C-terminal peptides coassemble into A?42
oligomers and protect neurons against
Erica A. Fradinger*†, Bernhard H. Monien*‡, Brigita Urbanc§¶, Aleksey Lomakin?, Miao Tan**, Huiyuan Li*,
Sean M. Spring*, Margaret M. Condron*, Luis Cruz§¶, Cui-Wei Xie**††, George B. Benedek?‡‡§§, and Gal Bitan*††§§¶¶
Departments of *Neurology and **Psychiatry and Biobehavioral Sciences, David Geffen School of Medicine,††Brain Research Institute,
and¶¶Molecular Biology Institute, University of California, Los Angeles, CA 90095;§Center for Polymer Studies, Department of Physics,
Boston University, Boston, MA 02215; and?Center for Material Science and Engineering, Material Processing Center,
and‡‡Department of Physics, Massachusetts Institute of Technology, Cambridge, MA 02139
Contributed by George B. Benedek, July 23, 2008 (sent for review June 18, 2008)
Alzheimer’s disease (AD) is an age-related disorder that threatens
to become an epidemic as the world population ages. Neurotoxic
oligomers of A?42 are believed to be the main cause of AD;
therefore, disruption of A? oligomerization is a promising ap-
proach for developing therapeutics for AD. Formation of A?42
oligomers is mediated by intermolecular interactions in which the
C terminus plays a central role. We hypothesized that peptides
derived from the C terminus of A?42 may get incorporated into
toxicity. We tested this hypothesis using A? fragments with the
general formula A?(x?42) (x ? 28–39). A cell viability screen
identified A?(31–42) as the most potent inhibitor. In addition, the
shortest peptide, A?(39–42), also had high activity. Both A?(31–
42) and A?(39–42) inhibited A?-induced cell death and rescued
disruption of synaptic activity by A?42 oligomers at micromolar
concentrations. Biophysical characterization indicated that the
action of these peptides likely involved stabilization of A?42 in
by which the fragments coassembled with A?42 to form heteroo-
ligomers. Thus, A?(31–42) and A?(39–42) are leads for obtaining
mechanism-based drugs for treatment of AD using a systematic
Alzheimer’s disease ? amyloid ?-protein ? inhibitor design
people. It is estimated that there are currently ?27 million
people suffering from AD worldwide (1). Because the world
population is aging rapidly, if no cure is found in the near future
AD will become an epidemic (2).
The amyloid cascade hypothesis proposed that amyloid ?-pro-
tein (A?) fibrils—an aggregated form of A? found in amyloid
plaques in the brains of patients with AD—were the neurotoxic
agents causing AD (3). However, in recent years, multiple lines
of evidence have led to a revision of this view, and today the
primary toxins causing AD are believed to be early-forming A?
oligomers rather than A? fibrils (4, 5). This paradigm shift
suggests that efforts toward development of therapeutic agents
targeting A? assembly should be directed at A? oligomers rather
than fibrils. In particular, genetic, physiologic, and biochemical
data indicate that oligomers of the 42-aa form of A?, A?42, are
most strongly linked to the etiology of AD (6–9) and therefore
are a particularly attractive target for inhibitor design.
Several groups have reported small-molecule inhibitors of A?
oligomerization (10–13). The importance of understanding the
mechanism of inhibition recently has been highlighted (14) after
findings that many small-molecule inhibitors of fibrillogenesis
may act nonspecifically, likely making them unsuitable for
treating amyloid-related disorders (15). In addition, inhibition of
fibril formation may actually lead to stabilization of toxic
lzheimer’s disease (AD) is the predominant cause of de-
oligomers (16). Interestingly, when oligomers are stabilized by
interaction with inhibitors or modulators, the toxicity of the
resulting oligomers depends on the stabilizing molecule. For
example, certain inositol derivatives, which were reported to
inhibit A?-induced toxicity (17), presumably stabilize nontoxic
A? oligomers (18). Nonetheless, to date, A? oligomerization
inhibitors have been found empirically with limited mechanistic
understanding of how they work, and currently mechanism-
based inhibitor design targeting A? oligomerization is lacking.
A substantial body of work suggests that the C terminus of
A?42 is a key region controlling A?42 oligomerization. Several
studies of prefibrillar A? have suggested that the C terminus of
A?42 is more rigid than the C terminus of the more abundant
and less toxic A?40 (19–22). The increased rigidity has been
attributed to interactions involving the C-terminal residues
I41–A42, which stabilize a putative turn conformation (23). The
higher conformational stability in the C terminus of A?42
correlates with formation oligomer populations distinct from
those of A?40 (8, 23, 24) and with higher neurotoxicity (7, 9).
Based on these data we hypothesized that molecules that possess
high affinity for the C terminus of A?42 may disrupt oligomer
formation and inhibit A?42-induced neurotoxicity. Because
homotypic intermolecular interactions in the C terminus appear
to be particularly important for A?42 self-assembly, we reasoned
that peptides derived from this region might act as such inhib-
itors [supporting information (SI) Fig. S1]. We therefore pre-
pared a series of A?42 C-terminal fragments (CTFs) (Table 1)
and tested their capability of inhibiting A?42 toxicity and
Solubility of CTFs. Being highly hydrophobic peptides, the CTFs
were expected to be poorly soluble and to aggregate in aqueous
solutions. To assess CTF solubility, peptide solutions were
Author contributions: E.A.F. and B.H.M. contributed equally to this work; E.A.F., B.H.M.,
B.U., and G.B. designed research; E.A.F., B.H.M., B.U., A.L., M.T., H.L., S.M.S., M.M.C., and
G.B. performed research; L.C. contributed new reagents/analytic tools; E.A.F., B.H.M., B.U.,
A.L., M.T., H.L., S.M.S., C.-W.X., G.B.B., and G.B. analyzed data; and E.A.F., B.H.M., B.U.,
G.B.B., and G.B. wrote the paper.
The authors declare no conflict of interest.
Whittier, CA 90608.
‡Present address: Deutsches Institut fu ¨r Erna ¨hrungsforschung Potsdam-Rehbru ¨cke, Abt.
Erna ¨hrungstoxikologie, Arthur-Scheunert-Allee 114–116, 14558 Nuthetal, Germany.
¶Present address: Department of Physics, Drexel University, Philadelphia, PA 19104.
whom correspondencemaybeaddressed.E-mail:firstname.lastname@example.org or
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2008 by The National Academy of Sciences of the USA
September 16, 2008 ?
vol. 105 ?
no. 37 ?
prepared by initial dissolution in dilute NaOH (25), followed by
dilution in phosphate buffer at physiologic pH and filtration
through 20-nm cutoff filters. The concentration of each sample
was then measured by amino acid analysis (AAA) (Table 1).
CTFs up to 10 aa long could be dissolved at concentrations
between 100 and 200 ?M. Longer peptides had low solubility,
but, except for A?(28–42), the solubility was sufficient for
evaluation of neurotoxicity inhibition. Measurement of particle
size by dynamic light scattering (DLS), ?-sheet content by CD
spectroscopy, and peptide morphology by EM indicated that,
upon incubation in aqueous buffer at pH 7.4, CTFs longer than
5 aa aggregated at rates that ranged from a few hours to a few
days depending on peptide length and sequence (data not
shown). Direct comparison of aggregation rates was difficult
because of the different solubility of the peptides.
Evaluation of CTF Toxicity. As peptides derived from A?42, the
CTFs may have been neurotoxic themselves. To test for self-
toxicity, CTFs were solubilized initially in DMSO and then
for 5 min at 16,000 ? g to remove preformed large aggregates.
The supernatant was added to differentiated PC-12 cells at the
desired concentration. All of the solutions were clear to the eye
when added to the cells, and the media remained clear at the end
of the incubation period. Most of the CTFs showed no toxicity
to neuronal cells up to the highest concentration used as assessed
by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro-
mide (MTT) cell-metabolism assay (26) (Fig. 1A), suggesting
that they could be tested for inhibition of A?42-induced toxicity.
An exception was A?(28–42), which was highly toxic (Fig. 1A),
possibly because of the presence of K at the N terminus, which
increases the positive charge at physiologic pH relative to the
Screening of CTFs for Inhibitory Activity. To evaluate the CTFs for
inhibition of A?42-induced neurotoxicity, A?42 was dissolved in
DMSO and diluted into cell culture medium. CTFs then were
dissolved in DMSO and mixed with A?42 at an A?42:CTF
concentration ratio of 1:10, respectively. The solution was cen-
trifuged for 5 min at 16,000 ? g to remove preformed aggregates
and then added to differentiated PC-12 cells and incubated for
15 h. Cell viability was assessed by using the MTT assay.
All 12 CTFs were found to protect the cells to some degree
from A?42-induced toxicity (Fig. 1B). Among them, A?(31–42)
showed the highest inhibitory activity, fully rescuing the cells
from A?42-induced toxicity. A?(39–42), the shortest CTFs used
(only four amino acid residues), also showed high inhibitory
activity (Fig. 1B). We therefore focus further discussion on these
two peptides. Although A?(30–42) showed an activity level
similar to that of A?(39–42), it was a less interesting peptide to
study because it is structurally similar to, but less active than,
Further Evaluation of A?(31–42) and A?(39–42) as Inhibitors of
A?-Induced Neurotoxicity. To study the effectiveness of A?(31–
42) and A?(39–42) as inhibitors of A?42-induced toxicity, dose
dependence curves were generated. A?(31–42) and A?(39–42)
yielded IC50values of 14 ? 2 and 16 ? 5 ?M in the MTT assay
(Fig. S2A). The MTT assay measures cell metabolism rather
than cell viability per se; however, because of the relatively short
period required for this assay, it is a standard assay for investi-
gations of A? toxicity (26, 27). In addition, A?(31–42) and
A?(39–42) yielded IC50 values of 20 ? 4 and 47 ? 14 ?M,
respectively, in the lactate dehydrogenase (LDH) release assay
(Fig. S2B), a direct measurement of cell death (28).
Synaptic failure has been postulated to be the primary event
leading to the development of AD (5, 29). A decrease in the
frequency of spontaneous miniature excitatory postsynaptic
excitatory synapses or a reduction in presynaptic release prob-
ability. A? has been shown to inhibit synaptic function and
decrease mEPSC frequency (30, 31). Here we used A?-induced
attenuation of mEPSC frequency in primary mouse hippocam-
pal neurons to evaluate the ability of A?(31–42) and A?(39–42)
to rescue A?42-mediated synaptic toxicity.
A?42 and CTF mixtures were prepared in a manner similar to
that used for cell viability assays, except that perfusion buffer
(vehicle) was used instead of cell culture medium. After estab-
Table 1. Sequence and solubility of CTFs used in this study
CTFSequence Solubility, ?M
22 ? 9
11 ? 3
62 ? 18
52 ? 24
134 ? 37
132 ? 29
149 ? 33
134 ? 20
143 ? 27
156 ? 33
141 ? 30
The solubility values are average concentrations (?SE) measured by AAA
for filtered solutions of each CTF in four to seven independent experiments.
concentrations of 0.1–20 ?M or mixtures of A?42:CTF at a 1:10 concentration
ratio, respectively, were incubated with differentiated PC-12 cells. In A, A?42
(black squares) is shown for comparison. In B, the nominal concentration of
MTT assay. Cell culture medium containing DMSO in the same concentrations
as used for peptide solubilization was used as a negative control, and 1 ?M
SD from at least three independent experiments, each performed with six
wells per condition.
Evaluation of CTF effect on neuronal cultures. CTFs at final nominal
www.pnas.org?cgi?doi?10.1073?pnas.0807163105Fradinger et al.
lishing a stable baseline recording for 5 min, cells were perfused
with vehicle, A?42, or A?42:CTF mixtures at either 1:10 (both
CTFs) or 1:1 [A?(31–42) only] concentration ratios, respec-
tively, for 20 min, and washed for 15 min after perfusion. At 3
?M, A?42 was found to induce robust inhibition of mEPSCs,
reducing spike frequency by 60–70% relative to baseline levels
within 20 min (Fig. 2). This effect persisted after a 15-min
washing period. Significant inhibition of the toxic effect of A?42
was observed at a 1:1 A?42:A?(31–42) concentration ratio, and
at 10-fold excess A?(31–42) rescued mEPSC deficits completely
viability, but also protected synaptic function from toxic insults
by A?42 oligomers. A?(39–42) showed a somewhat lower, yet
significant (P ? 0.05), inhibitory effect at 10-fold excess relative
to A?42 and was not studied at lower concentration ratios (Fig.
or A?42:CTF mixtures relative to vehicle were not significant.
CTF Effect on A?42 Assembly. To gain insight into the mechanism
by which CTFs inhibit A?42-induced toxicity we studied the
interaction between the CTFs and A?42 during assembly using
DLS, photo-induced cross-linking of unmodified proteins (PI-
CUP), and discrete molecular dynamics (DMD), methods that
have been useful for study of A? assembly (32–34).
For DLS experiments, mixtures of A?42:CTF at 30 ?M
nominal concentration each were prepared in 10 mM sodium
phosphate (pH 7.4) and compared with A?42 alone. The actual
concentration was determined post facto for each experiment by
AAA. In the absence of CTFs, A?42 comprised predominantly
designated as population 1 (P1, Fig. 3A, white bars). A minor
population of larger particles, P2, with RH2? 20–60 nm was
observed in some, but not all, measurements. Both A?(31–42)
of P2 particles (Fig. 3A Top and Middle, gray bars). In addition,
A?(39–42) caused compaction of P1 particles to RH1of ?4–9
absence of CTFs, A?42 formed particles of RH? 500–1,000 nm
(Fig. 3A Bottom, gray bars). Over a similar time period, slow
growth of P2 up to an RH2 of ?300 nm was observed in
A?42:CTF mixtures. Quantitative analysis showed that the
growth rate of P2 particles, dRH2/dt, was decreased substantially
in the presence of both CTFs relative to A?42 alone (Fig. 3B).
It is important to note that, even though CTFs increase the
abundance of P2 oligomers, the fraction of these oligomers is
overrepresented in the DLS experiments because scattering
from large particles is magnified proportionally to the square of
their mass. Thus, P2 assemblies account for no more than a few
percent of the total A? population.
In control DLS experiments using CTFs in the absence of
full-length A?42 we observed a behavior different from the one
particle size of ?100 nm after several days. Distinct oligomer
populations similar to P1 or P2 were not observed. A?(39–42)
showed no aggregation at concentrations up to 140 ?M.
As a complementary method for evaluating aggregation in
A?42:CTF mixtures, we measured the average frequency of
intensity spikes that occur when large particles, presumably
fibrils, cross the DLS instrument’s laser beam during the first 3
days of incubation (Fig. 3C). Both A?(31–42) and A?(39–42)
showed substantial inhibition of fibril growth relative to A?42.
Next we investigated the effect of CTFs on formation of small
oligomers using PICUP. When low-molecular-weight (LMW)
A?42 (35) is subjected to PICUP and analyzed by SDS/PAGE,
the most abundant oligomers observed are pentamers and
hexamers, which self-assemble to form larger oligomers and
therefore have been termed paranuclei (8). Paranucleus forma-
tion requires no incubation—these oligomers are observed im-
mediately after dissolution and cross-linking of A?42. To study
the effect of CTFs on these early-forming A? oligomers, ?30
?M LMW A?42 was mixed with CTFs and cross-linked imme-
diately. Importantly, the CTFs contain only residues that have
little or no reactivity in PICUP chemistry (33). Therefore,
cross-linking of CTFs to A?42 or to themselves was not ob-
served, facilitating unhindered analysis of A?42 oligomer size
distributions. A?(31–42) was found to cause a dose-dependent
decrease in the formation of A?42 paranuclei at concentrations
between ?3 and 35 ?M (Fig. S3) whereas A?(39–42) did not
show this effect at a concentration as high as 155 ?M (Fig. S3B).
These data suggest that as the PICUP-inert A?(31–42) mole-
cules coassemble with A?42, they spatially separate and ‘‘dilute’’
the A?42 monomers, preventing cross-linking. A?(39–42) might
have induced a similar effect at a higher concentration if such
high solubility could have been achieved. Alternatively, A?(39–
42) may interfere with A?42-induced toxicity by a distinct
mechanism that does not affect cross-linking.
Because of their noncrystalline and metastable nature, A?
oligomers are not amenable to structural investigation using
high-resolution experimental techniques, such as x-ray crystal-
lography or solution-state NMR. To study the interactions
between A?42 and CTFs during oligomerization in high reso-
lution, we used computer simulations that combine DMD and a
simplified protein structure. This approach, unlike traditional
molecular dynamics using all-atom models, enables modeling of
large molecular ensembles within relatively short times (34).
Previously, this modeling strategy was used to study the oli-
8), 1:1 A?42:A?(31–42) (n ? 9), 1:10 A?42:A?(31–42) (n ? 10), or 1:10
were measured. (A) Representative recording traces collected before (0 min)
perfused with vehicle for 5 min to establish baseline, and then with peptide
solutions for an additional 20 min, and allowed to recover in vehicle solution
for 15 min. The curves show the time dependence of mEPSC frequency after
exposure to A?42 in the absence or presence of CTFs over 40 min.
CTFs rescue mEPSCs in A?42-treated hippocampal neurons. Mouse
Fradinger et al.PNAS ?
September 16, 2008 ?
vol. 105 ?
no. 37 ?
gomerization processes of A?40 and A?42 (23, 24), yielding
oligomer size distributions in good agreement with experimental
findings (8, 36).
Here we modeled the self-assembly of A?42 in the presence
of A?(31–42) or A?(39–42), each at A?42:CTF number con-
centration ratios ranging from 1:1 to 1:8. In all cases, we found
that A?42 and the CTF molecules coassembled into ‘‘heteroo-
ligomers.’’ An example is shown in Fig. 4A. Formation of
heterooligomers of A?42 and A?(31–42) was observed already
after 105simulation steps, and by 107steps all of the molecules
associated into one large heterooligomer. Movie S1 shows the
time evolution of the heterooligomers. This behavior was ob-
served for the A?42:A?(31–42) system at 1:2 and higher ratios,
whereas in the A?42:A?(39–42) system a 1:8 ratio was necessary
for the coassembly of all of the molecules into one heterooli-
gomer. Within the heterooligomers, intermolecular interactions
among A?42 monomers were inhibited. A?(31–42) was found to
inhibit these intermolecular interactions substantially more ef-
ficiently than A?(39–42) (Fig. 4B).
We have used an approach for developing A?42 oligomerization
inhibitors based on putative homotypic association of peptide
sequences in the C terminus of A?42. Peptides derived from the
C terminus of A?42 were found to disrupt the assembly and
inhibit the neurotoxicity of A?42 oligomers. This proof-of-
concept study using A?42 CTFs has yielded two lead peptide
inhibitors of A?42 assembly and neurotoxicity, A?(31–42) and
A?(39–42). The higher inhibitory activity of A?(31–42) and
A?(39–42) relative to other CTFs suggests that the inhibition is
specific rather than based on generic hydrophobic association.
In our initial screen, in which A?42 was mixed with each CTF at
a 1:10 ratio, respectively, A?(31–42) was the only CTF that
completely rescued the cells from A?42-induced toxicity. It was
followed by A?(30–42) and A?(39–42), each of which attenuated
A?42 toxicity by ?80% (Fig. 1B). When the inhibitory activity is
plotted versus CTF length, A?(31–42) gives rise to an inhibitory
activity peak (Fig. 1B). The high activity of A?(30–42) was inter-
preted as resulting from its close similarity to A?(31–42). In
contrast, the high activity of A?(39–42) was surprising given its
small size and presumed absence of stable conformation.
In the three biological tests applied, cell death (LDH assay),
mitochondrial integrity (MTT assay), and synaptic function
(mEPSC assay), A?(31–42) consistently showed higher potency
as an inhibitor of A?42-induced toxicity than A?(39–42). Struc-
tural studies of A?42 have suggested the existence of a quasi-
stable conformation in the C terminus (19–22), likely a turn
centered at G37–G38 (23, 24, 37). We conjecture that this
conformation is important for intermolecular interaction among
the C termini of A?42 that lead to oligomerization. A similar
putative structure in A?(31–42) may account, at least partially,
for the high inhibitory activity of this peptide. In contrast,
A?(39–42) is not expected to have a stable conformation. These
as A?(39–42), suggest that the two CTFs may act by different
We anticipated that CTFs would disrupt A?42 oligomeriza-
tion by incorporating into a putative hydrophobic core of A?42
oligomers (Fig. S1), in which the C terminus was predicted to be
an important component. Our physicochemical studies suggest
that the CTFs indeed interact with A?42 molecules and get
incorporated into oligomers. DLS data (Fig. 3A) show two initial
oligomeric populations of A?42, high-abundance, small oli-
next day (Center), and after 7 or 9 days (Right). White bars represent P1 particles. Gray bars represent P2 or larger particles (in the case of A?42 alone). Days of
measurement and the total scattering intensities in counts per second are shown in the upper left corner of each panel. Only intensities within the same row
are directly comparable with each other. (B) Growth rates of P2 particles (dRH2/dt) in the absence or presence of CTFs. (C) Average number of intensity spikes
per hour during the first 3 days of measurement in the absence or presence of CTFs.
CTF effect on A?42 assembly. (A) Representative distributions of A?42 in the absence or presence of CTFs immediately after preparation (Left), on the
www.pnas.org?cgi?doi?10.1073?pnas.0807163105 Fradinger et al.
size oligomers of RH ? 20–60 nm (P2). In the presence of
growth is attenuated (Fig. 3 A and B). In addition, both CTFs
inhibit formation of intensity spikes in DLS experiments (Fig.
3C), suggesting inhibition of fibril formation. In correlation with
the higher inhibitory activity observed for A?(31–42) in the
MTT, LDH, and mEPSC assays, it was found to inhibit both the
increase in size of P2 particles and the average number of
intensity spikes per hour with higher potency than A?(39–42)
(Fig. 3 B and C). In addition, CTFs that showed low inhibition
of toxicity had little effect on particle growth (data not shown),
demonstrating an overall good agreement between inhibition of
particle growth and inhibition of toxicity.
In support of different mechanisms of toxicity inhibition by
A?(31–42) and A?(39–42), only A?(39–42) was found to
reduce the size of the P1 oligomer population to ?4–9 nm,
suggesting that interaction with A?(39–42) altered the tertiary
and/or quaternary structure of A?42 within P1 oligomers. An-
other important difference between the two CTFs was found in
PICUP experiments, in which A?(31–42) was found to inhibit
paranucleus formation dose-dependently (Fig. S3), whereas
A?(39–42) did not show such inhibition at the highest concen-
tration tested (Fig. S3B).
The observed differences between the behaviors of A?(31–
42) and A?(39–42) in both the PICUP and the DLS experiments
correlated qualitatively with the simulation findings. In agree-
ment with the PICUP data, the model predicted more efficient
A?(31–42) than by A?(39–42) (Fig. 4B). The computer simu-
lations also help explaining, qualitatively, how CTFs can both
disrupt paranucleus formation and promote formation of P2
oligomers. In the model, relatively large heterooligomers are
observed at high numbers of simulation steps (Fig. 4A). Interrup-
tion of intermolecular contacts within these heterooligomers by
because the cross-linking is ‘‘zero length’’; i.e., it requires direct
intermolecular interactions between A?42 monomers.
Taken together, the data indicate that the CTFs inhibit
A?42-induced toxicity by formation of nontoxic heterooli-
gomers, similar to the mechanism proposed for the inhibitory
activity of inositols (17, 18) and for the green tea-derived
polyphenol epigallocatechin gallate (38). The observation that
highly hydrophobic peptides are acting by a mechanism similar
to that of polyols is interesting and suggests that stabilization of
nontoxic oligomers may be a general mechanism for compounds
that inhibit the toxic effects of amyloidogenic proteins. Using
peptides derived from the C terminus of A?42, rather than
carbohydrate-based inhibitors, allows delineating the relation-
ship between inhibitor structure and bioactivity, providing a
framework for development of future derivatives. An advantage
of using CTFs as inhibitors is that the hydrophobic nature of
these peptides may facilitate penetration through biological
barriers, such as the plasma membrane and the blood–brain
barrier. Our findings provide a foundation for lead optimization by
systematic structure–activity relationship studies. A?(31–42) is a
methods, such as alanine scanning and introduction on nonnatural
amino acids. A?(39–42) is a somewhat weaker inhibitor, but its
small size may facilitate transformation into peptidomimetics lead-
ing to novel, disease-modifying drugs for AD.
details, see SI Text.
Cell Culture. Rat pheochromocytoma (PC-12) cells were used 48 h after differ-
entiation. Primary embryonic hippocampal cultures were maintained for 2
weeks before initiation of experiments. For additional details, see SI Text.
Cell Viability Assays. The biological activity of the CTFs themselves and of
A?42:CTF mixtures was assessed by the CellTiter 96 Cell Proliferation Assay
(MTT assay; Promega) and CytoTox-ONE Homogenous Membrane Integrity
Assay (LDH assay; Promega). For additional details, see SI Text.
Electrophysiological Studies. Spontaneous mEPSCs were recorded at a holding
potential of ?70 mV by using an Axopatch 200A patch-clamp amplifier (Axon
Instruments). For additional details, see SI Text.
DLS. A?42:CTF mixtures prepared at 30 ?M (nominal concentration) of each
peptide were studied by using an in-house-built system with a He-Ne laser
additional details, see SI Text.
and subjected immediately to PICUP as described previously (33). For addi-
tional details, see SI Text.
DMD. DMD simulations were performed by using a four-bead protein model
with backbone hydrogen bonding and effective amino acid-specific interac-
tions due to hydropathy, as described previously (23, 24). For additional
details, see SI Text.
ACKNOWLEDGMENTS. We express our gratitude for financial support from
National Institutes of Health/National Institute on Aging Grants AG027818 and
AG023661, Grant 2005/2E from the Larry L. Hillblom Foundation, a private do-
gomerization. (A) Configurations of 16 A?42 and 128 A?(31–42) molecules at
different time frames measured at t simulation steps. CTFs are displayed in
dark blue, and A?42 molecules are represented by their secondary structure:
yellow ribbons, ?-strands; blue tubes, turns; silver tubes, random coil. (B)
Intermolecular contact maps of A?42 in the absence or presence of CTFs
The contact maps are oriented such that the contact strength between pairs
of N-terminal residues is displayed at the top left corner and the contact
strength between pairs of C-terminal residues is at the bottom right corner.
The strength of the contact between two amino acids is color-coded from 0.0
(blue) to a maximal strength (red), corresponding to 30 contacts.
Simulation of the interaction between A?42 and CTFs during oli-
Fradinger et al. PNAS ?
September 16, 2008 ?
vol. 105 ?
no. 37 ?
1. BrookmeyerR,JohnsonE,Ziegler-GrahamK,ArrighiHM(2007)Forecastingtheglobal Download full-text
burden of Alzheimer’s disease. Alzheimer’s Dementia 3:186–191.
2. Alzheimer Association (2008) 2008 Alzheimer’s disease facts and figures. Alzheimer’s
3. Hardy JA, Higgins GA (1992) Alzheimer’s disease: The amyloid cascade hypothesis.
4. Kirkitadze MD, Bitan G, Teplow DB (2002) Paradigm shifts in Alzheimer’s disease and
other neurodegenerative disorders: The emerging role of oligomeric assemblies.
J Neurosci Res 69:567–577.
5. Walsh DM, Selkoe DJ (2007) A? oligomers—A decade of discovery. J Neurochem
6. Borchelt DR, et al. (1996) Familial Alzheimer’s disease-linked presenilin 1 variants
elevate A?1–42/1–40 ratio in vitro and in vivo. Neuron 17:1005–1013.
7. Dahlgren KN, et al. (2002) Oligomeric and fibrillar species of amyloid-? peptides
differentially affect neuronal viability. J Biol Chem 277:32046–32053.
8. Bitan G, et al. (2003) Amyloid ?-protein (A?) assembly: A?40 and A?42 oligomerize
through distinct pathways. Proc Natl Acad Sci USA 100:330–335.
9. McGowan E, et al. (2005) A?42 is essential for parenchymal and vascular amyloid
deposition in mice. Neuron 47:191–199.
10. Wang Z, Chang L, Klein WL, Thatcher GR, Venton DL (2004) Per-6-substituted-per-6-
deoxy ?-cyclodextrins inhibit the formation of ?-amyloid peptide derived soluble
oligomers. J Med Chem 47:3329–3333.
11. Walsh DM, et al. (2005) Certain inhibitors of synthetic amyloid ?-peptide (A?) fibril-
logenesis block oligomerization of natural A? and thereby rescue long-term potenti-
ation. J Neurosci 25:2455–2462.
12. Yang F, et al. (2005) Curcumin inhibits formation of amyloid ? oligomers and fibrils,
binds plaques, and reduces amyloid in vivo. J Biol Chem 280:5892–5901.
indicate that amyloid ? oligomerization and fibrillization pathways are independent
and distinct. J Biol Chem 282:10311–10324.
14. Research Highlights (2008) Chemical biology: Aggravating aggregating. Nature
15. Feng BY, et al. (2008) Small-molecule aggregates inhibit amyloid polymerization. Nat
Chem Biol 4:197–199.
16. Martins IC, et al. (2008) Lipids revert inert A? amyloid fibrils to neurotoxic protofibrils
that affect learning in mice. EMBO J 27:224–233.
17. McLaurin J, et al. (2006) Cyclohexanehexol inhibitors of A? aggregation prevent and
reverse Alzheimer phenotype in a mouse model. Nat Med 12:801–808.
18. McLaurin J, Golomb R, Jurewicz A, Antel JP, Fraser PE (2000) Inositol stereoisomers
stabilize an oligomeric aggregate of Alzheimer amyloid ? peptide and inhibit A?-
induced toxicity. J Biol Chem 275:18495–18502.
19. Murakami K, et al. (2005) Formation and stabilization model of the 42-mer A? radical:
20. RiekR,GuntertP,Do ¨beliH,WipfB,Wu ¨thrichK(2001)NMRstudiesinaqueoussolution
fail to identify significant conformational differences between the monomeric forms
of two Alzheimer peptides with widely different plaque-competence, A?(1–40)oxand
A?(1–42)ox. Eur J Biochem 268:5930–5936.
22. Lazo ND, Grant MA, Condron MC, Rigby AC, Teplow DB (2005) On the nucleation of
amyloid ?-protein monomer folding. Protein Sci 14:1581–1596.
Proc Natl Acad Sci USA 101:17345–17350.
formation: A discrete molecular dynamics study. Biophys J 92:4064–4077.
25. Fezoui Y, et al. (2000) An improved method of preparing the amyloid ?-protein for
fibrillogenesis and neurotoxicity experiments. Amyloid 7:166–178.
with the MTT assay on differentiated neuroblastoma cells. Brain Res Bull 62:223–229.
27. Shearman MS, Ragan CI, Iversen LL (1994) Inhibition of PC12 cell redox activity is a
specific, early indicator of the mechanism of ?-amyloid-mediated cell death. Proc Natl
Acad Sci USA 91:1470–1474.
28. Korzeniewski C, Callewaert DM (1983) An enzyme-release assay for natural cytotox-
icity. J Immunol Methods 64:313–320.
29. Haass C, Selkoe DJ (2007) Soluble protein oligomers in neurodegeneration: Lessons
from the Alzheimer’s amyloid ?-peptide. Nat Rev Mol Cell Biol 8:101–112.
30. Chin JH, Ma L, MacTavish D, Jhamandas JH (2007) Amyloid ? protein modulates
glutamate-mediated neurotransmission in the rat basal forebrain: Involvement of
presynaptic neuronal nicotinic acetylcholine and metabotropic glutamate receptors.
J Neurosci 27:9262–9269.
31. Parameshwaran K, et al. (2007) Amyloid ?-peptide A?(1–42) but not A?(1–40) atten-
uates synaptic AMPA receptor function. Synapse 61:367–374.
32. Lomakin A, Teplow DB (2006) Quasielastic light scattering study of amyloid ?-protein
fibril formation. Protein Pept Lett 13:247–254.
33. Bitan G (2006) Structural study of metastable amyloidogenic protein oligomers by
photo-induced cross-linking of unmodified proteins. Methods Enzymol 413:217–
34. Urbanc B, Borreguero JM, Cruz L, Stanley HE (2006) Ab initio discrete molecular
dynamics approach to protein folding and aggregation. Methods Enzymol 412:314–
35. Bitan G, Teplow DB (2005) Preparation of aggregate-free, low molecular weight
amyloid-? for assembly and toxicity assays. Methods Mol Biol 299:3–9.
36. Bernstein SL, et al. (2005) Amyloid ?-protein: Monomer structure and early aggrega-
tion states of A?42 and its Pro19 alloform. J Am Chem Soc 127:2075–2084.
37. Takano K, et al. (2006) Structure of amyloid ? fragments in aqueous environments.
FEBS J 273:150–158.
38. Ehrnhoefer DE, et al. (2008) EGCG redirects amyloidogenic polypeptides into unstruc-
tured, off-pathway oligomers. Nat Struct Mol Biol 15:558–566.
and growth of amyloid ?-protein fibrils: Detection of nuclei and quantitation of rate
constants. Proc Natl Acad Sci USA 93:1125–1129.
40. Condron MM, Monien BH, Bitan G (2008) Synthesis and purification of highly hydro-
phobic peptides derived from the C-terminus of amyloid ?-protein. Open Biotechnol
www.pnas.org?cgi?doi?10.1073?pnas.0807163105Fradinger et al.