MICROSCOPY RESEARCH AND TECHNIQUE 67:164174 (2005)

Amyloid-b Aggregates Formed at PolarNonpolar Interfaces

Differ From Amyloid-b Protobrils Produced in
Aqueous Buffers
MICHAEL R. NICHOLS,1 MELISSA A. MOSS,1 DANA KIM REED,1 JAN H. HOH,2
1
AND TERRONE L. ROSENBERRY *
1

Department of Neuroscience, Mayo Clinic College of Medicine, Jacksonville, Florida 32224

Department of Physiology, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205

KEY WORDS

Alzheimers disease; peptide b-structure; brils; stability; hexauoroisopropanol

ABSTRACT
The deposition of aggregated amyloid-b (Ab) peptides in the brain as senile
plaques is a pathological hallmark of Alzheimers disease (AD). Several lines of evidence indicate that
brillar and, in particular, soluble aggregates of these 40- and 42-residue peptides are important in
the etiology of AD. Recent studies also stress that amyloid aggregates are polymorphic and that a
single polypeptide can fold into multiple amyloid conformations. Here we review our recent reports
that Ab(1-40) in vitro can form soluble aggregates with predominant b-structures that differ in
stability and morphology. One class of aggregates involved soluble Ab protobrils, prepared by vigorous overnight agitation of monomeric Ab(1-40) in low ionic strength buffers. These aggregates
were quite stable and disaggregated to only a limited extent on dilution. A second class of soluble
Ab aggregates was generated at polarnonpolar interfaces. Aggregation in a two-phase system of
buffer over chloroform occurred more rapidly than in buffer alone. In buffered 2% hexauoroisopropanol (HFIP), microdroplets of HFIP were formed and the half-time for aggregation was less than
10 minutes. Like Ab protobrils, these interfacial aggregates showed increased thioavin T uorescence and were rich in b-structure by circular dichroism. However, electron microscopy and atomic
force microscopy revealed very different morphologies. The HFIP aggregates formed initial globular clusters that progressed over several days to soluble brous aggregates. When diluted out of
HFIP these aggregates initially were very unstable and disaggregated completely within 2 minutes.
However, their stability increased as they progressed to bers. It is important to determine
whether similar interfacial Ab aggregates are produced in vivo. Microsc. Res. Tech. 67:164174,
2005. ' 2005 Wiley-Liss, Inc.V 2004 Wiley-Liss, Inc.
C

INTRODUCTION
The brains of patients with Alzheimers disease (AD)
contain large numbers of amyloid deposits in the form
of senile plaques (Cohen and Calkins, 1959). The amyloid core of the plaques consists of interwoven brils,
each 79 nm in diameter (Terry et al., 1964), that can
be visualized by light microscopy after staining with
Congo Red or thioavin S (Terry, 1985). The brils are
composed of 40- and 42-residue peptides (Glenner and
Wong, 1984; Miller et al., 1993), denoted Ab(1-40) and
Ab(1-42). These peptides are produced by cleavage of
cellular amyloid precursor protein (APP) by two proteases called b- and g-secretase (reviewed in Selkoe,
2001). As originally suggested by the amyloid cascade
hypothesis (Hardy and Higgins, 1992), it appears likely
that Ab aggregates are important in the etiology of AD.
The most striking evidence supporting this hypothesis
comes from the identication of numerous mutations
linked to early-onset familial AD (FAD) (Selkoe and
Podlisny, 2002). These mutations are located within
the APP gene or the genes for presenilins 1 and 2,
which play an integral role in g-secretase activity. All
FAD mutations reported thus far increase either the
level of the more amyloidogenic Ab(1-42) peptide
C
V

2005 WILEY-LISS, INC.

(reviewed in Selkoe and Podlisny, 2002) or the propensity of a mutated Ab peptide to form amyloid aggregates (Nilsberth et al., 2001).
The amyloid hypothesis for the etiology of AD is also
supported by early results from therapeutic strategies
that reduce the amount of aggregated Ab in the brain
(Lansbury, 1997). An extremely promising example is a
vaccine comprised of Ab(1-42) (Schenk et al., 1999).
While a clinical trial with this vaccine was halted
because of adverse side effects in
5% of individuals
(Schenk, 2002), the antibodies generated by patients in

Abbreviations: AD, Alzheimers Disease; AFM, atomic force microscopy; APP,

amyloid precursor protein; CD, circular dichroism; DLS, dynamic light scattering;
EM, electron microscopy; FAD, early-onset familial AD; HFIP, hexauoroisopropanol; MALS, multi-angle light scattering; SEC, size exclusion chromatography.
*Correspondence to: T.L. Rosenberry, Department of Neuroscience, Mayo
Clinic College of Medicine, 4500 San Pablo Road, Jacksonville, FL 32224.
E-mail: rosenberry@mayo.edu
Contract grant sponsor: American Heart Association, Florida/Puerto Rico
Afliate (to M.R.N. and M.A.M.)
Received 21 January 2005; accepted in revised form 21 March 2005
DOI 10.1002/jemt.20189
Published online in Wiley InterScience (www.interscience.wiley.com).

AMYLOID-b AGGREGATION AT NONPOLAR INTERFACES

this trial appeared to be effective. Twenty patients generated antibodies against Ab, as determined by tissue
amyloid plaque immunoreactivity assay. These
patients showed signicantly slower rates of decline of
cognitive functions and activities of daily living than
those observed with patients who did not produce such
antibodies (Hock et al., 2003).
Early evidence suggested that brillar deposits of Ab
initiate a cascade of events that result in neuronal cell
death and lead to the cognitive decline characteristic of
AD (Yankner, 1996). However, concerns were raised
because the number of amyloid deposits detected by
neuropathological analysis of postmortem brains does
not correlate well with the degree of cognitive impairment experienced by AD patients prior to death (Hardy
and Selkoe, 2002). A number of investigators now
propose that soluble aggregates of Ab (also called
oligomers or protobrils), rather than monomers or
insoluble amyloid brils, may be responsible for synaptic dysfunction in the brains of AD patients and AD
animal models (Hardy and Selkoe, 2002; Hartley et al.,
1999; Klein et al., 2001; Westerman et al., 2002;
Kawarabayashi et al., 2004). This proposal is supported by observations that soluble aggregates generated in vitro from synthetic Ab(1-40) and (1-42)
induced toxicity in cultured cells (Hartley et al., 1999;
Lambert et al., 1998), that soluble Ab aggregates produced in cell culture markedly inhibited hippocampal
long-term potentiation in rats in vivo (Walsh et al.,
2002), and that transgenic mice expressing human Ab
show functional decits that precede extracellular deposition of brillar Ab (Hsia et al., 1999; Westerman
et al., 2002).
A mechanistic understanding of the brillogenesis
process would be very useful in identifying individual
steps as therapeutic targets (Teplow, 1998). Since this
understanding is difcult to attain in the complex environment of living cells, it is fortunate that synthetic
Ab(1-40) and Ab(1-42) form both soluble aggregates
and amyloid brils in vitro that share many features
with the amyloid in AD plaques as well as the amyloid
brils formed by other proteins in a number of diseases. These brils contain peptide segments that align
to form b-sheets with extended peptide strands perpendicular to and interstrand H-bonds parallel to the bril
axis (Sunde et al., 1997; Serpell et al., 2000). This
cross-b structure may underlie common properties,
including the binding of dyes like thioavin T and
Congo Red and of antibodies that react with a common
epitope in several amyloidogenic proteins (ONuallain
and Wetzel, 2002; Kayed et al., 2003). However, recent
reports also indicate a fascinating degree of amyloid
polymorphism at the molecular level. This feature was
rst noted with mammalian and yeast prion proteins
when it was shown that a single polypeptide can misfold into multiple amyloid conformations (Chien et al.,
2004). Specically, the yeast Sup35p prion protein was
aggregated at different temperatures into amyloid conformations that could be distinguished by thermal
stability and EPR spectroscopy, and infection of yeast
with these different conformations led to different
propagating yeast [PSI] strains (Tanaka et al., 2004;
King and Diaz-Avalos, 2004). Ab brils also show

165

molecular diversity. Solid-state NMR measurements

revealed that Ab(1-40), Ab(1-42), and Ab(10-35) brils
contained in-register, parallel b-sheets (Tycko, 2003;
Burkoth et al., 2000), while brils formed by the
shorter peptides Ab(16-22), Ab(34-42), and Ab(11-25)
adopted anti-parallel b-strand alignments (Balbach
et al., 2000; Lansbury et al., 1995; Petkova et al.,
2004). Very recently, two types of amyloid brils were
formed by Ab(1-40) following aggregation under mildly
agitated or quiescent conditions, and chemical shift
and line-width data from solid-state NMR for 33 of the
40 residues indicated different underlying structures
(Tycko, 2004; Petkova et al., 2005).
The emergence of amyloid polymorphism underscores the need for measures that would distinguish
amyloid aggregates formed under different conditions.
The cellular environment provides a variety of interfaces that are likely to dictate the formation of particular amyloid structures, and it is important to develop
measures that eventually can be applied to in situ Ab
aggregates in tissue samples. In this review we incorporate images obtained by EM and AFM in three of our
recent reports (Nichols et al., 2002, 2005a,b) to enhance
our characterization of soluble Ab(1-40) aggregates
generated either in a homogeneous buffer or in twophase systems at a polarnonpolar interface. While
these aggregates all showed enhanced uorescence
with thioavin T and enrichment in b-structure by CD,
they differed strikingly in their stabilities on dilution
and their morphological features.
MATERIALS AND METHODS
Preparation of Monomeric Ab(1-40)
Ab(1-40) peptide was obtained from QCB (Hopkinton, MA), rPeptide (Athens, GA), or the protein and
peptide core facility at the Mayo Clinic (Rochester,
MN). Samples were routinely dissolved in water and
any preformed aggregates were removed from stock
solutions by size exclusion chromatography (SEC) on a
1  30 cm Superdex 75 HR 10/30 column (Amersham
Biosciences, Piscataway, NJ) (see Teplow, 1998). A single elution peak was observed at a partially included
volume (Nichols et al., 2002), and translational diffusion measurements by NMR (Tseng et al., 1999) as well
as direct Mw determination by static multiangle light
scattering (MALS) (Nichols et al., 2005a) indicated that
this peak corresponds to monomeric Ab(1-40) (4.3
4.5 kDa).
Isolation of Ab(1-40) Protobrils
Monomeric Ab(1-40) was aggregated with vigorous
agitation in Tris-EDTA buffers (550 mM Tris-HCl,
5 mM EDTA at pH 8.0) containing 0150 mM NaCl at
238C. Aggregation was monitored with thioavin T
(Nichols et al., 2002), a uorophore that shows greatly
enhanced uorescence on binding to amyloid brils
(LeVine, 1993) as well as protobrils (Walsh et al.,
1999). Samples were microfuged for 10 minutes in a
tabletop centrifuge (Beckman Coulter, Fullerton, CA)
at 18,000g. The supernatant was fractionated by SEC
on Superdex 75, and Ab eluting in the void volume was
dened as the protobril fraction (Nichols et al., 2002).
Superdex columns were routinely pretreated with a

166

M.R. NICHOLS ET AL.

bolus of bovine serum albumin (50 mg) in running buffer to block nonspecic binding of Ab protobrils to the
resin and occasionally were washed with 1 N NaOH.
Light-Scattering Measurements
Particle hydrodynamic radii (RH) were measured by
dynamic light scattering (DLS) at room temperature
with a DynaPro MSX instrument (Wyatt Technology,
Santa Barbara, CA). Total light scattering intensity at
a 908 angle in kilocounts/sec (kcts) was collected using
a 5-second acquisition time (Nichols et al., 2002).
Weight average molecular masses (Mw) were estimated
by MALS with a DAWN EOS instrument (Wyatt Technology, Santa Barbara, CA) in-line with a Superdex 75
SEC column (Nichols et al., 2002). This SEC-MALS
procedure is attractive because, in contrast to DLS, Mw
estimates can be made without assumptions of molecular or aggregate shape.
Circular Dichroism (CD)
Spectra were obtained on an Aviv Model 215 Circular
Dichroism Spectrometer with a 0.1 cm pathlength
quartz cuvette (Hellma, Mullheim, Germany). Buffer
control spectra were averaged and subtracted from the
average of triplicate scans of each Ab sample spectra,
and each resulting point ([y]obs, deg) was converted to
mean residue ellipticity ([y], deg cm2 dmol1) (Nichols
et al., 2005a).
Electron Microscopy (EM)
Samples of Ab aggregates were applied to 200-mesh
formvar-coated copper grids (Ernest F. Fullam,
Latham, NY) and incubated for 1015 minutes at room
temperature. The sample was then wicked off with lens
paper, washed briey by placing the grid face-down on
a wash droplet, and stained by transferring the grid
face-down to a droplet of 2% uranyl acetate (Polysciences, Warrington, PA) for 510 minutes before wicking off the solution and air drying. With samples generated in 2% HFIP, all wash and treatment solutions
applied to the grids also contained 2% HFIP. Grids
were visualized in a Philips EM208S transmission electron microscope.
Atomic Force Microscopy (AFM)
Samples were applied to freshly cleaved mica that
had been modied with 30 -aminopropyl-triethoxysilane (APTES) and incubated for 15 minutes (Nichols
et al., 2002, 2005a). The residual sample liquid was
aspirated off and the disk was then rinsed gently with
water and blown dry with compressed air. With aggregates generated in HFIP, the rinse included 2% HFIP.
A Nanoscope III controller with a Multimode AFM
(Digital Instruments, Santa Barbara, CA) was used for
imaging by ambient tapping mode. Images were
obtained in either amplitude mode or height mode,
where increasing brightness indicates greater damping
of cantilever oscillation (Harper et al., 1999) or increasing feature height, respectively. Height mode images
were attened prior to measurement of particle
height distributions with NanoScope (R) III software
(v. 5.13r5, Digital Instruments).

RESULTS AND DISCUSSION

Formation and Growth of Ab(1-40) Protobrils in
Aqueous Buffers
The term protobril was introduced to describe
soluble Ab aggregates prepared in vitro and isolated by
SEC. These aggregates exhibited Stokes radii of 10
50 nm as measured by DLS (Walsh et al., 1997),
showed enhanced uorescence with thioavin T, and
gave a CD spectrum enriched in b-structure (Walsh
et al., 1999). Other groups have generated similar soluble Ab aggregates, sometimes called oligomers, but the
morphology of the aggregates has varied somewhat.
Initial EM and AFM reports showed mixtures of
roughly spherical globules and short, curly bers with
lengths of 10200 nm (Harper et al., 1997, 1999; Walsh
et al., 1997), but other groups have observed similar
globules together with short rods, also with lengths of
10200 nm (Goldsbury et al., 2000; Huang et al., 2000;
Dahlgren et al., 2002). These morphological differences
probably arise from variations in pH, ionic strength,
cosolvent, and sample preparation. We prepared Ab(140) protobrils by aggregation of monomeric Ab at pH
8.0 (Nichols et al., 2002). The reaction was monitored
by thioavin T uorescence and proceeded with a
delay, or lag time, that is well known for amyloidogenic
proteins (Jarrett and Lansbury, 1993) (see Fig. 2A).
Samples were periodically microcentrifuged for 10
minutes at 18,000g and the soluble aggregates that
remained in the supernatant were designated protobrils and the sedimented aggregates as brils. The
yield of protobrils depended on the ionic strength,
increasing from less than 20% of the total thioavin T
uorescence in 150 mM NaCl to more than 80% at or
below 40 mM NaCl. Our protobrils isolated by SEC
shared several features of protobrils and soluble Ab
oligomers reported by others. DLS measurements indicated average hydrodynamic radii RH of 3070 nm, and
AFM revealed small aggregates with a globular
appearance, rods (with lengths of 20400 nm), and
globular aggregates with rods emanating from them
(Fig. 1A). Protobril heights obtained from our AFM
images were about 3 nm, consistent with those
reported previously for Ab(1-40) protobrils (Harper
et al., 1997, 1999). Mw values determined by SECMALS for several protobril preparations ranged from
732  103 kDa (Nichols et al., 2002). These protobrils
thus are quite large, containing at least 1,500 monomers. Although DLS and MALS provide very sensitive
measures of aggregation, we have not been able to
detect the smaller Ab(1-40) aggregates reported by
others (Lashuel et al., 2003). Smaller aggregation
intermediates under our conditions thus must occur at
very low steady-state levels.
Since mature Ab brils are longer and have a larger
diameter than Ab protobrils, several possible mechanisms of protobril growth may exist. Studies of Ab
bril formation have examined Ab monomer aggregation (Jarrett et al., 1993) and Ab bril extension by
monomer deposition (Naiki and Nakakuki, 1996; Esler
et al., 1996, 2000). However, few reports have focused
on protobril growth in solution. We distinguished two
growth processes, protobril extension by monomer
deposition, a process we termed elongation, and direct
protobrilprotobril association (Nichols et al., 2002).

AMYLOID-b AGGREGATION AT NONPOLAR INTERFACES

167

Fig. 1. Protobril growth arising from elongation by added monomer or direct protobril association as monitored by AFM. Ab(1-40)
protobrils were isolated by SEC in 5 mM Tris-EDTA and diluted to
1 mM (in Ab monomer units) in 50 mM Tris-EDTA. The dilution for
the elongation reaction included 30 mM Ab(1-40) monomer and after
1 hour the mixture was fractionated by SEC to isolate a subpopulation of elongated protobrils. The dilution for the association reaction

included 150 mM NaCl, but low recoveries precluded further fractionation by SEC. Aliquots (100 ml) from each of the three samples were
applied to mica stubs and analyzed by AFM. Initial (A), elongated
(B), and associated (C) protobrils are shown as 2.5  2.5 mm images
(cropped from 5  5 mm elds). Images are presented in amplitude
mode. Slightly different images from the same 5  5 mm elds were
previously presented in height mode (Nichols et al., 2002).

Elongation occurred immediately upon addition of

puried Ab monomer to puried protobrils and could
be conveniently followed by increases in thioavin T
uorescence or RH. This growth occurred at protobril
ends. Both AFM and EM indicated a higher proportion
of aggregates corresponding to rods with extended
lengths of several mm after elongation. Fractionation of
the elongation reaction by SEC yielded a subpopulation
of shorter rods that could be eluted from the column
(Fig. 1B), and MALS analyses of this fraction showed
consistent decreases in mass per unit length (MPL)
after the protobrils were elongated (Nichols et al.,
2002). Protobril growth also occurred in the absence
of added Ab monomer, as observed previously by DLS
(Walsh et al., 1999) and AFM (Harper et al., 1999). This
growth involved association of the initial protobrils
and was very sensitive to ionic strength. It occurred at
a very slow rate, if at all, in the absence of NaCl. Addition of 150 mM NaCl to an aliquot of isolated protobrils more than doubled the average RH over a 2-hour
period with little change in thioavin T uorescence
(Nichols et al., 2002). Values of MPL from SEC-MALS
for these associated protobrils were about 3 times
larger than those for the elongated protobrils, and
AFM showed that association involved clustering of
the initial protobrils in a relatively disordered array
with no extension of tendrils (Fig. 1C).

nucleation in vivo, and anionic phospholipid vesicles

(Terzi et al., 1997) as well as vesicles containing GM1
ganglioside (Choo-Smith et al., 1997; Choo-Smith and
Surewicz, 1997) promoted Ab binding and b-sheet content in vitro. Ab interactions with these vesicles
appeared restricted to the lipid polar head groups
(Terzi et al., 1997; Matsuzaki and Horikiri, 1999; Yip
and McLaurin, 2001). Other reports have indicated
that amphiphilic interactions may be important and
suggested that micellar-like Ab structures may play a
role in nucleation (Lomakin et al., 1996; Tjernberg
et al., 1999; Yong et al., 2001). Amphiphilic molecules
tend to accumulate at interfaces between water and air
or nonpolar liquids (Pratt and Pohorille, 2002), and an
Ab monolayer at an air/water interface did undergo a
rapid and substantial increase in interfacial b-sheet
content compared to that of a similar Ab incubation in
bulk solution (Schladitz et al., 1999).
We have identied two systems in which amphiphilic
interfacial interactions appear to promote Ab aggregation into structures rich in b-structure. The rst
involves addition of monomeric Ab(1-40) to the upper
aqueous phase of a buffer/chloroform two-phase system
(Nichols et al., 2005b). Ab aggregates formed within
hours, in contrast to the formation of Ab protobrils
from the same 50100 mM concentrations of Ab monomers in dilute aqueous buffer, a process that typically
involves a lag time of days in a quiescent one-phase
system. The liquid/liquid interface was necessary for
this effect, as buffer saturated with chloroform failed to
show any acceleration of aggregation. The second system consists of dilute HFIP in buffer. Our study of this
system was prompted by a report that 14% (v/v) HFIP
strongly accelerated amyloid ber formation by islet
amyloid polypeptide (IAPP) (Padrick and Miranker,
2002). This is less clearly a two-phase system, but we
present evidence below to conrm a second phase composed of HFIP microdroplets. Rates of Ab aggregation
in aqueous buffer alone and in dilute HFIP are compared in Figure 2. The formation of protobrils from
monomeric Ab(1-40) during vigorous agitation in dilute

Interfacial Ab Aggregation
The concept that Ab brillogenesis is a nucleationdependent polymerization process (Jarrett et al., 1993;
Walsh et al., 1997) has provided a useful framework for
in vitro Ab aggregation studies. Nucleus formation is
considered rate-limiting in this process, but this rate
(as measured by the delay or lag time to the maximal
rate of aggregation) is highly variable even in homogeneous aqueous solutions and is sensitive to many conditions including pH, ionic strength, temperature, and
agitation (Wood et al., 1996; Harper et al., 1999; Nichols et al., 2002). Interfaces including phospholipid
membranes may play important roles in promoting

168

M.R. NICHOLS ET AL.

lag time for Ab aggregation was abolished (Fig. 2B).

The aggregation did not require a particular mixing
sequence, as similar rates were observed when stock
Ab in 50% HFIP was diluted or a small volume of neat
HFIP was added to diluted Ab. HFIP at higher concentrations is widely known as an effective solvent for Ab
peptides (Zagorski et al., 1999), and we observed that
Ab(1-40) protobrils solubilized at HFIP concentrations above 20% (v/v) were converted to monomeric Ab
as determined by SEC. CD spectra were compared to
document these differences. In agreement with previous reports (Terzi et al., 1997; LeVine, 2002; Walsh
et al., 1999), SEC-puried Ab(1-40) monomer was
largely random coil in buffer alone (a single minimum
at 197 nm) and a-helical in 30% HFIP (double minima
at 208 and 222 nm), while isolated protobrils showed
predominantly b-structure (a minimum at 216 nm and
a maximum at 195 nm) (Fig. 3A). The CD spectrum of
Ab(1-40) aggregates in the aqueous phase of the chloroform/buffer system also showed a single minimum at
216 nm, indicating considerable b-structure (Fig. 3B).
Addition of monomeric Ab(1-40) to fresh 2% HFIP gave
a spectrum consistent with signicant b-structure,
while addition of the same Ab to 2% HFIP that had
been centrifuged to reduce the number of apparent
HFIP microdroplets (see below) gave a smaller shift at
216 nm in the corresponding CD spectrum (Fig. 3C).
Inclusion of the random coil spectrum for control monomeric Ab(1-40) in Figure 3C revealed an isodichroic
point at 209 nm for these three solutions, indicating a
two-component mixture with a rapid equilibrium
between the random coil and b-structures. Therefore,
these CD spectral analyses demonstrated clear differences in Ab structure at high and low concentrations of
HFIP.

Fig. 2. Aggregation of Ab(1-40) in one- and two-phase systems as

monitored by thioavin T uorescence (F). A: (Nichols et al., 2002)
Monomeric Ab(1-40) (70 mM) in 50 mM Tris-EDTA and 40 mM (D) or
150 mM (O) NaCl was shaken continuously at
600 rpm. Aliquots
(10 ml) were diluted into 200 ml of 5 mM thioavin T in buffer at the
indicated times. B: (Nichols et al., 2005a) Indicated concentrations of
HFIP (% v/v) were included in quiescent solutions containing monomeric Ab(1-40) (20 mM) and 5 mM thioavin T in 30 mM Tris-HCl (pH
8.0), and uorescence was recorded continuously. A control with buffer and 5 mM thioavin T in the absence of Ab(1-40) gave a uorescence value of 1. Addition of 2% HFIP increased the value to 3.

buffer showed a lag time of about 6 hours in Figure 2A.

Increasing the NaCl concentration to 150 mM further
decreased the lag time to less than 4 hours but resulted
in a greater yield of brils than of protobrils (Wood
et al., 1996; Harper et al., 1999; Nichols et al., 2002), as
noted above. In contrast, over a narrow range of HFIP
concentrations from 14%, aggregation of Ab(1-40)
without agitation occurred in minutes and the usual

Stabilities of Ab Aggregates
The stability of soluble Ab aggregates has received
much less attention than their assembly pathways.
Virtually the only quantitative estimate of an in vitro
protobril disaggregation rate involved Ab(1-40) protobrils trace-labeled with 125I-Ab(1-40). The dialyzed
protobrils released only about 30% of their radioactivity, mostly over the rst 2 days of dialysis (Walsh et al.,
1999). We measured the stability of our Ab(1-40) protobril preparations by diluting Ab(1-40) aggregation
reactions that were optimized for Ab protobril formation as outlined above. Large dilutions were necessary
to minimize further monomer deposition and reveal
disaggregation, indicated by a progressive decrease in
thioavin T uorescence (Fig. 4A). The uorescence
decrease occurred in at least two kinetic phases, with
initial disaggregation followed by a much slower secondary phase. Less than half of the initial protobrils disaggregated over a 16-hour period, an observation quite
compatible with the extent of disaggregation reported
for dialyzed 125I-Ab(1-40) protobrils (Walsh et al.,
1999). Continued incubation of the aggregation reactions revealed a progressive stabilization of the Ab protobrils, as demonstrated by decreased disaggregation
rate constants (Nichols et al., 2005a) and increased
percentages of stable aggregates after dilution at longer incubation times. In Figure 4A, no disaggregation
could be detected after 5 days of incubation. Ab(1-40)
protobrils isolated after SEC also were completely

AMYLOID-b AGGREGATION AT NONPOLAR INTERFACES

169

Fig. 3. CD spectra of Ab monomers and aggregates. A: (Nichols

et al., 2005a) CD spectra for monomeric Ab(1-40) (30 mM) in 5 mM
Tris-HCl (pH 8.0) (solid) or after addition of 30% HFIP (dots). Ab(140) protobrils were prepared as in Figure 2A and isolated by SEC
(dashed). B: (Nichols et al., 2005b) Ab(1-40) (50 mM) in 5 mM Tris-HCl
(pH 8.0) was incubated over chloroform for 24 hours and a CD spec-

trum was recorded for an aliquot of the aqueous phase (dashed). A

control spectrum of stock monomeric Ab(1-40) (solid) was also recorded.
C: (Nichols et al., 2005a) CD spectra for monomeric Ab(1-40) (30 mM)
in 5 mM Tris-HCl (solid) and after aggregation for 24 hours with 2%
HFIP. The HFIP was added either neat (dots) or from 4% HFIP/water
after a 10-minute centrifugation at 18,000g (dashes).

Fig. 4. Stabilities of Ab protobrils and interfacial aggregates as

monitored by continuous thioavin T uorescence (F) or dynamic
light scattering intensity (LS). SEC-puried monomeric Ab(1-40) was
aggregated as described below and incubated at 258C without further
manipulation. At the indicated times samples were diluted to initiate
disaggregation. A: (Nichols et al., 2005a) Ab (100 mM) was aggregated
with agitation for 22 hours in 5 mM Tris-EDTA and aliquots were
diluted 40-fold into 5 mM Tris-EDTA buffer containing 5 mM thioavin T. B: (Nichols et al., 2005b) Ab (30 mM) in 50 mM Tris-HCl (pH

8.0) was incubated over chloroform for 30 hours and aliquots were
removed from the aqueous phase and either diluted 15-fold into the
same buffer containing 5 mM thioavin T (left) or analyzed without
dilution (right). C: (Nichols et al., 2005a) Ab (20 mM) was aggregated
for 30 minutes without agitation in a solution containing 2% HFIP
and 5 mM thioavin T in 5 mM Tris-EDTA and aliquots were diluted
15-fold into the same thioavin T and buffer without HFIP. The
9 hours disaggregation curve from A (PF) is reproduced to allow comparison with Ab(1-40) protobrils.

stable. However, chemical modication of the free

amino groups in Ab allowed a wider range of stability
measurements. The somewhat less stable protobrils
prepared from reductively radiomethylated Ab(1-40)
showed progressive stabilization on dilution of isolated
protobrils as well as aggregation reactions (Nichols
et al., 2005a).
A striking feature of the interfacial Ab aggregates
generated both in chloroform/buffer and in dilute HFIP
was their instability on dilution. Ab aggregates
removed from the aqueous phase after a 1-day incubation in the two-phase system disaggregated rapidly following dilution into buffer containing thioavin T,
declining to background in less than 10 minutes (Fig.
4B, left). Furthermore, simply removing an aliquot of

the aqueous phase from the interface was sufcient to

induce essentially complete disaggregation (Fig. 4B,
right). The HFIP aggregates also were initially very
unstable. Fifteen-fold dilution of the aggregates formed
after 30 minutes in 2% HFIP into buffer with thioavin
T resulted in a uorescence decrease too fast to measure. The speed of the disaggregation resulted from loss
of the HFIP subphase, because maintenance of 2%
HFIP in the dilution buffer resulted in a much slower
uorescence decrease (Nichols et al., 2005a). As
observed in Figure 4A with Ab protobrils, continued
incubation of Ab(1-40) in dilute HFIP resulted in a
marked stabilization of the aggregates without an
increase in the overall thioavin T binding (Fig. 4C).
Disaggregation rate constants on dilution into buffer

170

M.R. NICHOLS ET AL.

Fig. 5. Comparison of Ab aggregates induced at the chloroform/

buffer interface with those of Ab protobrils by AFM (Nichols et al.,
2005b). Images are presented in both amplitude mode (AC) and
height mode (DF). A,D: Ab(1-40) protobrils were prepared by vigorous continued vortexing of monomeric Ab (180 mM) in 5 mM TrisEDTA overnight at 258C. The protobrils were elongated by incubation with monomeric Ab (30 mM) for 0.5 hours and isolated a second
time by SEC. A 100-ml sample (1 mM in Ab monomer units) was

applied to the mica surface. B,C,E,F: Ab aggregates were induced by

24-hour incubation in buffer containing 50 mM Ab(1-40) in 50 mM
Tris-HCl (pH 8.0) over chloroform. A 20-ml aliquot of the aqueous
phase was diluted 5-fold with 50 mM Tris-HCl (pH 8.0) directly on the
mica surface. Images are 2.5  2.5 mm (A,B,D,E; insets are the same
scale) or 10  10 mm (C,F) and were taken from the same sample disk.
Arrows in F indicate areas of ber discontinuity.

without HFIP were decreased and, after nearly 10

days, approached the rate constant obtained for disaggregation of Ab protobrils formed in the absence of
HFIP (Fig. 4C). The HFIP-induced aggregates
remained soluble despite this stabilization, as indicated by almost complete retention of thioavin T uorescence in the supernatant following centrifugation of
the dilutions at 18,000g for 10 minutes.

clear: the two-phase aggregates included no rod-like laments, and the elongated protobrils showed no
chain-like alignment of globules. Furthermore, Figure
5F showed discontinuities within central sections of
several bers that appeared to reect disaggregation of
individual globular components of the bers after
adsorption of the diluted sample and washing of the
AFM disk. Such rapid disaggregation is consistent with
the ber instability in the aqueous phase shown in
Figure 4B. One limitation of the two-phase chloroform/
buffer system in our analysis of interfacial Ab aggregation was the continued growth of the aggregates. Maximum aggregate levels were obtained after 1-4 days,
and this peak was followed by a rapid decline in aggregate concentration in the aqueous phase and formation
of a cloudy white precipitate hovering at and just above
the aqueous/chloroform interface (Nichols et al.,
2005b). This precipitation prevented us from examining long-term changes in the stability of the aggregates. In contrast, HFIP-induced Ab aggregates
remained soluble for several days, allowing the progressive stabilization of the aggregates to be discerned
(Fig. 4C).
EM analyses, conducted following negative staining
of samples with uranyl acetate, initially were useful in
clarifying the disperse state of dilute HFIP itself. Control samples of freshly diluted 2% HFIP showed no features in EM images, but addition of bovine serum albumin to the diluted HFIP resulted in the appearance of
circular features (Fig. 6A). The number of these fea-

Morphologies of Interfacial Ab Aggregates

We applied AFM and EM imaging to compare the
morphologies of Ab aggregates formed in these interfacial systems with those of Ab protobrils. For comparison with the two-phase aggregates, protobrils were
rst elongated by brief deposition of Ab monomers. The
subset of shorter elongated protobrils was isolated by
SEC and examined by AFM (Fig. 5A,D). They consisted
of rod-like laments with an average height of 3.1 nm,
often emanating from globular cores that were typical
of the protobrils prior to elongation (see Fig. 1A,B).
AFM images of the two-phase aggregates (Fig. 5B,E;
C,F) revealed structures that we denote exible bers,
with lengths of much greater than 1 mm, as well as
numerous species that appeared globular. The bers
appeared to be composed of a collection of the globular
species aligned roughly along a longitudinal axis. The
measured heights for both bers and globules in Figure
5E were similar (45 nm). Distinctions between the
elongated protobrils in Figure 5A,D and the twophase aggregate sample in Figure 5B,E and C,F are

AMYLOID-b AGGREGATION AT NONPOLAR INTERFACES

171

Fig. 6. Electron micrographs of HFIP-induced Ab(1-40) aggregates reveal progressive changes in morphology (Nichols et al.,
2005a). Bovine serum albumin (0.25 mg/ml) (A) or SEC-puried
monomeric Ab(1-40) (40 mM) (BE) were incubated in 5 mM thioavin
T, 5 mM Tris-HCl (pH 8.0), and 2% HFIP (neat) at room temperature.
After 0.1 hours a 10-ml sample of the albumin solution was removed
and at the indicated times aliquots of the Ab(1-40) solution were centrifuged at 18,000g for 10 minutes and supernatant samples (10 ml)

were removed. These samples were processed for negative staining

and EM. F: SEC-puried monomeric Ab(1-40) (100 mM) in 5 mM TrisEDTA was incubated with agitation for 22 hours. An aliquot was centrifuged (18,000g for 10 minutes) and diluted 10-fold and the protobrils in the supernatant were elongated by addition of monomeric
Ab(1-40) (40 mM) for 15 minutes. A 10-ml sample was removed for negative staining and EM as in panels AE. Images are shown relative to
a calibration bar of 200 nm.

tures was greatly reduced when the 2% HFIP solutions

were centrifuged at 18,000g for 10 minutes before and
after addition of albumin (data not shown), indicating
that the albumin had slightly adsorbed to spherical
HFIP microdroplets. This conclusion was supported by
observations that centrifugation of 24% HFIP stocks
removed particles responsible for 90% of the DLS
intensity from the supernatant and decreased the rate
and extent of aggregation when Ab was introduced
(Nichols et al., 2005a; Fig. 3C). It is also consistent
with a report of hydrated HFIP oligomers in dilute
HFIP solutions (Yoshida et al., 2003).
To determine whether the stabilization of Ab aggregates formed in dilute HFIP was accompanied by morphological changes, we examined the aggregation reactions by EM and AFM. A solution of Ab(1-40) was prepared in freshly diluted 2% HFIP and aliquots were
removed at several time points, centrifuged at 18,000g
for 10 minutes, and examined by EM. After 1 hour of
incubation, clustered globular structures were the predominant negatively stained feature (Fig. 6B). Some
individual globular structures were also observed (Fig.
6B inset). The diameter of the globules extended over a
range of 20200 nm and punctate deposits could be
seen on several of the larger globules. The globules presumably represented HFIP microdroplets upon which
Ab had deposited and begun to aggregate. However,
based on their uorescence with thioavin T, these Ab

deposits differed from the albumin-coated microdroplets by remaining soluble after centrifugation at
18,000g. Two days later the clustered globular structures were incorporated into a mesh or lattice of berlike elements (Fig. 6C), and after 9 days more distinct
bers were apparent (Fig. 6D). These included long
bers from which many short bers branched and, very
rarely, structures resembling initial microdroplets
on which bers had formed (Fig. 6D, inset). After 23
days (Fig. 6E) the HFIP-induced bers appeared similar to control brils produced by elongation of Ab(1-40)
protobrils (Fig. 6F), but numerous short branches
remained that were not found on the control brils.
The bers at this time still remained soluble, as they
failed to sediment during 18,000g centrifugation.
AFM images supported and extended the features of
HFIP-induced Ab aggregates observed by EM. The
same mixture of clustered globular structures and individual globules was observed 1 hour after adding Ab(140) to dilute HFIP (Fig. 7A and inset). The heights of
the clustered structures typically ranged from 10
15 nm but in some cases extended to over 50 nm. After
23 days, bers with heights of 45 nm had appeared
and numerous very short rods with heights of 2.5
4 nm had emerged that were barely evident in the 1hour sample (Fig. 7B). These short rods were less clear
in the EM images, perhaps because of differences in
retention on the EM grid and AFM mica surfaces or

172

M.R. NICHOLS ET AL.

well-formed bers on spherical structures that resemble microdroplets (Fig. 6D, inset) suggest that some
reorganization of the initial b-structured aggregates
can occur directly on the microdroplet surface.

Fig. 7. AFM images of HFIP-induced Ab(1-40) aggregates also

show changes in morphology with time (Nichols et al., 2005a). Samples were taken at the indicated times from the same incubated solution of Ab(1-40) in 2% HFIP described in Figure 6 and diluted 25-fold
with 2% HFIP. Aliquots (100 ml) were applied directly to the mica surface. Images, presented in amplitude mode, are 2.5  2.5 mm, and
insets are on this same scale.

because they showed little negative staining. A 9-day

image showed a ber-like structure with several
branching arms (Fig. 7B, inset), consistent with the
branched bers observed by EM.
Interfaces between aqueous and nonpolar phases
promote Ab aggregation by concentrating amphiphilic
Ab monomers at the interface in a conformation that
favors formation of a nucleus. Polymerization then
occurs by addition of monomer from the aqueous phase
and/or from a monomer pool already associated at the
interface. The process by which the initial Ab aggregates formed on HFIP microdroplets convert to bers
may involve additional steps. Because of the initial
instability of these HFIP-induced aggregates, the
growth of bers may reect monomer dissociation from
the globular clusters and redeposition on more stable
nascent bers. However, the rare cases of relatively

Ab Aggregate Polymorphism
Since the interfacial Ab aggregates in this report differ from Ab protobrils in morphology and stability,
their molecular structures must differ at least in subtle
ways. One measure of this conformational specicity is
the efciency with which these aggregates acted as
nuclei or seeds for elongation by monomeric Ab
(ONuallain et al., 2004). HFIP-induced Ab aggregates
that had stabilized for 3 days and Ab(1-40) protobril
controls were diluted to equivalent concentrations
(
1 mM) and incubated with 60 mM Ab(1-40). The net
elongation rate of the protobrils as measured by thioavin T uorescence was almost 30 times faster than
that of the HFIP-induced Ab aggregates (Nichols et al.,
2005a). The difference in elongation rates was not due
to a difference in aggregate and protobril sizes, as the
RH values measured by DLS were similar. We concluded that the seeding efciency of the HFIP-induced
aggregates was much lower than that of protobrils, at
least in aqueous buffers without HFIP.
Further studies are required to determine the molecular interactions in these aggregates, but models of
amyloid brils suggest ways in which differences could
arise. In one structural model for Ab(1-40) based on
solid-state NMR data, a cross-b unit is composed of two
b-strand segments involving residues 12-24 and 30-40
separated by a segment with a bend angle of 1808,
allowing interpeptide hydrogen bonding and separate
parallel b-sheet formation from both b-strands (Tycko,
2003). In contrast, low-angle X-ray diffraction analysis
of Ab(11-25) brils indicated an antiparallel alignment
of fully extended b-strands (Sikorski et al., 2003). In
these brils the b-sheets appeared to stack by slipping
relative to each other by the length of two amino acid
units (Sikorski et al., 2003), and the registry of the
hydrogen bonding appeared to vary with pH (Tycko,
2003). It is possible then that Ab(1-40) protobrils and
HFIP-induced aggregates differ by their b-strand
alignments or the registry of their b-sheets or hydrogen
bonding, and solid-state NMR measurements will be
required to resolve this issue.
It will also be of interest to determine whether
agents or conditions that favor rapid interfacial formation of unstable Ab aggregates occur in vivo. Analogs of
biological membranes, namely, GM1 ganglioside
micelles and articial lipid rafts that contain GM1
(Kakio et al., 2002), as well as lipoprotein particles
(Chan et al., 1996) have been reported to bind Ab peptides in a saturable manner and to convert the peptide
conformations to b-structures in the complexes. The
stability of these aggregates has not been investigated.
A recent study has also found that inhaled anesthetics
promote Ab aggregation (Eckenhoff et al., 2004). The
haloalkane halothane and the haloether isourane,
like HFIP, are highly uorinated, and both increased
the rates of Ab aggregation as measured by thioavin
T uorescence. Furthermore, an EM image of the Ab
aggregates induced by halothane at concentrations

AMYLOID-b AGGREGATION AT NONPOLAR INTERFACES

achieved in routine clinical anesthesia (<1 mM) was

strikingly similar to those in Figure 6C,D here. The
authors suggested that cytotoxicity induced by these
aggregates may contribute to persistent postoperative
cognitive problems that occur in the elderly.
CONCLUSIONS
A number of investigators now propose that soluble
aggregates of Ab peptides play a major role in the
development of AD. The low levels of these aggregates
in vivo have prevented virtually any characterization
of their structural features. Insoluble Ab brils generated in vitro show variations in underlying molecular
structure that depend on the aggregation conditions
(Petkova et al., 2005), and in this report we suggest
that similar variations can arise among soluble Ab
aggregates. In particular, we demonstrate that soluble
Ab aggregates produced in buffer differ in structure
from those formed at polarnonpolar interfaces. While
both classes of soluble aggregates shared certain features with insoluble brils, including a predominance
of b-structure by CD, an enhanced uorescence of
bound thioavin T, and ber-like morphologies in EM
and AFM, they also showed important differences. The
interfacial aggregates disaggregated more rapidly and
showed morphologies that were distinct from the soluble protobrils produced in buffer. Furthermore, the
protobrils but not the interfacial aggregates readily
seeded Ab monomer deposition and grew into brils.
Further studies are needed to identify techniques that
will allow comparison of the structural features of soluble Ab aggregates produced in vivo with those of the in
vitro aggregates described here. Such a comparison
may inform us about the diversity of soluble in vivo
aggregates and their sites of formation and eventually
allow us to distinguish which soluble aggregates
induce the neurotoxicity that leads to AD.
ACKNOWLEDGMENTS
We thank Stephanie Cratic-McDaniel in the laboratory of Professor Jan H. Hoh at Johns Hopkins University for assistance in atomic force microscope imaging
of our samples, and Dr. Wen-Lang Lin at Mayo Clinic
Jacksonville for helpful suggestions on electron microscopy during the course of these studies.
REFERENCES
Balbach JJ, Ishi Y, Antzutkin ON, Leapman RD, Rizzo NW, Dyda F,
Reed J, Tycko R. 2000. Amyloid bril formation by Ab16-22, a sevenresidue fragment of the Alzheimers b-amyloid peptide, and structural characterization by solid state NMR. Biochemistry 39:13748
13759.
Burkoth TS, Benzinger TLS, Urban V, Morgan DM, Gregory DM,
Thiyagarajan P, Botto RE, Meredith SC, Lynn DG. 2000. Structure
of the b-amyloid(10-35) bril. J Am Chem Soc 122:78837889.
Chan W, Fornwald J, Brawner M, Wetzel R. 1996. Native complex formation between apolipoprotein E isoforms and the Alzheimers disease peptide Ab. Biochemistry 35:71237130.
Chien P, Weissman JS, DePace AH. 2004. Emerging principles of conformation-based prion inheritance. Annu Rev Biochem 73:617656.
Choo-Smith LP, Garzon-Rodriguez W, Glabe CG, Surewicz WK.
1997a. Acceleration of amyloid bril formation by specic binding
of Ab-(1-40) peptide to ganglioside-containing membrane vesicles.
J Biol Chem 272:2298722990.

173

Choo-Smith LP, Surewicz WK. 1997b. The interaction between Alzheimer amyloid b(1-40) peptide and ganglioside GM1-containing
membranes. FEBS Lett 402:9598.
Cohen AS, Calkins E. 1959. Electron microscopic observation on a
brous component in amyloid of diverse origins. Nature 183:1202
1203.
Dahlgren KN, Manelli AM, Stine Jr WB, Baker LK, Krafft GA, LaDu
MJ. 2002. Oligomeric and brillar species of amyloid-b peptides differentially affect neuronal viability. J Biol Chem 277:3204632053.
Eckenhoff RG, Johansson JS, Wei H, Carnini A, Kang B, Wei W, Pidikiti R, Keller JM, Eckenhoff MF. 2004. Inhaled anesthetic enhancement of amyloid-b oligomerization and cytotoxicity. Anesthesiology
101:703709.
Esler WP, Stimson ER, Ghilardi JR, Vinters HV, Lee JP, Mantyh PW,
Maggio JE. 1996. In vitro growth of Alzheimers disease b-amyloid
plaques displays rst-order kinetics. Biochemistry 35:749757.
Esler WP, Stimson ER, Jennings JM, Vinters HV, Ghilardi JR, Lee JP,
Mantyh PW, Maggio JE. 2000. Alzheimers disease amyloid propagation by a template-dependent dock-lock mechanism. Biochemistry 39:62886295.
Glenner GG, Wong CW. 1984. Alzheimers disease: initial report of
the purication and characterization of a novel cerebrovascular
amyloid protein. Biochem Biophys Res Commun 120:885890.
Goldsbury CS, Wirtz S, Muller SA, Sunderji S, Wicki P, Aebi U, Frey
P. 2000. Studies on the in vitro assembly of Ab 1-40: Implications
for the search for Ab bril formation inhibitors. J Struct Biol
130:217231.
Hardy JA, Higgins GA. 1992. Alzheimers disease: the amyloid cascade hypothesis. Science 256:184185.
Hardy J, Selkoe DJ. 2002. The amyloid hypothesis of Alzheimers disease: progress and problems on the road to therapeutics. Science
297:353356.
Harper JD, Wong SS, Lieber CM, Lansbury Jr PT. 1997. Observation
of metastable Ab amyloid protobrils by atomic force microscopy.
Chem Biol 4:119125.
Harper JD, Wong SS, Lieber CM, Lansbury Jr PT. 1999. Assembly of
Ab amyloid peptides: an in vitro model for a possible early event in
Alzheimers disease. Biochemistry 38:89728980.
Hartley DM, Walsh DM, Ye CP, Diehl T, Vasquez S, Vassilev PM,
Teplow DB, Selkoe DJ. 1999. Protobrillar intermediates of amyloid
b-protein induce acute electrophysiological changes and progressive
neurotoxicity in cortical neurons. J Neurosci 19:88768884.
Hock C, Konietzko U, Streffer JR, Tracy J, Signorell A, Muller-Tillmanns B, Lemke U, Henke K, Moritz E, Garcia E, Wollmer MA,
Umbricht D, de Quervain DJ, Hofmann M, Maddalena A, Papassotiropoulos A, Nitsch RM. 2003. Antibodies against b-amyloid slow
cognitive decline in Alzheimers disease. Neuron 38:547554.
Hsia A, Masliah E, McConlogue L, Yu G, Tatsuno G, Hu K, Kholodenko D, Malenka RC, Nicoll RA, Mucke L. 1999. Plaque-independent disruption of neural circuits in Alzheimers disease mouse models. Proc Natl Acad Sci U S A 96:32283233.
Huang TH, Yang DS, Plaskos NP, Go S, Yip CM, Fraser PE, Chakrabartty A. 2000. Structural studies of soluble oligomers of the Alzheimer beta-amyloid peptide. J Mol Biol 297:7387.
Jarrett JT, Lansbury Jr PT. 1993. Seeding one-dimensional crystallization of amyloid: a pathogenic mechanism in Alzheimers disease
and scrapie? Cell 73:10551058.
Jarrett JT, Berger EP, Lansbury Jr PT. 1993. The carboxy terminus of
the b amyloid protein is critical for the seeding of amyloid formation: implications for the pathogenesis of Alzheimers disease. Biochemistry 32:46934697.
Kakio A, Nishimoto S, Yanagisawa K, Kozutsumi Y, Matsuzaki K.
2002. Interactions of amyloid b-protein with various gangliosides in
raft-like membranes: importance of GM1 ganglioside-bound form
as an endogenous seed for Alzheimer amyloid. Biochemistry
41:73857390.
Kawarabayashi T, Shoji M, Younkin LH, Lin WL, Dickson DW, Murakami T, Matsubara E, Abe K, Ashe KH, Younkin SG. 2004. Dimeric
amyloid b protein rapidly accumulates in lipid rafts followed by apolipoprotein E and phosphorylated tau accumulation in the Tg2576
mouse model of Alzheimers disease. J Neurosci 24:38013809.
Kayed R, Head E, Thompson JL, McIntire TM, Milton SC, Cotman
CW, Glabe CG. 2003. Common structure of soluble amyloid oligomers implies common mechanism of pathogenesis. Science 300:486
489.
King CY, Diaz-Avalos R. 2004. Protein-only transmission of three
yeast prion strains. Nature 428:319323.
Klein WL, Krafft GA, Finch CE. 2001. Targeting small Ab oligomers:
the solution to an Alzheimers disease conundrum? Trends Neurosci
24:219224.

174

M.R. NICHOLS ET AL.

Lambert MP, Barlow AK, Chromy BA, Edwards C, Freed R, Liosatos

M, Morgan TE, Rozovsky I, Trommer B, Viola KL, Wals P, Zhang C,
Finch CE, Drafft GA, Klein WL. 1998. Diffusible, nonbrillar
ligands derived from Ab1-42 are potent central nervous system neurotoxins. Proc Natl Acad Sci U S A 95:64486453.
Lansbury Jr PT. 1997. Inhibition of amyloid formation: a strategy to
delay the onset of Alzheimers disease. Curr Opin Chem Biol 1:260
267.
Lansbury Jr PT, Costa PR, Grifths JM, Simon EJ, Auger M, Halverson KJ, Kocisko DA, Hendsch ZS, Ashburn TT, Spencer RG, Tidor
B, Grifn RG. 1995. Structural model for the b-amyloid bril based
on interstrand alignment of an antiparallel-sheet comprising a Cterminal peptide. Nat Struct Biol 2:990998.
Lashuel HA, Hartley DM, Petre BM, Wall JS, Simon MN, Walz T,
Lansbury Jr PT. 2003. Mixtures of wild-type and a pathogenic
(E22G) form of Ab40 in vitro accumulate protobrils, including
amyloid pores. J Mol Biol 332:795808.
LeVine H. 1993. Thioavine T interaction with synthetic Alzheimers
disease b-amyloid peptides: detection of amyloid aggregation in solution. Protein Sci 2:404410.
LeVine 3rd H. 2002. 4,40 -Dianilino-1,10 -binaphthyl-5,50 -disulfonate:
report on non-b-sheet conformers of Alzheimers peptide b(1-40).
Arch Biochem Biophys 404:106115.
Lomakin A, Chung DS, Benedek GB, Kirschner DA, Teplow DB.
1996. On the nucleation and growth of amyloid b-protein brils:
detection of nuclei and quantication of rate constants. Proc Natl
Acad Sci U S A 93:11251129.
Matsuzaki K, Horikiri C. 1999. Interactions of amyloid b-peptide(140) with ganglioside-containing membranes. Biochemistry 38:
41374142.
Miller DL, Papayannopoulos IA, Styles J, Bobin SA, Lin YY, Biemann
K, Iqbal K. 1993. Peptide compositions of the cerebrovascular and
senile plaque core amyloid deposits of Alzheimers disease. Arch
Biochem Biophys 301:4152.
Naiki H, Nakakuki K. 1996. First-order kinetic model of Alzheimers
b-amyloid bril extension in vitro. Lab Invest 74:374383.
Nichols MR, Moss MA, Reed DK, Lin W-L, Mukhopadhyay R, Hoh
JH, Rosenberry TL. 2002. Growth of b-amyloid(1-40) protobrils by
monomer elongation and lateral association. Characterization of
distinct products by light scattering and atomic force microscopy.
Biochemistry 41:61156127.
Nichols MR, Moss MA, Reed DK, Cratic-McDaniel S, Hoh JH, Rosenberry TL. 2005a. Amyloid-b protobrils differ from amyloid-b aggregates induced in dilute hexauoroisopropanol in stability and morphology. J Biol Chem 280:24712480.
Nichols MR, Moss MA, Reed DK, Hoh JH, Rosenberry TL. 2005b.
Rapid assembly of amyloid-b peptide at a liquid/liquid interface produces unstable b-sheet bers. Biochemistry 44:165173.
Nilsberth C, Westlind-Danielsson A, Eckman CB, Condron MM, Axelman K, Forsell C, Stenh C, Luthman J, Teplow DB, Younkin SG,
Naslund J, Lannfelt L. 2001. The Arctic APP mutation (E693G)
causes Alzheimers disease by enhanced Ab protobril formation.
Nat Neurosci 4:887893.
ONuallain B, Wetzel R. 2002. Conformational Abs recognizing a
generic amyloid bril epitope. Proc Natl Acad Sci U S A 99:1485
1490.
ONuallain B, Williams AD, Westermark P, Wetzel R. 2004. Seeding
specicity in amyloid growth induced by heterologous brils. J Biol
Chem 279:1749017499.
Padrick SB, Miranker AD. 2002. Islet amyloid: phase partitioning
and secondary nucleation are central to the mechanism of brillogenesis. Biochemistry 41:46944703.
Petkova AT, Buntkowsky G, Dyda F, Leapman RD, Yau WM, Tycko R.
2004. Solid state NMR reveals a pH-dependent antiparallel b-sheet
registry in brils formed by a b-amyloid peptide. J Mol Biol
335:247260.
Petkova AT, Leapman RD, Guo Z, Yau WM, Mattson MP, Tycko R.
2005. Self-propagating, molecular-level polymorphism in Alzheimers b-amyloid brils. Science 307:262265.
Pratt LR, Pohorille A. 2002. Hydrophobic effects and modeling of biophysical aqueous solution interfaces. Chem Rev 102:26712692.
Schenk D. 2002. Opinion: amyloid-b immunotherapy for Alzheimers
disease: the end of the beginning. Nat Rev Neurosci 3:824828.
Schenk D, Barbour R, Dunn W, Gordon G, Grajeda H, Guido T, Hu K,
Jiping-Huang, Johnson-Wood K, Khan K, Kholodenko D, Lee M,
Liao Z, Lieberburg I, Motter R, Mutter L, Soriano F, Shopp G, Vasquez N, Vandevert C, Walker S, Wogulis M, Yednock T, Games D,

Seubert P. 1999. Immunization with amyloid-b attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature
400:173177.
Schladitz C, Vieira EP, Hermel H, Mohwald H. 1999. Amyloid-b-sheet
formation at the air-water interface. Biophys J 77:33053310.
Selkoe DJ. 2001. Alzheimers disease: genes, proteins, and therapy.
Physiol Rev 81:741766.
Selkoe DJ, Podlisny MB. 2002. Deciphering the genetic basis of Alzheimers disease. Annu Rev Genomics Hum Genet 3:6799.
Serpell LC, Sunde M, Benson MD, Tennent GA, Pepys MB, Fraser
PE. 2000. The protolament substructure of amyloid brils. J Mol
Biol 300:10331039.
Sikorski P, Atkins ED, Serpell LC. 2003. Structure and texture of
brous crystals formed by Alzheimers Ab(11-25) peptide fragment.
Structure (Camb) 11:915926.
Sunde M, Serpell LC, Bartlam M, Fraser PE, Pepys MB, Blake CC.
1997. Common core structure of amyloid brils by synchrotron Xray diffraction. J Mol Biol 273:729739.
Tanaka M, Chien P, Naber N, Cooke R, Weissman JS. 2004. Conformational variations in an infectious protein determine prion strain
differences. Nature 428:323328.
Teplow DB. 1998. Structural and kinetic features of amyloid b-protein
brillogenesis. Amyloid: Int J Exp Clin Invest 5:121142.
Terry RD. 1985. Textbook of Neuropathology. Baltimore: Williams &
Wilkins. p 824841.
Terry RD, Gonatas NK, Weiss M. 1964. Ultrastructural studies in Alzheimers presenile dementia. Am J Pathol 44:269297.
Terzi E, Holzemann G, Seelig J. 1997. Interaction of Alzheimer b-amyloid peptide(1-40) with lipid membranes. Biochemistry 36:14845
14852.
Tjernberg LO, Pramanik A, Bjorling S, Thyberg P, Thyberg J, Nordstedt C, Berndt KD, Terenius L, Rigler R. 1999. Amyloid b-peptide
polymerization studied using uorescence correlation spectroscopy.
Chem Biol 6:5362.
Tseng BP, Esler WP, Clish CB, Stimson ER, Ghilardi JR, Vinters HV,
Mantyh PW, Lee JP, Maggio JE. 1999. Deposition of monomeric, not
oligomeric, Ab mediates growth of Alzheimers disease amyloid
plaques in human brain preparations. Biochemistry 38:1042410431.
Tycko R. 2003. Insights into the amyloid folding problem from solidstate NMR. Biochemistry 42:31513159.
Tycko R. 2004. Progress towards a molecular-level structural understanding of amyloid brils. Curr Opin Struct Biol 14:96103.
Walsh DM, Lomakin A, Benedek GB, Condron MM, Teplow DB. 1997.
Amyloid b-protein brillogenesis: detection of a protobrillar intermediate. J Biol Chem 272:2236422372.
Walsh DM, Hartley DM, Kusumoto Y, Fezoui Y, Condron MM, Lomakin A, Benedek GB, Selkoe DJ, Teplow DB. 1999. Amyloid b-protein
brillogenesis: structure and biological activity of protobrillar
intermediates. J Biol Chem 274:2594525952.
Walsh DM, Klyubin I, Fadeeva JV, Cullen WK, Anwyl R, Wolfe MS,
Rowan MJ, Selkoe DJ. 2002. Naturally secreted oligomers of amyloid b protein potently inhibit hippocampal long-term potentiation
in vivo. Nature 416:535539.
Westerman MA, Cooper-Blacketer D, Mariash A, Kotilinek L, Kawarabayashi T, Younkin LH, Carlson GA, Younkin SG, Ashe KH. 2002.
The relationship between Ab and memory in the Tg2576 mouse
model of Alzheimers disease. J Neurosci 22:18581867.
Wood SJ, Maleeff B, Hart T, Wetzel R. 1996. Physical, morphological
and functional differences between pH 5.8 and 7.4 aggregates of the
Alzheimers amyloid peptide Ab. J Mol Biol 256:870877.
Yankner BA. 1996. New clues to Alzheimers disease: unraveling the
roles of amyloid and tau. Nat Med 2:850852.
Yip CM, McLaurin J. 2001. Amyloid-b peptide assembly: a critical
step in brillogenesis and membrane disruption. Biophys J
80:13591371.
Yong W, Lomakin A, Kirkitadze MD, Teplow DB, Chen SH, Benedek
GB. 2001. Structure determination of micelle-like intermediates in
amyloid b-protein bril assembly by using small angle neutron scattering. Proc Natl Acad Sci U S A 99:150154.
Yoshida K, Yamaguchi T, Adachi T, Otomo T, Matsuo D, Takamuku T,
Nishi N. 2003. Structure and dynamics of hexauoroisopropanolwater mixtures by x-ray diffraction, small-angle neutron scattering,
NMR spectroscopy, and mass spectrometry. J Chem Phys 119:6132
6142.
Zagorski MG, Yang J, Shao H, Ma K, Zeng H, Hong A. 1999. Methodological and chemical factors affecting amyloid beta peptide amyloidogenicity. Methods Enzymol 309:189204.