Solvent exposure of Tyr10 as a probe of structural differences between monomeric and aggregated forms of the amyloid-β peptide

Aggregation of amyloid-β (Aβ) peptides is a characteristic pathological feature of Alzheimer's disease. We have exploited the relationship between solvent exposure and intrinsic fluorescence of a single tyrosine residue, Tyr10, in the Aβ sequence to probe structural features of the monomeric, oligomeric and fibrillar forms of the 42-residue Aβ1-42. By monitoring the quenching of Tyr10 fluorescence upon addition of water-soluble acrylamide, we show that in Aβ1-42 oligomers this residue is solvent-exposed to a similar extent to that found in the unfolded monomer. By contrast, Tyr10 is significantly shielded from acrylamide quenching in Aβ1-42 fibrils, consistent with its proximity to the fibrillar cross-β core. Furthermore, circular dichroism measurements reveal that Aβ1-42 oligomers have a considerably lower β-sheet content than the Aβ1-42 fibrils, indicative of a less ordered molecular arrangement in the former. Taken together these findings suggest significant differences in the structural assembly of oligomers and fibrils that are consistent with differences in their biological effects.


Introduction
Alzheimer's disease (AD) and a range of related disorders are associated with the self-assembly, aggregation, and fibril formation of disease-specific peptides and proteins [1]; in the case of AD such processes are involved with aggregation of the amyloid-b (Ab) peptide [2,3]. In its fibrillar form, this peptide is the main proteinaceous component of the extracellular plaque deposits that are characteristic of AD pathology [4,5], but in the brain Ab also exists in a variety of monomeric and oligomeric forms [6]. It has been reported that soluble Ab concentrations correlate more closely with dementia than the amount of amyloid plaques [7,8], and indeed soluble Ab oligomers are now thought to be the main culprits in the pathogenesis of AD and related conditions [9], particularly since they have been associated with impaired cognitive function [10e12] and have been shown to induce cellular toxicity [13e17].
Unfortunately, the small size and low abundance of Ab oligomers, in combination with their considerable heterogeneity and high sensitivity to environmental changes, has rendered them challenging to characterise. Although a detailed molecular model of a stabilised Ab 1-42 protofibril has been recently reported [18], little is known in detail how monomers arrange to build up the smaller, globular oligomers that are populated during Ab aggregation reactions and which can form in the AD brain [19].
In the present paper we have used fluorescence and circular dichroism spectroscopies to examine structural differences, particularly related to the microenvironment around the N-terminal region, to compare soluble, globular Ab 1-42 oligomers, prepared in vitro, with monomers and fibrils formed by both Ab 1-40 and Ab  . The oligomers were prepared using established methods for producing stable oligomers of a type that is often referred to as amyloid-derived diffusible ligands (ADDLs) [19]. These ADDLs have been reported to be neurotoxic [20] and it has been found that antibodies raised against in vitro prepared oligomers of this type also recognise oligomeric species that are elevated in AD brains [21], suggesting their resemblance to naturally occurring forms. Although it has been proposed that oligomers of this type could be of a fibrillar nature [22,23], it is not known how monomers are present within these oligomer assemblies or to what extent the microscopic architecture resembles that found in mature fibrils. To address this issue, we have taken advantage of the fact that the Ab peptide contains only a single intrinsically fluorescent residue, tyrosine at position 10 (Tyr 10 ), which is located in between the bcore and the flexible N-terminus of fibrillar Ab 1-40 and Ab  [24e28] (Fig. 1A). This residue, therefore, has the potential to report on the participation of this region of the Ab peptide in the different aggregated states. In this study, we have exploited the intrinsic fluorescent properties of Tyr 10 and also its susceptibility to fluorescence quenching by water-soluble acrylamide to examine the characteristics of the monomeric, oligomeric and fibrillar forms of Ab. Our results indicate that Ab 1-42 oligomers show the presence of a degree of b-sheet structure, but that they are distinctly less ordered than fibrils, and this is confirmed by a substantially higher degree of solvent exposure around Tyr 10 . Taken together these results highlight significant differences in the molecular architecture of the oligomeric versus the fibrillar forms of the Ab peptides which may offer new insights into the differential biological activities.

Materials
Synthetic Ab 1-40 and Ab 1-42 peptides were acquired as lyophilised powders from Anaspec EGT (Fremont, USA) and were prepared for use as described below. All other reagents were purchased from SigmaeAldrich (Dorset, UK).

Sample preparation
The Ab peptide powders were dissolved in ice-cold trifluoroacetic acid, sonicated (30 s, on ice), flash frozen and again lyophilised.
The samples were redissolved in ice-cold hexafluoroisopropanol (1 mL) and the solutions were kept on ice (10 min) then divided into aliquots (50 mL) whilst working at 4 C, and dried by rotary evaporation. The peptide concentration was determined by amino acid analysis and all experiments were performed in 50 mM sodium phosphate buffer (pH 7.4). The monomeric form of the Ab peptides was obtained by dissolving a fresh peptide aliquot to a final concentration of 5 mM directly in buffer.
The solution was analysed immediately to minimise the formation of aggregates. Fibrillar Ab samples were prepared by incubation at room temperature for 48 h under shaking conditions (1400 rpm in a Titramax 100 shaker, Heidolph Instruments GmbH, Schwabach, Germany). Oligomers of Ab 1-42 were prepared by resuspending peptide aliquots in DMSO (2 mL), followed by dilution in ice-cold buffer (10 mM NaCl, 10 mM sodium phosphate, pH 7.4), to a final peptide concentration of 100 mM followed by incubation (overnight, 4 C under quiescent conditions) [20]. The solutions were diluted in 50 mM sodium phosphate buffer (pH 7.4) (5 mM based on monomer concentration) prior to experimental analysis. The oligomer yield was determined by amino acid analysis following separation of the oligomers from residual monomer by ultracentrifugation (1 h, 90,000 g) using an Optima TLX ultracentrifuge (Beckman Coulter Inc, Brea, USA).

Fluorescence spectroscopy
Fluorescence spectra were recorded on a Cary Eclipse fluorimeter (Agilent Technologies, Stockport, UK) using a reduced pathlength quartz cuvette (4 mm excitation/10 mm emission). The excitation wavelength was 275 nm and emission spectra were recorded at 1 nm increments between 290 and 350 nm, with excitation and emission slit widths of 5 nm and 10 nm, respectively, and a scan rate of 60 nm/min. Samples were subjected to an acrylamide gradient following additions of a 0.2 M stock solution (10 mL aliquots), the fluorescence spectra being recorded immediately after the addition of the acrylamide aliquots. The recorded spectra were corrected for the background contributions and the fluorescence intensity in each case was taken as the sum of the intensities in a 6 nm range centred around the emission maxima, to increase signal-to-noise. Data were analysed using the Stern-Volmer equation [29].
where F 0 and F are the fluorescence intensities in the absence and presence of quencher, [Q] is the concentration of quencher, and K SV the SterneVolmer constant. All experiments were performed in triplicate and are reported as mean ± SD. As acrylamide absorbs significantly at the excitation wavelength of tyrosine and therefore acts as an inner filter, corrections were made using separate experiments in which a non-quenching molecule (here DNA) was titrated into solutions of tyrosine at the same absorbance increments in order to determine correction factors for primary inner filter effects.

Circular dichroism (CD) spectroscopy
CD spectra were recorded on a JASCO J-810 spectropolarimeter (JASCO Inc. Tokyo, Japan) between 190 and 250 nm using a 1 mm quartz cuvette. 10 scans were recorded and averaged using a bandwidth of 2 nm and a scan speed of 50 nm/min. The peptide concentration was 30 mM and all spectra were corrected for background contributions by subtracting buffer blanks.

Transmission electron microscopy (TEM)
Ab samples were adsorbed (2 min) onto carbon-coated copper grids (Taab Laboratories Equipment Ltd, Berks, UK). The grids were blotted, washed with milliQ water (2X) and negatively stained with 2% (w/v) uranyl acetate. Samples were imaged on a FEI Tecnai G2 transmission electron microscope (Eindhoven, Netherlands) and images were analysed using the SIS Megaview II Image Capture system (EMSIS GmbH, Muenster, Germany).

Atomic force microscopy (AFM)
AFM measurements were performed using a NanoWizard AFM system (JPK Instruments AG, Berlin, Germany). Samples were diluted using dH 2 O and deposited onto freshly cleaved mica surfaces and slowly dried before imaging. AFM imaging was carried out in the intermittent (air) contact mode using a silicon nitride cantilever (mmasch, NSC36/No Al, 65e130 kHz, 0.6e2 N/m). AFM images were analysed using Gwyddion software package (http:// gwyddion.net/).

Results and discussion
The intrinsic fluorescence of Tyr 10 in combination with circular dichroism spectroscopy has been used to examine five different preparations of Ab 1-40 and Ab 1-42 ; monomers (mAb 1-40 and mAb , fibrils (fAb 1-40 fAb 1-42 ) and Ab 1-42 oligomers (oAb  ) in order to assess the conformational differences between the various forms of Ab  , and also to explore potential differences between Ab 1-40 and Ab 1-42 fibrils.
Prior to spectroscopic characterisation the morphology of the different Ab species was examined (Fig. 1BeG). Analysis of the oligomeric oAb 1-42 samples by AFM (Fig. 1B) and TEM (Fig. 1CeD) showed that they contain a relatively homogeneous population of small aggregates with approximately spherical morphology; analysis of the TEM images indicates that their approximate diameters were 10e20 nm. The fibrillar fAb 1-40 and fAb 1-42 samples contained typical amyloid fibrils; importantly no such fibrillar structures were detected in the oligomeric samples used in this study and no discernible aggregates were observed in the monomeric samples (Fig. 1G,).
Next, the secondary structure content in the different Ab preparations was assessed (Fig. 2); the monomeric samples displayed CD spectra typical of random coils with negative peaks centred at 195 nm, consistent with their intrinsically disordered nature and confirming the absence of amyloid aggregates in the monomer preparations. The oligomeric and fibrillar samples, by contrast, exhibited characteristic b-sheet features with CD spectra displaying a negative peak at~218 nm and a positive peak at~196 nm. The CD signal from the oligomers, however, was considerably weaker than that of the fibrils indicating that they have significantly less b-sheet content.
The fluorescence spectra of the various Ab samples were recorded to monitor the intrinsic emission from Tyr 10 . Fig. 3A shows the spectra of the five different Ab preparations along with corresponding spectra for free tyrosine and for a tyrosinecontaining tri-peptide (VYV), the latter was used to determine the generic effects on tyrosine emission due to its incorporation into a polypeptide sequence. It was observed that the tyrosine emission intensity is decreased by as much as a factor of 2 upon incorporation into a peptide sequence, consistent with previous observations where the quenching of tyrosine in proteins has been attributed to photoinduced electron transfer to the peptide bond in presence of electron withdrawing groups [30]. All the Ab conformers examined, except fAb 1-42 , were observed to exhibit even lower intrinsic tyrosine fluorescence intensities than the VYV control peptide, indicating the existence of additional quenching interactions within the Ab sequence, attributable to interactions with neighbouring side-chains. Spectral broadening of the tyrosine emission peak was observed in all the Ab samples (Fig. 3B); the broadening on the blue edge of the emission peak is particularly apparent for fAb 1-40 , fAb 1-42 and oAb  and can be explained by light scattering by the aggregates present in these solutions, whereas the red edge broadening, which is mainly apparent for the monomeric Ab peptides, may be indicative of tyrosinate formation in the excited state [31,32].
Acrylamide quenching experiments were conducted to explore the extent to which tyrosine is exposed to solvent in its free form (Fig. 4A) compared to its exposure when it is incorporated as Tyr 10 in the different monomeric and aggregated forms of Ab. The resulting Stern-Volmer plots (Fig. 4B) were linear in all cases, an observation consistent with acrylamide being a predominately collisional quencher [33]. Fig. 4C summarises the calculated SterneVolmer quenching constants (K SV ) and shows that incorporation of tyrosine into a peptide sequence, even into the short VYV model peptide which displays a~30% reduction of the K SV value, results in shielding from the polar but non-charged water soluble acrylamide quencher. This effect is likely to be due to steric hindrance imparted by the neighbouring residues, which limits the number of possible quenching interactions but does not reflect shielding caused by secondary structure constraints. No appreciable differences in the solvent exposure of tyrosine in the monomeric mAb 1-40 and mAb 1-42 forms compared to the VYV peptide were observed. This finding is consistent with other observations showing that the Ab chain is highly unfolded in solution [34,35], at least on timescales in the order of the excited state lifetime of tyrosine (3.4 ns [36]). Analysis of the Stern-Volmer plots showed, however, that Tyr 10 in fAb 1-40 and fAb 1-42 is considerably less exposed to acrylamide quenching than in mAb 1-40 and mAb  , suggesting that the incorporation of the Ab molecules into a fibrillar structure protects this residue from solvent even though existing structural models [24e28] suggest that Tyr 10 does not participate directly in the cross-b core (see summary in Fig. 1). This effect can, however, be explained by the close packing of monomer units in the fibril protofilaments [37], which sterically hinders the acrylamide-Tyr 10 collisions, or it may also be a consequence of a fraction of the Ab N-termini becoming buried within the mature fibril; indeed, limited proteolysis data for fAb  suggests that at least 20% of Ab monomers have an N-terminal segment that is protected within the fibril structure [38]. The considerable similarity observed for fAb  and fAb  is interesting in relation to hydrogen/deuterium-exchange rates measured by NMR, which indicate significant differences between the two fibril types with respect to the amide solvent protection of the N-terminus, whereby the amide group of Tyr 10 is more solvent accessible in fAb 1-42 than in fAb 1-40 [26].
Interestingly, the oligomeric oAb 1-42 samples exhibit K SV values that are significantly higher than those observed for the fibrils (Fig. 4C). This observation suggests that Tyr 10 is considerably less shielded in these globular oligomers than in Ab fibrils and is consistent with our finding that oAb 1-42 species have a lower bsheet content than fAb   (Fig. 2A). More surprisingly, we also find that the K SV values of the oligomers are similar, within experimental error, to those of monomeric Ab implying that Tyr 10 is exposed to a similar extent as in a random coil monomer. To ensure that the high K SV value truly reflects the structure of the oAb 1-42 species and is not attributable to sample heterogeneity, and as it has been suggested previously that oligomers of the type we examine here can exist in binary mixtures with monomers, a further control experiment was performed [39]. Quantitative amino acid analysis following ultracentrifugation to separate the oAb 1-42 species from residual monomers showed that, in our samples, nearly 90% of the Ab monomers were incorporated into oligomers (Supplementary Information). This result is comparable to our previous report of the yield in fibril forming reactions with Ab 1-40 and Ab 1-42 [40]. Furthermore, these control experiments confirm that the difference in b-sheet content indicated in the oAb 1-42 and fAb 1-42 CD spectra, is indeed related to the oAb 1-42 species exhibiting some random coil nature, rather than due to the presence of significant quantities of unstructured monomers in the samples. Therefore, this study shows that the Ab 1-42 oligomers differ from fAb  , not only in size and shape, but also in secondary structure content and internal architecture. Our finding that Tyr10 in the oligomers is significantly solvent exposed, and thus likely to be present on the surface of the oligomers is in agreement with a previous study of a disc shaped Ab 1-42 pentamer [41]. These observations are also consistent with the conclusion that the Ab  peptides in these oligomers do not adopt the same b-hairpin arrangement as that formed in the mature Ab fibrils. This suggestion supports the idea that oligomers of this type may accumulate because their structural properties are such that they are not able to convert into fibrils [42,43].
In conclusion, this study has set out to examine the fluorescent properties and the degree of solvent exposure of Tyr 10 in Ab to gain insights into structural similarities and dissimilarities between different Ab species. We report significant differences between monomers and fibrils, suggesting that even though Tyr 10 is not directly part of the cross-b core, the side-chain is, on average, well shielded within the fibrils. In addition, we have shown that Ab  oligomers, due to their lower b-sheet content and extensive exposure of Tyr 10 to solvent, have significantly different structural properties from those of Ab fibrils.