The Peculiar Role of the A2V Mutation in Amyloid-β (Aβ) 1–42 Molecular Assembly*

Background: A2V mutation is associated with early onset AD-type dementia in homozygous individuals. Results: A2V mutation leads to a peculiar kinetics of Aβ oligomerization. Conclusion: The Aβ N-terminal region plays an important role in the molecular assembly. Significance: in the homozygous condition the A2V mutation led to aggregation, whereas in the heterozygous state the evolution and kinetics of the aggregation process was hindered. We recently reported a novel Aβ precursor protein mutation (A673V), corresponding to position 2 of Aβ1–42 peptides (Aβ1–42A2V), that caused an early onset AD-type dementia in a homozygous individual. The heterozygous relatives were not affected as an indication of autosomal recessive inheritance of this mutation. We investigated the folding kinetics of native unfolded Aβ1–42A2V in comparison with the wild type sequence (Aβ1–42WT) and the equimolar solution of both peptides (Aβ1–42MIX) to characterize the oligomers that are produced in the early phases. We carried out the structural characterization of the three preparations using electron and atomic force microscopy, fluorescence emission, and x-ray diffraction and described the soluble oligomer formation kinetics by laser light scattering. The mutation promoted a peculiar pathway of oligomerization, forming a connected system similar to a polymer network with hydrophobic residues on the external surface. Aβ1–42MIX generated assemblies very similar to those produced by Aβ1–42WT, albeit with slower kinetics due to the difficulties of Aβ1–42WT and Aβ1–42A2V peptides in building up of stable intermolecular interaction.

We recently reported a novel A␤ precursor protein mutation (A673V), corresponding to position 2 of A␤1-42 peptides (A␤1-42 A2V ), that caused an early onset AD-type dementia in a homozygous individual. The heterozygous relatives were not affected as an indication of autosomal recessive inheritance of this mutation. We investigated the folding kinetics of native unfolded A␤1-42 A2V in comparison with the wild type sequence (A␤1-42 WT ) and the equimolar solution of both peptides (A␤1-42 MIX ) to characterize the oligomers that are produced in the early phases. We carried out the structural characterization of the three preparations using electron and atomic force microscopy, fluorescence emission, and x-ray diffraction and described the soluble oligomer formation kinetics by laser light scattering. The mutation promoted a peculiar pathway of oligomerization, forming a connected system similar to a polymer network with hydrophobic residues on the external surface. A␤1-42 MIX generated assemblies very similar to those produced by A␤1-42 WT , albeit with slower kinetics due to the difficulties of A␤1-42 WT and A␤1-42 A2V peptides in building up of stable intermolecular interaction.
Alzheimer disease (AD) 2 is the most common form of dementia in the elderly accounting for up to 30 million cases worldwide, a figure that is predicted to double in 20 years (1). AD neurodegeneration is characterized by extensive neuronal atrophy especially in hippocampus and cerebral cortex, whereas neuropathology detects neuronal and synapse loss in association with the deposition of amyloid plaques and neurofibrillary tangles (2).
Amyloid plaque presents a core composed of misfolded amyloid ␤ (A␤) peptides of 37-43 amino acid lengths in the form of oligomers and amyloid fibrils. One of the pathogenic hypotheses to explain AD considers these aggregated A␤ species, particularly the soluble oligomers of A␤ but not monomers or insoluble amyloid fibrils (3)(4)(5)(6), to be the ultimate molecular triggers of a cascade of events (amyloid cascade) leading to synaptotoxicity and causing the observed neuronal loss (7). In fact there are several pieces of evidence that correlate A␤ peptide with the pathological mechanism of AD, suggesting that A␤ occupies a crucial position in the etiopathology. The most abundant peptides are A␤1-40 and A␤1-42, the first being the prevalent fragment and the second the most amyloidogenic (8). The aggregation of A␤ peptides starts with changes in their secondary structure leading to ␤-sheet formation, it progresses with aggregation of the misfolded peptides into oligomers, and it culminates in the production of amyloid fibers that precipitate into the brain forming amyloid plaques. Synthetic A␤ peptides are used to reproduce in vitro oligomeric structures, thus enabling the study of their features. The oligomers that have been described so far are paranucleus (5-nm diameter) (9), A␤-derived diffusible ligands (ϳ53 kDa), synthetic analog of A␤*56 (3,10), A␤O (ϳ90 kDa, 15-20-mer) (11), protofibrils (24 -700-mer) (8), annular assemblies (150 -250 kDa) (12,13), amylospheroid (ϳ150 -700 kDa) (14), and ␤amyball (50 -100-m diameter spheroids) (15). These oligomeric species differ by size and shape, and they can be both on-or off-pathway intermediates (15); however, all of them are able to dynamically assemble and progress to more aggregated states contributing to the growth and maturation of amyloid fibers.
A large body of literature confirms the importance of the A␤ sequence region spanning residues 21-30 in the molecular assembly (16 -24). This is a central hydrophobic core resistant to protease degradation, and the prediction of its importance in the determination of the aggregation tendency of A␤ peptides has been confirmed by the experimental analysis of A␤ peptides containing the Arctic (E22G), Dutch E22Q), and Iowa (D23N) mutations, all characterized by high propensity to form amyloid fibrils (20,25). Recently Scheidt et al. (26) demonstrated that the three-four amino acid residues at the N terminus of the A␤ region also play an important role in the formation of a stable ␤-sheet secondary structure in the A␤ peptide. Familial AD forms are linked to mutations in presenilin 1, presenilin 2, or in amyloid precursor protein genes and usually show an autosomal pattern of inheritance with total penetrance (27).  we described a new amyloid precursor protein mutation (A673V) that causes early onset AD when in homozygosity. The missense mutation consists of a C-to-T transition resulting in an alanine-to-valine substitution at position 673 of amyloid precursor protein that corresponds to position 2 of A␤1-40 and A␤1-42 peptides (A␤ A2V peptides). Notably, heterozygous individuals do not develop AD even in advanced age. In fact, five A673V heterozygous performed well on the neuropsychological assessments, and in particular, the 88-year-old aunt of the proband showed excellent performance on all the tests despite the fact that she was non-educated. The amyloid plaques and neurofibrillary tangles, the cardinal features of AD, were thought to underlie this chronic neurological disorder. Even today, after several years of research on AD, the A␤ peptides play a central role in the onset, development, and exacerbation of the AD in all of its forms of aggregation. However, the emerging soluble A␤ oligomers are now widely recognized as key pathogenic structures in AD (28,29). In fact, in light of recent findings and the realization that the amyloid cascade theory is insufficient to explain Alzheimer pathology, the amyloid hypothesis has been updated, as fibrils were considered the first and only species leading to AD pathogenesis (30). Recently, A␤ oligomers were identified as the main cause of synaptic dysfunction leading to alterations in both neuronal activity and cognitive function (31).
In this study we generated A␤1-42 oligomers from peptides spanning the wild type or the A2V sequences. Oligomers were produced from solutions of pure A␤1-42 WT , A␤1-42 A2V , and the equimolar solution of both (A␤1-42 MIX ). This enabled us to investigate the structure and the formation of toxic oligomeric species. Moreover, we conducted a comparative chemico-physical study on A␤1-42 WT , A␤1-42 A2V , and A␤1-42 MIX molecular assembly to describe features of different oligomeric populations produced in the early phase of the oligomerization process. This revealed the influence of the A2V mutation in A␤1-42 folding and produces evidence for the toxicity of A␤1-42 and the protective effect displayed by the A␤1-42 MIX . Finally, the position of the A2V mutation in the N terminus of A␤ peptides definitely strengthens the relevance of this region for peptide structure and spatial shape re-arrangement.

EXPERIMENTAL PROCEDURES
Peptide Synthesis and Sample Preparation-A␤1-42 peptides were synthesized using depsipeptide method as previously described (32)(33)(34)(35). A␤1-42 WT and A␤1-42 A2V were stored in acidic solution (water:trifluoroacetic acid, 0.02%) at a concentration of ϳ200 M. The depsipeptide method is a specific technique of synthesis used for amyloidogenic difficult sequences, and it allow us to obtain a batch with a low degree/level of aggregation free of either highly folded structures or fibrils and aggregates (seeds free) and as close as possible to monomeric conditions. In the case of A␤1-42, the method consists of introducing an O-acyl isopeptide structure into the Gly-25-Ser-26 sequence, stable at acidic pH and able to inhibit the self-aggregation. Upon a change to basic pH (switching procedure), the peptide is converted to the A␤1-42 native sequence. Before the switching procedure to minimize the pre-aggregated species and to obtain the best reproducibility, peptides were dissolved in acidic solution (water, 0.02% trifluoroacetic acid) and clarified overnight (16 -18 h) at 55,000 rpm to obtain seedsfree samples, filtered on a Microcon (centrifugal filter devices, cutoff 10 kDa, Millipore), and finally, concentrated on a Microcon (centrifugal filter devices, c.o. 3 kDa, Millipore) up to a concentration of Ն200 M. The switching procedure of depsipeptide A␤ was carried out at basic pH; a mix of sodium hydroxide (NaOH) and ammonium hydroxide (NH 4 OH) (ratio 3:1) was added to the peptide solutions (final pH of ϳ10) and incubated on ice for 10 -15 min. The preparation of MIX was done by adding an equimolar solution of both peptides before the switching procedure to obtain a A␤1-42 peptide mixture with a concentration of ϳ200 M constituted by 1 ⁄ 2 A␤1-42 WT and 1 ⁄ 2 A␤1-42 A2V . Oligomers were prepared using the following procedure; after the switching procedure, A␤1-42 solutions were brought to a final concentration of 100 M in 50 mM phosphate buffer, pH 7.4, and incubated for 24 h at 4 or 22°C to obtain oligomer-rich preparations (36,37).
Electron Microscopy (EM)-EM was used to investigate the structure of the peptide aggregates. 10 l of A␤1-42 oligomer preparations (A␤1-42 WT , A␤1-42 A2V , and A␤1-42 MIX ) were dropped onto 300-mesh Formvar/carbon nickel grids (Electron Microscopy Science), and after 5 min the solution was removed. Samples were counterstained for 5 min with saturated solution of uranyl acetate, washed with MilliQ water to eliminate excess uranyl acetate, and allowed to air dry (38,39). EM analyses were performed with a Libra 120 apparatus operating at 120 kV equipped with a Proscan Slow Scan CCD camera (Carl Zeiss).
Atomic Force Microscopy (AFM)-AFM was carried out on a Multimode AFM with a Nanoscope V system operating in tapping mode using standard phosphorus-doped silicon probes (thickness range, 3.5-4.5 m; length, 115-135 m; width, 30 -40 m; spring constant, 20 -80 newtons/m, Veeco/Digital Instruments) with a scan rate in the 0.5-1.2 Hz range, proportional to the area scanned. Freshly cleaved muscovite mica discs (Veeco/Digital Instruments) were used for deposition of peptide samples. A␤1-42 oligomeric solutions were added to freshly cleaved mica at room temperature for 1 min and, then the samples were washed and dried under gentle nitrogen flow. AFM images of A␤1-42 samples were analyzed for diameter and height with the Scanning Probe Image Processor (SPIP Version 5.1.6 (released April 13, 2011) data analysis package to describe oligomer structures. To exclude the interference of possible artifacts, extra control samples, such as freshly cleaved mica and freshly cleaved mica soaked with ultra-pure water, were also used. All the topographic patterns and SPIP characterization described in the text were confirmed by additional measurements in a minimum of 10 different, well separated areas.
Laser Light Scattering (LLS)-A␤1-42 WT , A␤1-42 A2V , and A␤1-42 MIX solutions were analyzed at final concentrations of 100 M in 50 mM phosphate buffer, pH 7.4, at 22°C by parallel and independent static and dynamic laser light scattering (SLS and DLS). The homemade LLS apparatus is described elsewhere (40). The average scattered intensity (SLS) readily reveals the emergence and growth of aggregates in solution, starting from monomers, as it is proportional to the square of the molecular mass of the scattering particles. The correlation function of the scattered intensity (measured by DLS) yielded the translational diffusion coefficients of particles in solution and then their average hydrodynamic diameter via the Stokes-Einstein relation. DLS data analysis was carried out using the method of cumulants, suitable to detect the evolution of the weight-average hydrodynamic size of particles in solution, and the non-negative least squares method (41), to determine their size distribution at different incubation times.
Small Angle X-ray Scattering (SAXS)-To obtain information on size, homogeneity, and shape of the A␤ peptide oligomers in solution on a local scale, we employed SAXS. Measurements were performed at the high brilliance ID02 beamline of European Synchrotron Radiation Facility (Grenoble, France) with a beam cross-section of 0.3 ϫ 0.8 mm and wavelength of 0.1 nm in the region of momentum transfer, q ϭ (2/)⅐sin(/2), 0.017 nm Ϫ1 Յ q Յ 4.65 nm Ϫ1 , where is the scattering angle. Plastic capillaries (KI-beam, ENKI) were mounted horizontally onto a six-place sample holder allowing for nearly simultaneous measurements on sample and reference cells in the same environmental conditions. Samples were prepared as described for LLS analysis, and all measurements were performed at 22°C. The exposure time of each measurement was very short, 0.1 s, to minimize any possible radiation damage. Several frames were collected on each sample, with 1-s sleeping times, carefully compared, and mediated if superimposable within experimental error. The measured SAXS profiles report the scattered radiation intensity as a function of the momentum transfer, q. Several spectra relative to the empty cells and the solvent were taken, carefully compared, and subtracted from each sample spectrum. To investigate a wide q-region, spectra relative to different q-ranges were compared and joined.
Fluorescence Emission Spectroscopy-Fluorescence spectroscopy was carried out using two specific probes, 1-anilino-8-naphthalene-sulfonate (ANS) and 4,4Ј-dianilino-1,1Ј-binaphthyl-5,5Јdisulfonic acid (Bis-ANS) (Sigma), to detect hydrophobicity in oligomer preparation at neutral pH, to probe high hydrophobic sites, and to monitor conformational changes. Fluorescence measurements were carried out on an LS50B Luminescence Spectrometer (PerkinElmer Life Sciences) using a quartz cuvette with a 1-cm light path. Thirty l of each oligomeric solution was added to 300 l of 30 mM citrate buffer, pH 2.4, containing 25 M ANS or Bis-ANS, and the fluorescence intensity was immediately recorded in the range of 400 -600-nm emission wavelengths and with an excitation wavelength of 386 nm (42). The analysis was performed on three/four replicates for each sample. Fluorescence was also acquired in the absence of A␤1-42 oligomers.
Circular Dichroism (CD)-A␤1-42 oligomers were analyzed immediately after their dilution to the final concentration of 25 M in M phosphate buffer, pH 7.4, to avoid signal saturation in the spectra. The CD spectra were recorded on a Jasco J-815 spectropolarimeter (Jasco, Easton, MD) at 4°C from 190 to 260 nm (1.0-nm bandwidth and 0.1-nm resolution) using a 0.1-cm path length quartz cell. Generally, a sensitivity of 100 millidegrees, a response of 4 s, a scan speed of 50 nm/min, and 5 accumulations were used. Spectrum of appropriate buffer was subtracted from the A␤1-42 spectra, and CD spectra were expressed as mean molar ellipticity (⌽).
Statistical Analysis-Means with standard error or standard deviation and one-way analysis of variance followed by Tukey's analysis were performed using Prism GraphPad software, Version 6.01 (GraphPad Software, Inc.).

Electron Microscopy Analysis of A␤1-42 Oligomeric
Assemblies-The temporal window for the formation of oligomers during the aggregation process was analyzed. Comparative EM analyses between A␤1-42 WT and A␤1-42 A2V revealed the prevalence of two different structures. A␤1-42 WT showed a predominant presence of globulomers with a size in the range of 15-40 nm and few annular structures with a size of about 60 nm (Fig. 1, A and D), whereas A␤1-42 A2V was highly enriched with annular structures in the range of 7-70 nm ( Fig.  1B and E), suggesting that the presence of the A2V mutation promoted a distinctive oligomerization pathway. As shown at higher magnification (Fig. 1E), A␤1-42 A2V annular aggregates had an electron-dense core that may suggest the formation of a pore where uranyl acetate solution was accumulated. The annular structures formed by A␤1-42 A2V continued the oligomerization process, as shown by the neoformed extension (see the inset of Fig. 1E). The co-incubation of A␤1-42 WT and A␤1-42 A2V was characterized by a morphology highly resembling the A␤1-42 WT peptide alone, with the presence of many globulomers and annular structures with smaller dimensions in the range of 9 -25 nm (Fig. 1, C and F).
Atomic Force Microscopy Analysis of A␤1-42 Oligomers-Peptides were analyzed immediately after the switching procedure and after incubation for 24 h at 4°C. Freshly prepared solutions contained only monomeric assemblies, whereas samples that underwent incubation disclosed the presence of small oligomers of different sizes (data not shown). SPIP software was used to analyze the distribution of oligomer population in terms of diameters ( Fig. 2A) and heights (Fig. 2, B-D). This software enables the elaboration of AFM images, and it specifically takes into account the features of the tips and the tapping mode. Therefore, it is able to obtain very accurate data on the height and diameter of the molecular assemblies formed by A␤ peptides (43). The cumulative frequency graph ( Fig. 2A) reports the diameter distribution. A␤1-42 WT produced a family of oligomers with a range of highly defined dimension, the majority (65%) being between 5 and 20 nm in diameter, and no oligomeric aggregates larger than 60 nm were detected. A␤1-42 A2V produced an evenly distributed population of oligomers (Ͼ90%) in the range of 20 -70 nm with scattered structures reaching 140 -180 nm. The co-incubation of A␤1-42 WT and A␤1-42 A2V did not generate aggregated structures larger than 60 nm, and most oligomers were in the range of 10 -50 nm. This indicates a much lower propensity to oligomerization for the A␤1-42 MIX than for A␤1-42 WT or A␤1-42 A2V (Fig. 2A, Table 1).
The same samples were also analyzed to determine oligomer height distribution (Fig. 2, B-D, and Table 1). As reported in Table 1, the means of the oligomer heights were similar for all samples. However, as reported in Figs. 2, B-D, it can be noticed that the main peak of the frequency distribution is sharper for A␤1-42 WT assemblies (90% Յ1 nm in heights).
Fluorescence Spectral Characterization of A␤ Oligomer Conformation-All oligomeric preparations were examined for their pattern of exposure of hydrophobic and hydrophilic residues using an ANS assay. This dye enables the orientation of aromatic side chains or the definition of formation and disruption of organized hydrophobic patches and clefts to be defined. It is well known that the binding of ANS to the exposed hydrophobic clusters of a protein results in both an increased intensity of the fluorescence emission of the dye and a blue shift in the maximum emission wavelength. Significant differences in fluorescence emission was only seen between A␤1-42 WT and A␤1-42 A2V as an indication of the presence of more hydrophobic residues on the external surface of A␤1-42 A2V (Fig. 3, A and  B). Moreover, A␤1-42 A2V showed a shift in its maximum emission wavelength from ϳ500 nm (referred to A␤1-42 WT ) to ϳ491 nm (Fig. 3C).
The same samples were then tested for the presence of soluble oligomeric forms with a ␣-helical or random coil/mixed conformers by Bis-ANS assay (Fig. 3A). This probe does not emit fluorescence in the presence of fibrillar structures. A␤1-42 A2V produced a significantly higher fluorescence signal than A␤1-42 WT or A␤1-42 MIX , indicating a greater amount of oligomeric assemblies conformers with a ␣-helical or random coil/mixed. It is important to note that the co-incubation of A␤1-42 WT and A␤1-42 A2V led to molecular assemblies closely resembling those of A␤1-42 WT . We also determined the absence of fibrillar assemblies in all three experimental groups using the classical thioflavin T assay (data not shown). As confirmatory information, a CD technique was used to determine the secondary structure of A␤ oligomers. CD spectra of A␤1-42 WT , A␤1-42 A2V , and A␤1-42 MIX oligomeric solutions did not show significant differences, and they evidenced a predominant random-coil conformation with a negative peak in the signal around 195-197 nm (Fig. 4).
Short-time Kinetics of A␤ Assembly Determined via Laser Light Scattering-The initial stages of aggregation of A␤ peptides were also followed by SLS and DLS measurements. The change in molecular assemblies in 100 M peptide solutions kept at 22°C was followed for 24 h.
Because the scattered intensity is proportional to the square of the particle mass, this technique is sensitive to the presence of preformed seeds with high molecular weight. If non-negligible in number, the contribution of seeds to the total scattered intensity dominates and hides the contribution of the small particles. In our case, immediately after the switching procedure, the initial states of A␤1-42 peptides were distinctly monomeric, allowing the first steps of aggregation to be followed. Experimental results are shown in Figs. 5 and 6 reporting, respectively, SLS and DLS observations.
For all A␤1-42 species, aggregation started immediately in solutions, leading to the rapid formation of oligomers in coexistence with a population of monomers with average hydrodynamic radii of 2-3 nm. In fact, looking at Fig. 6 one can appreciate that the size distribution, mainly monomeric at the beginning (0.1 h), has evolved to include a population of oligomers after 3 h. The oligomeric aggregates were slightly different in size, being larger for A␤1-42 WT (about 35 nm) than for A␤1-42 MIX and A␤1-42 A2V (average hydrodynamic radius of 25 nm). Moreover, although the earliest kinetics of oligomer formation and their initial size distribution were quite similar for the three peptides, remarkable differences were observed in the following 24 h. During the 24-h time-course, the aggregation of A␤1-42 WT and A␤1-42 MIX was characterized by a two-step process: (i) the prompt formation of early oligomers and (ii) a delayed slower aggregation, starting after a time lag of several hours (ϳ6 h), as shown in panel B of Fig. 5. The fitting curves for the second aggregation step, also shown in Fig. 5B, were obtained by I(t) ϭ I final (1 Ϫ e Ϫt/ ), where I final is the asymptotic value of the scattered intensity, and is the characteristic time. A␤1-42 MIX displayed a slightly longer lag time (t MIX ϭ 8.6 h) with respect to A␤1-42 WT (t WT ϭ 6.7 h) and a slower characteristic time ( WT ϭ 1.7 h, MIX ϭ 3 h). This indicates that the aggregation process proceeded more slowly for the A␤1-42 MIX . Moreover the asymptotic intensity value I final is lower for the A␤1-42 MIX (Fig. 5C). The increase in scattered intensity for both solutions could not be attributed to a simple increase in the number of oligomers because the average hydrodynamic radius of particles also migrated toward larger values. Monomers are progressively hidden by the overwhelming scattered intensity contributed by oligomers. In fact, they progressively increased in number over 24 h. This is shown in Fig. 6 where the size distribution after 24 h is reported. Notably, at this interval of time, all oligomeric species were still soluble.
Reversely, in the A␤1-42 A2V solution, the oligomer population readily appeared just after dissolution as shown in Fig. 6, where the initial distribution (0.1 h) reveals the presence of a small fraction of aggregates. All over, A␤1-42 A2V follows a different path for aggregation as compared with A␤1-42 WT and A␤1-42 MIX . In fact, after Ͻ2 h, an additional population appeared with a much larger size, as readily revealed by SLS, showing a huge spike in the scattered intensity as reported in panel A of Fig. 5. The sudden rise in intensity is connected to the formation of fibrils, very few in numbers and tending to precipitate and adhere to cell walls. Meanwhile, oligomeric species were also still present as the majority. Their time evolution was difficult to follow; nonetheless they preserved their solubility even after 24 h as seen in Fig. 6. The mutated peptide sequence had an intrinsic tendency to form rapidly structured oligomers.
Small Angle X-ray Scattering: Structural Elucidation of A␤1-42 Aggregates-To obtain information on the structure of A␤1-42 aggregates on a local scale, we performed small angle x-ray scattering measurements just after switching and after 3 h at 22°C. This late peptide solution displays an oligomer composition similar to that found after 24 h at 4°C (data not shown). Fig. 7 reports the SAXS intensity spectra obtained at 0.1-and 3-h delays in the momentum transfer range 0.017 nm Ϫ1 Յ q Յ 4.65 nm Ϫ1 for the three peptides.
In the high q-region, the spectra of the three different systems are superimposable and can be fitted with the form factor of very small nuclei of condensation of about 1.3 nm in size. The size of small particles is compatible with the presence of A␤1-42 monomers in all samples at both delays.
In the low-q region, the short-delay spectra can be fitted with the form factor of rod-like structures, as usually found for polymers and peptides, for the three peptides. This may indicate the onset of self-assembly of monomers up to a persistence length of tens of nm in agreement with laser light scattering measurements. The spectra collected 3 h later indicate that, besides monomers, more structured forms are present in solutions at that delay. In the case of A␤1-42 A2V SAXS spectrum, the characteristic slope of a packed polymer network (q Ϫ1.7 ) (44) indicates that the A␤1-42 A2V aggregates are not compact globular or rod-like structures but rather look like disperse assembly of particles with branched-type features. After 3 h, A␤1-42 WT and A␤1-42 MIX also showed connected structures, which did not recruit all the material, as can be inferred by the milder slopes of the intensity decays reported in Fig. 7. Notably, A␤1-42 MIX displayed the lowest degree of supramolecular complexation.

DISCUSSION
A673V mutation differs from other genetic alterations in the amyloid precursor protein sequence because of its recessive inheritance traits. An A673V homozygous carrier presented with early onset dementia characterized by an aggressive cortico-subcortical atrophy and subcortical white matter changes (45). Distinctive neuropathological features were the morphology, composition, and topology of A␤ deposits, which were of large size, mostly perivascular, and exhibited a complete correspondence between the pattern elicited by amyloid staining and the labeling obtained with anti A␤ antibodies (46). The amyloid deposits were predominantly composed by A␤1-40 and were also abundant in the cerebellum, at variance with sporadic AD (46). The A673V mutation enhances A␤ production and significantly increases the fibrillogenic properties of A␤. However, the interaction of A␤1-42 WT -and A␤1-42 A2V -mutated peptides inhibits A␤ folding (45,47). These findings are consistent with the observation that the A673V heterozygous carriers do not develop the disease and offer grounds for the development of a novel therapy for sporadic AD based on modified peptides homologous to residues 1-6 of A␤ carrying the A2V substitution (47). The strength of this approach, as compared with strategies based on purely theoretical grounds or large screening strategies, is that the A2V A␤ variant occurs in humans and prevents the development of disease when present in the heterozygous state (47).
We focused our interests on the molecular assembly of A␤1-42 peptides to understand the biochemical reasons of the influence of A673V mutation in A␤1-42 folding, the higher toxicity, and the protective effect seen in the heterozygous state. We investigated the early stages of the A␤ folding process after the formation of oligomeric structures, which are considered key toxic species in the onset and progression of AD pathogenesis.
We report here a qualitative and quantitative analysis of A␤ oligomer formation by A␤1-42 WT , A␤1-42 A2V , and A␤1-42 MIX to unveil the features of different oligomer populations formed over time. Morphological structural analysis of A␤1-42 A2V revealed the prevalent formation of annular structures. Lashuel et al. (13) showed that mutant amyloid proteins associated with familial AD (Arctic mutation) and familial Parkin-son disease (␣-synuclein mutants A53T and A30P) form morphologically indistinguishable annular assemblies that resemble a class of pore-forming bacterial toxins, suggesting that inappropriate membrane permeabilization might be the cause of cell dysfunction and even cell death in amyloid diseases (13). Under our experimental conditions, the oligomers formed by the A␤1-42 A2V also showed the presence of annular structures as deduced by the presence of uranyl acetate solution in the sample analyzed by EM. In fact, unlike A␤1-42 WT , A␤1-42 A2V formed annular structures wrapped around a wettable trapping central spot. This suggests that the hydrophilic residues of A␤1-42 A2V may be preferentially located in the center, with a corresponding disposition of the hydrophobic residues on the external rim or on the top and bottom sides of the annulus. These structures were almost absent in the case of A␤1-42 WT where globulomers were the most abundant population. The co-incubation of A␤1-42 WT and A␤1-42 A2V produced an intermediate morphological condition consisting of both globulomers and annular structures. Interestingly, AFM characterization of A␤1-42 MIX assemblies showed the presence of oligomers with a smaller average size, suggesting that A␤1-42 MIX has a much lower propensity to oligomerization than A␤1-42 WT or A␤1-42 A2V alone.
Kinetic comparison between A␤1-42 A2V and A␤1-42 WT showed that the former proceeded along a different pathway of structured oligomer formation. The key role was played by the first step, which was the formation of early assemblies that led to the formation of A␤ assemblies after a very efficient dockand-lock mechanism (48 -50).
When comparing A␤1-42 MIX with A␤1-42 WT , results showed that the route leading to the formation of oligomers in A␤1-42 WT , characterized by the formation of "early" oligomers, followed by a slow additional aggregation was dissimilar to the one observed in the case of A␤1-42 MIX where the second process of aggregation was less extensive and even slower. This was confirmed by LLS analysis of A␤1-42 MIX that had a slightly longer lag time (t MIX ϭ 8.6 h) with respect to A␤1-42 WT (t WT ϭ 6.7 h) and a slower characteristic time ( WT ϭ 1.7 h versus MIX ϭ 3 h).
Spectral analysis by fluorescence probes showed that A␤1-42 MIX assemblies closely resembled to those of A␤1-42 WT (Bis-ANS assay) with an intermediary exposure of hydrophobic residues (ANS assay). A␤1-42 A2V oligomers were characterized by the maximum hydrophobicity, confirming the existence of a hydrophilic core and an increase of the hydrophobic resi-  (51,52) reported that the exposure of hydrophobic residues on the surface of aberrant protein oligomers increases the toxicity of oligomeric structures, enabling a major interaction with cell membranes.
In terms of local structural organization SAXS analysis indicated that, as expected for a point mutation, the individual structural unit is the same for the three peptides. Nonetheless, their spatial arrangement in the structured oligomers was different. The transition from rod-like to more structured aggregates was more extensive and prompts in A␤1-42 A2V and clearly occurred via the formation of interconnected networks. The A␤1-42 MIX resulted in aggregates with the lowest degree of supramolecular complexation with respect to A␤1-42 WT and A␤1-42 A2V , suggesting a negative interference in the supramolecular organization.
In conclusion, we demonstrated that the A2V mutation is able to promote a peculiar oligomerization process pathway of A␤1-42 that leads to the formation of annular structures with a higher hydrophobicity profile and, hence, toxicity as also observed in an in vivo model (53). When the heterozygous condition was reproduced, the aggregation effect of the A2V mutation was lost, confirming that its effect is present only when in homozygosity one. Interestingly, A␤1-42 MIX not only was less prone to aggregate when compared with mutated alone, but also it produced smaller aggregates when compared with the wild type sequence as well. This suggests that the mutation in the heterozygous state is able to hinder the aggregation process and generates unstable structures. This is in good agreement with our previous observations showing that aggregates formed by an equimolar mixture of A␤1-42 WT and A␤1-42 A2V were far more unstable than those generated by either A␤1-42 WT or A␤1-42 A2V (45). This is most likely the biochemical basis of the  protective effect shown by the A673V mutation in the heterozygous carrier.
The second remarkable point evidenced in this work was the critical role played by the N-terminal region in the A␤ molecular assembly. In fact, in the homozygous condition the A2V mutation led to aggregation while on the other hand in the heterozygous state the evolution and kinetics of the aggregation process was hindered. This latter phenomenon was driven by the A␤1-42 A2V peptide difficulty in forming stable intermolecular interactions with the wild type sequence.