwise alter aggregates due to shear and 0.1% NaN3 at pH 7.4. The samples were centrifuged at
and adsorption. Clodfelter and coworkers demonstrated that CE 50,000g for a minimum of 10 h at 4 °C. The supernatant was incu-
is a more gentle technique relative to SEC for separating protein bated at 37 °C for 7 days until there was no indication of further
aggregates of another aggregate-forming peptide, C8GLIP [28]. aggregation based on ThT fluorescence and high-performance li-
Laser-induced fluorescence (LIF) is a common on-column detec- quid chromatography (HPLC) analysis of monomer concentration.
tion technique for CE, and CE–LIF has been used to study Ab pep- The Ab (1–42) monomer and fibril samples were prepared using
tides [23]. Laser-induced fluorescence anisotropy (LIFA) is a less the same procedure with the following exceptions: the Ab (1–42)
frequently used detection method for CE, and CE–LIFA has been monomer sample was centrifuged for 30 min at 20,000g and 4 °C,
applied to study protein–protein and protein–nucleic acid interac- and the Ab (1–42) fibril sample was incubated for 2 days to form
tions [29–32]. Fluorescence anisotropy is a fluorescence phenome- mature fibrils. These changes were due to the faster aggregation
non that is dependent on the size of the fluorescent entity being kinetics of the Ab (1–42) peptide compared with that of the Ab
investigated. Sabaté and Saupe demonstrated the use of bulk ThT (1–40) peptide.
anisotropy to monitor aggregation kinetics of the fungal prion pro- Fibril growth was monitored using ThT fluorescence and TEM as
tein HET-s(218–289) [33]. Allsop and coworkers demonstrated described previously [12,27]. Depletion of Ab monomer during
that bulk time-resolved fluorescence anisotropy can be used for fibril formation was monitored using a Shimadzu HPLC–UV (ultra-
monitoring early stages of Ab aggregation using a fluorescein-la- violet) instrument with detection at 215 nm as described
beled Ab peptide [34]. CE with LIFA detection has the potential to previously [26,27,35]. Samples were characterized by CE–UV. The
be developed into a powerful tool for studying Ab aggregation. CE–UV analysis was performed using a Beckman Coulter P/ACE
CE is capable of resolving different Ab aggregate species, and LIFA MDQ equipped with a diode array detector, and the absorbance
could provide a sensitive on-line detection method that also pro- was monitored at 190 nm as described previously [27]. The capil-
vides information about the sizes of the separated species. To the lary (i.d. = 50 lm, o.d. = 366 lm) used for CE–UV was cut to
best of our knowledge, there are no reports of the application of 63.0 cm total length with a window created 53.0 cm from the inlet
CE–LIFA to study protein aggregation. end using The Window Maker (MicroSolv Technology, Eatontown,
The goal of this work was to separate and detect individual Ab NJ, USA). The Ab samples were vortexed prior to every injection to
aggregates labeled with ThT using CE with LIF detection. Further- resuspend any aggregates that settled to the bottom of the sample
more, the potential of LIFA to characterize individual Ab aggregates vial.
labeled with ThT was explored. Prior to CE analysis, fibril samples were buffer exchanged from
PBS to 10.00 mM Tris electrophoresis buffer at pH 7.79 [6]. This
Materials and methods step eliminated conductivity and composition differences between
the sample buffer (PBS, higher conductivity) and the electrophore-
Chemicals sis buffer (Tris, lower conductivity). The PBS buffer was used for
aggregation to form fibrils because this buffer is broadly used for
All solutions were prepared in 18 MX water obtained from a studies of Ab peptide fibril formation. The higher conductivity of
Modulab water purification system (U.S. Filter, Palm Desert, CA, PBS would result in high electrophoretic current and poor results
USA) unless otherwise noted. Tris(hydroxymethyl) aminomethane due to excessive Joule heating if it were used for CE. The mono-
(Tris) and methanol (99.8%) were purchased from Fisher Scientific mer-equivalent concentrations of all samples were determined to
(Fair Lawn, NJ, USA). Tris buffer was prepared at 10.00 mM, and the be 5 to 30 lM by HPLC–UV as described previously [35]. For CE
pH was adjusted to 7.79 with HCl. This Tris buffer was used for all analysis, the total sample volume was 50.0 ll, and it was placed
experiments unless otherwise noted. The buffer was filtered in a 200-ll Thermowell polypropylene tube (Corning, Corning,
through a 0.02-lm filter (Whatman, Hillsboro, OR, USA). Mesityl NY, USA).
oxide (MO) was purchased from Alfa Aesar (Ward Hill, MA, USA),
and working solutions of MO were prepared in Tris buffer at a con- CE with LIFA detection
centration of 0.2% (v/v). Coumarin 334 was purchased from Acros
Organics (Morris Plains, NJ, USA), and a stock solution was pre- The Ab samples were first characterized by CE–UV and then
pared at 1.00 lM in methanol. Working solutions of coumarin analyzed with the CE–LIFA system on the same day. The samples
106 Analysis of Ab aggregates by CE–LIF / R.A. Picou et al. / Anal. Biochem. 425 (2012) 104–112

were not transferred to different vials when switching from the RC filters and collected by a National Instrument PCI-6024E DAQ
CE–UV instrument to the CE–LIFA instrument. A laboratory-con- board at a scan rate of 1000 Hz. The CE–LIFA data collection was
structed instrument was used for all CE–LIFA experiments. The controlled by a program written with LabView (version 5.0), and
CE part of the instrument was similar to an instrument described the data were analyzed using Origin Pro 7.5, Synaptosoft MiniAnal-
previously [36]. ysis Program (version 6.0.7), and Microsoft Excel 2007.
CE was performed in an open tubular capillary (i.d. = 50 lm, Eq. (1) was used to calculate the fluorescence anisotropy, r,
o.d. = 366 lm) cut to 60.0 cm total length. A detection window where Ik and I\ are the measured parallel and perpendicular fluo-
was created at 23.0 cm from the inlet end. A new capillary was ini- rescence intensities relative to the polarization of the excitation
tially conditioned with a manual syringe pump as follows: 400 ll source. The factor G corrects the anisotropy for the differences in
of 1.00 M NaOH, 400 ll of 18 MX water, and 2  400 ll of Tris buf- sensitivities of the two PMTs. For this work, G was empirically ad-
fer. The total rinse time was approximately 60 min. The capillary justed to 1 by optimizing the potentials applied to both PMTs so
was rinsed twice between different sample types with 400 ll of that the responses and sensitivities were equal for a small fluores-
Tris electrophoresis buffer using a manual syringe pump. The fluo- cent molecule, Lucifer Yellow (Molecular Probes, Eugene, OR, USA),
rescent neutral marker (coumarin 334) was injected for 1.0 s at from 250 nM to 1 lM (data not shown):
25.0 kV prior to the Ab sample injection for 5.0 s at 25.0 kV. This Ik  GI?
injection order minimized potential interactions between Ab and r¼ : ð1Þ
Ik þ 2GI?
the neutral marker because all forms of Ab detected migrated
slower than the neutral marker. The Ab samples were vortexed
prior to every injection. The electrophoretic potential was applied Microplate reader fluorescence
using a Spellman CZE1000R high-voltage power supply (Hauppa-
uge, NY, USA) at 25.0 kV (417 V/cm). The electrophoretic current All Ab samples were also analyzed by bulk ThT fluorescence
was 5 to 6 lA. This system was not temperature controlled, unlike using a BMG Lab Technologies FLUOstar 430 microplate reader
the commercial CE–UV system. (Offenbur, Germany). The instrument and data acquisition were
Fig. 1 presents a schematic of the detection system for the controlled by FLUOstar software (version 3.02-0). Here, 40 ll of
CE–LIFA instrument. The light source was a 445-nm diode laser each sample was transferred from the 200-ll polypropylene tubes
(model LDCU12/7532, Power Technology, Alexander, AR, USA). used for CE to separate wells of 96-well Corning 3650 black non-
The laser produced a continuous beam of polarized radiation, and transparent microplates. Consistent with the CE experiments, ThT
the measured power was 30.0 mW. The beam was attenuated with was added at 15 lM to each sample well. The excitation was set
a neutral density filter to 3.0 mW and focused onto the capillary to 440 ± 12 nm with a bandpass filter, and the emission was col-
window with a plano-convex lens (focal length = 25.4 mm). The lected at 485 ± 12 with a bandpass filter.
fluorescence emission was collected with a 20 microscope objec-
tive lens (numerical aperture = 0.4) at 90° relative to the excitation Results and discussion
beam. The emission was directed through a notch filter
(441.6 ± 10 nm, MK Photonics, Albuquerque, NM, USA) to remove The primary goal of this work was to separate and detect indi-
Rayleigh scattered light from the laser and a bandpass filter vidual Ab peptide aggregates by CE–LIF using ThT. ThT is a com-
(490 ± 10 nm, Omega Optical, Brattleboro, VT, USA) to selectively mon fluorescence probe used to study Ab aggregation because it
detect amyloid-bound ThT fluorescence. After being spectrally fil- is selective for amyloid aggregates, and its excitation/emission
tered, the emission was spatially filtered with a 1000-lm pinhole. maxima shift from 330/440 to 440/490 nm when it noncovalently
The emission was then separated into its parallel and perpendicu- binds to amyloid aggregates [11,12,37]. ThT was added to the elec-
lar polarized components (relative to the excitation source) with a trophoresis running buffer to bind to the Ab aggregates on-column
broadband polarizing cube beamsplitter (10FC16PB.3, Newport, [23]. This approach eliminates the potential for ThT to alter the Ab
Irvine, CA, USA). Both emission components were detected simul- aggregates’ sizes and shapes during the aggregation process while
taneously and equidistant from the capillary by two identical pho- still being used as an effective fluorescent label for their on-column
tomultiplier tubes (PMTs, H9306-04, Hamamatsu, Bridgewater, NJ, detection. At the excitation and emission wavelengths used in this
USA). The PMT collecting parallel emissions (PMTk) was at 1000 V, work, fluorescence interference from unbound ThT is minimal [11].
whereas the PMT collecting perpendicular emissions (PMT\) was The concentration of ThT (15 lM) was chosen based on previous
at 935 V. The outputs of the PMTs were filtered by 500-Hz lowpass publications that used 10 and 20 lM ThT [11,26] for Ab aggregate
detection. This is above the critical micelle concentration of free
ThT reported by Khurana and coworkers (3.8 ± 0.5 lM) [38].
k The work presented is distinct from the previous study by Kato
and coworkers [23] because the instrument used here was de-
signed to detect peaks resulting from individual aggregated Ab
j i
species, requiring the use of sufficiently fast data sampling and
electronic filters [39,40]. In addition, the ThT fluorescence mea-
h sured after CE separation was validated by measuring bulk ThT
fluorescence for the same samples.
e Detection of individual aggregates by CE–LIF

a The detection system was designed as a two-channel (polariza-

b c d tion) fluorescence detector to evaluate the potential of fluores-
cence anisotropy to size characterize individual aggregates (see
Fig.1. Schematic of the fluorescence anisotropy detection system for CE. Compo- Materials and Methods for details); however, for the purpose of
nents: (a) polarized laser (440 nm); (b) neutral density filter; (c) focusing lens; (d)
capillary; (e) microscope objective lens (20); (f) notch filter (440 ± 10 nm); (g)
separating and detecting individual ThT-labeled Ab aggregates by
bandpass filter (490 ± 10 nm); (h) 1000-lm pinhole; (i) polarizing beamsplitter CE–LIF, two-channel fluorescence detection is not required.
cube; (j) PMTk; (k) PMT\. Data from one of the two fluorescence channels are plotted in
Analysis of Ab aggregates by CE–LIF / R.A. Picou et al. / Anal. Biochem. 425 (2012) 104–112 107

Figs. 3–6 for clarity. The two-channel data are discussed later (see
‘‘Fluorescence anisotropy of individual Ab aggregates’’ section).
Four different Ab sample types were prepared and used for this
work: Ab (1–40) monomer, Ab (1–40) fibril, Ab (1–42) monomer,
and Ab (1–42) fibril. In nature, Ab (1–42) and Ab (1–42) are pro-
duced by cleavage of the same parent peptide at different sites.
Both Ab (1–40) and Ab (1–42) are known to be present in amyloid
plaques, and Ab (1–42) has been shown to aggregate much more
rapidly in vitro and is difficult to prepare in a purely monomeric
form [1,19,27]. Both peptides will aggregate to form amyloid fibrils
that fluoresce on ThT binding and are similar in appearance in TEM
images [1,13,19,27]. The two fibril samples were prepared to con-
tain large mature fibrils that would bind to ThT, resulting in
enhanced fluorescence. The Ab (1–40) monomer and Ab (1–42)
monomer samples were prepared to be free of aggregates and
serve as negative controls. A representative TEM of each sample
type is shown in Fig. 2. The Ab fibril samples (Fig. 2B and D) show
clear evidence of fibrillar aggregates. The Ab monomer samples do
not show fibrillar aggregates by TEM (Fig. 2A and C); therefore,
these samples were not expected to bind to ThT and cause ThT Fig.3. Electropherograms with LIF detection for Ab (1–40) monomer without ThT
fluorescence enhancement [11,12,37]. (A) and with 15 lM ThT (B) in the electrophoresis buffer (10.00 mM Tris at pH 7.79).
Fig. 3A and B show electropherograms with LIF detection for the The peak at 59 s in each electropherogram is the neutral marker (NM), coumarin
Ab (1–40) monomer sample without ThT and with ThT in the run- 334. Coumarin 334 was injected for 1.0 s at 25.0 kV prior to the Ab sample injection
for 5.0 s at 25.0 kV. The separation potential was 25.0 kV (417 V/cm). The dip in the
ning buffer, respectively. As expected, no peaks were observed in
baseline at 47 s in panel B is due to a ThT vacancy zone.
the absence of ThT (Fig. 3A). In contrast, in the presence of ThT, a
few small sharp peaks were detected for Ab monomer (Fig. 3B). ThT fluorescence similar to amyloid fibrils. We are currently
The migration times of these peaks appear to be random for consec- investigating ThT fluorescence in the presence of synthetic nano-
utive injections, and approximately the same numbers of peaks particles due to their potential to produce false positives in amyloid
were observed in electropherograms of the electrophoresis buffer fibril assays. The Ab (1–40) monomer sample also was analyzed by a
with ThT for which no Ab was injected (data not shown). These CE–UV method previously developed by Picou and coworkers [26].
peaks are hypothesized to be a result of unfiltered particulate A representative CE–UV electropherogram of the Ab (1–40) mono-
matter (non-Ab particles) in the running buffer that can enhance mer sample is available in the Supplementary material (Fig. S1). This

Fig.2. TEM images of Ab (1–40) monomer (A), Ab (1–40) fibril (B), Ab (1–42) monomer (C), and Ab (1–42) fibril (D). Samples were buffer exchanged into 10.00 mM Tris buffer
at pH 7.79 prior to imaging (scale bar = 500 nm).
108 Analysis of Ab aggregates by CE–LIF / R.A. Picou et al. / Anal. Biochem. 425 (2012) 104–112

experiment confirmed that the Ab (1–40) monomer sample does

contain the Ab monomer, and it does not indicate that Ab aggregates
are present.
The fluorescent dye, coumarin 334, was used as a neutral mar-
ker to determine the electoosmotic flow and was detected as a
large, relatively broad peak at approximately 60 s in Fig. 3A and
B. The coumarin 334 and Ab samples were injected consecutively
(electrokinetic injection) from separate vials to minimize potential
interaction of the neutral marker and Ab. For these experimental
conditions, Ab has a negative electrophoretic mobility. The couma-
rin 334 and Ab sample buffers did not contain ThT. As a result,
when they were injected into a capillary filled with ThT, a vacancy
zone was created and caused the fluorescence to decrease, as ob-
served at 47 s in Fig. 3B. No ThT vacancy zone is present in
Fig. 3A because ThT was not included in the running buffer. The va-
cancy peak in Fig. 3B gives an indication of the magnitude of the
background fluorescence due to unbound ThT and is consistent
with reports of ThT background fluorescence in the literature
Fig. 4A and B show representative electropherograms with LIF
detection for the Ab (1–40) fibril sample without ThT and with
ThT in the running buffer, respectively. The Ab (1–40) fibril sample
was prepared by allowing Ab (1–40) monomer to aggregate for
7 days to produce mature Ab (1–40) fibrils. All samples (monomer
and fibril) were vortexed briefly just before injection to ensure that
larger aggregates were suspended throughout the sample solution.
Many sharp peaks are observed near 110 s in both Fig. 4A and B.
The data presented in Table 1 show that the electropherograms
for the Ab (1–40) fibril sample with ThT in the separation buffer
contain approximately 100 of these peaks compared with 10 peaks
for the Ab (1–40) monomer sample (Fig. 3B). The peaks in Fig. 4A
and B are hypothesized to be due to individual Ab (1–40) aggre-
gates passing through the LIF detection zone. The inset of Fig. 4B
is an expanded view showing two individual aggregate peaks in
the Ab (1–40) fibril sample, illustrating their narrow peak widths.
It is extremely challenging to prove that the sharp peaks in the
CE–LIF electropherograms are due to individual amyloid fibrils. The
fundamental limitation is the lack of an analytical standard or sur-
rogate standard for Ab fibrils. Our laboratory has extensive experi-
ence with single particle detection [39,41] and Ab fibril preparation
and analysis [6,26,27]. The peak widths observed in Fig. 4A and B
are consistent with peaks due to single polymer particles of com-
parable size observed with a similar CE detector [39]. The expected
full width at half maximum (FWHM) of an individual aggregate
passing through a focused laser beam can be estimated by dividing
the calculated diameter of the focused laser spot by the measured
Fig.4. (A and B) Electropherograms with LIF detection of Ab (1–40) fibril without
migration rate of the aggregate [39]. In this case, the calculated
ThT (A) and with 15 lM ThT (B) in the electrophoresis buffer. The experimental
diameter of the focused laser beam is 5.7 lm, and the migration conditions for panels A and B are the same as in Fig. 3. The inset in panel B is an
rate for species detected at 105.03 s is 0.22 cm/s, giving a theoret- expanded view (0.2 s) of Ab (1–40) aggregate peaks. The same sample analyzed by
ical FWHM for a single aggregate peak of 3 ms. In the inset of CE with UV absorbance (190 nm) is shown in panel C. In this panel, the NM was
Fig. 4B, the peak at 105.03 s has a measured FWHM of 11 ms, injected for 2.0 s at 0.3 psi, followed by the Ab injection for 5.0 s at 0.5 psi. The
separation potential in panel C was 25.0 kV (397 V/cm). The peaks at 200 and 243 s
and the widths of peaks attributed to individual aggregates are due to the neutral marker (NM) and Ab monomer (M), respectively.
throughout the electropherogram ranged from 3 to 15 ms. Experi-
mental peak widths are expected to be larger than the calculated
FWHM of 3 ms due to refraction of the excitation beam at the
curved capillary surfaces that causes the actual spot size to be lar-
Table 1
ger than the calculated spot size [39]. In contrast, the measured Numbers of CE peaks detected with and without ThT.
FWHM for coumarin 334 is 1.2 s, which is typical for a CE peak
due to large numbers of identical fluorescent molecules. Sample Peaks detected

The presence of large Ab aggregates in the Ab (1–40) fibril sam- with ThT without ThT
ple was confirmed by CE–UV, TEM, and bulk ThT fluorescence mea- Ab (1–40) monomer 10 ± 3 1±1
surements. In the CE–UV electropherogram (Fig. 4C), the broad Ab (1–40) fibril 99 ± 23 13 ± 4
peak at 313 s and the sharp peaks from 300 to 400 s are typical Ab (1–42) monomer 176 ± 27 5±2
Ab (1–42) fibril 300 ± 46 300 ± 30
for Ab samples containing mature fibrils [27]. Bulk ThT fluores-
cence was higher for the Ab (1–40) fibril sample compared with Note: Peaks detected are averages with standard deviations from three injections.
the Ab (1–40) monomer sample (Table 2), consistent with the All peaks with signal/noise ratios greater than 3 were included.
Analysis of Ab aggregates by CE–LIF / R.A. Picou et al. / Anal. Biochem. 425 (2012) 104–112 109

presence of large Ab aggregates in the fibril sample [11,12,37]. Fi-

nally, the TEM of this sample (Fig. 2B) shows that mature Ab fibrils
were present in this sample.
The peaks in Fig. 4A (Ab (1–40) fibril without ThT) are hypoth-
esized to be due to scattered light from individual aggregates and
are not unexpected. We have previously shown that polystyrene
spheres as small as 110 nm in diameter can be detected individu-
ally by light scattering after CE separation [39] and that Ab fibril
samples produce similar peaks in instruments designed to detect
scattered light with CE (unpublished results). The peaks in
Fig. 4A were detected despite the use of a 440-nm notch filter
and a 490-nm bandpass filter to minimize detection of scattered
light and isolate ThT fluorescence. The data presented in Fig. 4A
suggest that aggregates in the mature Ab (1–40) fibril sample re-
sulted in a non-negligible light scattering signal in the absence of
ThT that contributed to the apparent fluorescence signal in the
presence of ThT. Comparing Fig. 4A and B, it is clear that most of
the observed signal in Fig. 4B is due to enhanced ThT fluorescence.
The microplate fluorescence measurements of Ab fibrils (Table 2,
without ThT) also exhibit a higher signal than microplate measure-
ments of Ab monomer samples both with and without ThT added.
It is important to recognize that the migration times and elec-
trophoretic mobilities for the aggregate peaks in the three electro-
pherograms in Fig. 4 cannot be compared directly. The differences
between the migration times of the aggregate peaks in the CE–LIF
and CE–UV electropherograms are due mainly to the difference in
capillary length from the inlet to the detection window, LD, for the
two instruments (23.0 cm, Fig. 4A and B; 53.0 cm, Fig. 4C). This will
not affect electrophoretic mobilities of the Ab aggregate peaks, but
ThT added to the running buffer in Fig. 4B can potentially alter the
electrophoretic mobilities of Ab aggregates compared with Fig. 4A
and C. Finally, the laboratory-constructed CE–LIF instrument used
for Fig. 4A and B was not thermostatted, but the commercial CE
instrument used for Fig. 4C does control the capillary temperature.
Even though the separation buffers used for all separations were
identical, changes in capillary temperature will affect the electro-
phoretic mobilities of the Ab species.
Ab (1–42) peptide samples also were prepared and investigated
by CE–LIF. Preparing an Ab (1–42) monomer sample to be com-
pletely aggregate free is extremely challenging owing to the fast
aggregation kinetics of the Ab (1–42) peptide relative to that of
the Ab (1–40) peptide [19]. Fig. 5A and B show representative elec-
tropherograms of the Ab (1–42) monomer sample without ThT and
with ThT in the running buffer, respectively. Fig. 5A shows that,
other than the coumarin 334 peak at 58 s, few peaks (<10) are de-
tected in the Ab (1–42) monomer sample when no ThT is included
in the running buffer (Table 1). In contrast, Fig. 5B shows an elec- Fig.5. (A and B) Electropherograms with LIF detection of Ab (1–42) monomer
tropherogram of the same Ab (1–42) monomer sample injected analyzed without ThT (A) and with 15 lM ThT (B) in the electrophoresis buffer.
with 15 lM ThT in the electrophoresis buffer, and more than 150 Panel C shows the same sample analyzed by CE–UV at 190 nm. The peaks at 213
peaks were detected (Table 1). These peaks have FHWMs ranging and 259 s are due to neutral marker (NM) and Ab monomer (M), respectively. The
experimental conditions for panels A to C are the same as in Fig. 4.
from 3 to 15 ms, similar to those attributed to individual aggre-
gates for the Ab (1–40) fibril sample. Characterization of this sam- contains substantial amounts of unaggregated Ab (1–42) monomer
ple by CE–UV also indicated that the sample contained aggregates [27]. The presence of these apparent aggregates in samples metic-
based on the presence of the broad peak at 343 s (Fig. 5C) [26,27]. ulously prepared to be aggregate free highlights the relatively ra-
The large UV peak at 259 s also demonstrates that this sample pid kinetics of Ab (1–42) aggregation [27].
Table 2
Comparison of total signals measured by CE and bulk fluorescence (plate reader).

Sample Capillary electrophoresis Plate reader

with ThT (RFU) without ThT (RFU) with ThT (RFU) without ThT (RFU)
Ab (1–40) monomer 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0
Ab (1–40) fibril 3.8 ± 0.1 0.9 ± 0.5 8.8 ± 0.1 2.1 ± 0.1
Ab (1–42) monomer 5.9 ± 0.6 0.2 ± 0.0 3.2 ± 0.1 0.2 ± 0.0
Ab (1–42) fibril 100.0 ± 26.3 12.3 ± 3.0 100.0 ± 0.5 18.9 ± 0.1

Note: All data are presented as averages with standard deviations from three injections or wells. Data are normalized to the values for the Ab (1–42) fibril sample. The CE
values were calculated by summing the integration of the peaks detected in the electropherograms.
110 Analysis of Ab aggregates by CE–LIF / R.A. Picou et al. / Anal. Biochem. 425 (2012) 104–112

In the presence of ThT, a broad peak (FWHM = 9 s) at 98 s that Fig. 6A and B present electropherograms of Ab (1–42) fibrils
was separated from most of the peaks attributed to individual without ThT and with ThT in the running buffer, respectively.
aggregates (Fig. 5B) was obtained for the Ab (1–42) monomer sam- Fig. 6A shows that analysis of the Ab (1–42) fibrils resulted in many
ple. Analysis of the Ab (1–40) fibril sample by CE–LIF with ThT also sharp peaks in the absence of ThT that are thought to be due to scat-
resulted in a broad peak that comigrated with most of the peaks tered light from individual aggregates, similar to the scattering ob-
attributed to individual aggregates (Fig. 4B). These broad peaks served for the Ab (1–40) fibril sample (Fig. 4A). Table 1 shows that an
in Figs. 4 and 5B could be due to oligomeric structures, which are average of 300 peaks were detected for the Ab (1–42) fibril sample in
too small to produce large distinct peaks for individual aggregates the absence of ThT, and the peaks were much larger than those ob-
but are large enough to bind ThT and enhance its fluorescence. served for Ab (1–40) fibrils. This is consistent with the microplate
Biancalana and Koide discussed minimal b-sheet requirements measurements for the samples without ThT (Table 2). The large peak
for binding and fluorescence enhancement of ThT using a heights for the Ab (1–42) fibrils relative to the Ab (1–40) fibrils,
peptide self-assembly mimic [37], but it is not certain exactly what without ThT, suggest that the Ab (1–42) fibrils are larger.
Ab aggregates (size or shape) meet the minimum b-sheet Adding ThT to the running buffer resulted in an increase in the
requirements. intensity of the peaks detected for the Ab (1–42) fibril sample.
Ab (1–42) fibrils also were analyzed by CE–LIF (Fig. 6) and CE– Most of the peaks were off-scale initially, and a neutral density
UV. The CE–UV electropherograms (Fig. S2 in Supplementary mate- (ND) filter (10.7% transmission) was used to attenuate the emission
rial), bulk ThT fluorescence (Table 2), and TEM data (Fig. 2D) from for the electropherogram shown in Fig. 6B. The inset of 6B high-
Ab (1–42) fibrils clearly indicate that large aggregates were present lights a few of the Ab (1–42) aggregate peaks, showing that most
in the sample. Unlike the Ab (1–40) fibrils (Fig. 4), the CE–UV data peaks are well resolved from each other. The average values for to-
of the Ab (1–42) fibrils showed no signs of Ab monomer (Fig. S2). tal peaks presented in Table 1 are the same for the Ab (1–42) fibril
The absence of a monomer peak for Ab (1–42) fibrils is due to sample with and without ThT, but the peak intensities are much
the fast aggregation kinetics and the smaller monomer equilibrium higher with ThT in the running buffer. The total signal from the
concentration of the Ab (1–42) peptide relative to that of the Ab Ab (1–42) fibril sample increases by a factor of 8 when analyzed
(1–40) peptide [19]. in the presence of ThT (Table 2).

Validating CE–LIF results by bulk measurement

Bulk ThT fluorescence measurements are widely used for study-

ing amyloid aggregation [11,12,19,26,37,42]. Although the studies
presented earlier (see ‘‘Detection of individual aggregates by
CE–LIF’’ section) and by Kato and coworkers [23] show that ThT
can be used as a label for LIF detection of Ab aggregates with CE,
it is not known whether the fluorescence detected as peaks by
CE is equivalent to the increase in bulk ThT experiments commonly
used to indicate increased Ab aggregation and to study Ab aggrega-
tion kinetics. To validate the CE results, the total fluorescence
detected in electropherograms (summed area of all peaks) has
been compared with bulk fluorescence measurements obtained
with a microplate fluorometer using the same Ab samples.
The results of this validation are presented in Table 2. All micro-
plate fluorometer data were normalized to the fluorescence for the
Ab (1–42) sample with ThT, and all CE–LIF data were normalized to
the total fluorescence for the Ab (1–42) fibril sample with ThT for
CE experiments. The data presented in Table 2 demonstrate that
the results from the two methods are equivalent. Both methods
show almost no signal for Ab (1–40) monomer both with and with-
out ThT added. Relative to Ab (1–40) monomer, both techniques
indicate a significant increase in signal for the Ab (1–40) fibril sam-
ple without the addition of ThT, suggesting that signal due to scat-
tered light for CE experiments (Fig. 4A) is also detected by the
microplate reader. On binding to ThT, the fluorescence signal for
Ab (1–40) fibrils increased by a factor of 4.2 using both techniques.
Similarly, both Ab (1–42) monomer and Ab (1–42) fibrils show
similar signal intensities using both techniques in the presence
and absence of ThT.
Overall, the data in Table 2 show that the CE–LIF data are con-
sistent with standard bulk fluorescence measurements widely
used for fundamental studies of Ab aggregation. The advantage of
the CE–LIF technique is that it separates aggregates, providing a
profile of mobility and intensity for the Ab aggregate population.

Fig.6. Electropherograms with LIF detection of Ab (1–42) fibrils analyzed without Fluorescence anisotropy of individual Ab aggregates
ThT (A) and with ThT (B) in the electrophoresis buffer. A neutral density filter was
used in panel B to attenuate the emission (10.7% transmission), and the y axis in
panel B was adjusted to account for this. The inset in panel B shows a 0.2-s segment
The results presented in the above two sections of Results and
of the electropherogram displaying several of the Ab (1–42) aggregate peaks. The discussion demonstrate that this CE–LIF method can separate Ab
experimental conditions for panels A and B are the same as in Fig. 3. aggregates, providing information about their numbers, electro-
Analysis of Ab aggregates by CE–LIF / R.A. Picou et al. / Anal. Biochem. 425 (2012) 104–112 111

phoretic mobilities, and fluorescence intensity with ThT labeling.

Although this approach provides a wealth of information about
the Ab aggregate population compared with bulk measurements
of ThT fluorescence, it does not allow one to determine the size
and shape of an individual aggregate based on CE–LIF peaks. For
particles of the size scale of Ab fibrils, electrophoretic mobility is
not a simple function of aggregate size and shape [39,43]. Stan-
dards or suitable surrogate standards representative of amyloid fi-
brils to determine these relationships do not exist. The relationship
between aggregate size, shape, and structure and ThT fluorescence
also is not well understood. This section investigates the potential
of characterizing individual aggregates by fluorescence anisotropy
with ThT fluorescence.
Fluorescence anisotropy is a technique capable of providing
size-related information about fluorescent analytes. Eq. (2) shows
how fluorescence anisotropy is related to the rotational diffusion
coefficient, D, through the Perrin equation, where r0 is the funda-
mental anisotropy and s is the fluorescence lifetime. The funda-
mental anisotropy is dependent on the angle between the
absorption and emission transition dipoles of the fluorophore
[44]. It was hypothesized that the differences in size and structure
of Ab aggregates detected by CE would result in differences in fluo-
rescence anisotropy, according to Eq. (2):
r¼ : ð2Þ
1 þ 6Ds
For this work, the fluorescence anisotropy detector used two
PMTs that simultaneously collected perpendicular and parallel
emission relative to the polarization of the excitation source
(Fig. 1). A pair of data points was collected at each time point,
and the emission intensities were used as described in Eq. (1) to
calculate the fluorescence anisotropy, r, as described in Materials
and methods.
Fig.7. Electropherograms of Ab (1–42) monomer for LIFA detection. Panels A to C
Fig. 7A–C show three representative 100-ms time segments of are three 0.1-s segments of the separation shown in Fig. 5B. The parallel and
the parallel and perpendicular fluorescence signals from the electro- perpendicular fluorescence emissions are shown as solid and dashed lines,
pherograms of the Ab (1–42) monomer sample. This sample was respectively. Panels D to F show fluorescence anisotropy electropherograms
corresponding to panels A to C, respectively. The anisotropy data in panels D to F
chosen to test the potential of fluorescence anisotropy to character-
were calculated using Eq. (1).
ize Ab aggregates because it produced on-scale peaks in the presence
of ThT and no significant signal in the absence of ThT (i.e., minimal cence anisotropy versus time for an Ab aggregate sample would re-
scattering signals). These are important attributes for this assess- sult in distinguishable fluorescence anisotropy peaks and that the
ment because peak intensities, which are used in the calculation of intensity of these peaks might provide information about the rela-
fluorescence anisotropy (see Eq. (1)), cannot be determined from tive sizes of the aggregates. Protein–protein affinity interaction
off-scale peaks. Moreover, it is known that scattered light from a studies by CE–LIFA included plots of fluorescence anisotropy ver-
polarized source is highly polarized and could have a significant im- sus time with distinct peaks [31]. Sabate and Saupe reported an in-
pact on the fluorescence anisotropy results [44]. crease in ThT fluorescence anisotropy due to aggregation of a
In Fig. 7A–C, the parallel channel data are plotted as solid lines fungal prion protein, HET-s(218–289) [33]. A study by Khurana
and the perpendicular channel data are plotted as dashed lines. and coworkers reported anisotropy values near 0.3 for ThT without
Note that the parallel channel data in Fig. 7A–C are from the same any protein aggregates present in the solution [38]. Although ThT
separation shown in Fig. 5B. It is clear from Fig. 7–C that the signal clearly produced enhanced fluorescence for Ab aggregates in this
detected in the parallel channel is higher than that detected in the study, fluorescence anisotropy did not appear to increase signifi-
perpendicular channel, indicating fluorescence anisotropy. Couma- cantly relative to background levels. To effectively use CE with
rin 334, which was used as a neutral marker to measure the elec- fluorescence anisotropy detection to characterize individual
troosmotic flow for these studies, was also used as a control to aggregates, the fluorescence anisotropy of Ab aggregates must be
assess the performance of the fluorescence anisotropy detection distinct from that of the background anisotropy.
system. The fluorescence anisotropy of coumarin 334 is expected
to be near zero because it has a low molecular weight (283.32 g/
mol) [44]. The fluorescence anisotropy of coumarin 334 that was Conclusions
injected with the Ab (1–42) monomer sample in Fig. 7 was calcu-
lated to be 0.007 (coumarin 334 data not shown). This verifies that This work shows that CE with LIF detection can be used to ana-
the detection system was performing properly. lyze samples of aggregated Ab (1–40) and Ab (1–42) using ThT, a
Fig. 7D–F show fluorescence anisotropy versus time calculated dye that fluoresces on binding to the amyloid fibrils, in the running
from the LIF data in Fig. 7A–C. Fig. 7D–F show that individual buffer. This study distinguishes itself from a previous study [23] by
aggregate peaks have non-zero anisotropies, but these anisotropy demonstrating the ability to separate and detect individual aggre-
values are indistinguishable from the anisotropy calculated from gates and by validating the CE–LIF data using direct comparison
the background signal (i.e., no fluorescence anisotropy peaks were with bulk ThT fluorescence measurements, which are widely used
observed in Fig. 7D–F). It was hypothesized that plots of fluores- to study aggregation kinetics of amyloidogenic peptides. Compared
112 Analysis of Ab aggregates by CE–LIF / R.A. Picou et al. / Anal. Biochem. 425 (2012) 104–112

with bulk ThT fluorescence measurements, CE–LIF provides a pro- [18] D.M. Hartley, D.M. Walsh, C.P. Ye, T. Diehl, S. Vasquez, P.M. Vassilev, D.B.
Teplow, D.J. Selkoe, Protofibrillar intermediates of amyloid b-protein induce
file of the aggregate population based on electrophoretic mobili-
acute electrophysiological changes and progressive neurotoxicity in cortical
ties, fluorescence intensities, and numbers of individual neurons, J. Neurosci. 19 (1999) 8876–8884.
aggregate peaks. The studies presented here focused on analysis [19] A. Jan, O. Gokce, R. Luthi-Carter, H.A. Lashuel, The ratio of monomeric to
of Ab aggregates, but they should be applicable to studies of other aggregated forms of Ab40 and Ab42 is an important determinant of amyloid-b
aggregation, fibrillogenesis, and toxicity, J. Biol. Chem. 283 (2008) 28176–
amyloidogenic peptides. 28189.
Initial studies of CE–LIFA to characterize individual Ab aggre- [20] P.J. Sweeney, J.G. Darker, W.A. Neville, J. Humphries, P. Camilleri,
gates yielded encouraging results, but significant improvements Electrophoretic techniques for the analysis of synthetic amyloid b-A4-related
peptides, Anal. Biochem. 212 (1993) 179–184.
must be made to apply this effectively to obtain relative size and [21] G. Thorsen, J. Bergquist, A. Westlind-Danielsson, B. Josefsson, Stereoselective
shape information for Ab aggregates. determination of amino acids in b-amyloid peptides and senile plaques, Anal.
Chem. 73 (2001) 2625–2631.
Acknowledgments [22] E. Varesio, S. Rudaz, K.-H. Krause, J.-L. Veuthey, Nanoscale liquid
chromatography and capillary electrophoresis coupled to electrospray mass
spectrometry for the detection of amyloid-b peptide related to Alzheimer’s
This work was supported by Louisiana State University, the Na- disease, J. Chromatogr. A 974 (2002) 135–142.
tional Science Foundation (CHE-0505972, S.D.G.), the Rosalinde [23] M. Kato, H. Kinoshita, M. Enokita, Y. Hori, T. Hashimoto, T. Iwatsubo, T.
Toyo’oka, Analytical method for b-amyloid fibrils using CE-laser induced
and Arthur Gilbert Foundation/AFAR Award (I.K.), and the Louisi-
fluorescence and its application to screening for inhibitors of b-amyloid
ana Board of Regents (I.K.). Ryan Picou was supported by a Louisi- protein aggregation, Anal. Chem. 79 (2007) 4887–4891.
ana Board of Regents Fellowship. Rebekah Cerqua was supported [24] A. Kishita, S. Nishino, T. Togashi, Y. Nishida, Formation of methionine sulfoxide
of amyloid b-peptide (1-40) by Cu(bdpe)/H2O2 system, Synth. React. Inorg.
by National Science Foundation Research Experiences for Under-
Met.–Org. Nano-Met. Chem. 35 (2005) 677–681.
graduates (REU, CHE-0648841). The authors thank the Louisiana [25] T. Frackowiak, T. Baczek, R. Kaliszan, B. Zbikowska, M. Glensk, I. Fecka, W.
State University School of Veterinary Medicine Microscopy Center Cisowski, Binding of an oxindole alkaloid from Uncaria tomentosa to amyloid
in the Department of Comparative Biomedical Sciences for TEM protein (Ab1–40), Z. Naturforsch. C 61 (2006) 821–826.
[26] R. Picou, J.P. Moses, A.D. Wellman, I. Kheterpal, S.D. Gilman, Analysis of
work. monomeric Ab (1–40) peptide by capillary electrophoresis, Analyst 135 (2010)
[27] R.A. Picou, I. Kheterpal, A.D. Wellman, M. Minnamreddy, G. Ku, S.D. Gilman,
Appendix A. Supplementary data Analysis of Ab (1–40) and Ab (1–42) monomer and fibrils by capillary
electrophoresis, J. Chromatogr. B 879 (2011) 627–632.
Supplementary data associated with this article can be found, in [28] D.K. Clodfelter, M.A. Nussbaum, J. Reilly, Comparison of free solution capillary
electrophoresis and size exclusion chromatography for quantitating non-
the online version, at http://dx.doi.org/10.1016/j.ab.2012.03.006.
covalent aggregation of an acylated peptide, J. Pharm. Biomed. Anal. 19 (1999)
References [29] Q.-H. Wan, X.C. Le, Studies of protein–DNA interactions by capillary
electrophoresis/laser-induced fluorescence polarization, Anal. Chem. 72
[1] D.M. Walsh, D.J. Selkoe, Deciphering the molecular basis of memory failure in (2000) 5583–5589.
Alzheimer’s disease, Neuron 44 (2004) 181–193. [30] R.J. Whelan, R.K. Sunahara, R.R. Neubig, R.T. Kennedy, Affinity assays using
[2] H. Levine III, Small molecule inhibitors of Ab assembly, Amyloid 14 (2007) fluorescence anisotropy with capillary electrophoresis separation, Anal. Chem.
185–197. 76 (2004) 7380–7386.
[3] T. Takahashi, H. Mihara, Peptide and protein mimetics inhibiting amyloid b- [31] P. Yang, R.J. Whelan, Y. Mao, A.W.-M. Lee, C. Carter-Su, R.T. Kennedy,
peptide aggregation, Acc. Chem. Res. 41 (2008) 1309–1318. Multiplexed detection of protein–peptide interaction and inhibition using
[4] T. Ban, Y. Goto, Direct observation of amyloid growth monitored by total capillary electrophoresis, Anal. Chem. 79 (2007) 1690–1695.
internal reflection fluorescence microscopy, Methods Enzymol. 413 (2006) 91– [32] L. Ye, X.C. Le, J.Z. Xing, M. Ma, R. Yatscoff, Competitive immunoassay for
102. cyclosporine using capillary electrophoresis with laser induced fluorescence
[5] C. Ha, C.B. Park, Ex situ atomic force microscopy analysis of b-amyloid self- polarization detection, J. Chromatogr. B 714 (1998) 59–67.
assembly and deposition on a synthetic template, Langmuir 22 (2006) 6977– [33] R. Sabaté, S.J. Saupe, Thioflavin T fluorescence anisotropy: an alternative
6985. technique for the study of amyloid aggregation, Biochem. Biophys. Res.
[6] I. Kheterpal, K.D. Cook, R. Wetzel, Hydrogen/deuterium exchange mass Commun. 360 (2007) 135–138.
spectrometry analysis of protein aggregates, Methods Enzymol. 413 (2006) [34] D. Allsop, L. Swanson, S. Moore, Y. Davies, A. York, O.M.A. El-Agnaf, I. Soutar,
140–166. Fluorescence anisotropy: a method for early detection of Alzheimer b-peptide
[7] A. Lomakin, D.B. Teplow, Quasielastic light scattering study of amyloid b- (Ab) aggregation, Biochem. Biophys. Res. Commun. 285 (2001)
protein fibril formation, Protein Pept. Lett. 13 (2006) 247–254. 58–63.
[8] J. Ryu, H.-A. Joung, M.-G. Kim, C.B. Park, Surface plasmon resonance analysis of [35] B. O’Nuallain, A.K. Thakur, A.D. Williams, A.M. Bhattacharyya, S. Chen, G.
Alzheimer’s b-amyloid aggregation on a solid surface. From monomers to Thiagarajan, R. Wetzel, Kinetics and thermodynamics of amyloid assembly
fully-grown fibrils, Anal. Chem. 80 (2008) 2400–2407. using a high-performance liquid chromatography-based sedimentation assay,
[9] S.A. Funke, E. Birkmann, F. Henke, P. Gortz, C. Lange-Asschenfeldt, D. Riesner, D. Methods Enzymol. 413 (2006) 34–74.
Willbold, Single particle detection of Ab aggregates associated with [36] K.F. Schrum, J.M. Lancaster III, S.E. Johnston, S.D. Gilman, Monitoring
Alzheimer’s disease, Biochem. Biophys. Res. Commun. 364 (2007) 902–907. electroosmotic flow by periodic photobleaching of a dilute, neutral
[10] R. Tycko, Characterization of amyloid structures at the molecular level by solid fluorophore, Anal. Chem. 72 (2000) 4317–4321.
state nuclear magnetic resonance spectroscopy, Methods Enzymol. 413 (2006) [37] M. Biancalana, S. Koide, Molecular mechanism of thioflavin-T binding to
103–122. amyloid fibrils, Biochim. Biophys. Acta 2010 (1804) 1405–1412.
[11] H. Levine III, Multiple ligand binding sites on Ab (1–40) fibrils, Amyloid 12 [38] R. Khurana, C. Coleman, C. Ionescu-Zanetti, S.A. Carter, V. Krishna, R.K. Grover,
(2005) 5–14. R. Roy, S. Singh, Mechanism of thioflavin T binding to amyloid fibrils, J. Struct.
[12] H. Naiki, F. Gejyo, Kinetic analysis of amyloid fibril formation, Methods Biol. 151 (2005) 229–238.
Enzymol. 309 (1999) 305–318. [39] Y.H. Rezenom, A.D. Wellman, L. Tilstra, C.D. Medley, S.D. Gilman, Separation
[13] S. Sabella, M. Quaglia, C. Lanni, M. Racchi, S. Govoni, G. Caccialanza, A. and detection of individual submicron particles by capillary electrophoresis
Calligaro, V. Bellotti, E. De Lorenzi, Capillary electrophoresis studies on the with laser-light-scattering detection, Analyst 132 (2007) 1215–1222.
aggregation process of b-amyloid 1–42 and 1–40 peptides, Electrophoresis 25 [40] H. Ahmadzadeh, R. Dua, A.D. Presley, E.A. Arriaga, Automated analysis of
(2004) 3186–3194. individual particles using a commercial capillary electrophoresis system, J.
[14] Y.-M. Kuo, M.R. Emmerling, C. Vigo-Pelfrey, T.C. Kasunic, J.B. Kirkpatrick, G.H. Chromatogr. A 1064 (2005) 107–114.
Murdoch, M.J. Ball, A.E. Roher, Water-soluble Ab (N-40, N-42) oligomers in [41] S.D. Gilman, A.G. Ewing, Analysis of single cells by capillary electrophoresis
normal and Alzheimer disease brains, J. Biol. Chem. 271 (1996) with on-column derivatization and laser-induced fluorescence detection, Anal.
4077–4081. Chem. 67 (1995) 58–64.
[15] Y.-F. Mok, J. Howlett Geoffrey, Sedimentation velocity analysis of amyloid [42] B. O’Nuallain, S. Shivaprasad, I. Kheterpal, R. Wetzel, Thermodynamics of Ab
oligomers and fibrils, Methods Enzymol. 413 (2006) 199–217. (1–40) amyloid fibril elongation, Biochemistry 44 (2005) 12709–12718.
[16] H.-W. Klafki, J. Wiltfang, M. Staufenbiel, Electrophoretic separation of bA4 [43] S.P. Radko, A. Chrambach, Separation and characterization of sub-lm- and
peptides (1–40) and (1–42), Anal. Biochem. 237 (1996) 24–29. lm-sized particles by capillary zone electrophoresis, Electrophoresis 23
[17] G. Bitan, Structural study of metastable amyloidogenic protein oligomers by (2002) 1957–1972.
photo-induced cross-linking of unmodified proteins, Methods Enzymol. 413 [44] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, Kluwer Academic/
(2006) 217–236. Plenum, New York, 1999.