The Off-rate of Monomers Dissociating from Amyloid-β Protofibrils*

Background: Protofibrils of the amyloid-β peptide (Aβ) are neurotoxic oligomers implicated in development and progression of Alzheimer disease. Results: The dissociation of Aβ protofibrils into their monomeric subunits is a slow process, occurring on the time scale of hours. Conclusion: Aβ protofibrils possess a high kinetic stability toward dissociation into monomers. Significance: The longevity of Aβ protofibrils permits sustained toxic effects. The interconversion of monomers, oligomers, and amyloid fibrils of the amyloid-β peptide (Aβ) has been implicated in the pathogenesis of Alzheimer disease. The determination of the kinetics of the individual association and dissociation reactions is hampered by the fact that forward and reverse reactions to/from different aggregation states occur simultaneously. Here, we report the kinetics of dissociation of Aβ monomers from protofibrils, prefibrillar high molecular weight oligomers previously shown to possess pronounced neurotoxicity. An engineered binding protein sequestering specifically monomeric Aβ was employed to follow protofibril dissociation by tryptophan fluorescence, precluding confounding effects of reverse or competing reactions. Aβ protofibril dissociation into monomers follows exponential decay kinetics with a time constant of ∼2 h at 25 °C and an activation energy of 80 kJ/mol, values typical for high affinity biomolecular interactions. This study demonstrates the high kinetic stability of Aβ protofibrils toward dissociation into monomers and supports the delineation of the Aβ folding and assembly energy landscape.


The interconversion of monomers, oligomers, and amyloid fibrils of the amyloid-␤ peptide (A␤) has been implicated in the pathogenesis of Alzheimer disease. The determination of the kinetics of the individual association and dissociation reactions is hampered by the fact that forward and reverse reactions to/from different aggregation states occur simultaneously.
Here, we report the kinetics of dissociation of A␤ monomers from protofibrils, prefibrillar high molecular weight oligomers previously shown to possess pronounced neurotoxicity. An engineered binding protein sequestering specifically monomeric A␤ was employed to follow protofibril dissociation by tryptophan fluorescence, precluding confounding effects of reverse or competing reactions. A␤ protofibril dissociation into monomers follows exponential decay kinetics with a time constant of ϳ2 h at 25°C and an activation energy of 80 kJ/mol, values typical for high affinity biomolecular interactions. This study demonstrates the high kinetic stability of A␤ protofibrils toward dissociation into monomers and supports the delineation of the A␤ folding and assembly energy landscape.
Amyloid fibrils composed of amyloid-␤ peptide (A␤) 3 are the main protein component of senile plaques found in the brains of Alzheimer disease patients (1,2). Genetics, cell culture studies, and animal models support a critical role of A␤ in Alzheimer disease pathogenesis, with the 42-amino acid A␤42 peptide being more aggregation-prone and neurotoxic than the more common 40-amino acid A␤40 variant. A␤ forms different types of prefibrillar oligomers, which, according to several lines of evidence, include the most toxic A␤ species (3,4). Neurotoxic high molecular weight oligomers referred to as protofibrils or A␤-derived diffusible ligands have been purified by size exclusion chromatography (SEC) of A␤ incubations (4 -10). A␤ protofibrils are heterogeneous in size (in the range of 50 -1500 kDa) and morphology, comprising spherical, annular, and curvilinear assemblies. The protofibrils are in equilibrium with both monomers and fibrils (7). The dynamics of the monomerprotofibril-fibril equilibrium is crucial for A␤ toxicity, as (i) the different association states exhibit different toxicities (2,3), and (ii) toxicity emanates from the polymerization reaction itself (11). It is thus valuable to gain kinetic information on the individual interconversion steps. Knowledge of the kinetic stability of the involved species furthermore aids in the delineation of the amyloid folding and assembly energy landscape (12,13).
In this study, we determined the off-rate of monomers dissociating from A␤ protofibrils by tryptophan fluorescence using a tryptophan-containing variant of the engineered A␤-binding protein ZA␤3. ZA␤3, which has been selected previously from an Affibody protein library, specifically binds monomeric A␤ with an affinity of K d ϭ 17 nM (14,15). ZA␤3 is a dimer of two identical subunits composed of 58 amino acids covalently linked by a disulfide bond. ZA␤3 inhibits A␤ oligomerization and aggregation by sequestering the aggregation-prone central and C-terminal sequence regions of monomeric A␤ (15,16). The addition of an excess of the tryptophan-containing variant ZA␤3W to SEC-purified A␤ protofibrils enabled the detection of monomeric A␤ dissociating from protofibrils while preventing the reverse reaction from monomers to protofibrils as well as the reaction to amyloid fibrils. The temperature-dependent kinetic stability of A␤ protofibrils toward dissociation into monomers could thus be assessed.

EXPERIMENTAL PROCEDURES
Cloning of ZA␤3W-Site-directed mutagenesis was achieved by back-to-back primer PCR (17). Plasmid pAY442 containing the ZA␤3 gene (14) was amplified with phosphorylated primers, one of which carried the mutation for the Y18W exchange at the 5Ј-end. The vector was religated, and the mutation was verified by sequencing.
Expression and Purification of ZA␤3W-Expression and purification were done as described previously (18) with minor modifications. For cell lysis, high pressure (2.9 kilobars; Constant Systems) was used. After affinity chromatography, SEC (Superdex 75 16/60, GE Healthcare) was performed in 20 mM sodium phosphate and 50 mM NaCl (pH 7.2).
Expression, Purification, and Protofibril Formation of A␤-A␤40 and A␤42 were produced with an N-terminal methionine by recombinant coexpression with ZA␤3 (19). For protein expression, unlabeled M9 medium without Celtone was used. The cell pellet was resuspended in 50 mM sodium phosphate, 200 mM NaCl, and 20 mM imidazole (pH 8.0). For cell disruption, high pressure (2.9 kilobars) was used. Affinity chromatography was performed on an Ä KTA Purifier system using a 5-ml HisTrap FF column (GE Healthcare). A␤ was eluted from the ZA␤3-A␤ complex bound on the column with 8 M urea and 20 mM sodium phosphate (pH 7.6). The eluted A␤ was washed over a 1-ml HisTrap FF column equilibrated in urea to remove any residual ZA␤3. Buffer exchange to 20 mM sodium phosphate (pH 7.0) was achieved by SEC (Superdex 75 16/60). The pH was adjusted to 10 to prevent aggregation during storage (4°C) and during concentration (Vivaspin 20 3000 MWCO PES, Sartorius AG). For preparation of protofibrils, the pH was titrated back to 7. At room temperature, A␤42 was incubated at a concentration of ϳ100 M for 1-4 h, whereas A␤40 was incubated at a concentration of ϳ500 M for 24 h. The formation of protofibrils was monitored by analytical SEC runs (Superdex 75 10/300; 20 mM sodium phosphate (pH 7.0)). Protofibrils were purified by SEC and immediately employed in the fluorescence assay. The A␤ concentration of the freshly eluted protofibril fractions was determined by UV absorption at 280 nm. A protofibril batch referred to in this study is defined as one particular protofibril preparation starting from an individual A␤ expression cell pellet. To test fibril dissociation, amyloid fibrils were prepared by incubation of 650 l of 380 M A␤42 in 20 mM sodium phosphate (pH 7.0) and 0.03% sodium azide for 1 week at 37°C in a 2-ml glass vial, with stirring at 300 rpm using a micro stir bar.
Fluorescence Assay-Protofibril dissociation samples with a volume of ϳ500 l were prepared in 5 ϫ 5-mm fluorescence cells (101.016-QS, Hellma) and sealed with Parafilm. An airwater interface was present at the top of the solutions. The samples were equipped with crown magnetic stirring bars (Jasco) and stirred at 500 rpm. A␤ concentrations between 4 and 35 M were employed in the dissociation assay, and ZA␤3W was added at an excess of ϳ25% to ensure removal of dissociated A␤ monomers from the monomer-protofibril equilibrium. The time elapsed between protofibril elution from SEC and the addition of ZA␤3W was ϳ15 min.
Tryptophan fluorescence was excited at 295 nm, and emission spectra were recorded from 330 to 360 nm on a Jasco FP-6500 spectrofluorometer. The excitation and emission bandwidths were 5 and 1 nm, respectively, and the data pitch was 0.2 nm. The fluorescence was repeatedly measured over ϳ12 h, during which time the sample temperature was kept constant using external water-jacketed cell holders. For fluorescence measurements, cells were transferred to the spectrofluorometer cell holder, which was thermostatted at 25°C.
The max values of the fluorescence emission spectra were determined by fitting the fluorescence intensity to an empirical fitting function (1) using a trust region reflective algorithm implemented in MATLAB 2011 (MathWorks) (Equation 1), where F 0 is the fluorescence intensity offset, F 0 ϩ ⌬F is the fluorescence intensity at max , and ⌫ is the peak width. The fraction of free ZA␤3W was determined from max as explained under "Results," employing the fit shown in Fig. 1D and complying with Equation 2.
The fraction of free ZA␤3W dependent on incubation time was plotted, and a monoexponential fit to Equation 3 was employed to obtain individual apparent k off values for each dissociation experiment using OriginPro 8.6 (OriginLab) (Equation 3).
A global fit of all A␤42 protofibril dissociation data sets recorded at 25°C to a triexponential decay with individual amplitudes A 1 , A 2 , and A 3 and offset y 0 but shared rate constants k off,1 , k off,2 , and k off,3 was performed according to Equation 4.
The fraction of A␤ in protofibrils was calculated according to Equation 5.
Transmission Electron Microscopy-A␤ samples were diluted to 1 M and incubated for 3 min on a Formvar/carbon-coated copper grid (S162, Plano). The grid was washed three times with H 2 O and one time with 2% aqueous uranyl acetate before incubation for 1 min with 2% aqueous uranyl acetate for negative staining, followed by drying overnight. The samples were examined with a Libra 120 electron microscope (Zeiss) operating at 120 kV.
Analytical Ultracentrifugation-Analytical ultracentrifugation (AUC) was performed in an Optima XL-A analytical ultracentrifuge (Beckman Coulter) with absorbance optics using an An-60 Ti rotor with aluminum 2-channel centerpiece cells. Sedimentation velocity centrifugation was done at 40,000 rpm and 20°C. The intensity at 230 nm was recorded (radial resolution, 0.03 cm; continuous mode; no replicate; one scan/3 min). The run duration was 5 h for the protofibril sample and 10 h for the samples containing ZA␤3W. Data were fitted using the continuous distribution (c(s)) Lamm equation model with a v of 0.738 cm 3 /g based on the A␤42 sequence for the protofibril sample and a v of 0.721 cm 3 /g based on the ZA␤3W sequence for the ZA␤3W samples in the software package Sedfit. An s-value resolution of 0.3 or 0.025 S was chosen for the fit of the protofibril sample or the ZA␤3W samples, respectively. The quality of the fits was confirmed by low root mean square deviations below A 230 ϭ 0.005. Specified amounts of species in the protofibril sample resulted from the c(s) distribution exclusive of the area below 0.6 S, which contains a base-line deconvolution artifact. The s-values determined were corrected for water at 20°C.

ZA␤3W Is a Tryptophan Fluorescence Probe for Monomeric
A␤-Binding of ZA␤3 to A␤ is accompanied by a decrease in tyrosine fluorescence (18), presumably due to altered fluorescence properties of Tyr-18 in both ZA␤3 subunits, which are located at the binding interface. To exploit the higher extinction coefficient, higher quantum yield, and greater environment sensitivity of tryptophan compared with tyrosine, a ZA␤3 variant termed ZA␤3W was generated by site-directed mutagenesis of Tyr-18 to Trp-18 (Fig. 1A). ZA␤3W bound monomeric A␤40 with an affinity of K d ϭ 20 nM as determined by isothermal titration calorimetry, which is close to the value of K d ϭ 17 nM obtained for ZA␤3 (15). Thus, the Y18W mutation did not significantly affect the affinity. The fluorescence emission spectrum of free ZA␤3W had a maximum at 348 nm (Fig. 1C), indicative of water exposure of the tryptophan side chains (20). Upon A␤ binding, the fluorescence intensity increased, and the emission maximum was blueshifted to a wavelength of max ϭ 340 nm, in agreement with a less polar environment of the tryptophan indole groups in the bound state. Because of the spectral differences between its free and bound states, ZA␤3W could be employed to detect and quantify monomeric A␤ by sequestering it into the ZA␤3W-A␤ complex. Both the fluorescence intensity and max could in principal be evaluated for this purpose. However, the fluorescence intensity decreased upon repeated measurements of free and bound ZA␤3W, indicative of photobleaching, whereas max remained constant. Therefore, max was chosen as the spectral property to evaluate for the detection of monomeric A␤. However, max does not linearly depend on the fraction of free/bound ZA␤3W (21). To derive the relationship between max and the fraction of free/bound ZA␤3W, the simulated spectra of mixtures of free and bound ZA␤3W were calculated from the spectra of free and bound ZA␤3W, and the resulting max values dependent on the fraction of free ZA␤3W were plotted (Fig. 1D). A fit to an exponential function was employed for the calculation of the fraction of free ZA␤3W from experimentally determined max values. In conclusion, ZA␤3W addition to test solutions permits the detection and quantification of monomeric A␤. In particular, the concentration of monomeric A␤ in a test solution corresponds to the concentration of bound ZA␤3W if an excess of ZA␤3W is added, ensuring a concentration of free ZA␤3W of [ZA␤3W] free Ͼ Ͼ K d ϭ 20 nM.
Exponential Decay Kinetics of A␤ Protofibril Dissociation-ZA␤3W was employed to study the dissociation kinetics of A␤ protofibrils as schematically shown in Fig. 1B. Monomers dissociating from protofibrils were captured by ZA␤3W. The ZA␤3W binding kinetics were sufficiently fast (time constant on the order of seconds (18)) not to obscure the slow protofibril dissociation kinetics. Sequestration of monomeric A␤ by ZA␤3W effectively inhibited fibril formation, which otherwise is a competing reaction of protofibril dissociation (11,16). Under the experimental conditions, the decrease in the fraction of free ZA␤3W was proportional to the decrease in the fraction of A␤ within protofibrils and to the amount of A␤ monomers dissociated from protofibrils. A␤ protofibrils were prepared by SEC of incubated samples of monomeric A␤ as described previously (4 -8, 10, 16). Fresh samples of monomeric A␤42 or A␤40 were incubated at 25°C in 20 mM sodium phosphate (pH 7.0), followed by injection of the solution onto a Superdex 75 10/300 SEC column and isolation of the protofibril fraction, which eluted close to the void volume of the column (Fig. 2A). The protofibrils exhibited mainly curvilinear but also spherical and annular morphologies in transmission electron microscopy (Fig. 2B), in agreement with previous studies (4 -9). AUC of the A␤42 protofibril fraction detected a monomer content of ϳ30% of total A␤, in accordance with dissociation of protofibrils on the time scale of AUC sample preparation and measurement. Of the remaining A␤, the majority (ϳ50% of total A␤) could be resolved by sedimentation velocity AUC at 40,000 rpm and 20°C, yielding a size distribution from 2 to 13 S with an average s 20,w ϭ 7.6 S (Fig.  2C). These values are in agreement with literature AUC data for the high molecular weight SEC fraction of A␤42 (10). An additional A␤ fraction (ϳ20% of total A␤) sedimented too fast for reliable determination of s-values, possibly comprising larger protofibrils as well as mature amyloid fibrils.
Protofibril dissociation samples containing between 4 and 35 M A␤ were prepared in fluorescence cells. The time of A␤ (16 -40) is shown in orange. B, scheme of protofibril dissociation monitored by binding of A␤ monomers to ZA␤3W. C, fluorescence emission spectra of ZA␤3W in the absence (red) and presence (blue) of a stoichiometric amount of A␤42 monomers and spectra of free A␤42 (magenta) and buffer (black). D, max of fluorescence emission spectra of simulated mixtures of free and bound ZA␤3W (F). The line represents a fit to an exponential function employed to calculate the fraction of free ZA␤3W from experimentally determined max values.
ZA␤3W addition was taken as the starting time of the protofibril dissociation reaction. Fluorescence emission spectra showed a blue shift of the tryptophan fluorescence with time, reporting the binding of ZA␤3W to monomeric A␤ dissociated from protofibrils on the time scale of minutes to hours (Fig. 3A). This is in agreement with the appearance of an A␤ monomer peak on this time scale after reinjection of the protofibril fraction onto an SEC column (Fig. 3B) as reported previously (16,22). At the end of the dissociation experiments, protofibrils or other protein aggregates were not detected by transmission electron microscopy (Fig. 3C). In contrast, a protofibril sample incubated for the same time without ZA␤3W contained amyloid fibrils. AUC performed at the end of protofibril dissociation experiment showed that protofibrils were dissociated and that A␤ was bound to ZA␤3W (Fig. 3E).
In contrast to protofibrils, A␤ amyloid fibrils did not cause a significant blue shift of ZA␤3W tryptophan fluorescence after prolonged incubation (Fig. 3D). This is in agreement with previous data obtained by NMR-detected binding of ZA␤3 to A␤, which showed a very high kinetic stability of A␤ fibrils prepared in vitro (16).
For the kinetic analysis of protofibril dissociation, the fraction of free ZA␤3W was calculated from the max values of the fluorescence emission spectra. The time traces of the fraction of free ZA␤3W revealed that protofibril dissociation was a rather slow reaction, occurring on the time scale of minutes to hours (Fig. 4). The dissociation kinetics were not affected by agitation, as expected for a dissociation reaction occurring in solution (Fig. 4). The time traces could be fitted to a monoexponential decay function, providing individual k off values for each protofibril dissociation sample (Figs. 4A and 5). The fits did not cover the complete amplitude of the decay of the fraction of A␤ in protofibrils, explainable by (i) the dissociation of protofibrils during the time elapsed between protofibril elution from SEC and addition of ZA␤3W and (ii) the fast dissociation of a fraction of A␤ monomers loosely bound to the protofibril surface, which has been detected before by dark-state exchange saturation transfer NMR (23).
The average k off at 25°C determined from 12 A␤42 protofibril dissociation experiments of six independent protofibril batches was (1.4 Ϯ 0.7) ϫ 10 Ϫ4 s Ϫ1 . The k off values obtained for different protofibril batches varied considerably, ranging from 0.8 ϫ 10 Ϫ4 to 3.5 ϫ 10 Ϫ4 s Ϫ1 (Figs. 4A and 5). The k off value did not correlate with the A␤ concentration in the protofibril preparation (Fig. 5). k off did also not vary with the protofibril maturation time, as protofibril fractions harvested at two different times from the same A␤ incubation did not exhibit different dissociation kinetics within the error of the experiment (Figs. 4 and 5). One possible explanation for the heterogeneity in k off values would be a high sensitivity of the energy barrier of a defined rate-limiting step of protofibril dissociation to subtle differences in the solution conditions of different protofibril preparations. Alternatively, protofibril batches might be variably composed of a set of protofibrillar structures, each associated with a distinct individual energy barrier of the rate-limiting step of dissociation. A global fit of the 12 A␤42 protofibril dissociation data sets recorded at 25°C to a triexponential decay with shared rate constants performed similarly to the individual monoexponential fits, yielding the rate constants k off,1 ϭ 0.77 ϫ 10 Ϫ4 s Ϫ1 , k off,2 ϭ 1.5 ϫ 10 Ϫ4 s Ϫ1 , and k off,3 ϭ 3.5 ϫ 10 Ϫ4 s Ϫ1 (Fig. 4B). Our data are thus compatible with the existence of a set of protofibrillar structures of different kinetic stabilities, with variable distribution in different protofibril batches.
A␤40 and A␤42 Protofibrils Possess Similar Kinetic Stabilities-The k off values of three protofibril dissociation experiments of an A␤40 protofibril batch were within the range of those observed for A␤42 protofibrils, with an average k off at 25°C of 1.2 ϫ 10 Ϫ4 s Ϫ1 for A␤40 (Fig. 5). A␤40 and A␤42 protofibrils thus possess similar kinetic stabilities.
Activation Energy of Protofibril Dissociation-The temperature dependence of the protofibril dissociation kinetics was investigated in the range of 19 -37°C for three independent A␤42 protofibril preparations (Fig. 5). The k off values were obtained from fits to a monoexponential decay function. The ln(k off ) versus 1/T plots could be fit linearly, in concordance with Arrhenius law behavior in this temperature range. The activation energy was obtained from the slope of the linear fits, yielding values between 75 and 89 kJ/mol, with an average of 80 kJ/mol.

DISCUSSION
This study introduces a fluorescence assay for the determination of A␤ protofibril dissociation kinetics, employing the engineered binding protein ZA␤3W. ZA␤3W is uniquely suited for this purpose because it requires the hydrophobic central region of A␤ comprising residues 17-36 to be accessible to coupled folding-binding (15,18). This region is buried in the hydrophobic core of A␤ oligomers and amyloid fibrils (24 -27), implying that ZA␤3W specifically binds monomeric A␤, a prerequisite for the dissociation assay.
The results from this study reveal that the dissociation of monomers from A␤ protofibrils follows exponential decay kinetics. This is remarkable, inasmuch as the protofibril fraction obtained from SEC includes particles of different sizes, with spherical as well as (curvi)linear morphologies. Exponential decay behavior is not in agreement with a rate-limiting role of processes occurring exclusively at the end of (curvi)linear assemblies. However, it would be compatible with dissociation of (curvi)linear assemblies, e.g. (i) by rate-limiting dissociation of subunits at a rate that is independent of the subunit position within the linear assembly (Fig. 6A) or (ii) through a pre-equilibrium of linear with nonlinear assemblies and a rate-limiting dissociation of the latter into monomers (Fig. 6B). Interestingly, both models require the postulation of an A␤ oligomer as an intermediate unit through which protofibril dissociation occurs. A␤ oligomers potentially fulfilling this role have been identified before, e.g. by ion mobility coupled with mass spectrometry (28). Exponential dissociation kinetics have also been reported for prefibrillar oligomers of the SH3 domain of PI3K (29).
A␤ protofibril dissociation is a slow process, with a time constant of ϳ2 h at 25°C. The low rate of dissociation corresponds to a high free energy barrier that has to be crossed in the ratelimiting step. The energy barrier has a considerable enthalpic component as evidenced by the high activation energy of dissociation of ϳ80 kJ/mol, a value typical for high affinity biomolecular interactions (30,31). This indicates that a significant number of interactions have to be broken to reach the transition state. A␤ protofibrils contain extended H-bonded and ␤-sheet structure as detected by hydrogen-deuterium exchange and CD spectroscopy (7,32,33). Destruction of this structure can be expected to be associated with a considerable energy barrier, explaining the slow kinetics and high activation energy of protofibril dissociation.
The apparent k off values of different protofibril batches varied considerably, indicating the existence of distinct protofibrillar structures with different kinetic stabilities. This is reminiscent of prion/amyloid strains, protein aggregates of different conformation and thermodynamic stabilities that are associated with different phenotypes (34,35).
Formation of protofibrils by A␤40 requires higher monomer concentrations and longer incubation times compared with A␤42, reflecting a marked difference in the association kinetics (4). In contrast, the dissociation kinetics of A␤40 and A␤42 protofibrils provided here are similar. This indicates that the main difference between A␤40 and A␤42 is the lower monomer solubility and increased oligomerization of the latter (36), whereas the kinetic stabilities of the formed protofibrils are alike.
In contrast to protofibrils, A␤ amyloid fibrils prepared in vitro do not dissociate effectively in the presence of ZA␤3 over weeks (16). This is in line with an even higher kinetic stability of fibrils compared with protofibrils, probably due to a more ordered and expanded ␤-sheet core (32).
The data gained by the ZA␤3W sequestration approach can be compared with previous data for the dissociation of different A␤ aggregates obtained from other techniques. The dissociation of radiolabeled A␤40 deposited onto an amyloid fibril template occurred with dissociation half-times of ϳ10 min or Ͼ Ͼ1000 min depending on the deposition time, providing evi-  . Models of A␤ protofibril dissociation into monomers compatible with exponential decay kinetics. A, rate-limiting dissociation of protofibril subunits at a rate that is independent of the subunit position within the linear assembly. B, pre-equilibrium of linear and nonlinear protofibrils and rate-limiting dissociation of the latter into monomers. For illustration purposes, only one of the protofibril subunits (shown in magenta) dissociates. The indicated dissociation time constant () was determined for A␤42 protofibrils at 25°C. dence for two different A␤ association states, a docked and a locked state (37). Dissociation of the weakly bound, docked A␤ is approximately an order of magnitude faster than protofibril dissociation. A␤40 fibrils released soluble species at a rate of ϳ1 ϫ 10 Ϫ4 s Ϫ1 as detected by two-color coincidence detection fluorescence (38), a value in the range of the rate of A␤ protofibril dissociation reported here. Analysis of hydrogen-deuterium exchange of A␤ fibrils employing a model of a recycling mechanism resulted in comparatively fast monomer off-rates of 0.6 ϫ 10 Ϫ2 and 1.0 ϫ 10 Ϫ2 s Ϫ1 for A␤40 and A␤42 fibrils, respectively (39). The comparability of the data sets is limited, however, inasmuch as ZA␤3W sequestration reports on the appearance of monomers, whereas the other techniques detect all soluble particles or hydrogen-deuterium exchangeable species.
This study demonstrates the applicability of a binding molecule obtained by protein engineering for the characterization of a key intermediate in amyloid formation. ZA␤3W can further be used to identify conditions and compounds that modulate A␤ aggregation, for example by destabilizing soluble oligomers (40).