Quenched hydrogen ⁄ deuterium exchange NMR characterization of amyloid-b peptide aggregates formed in the presence of Cu 2+ or Zn 2+ Anders Olofsson 1 , Malin Lindhagen-Persson 1 , Monika Vestling 1 , A. Elisabeth Sauer-Eriksson 2 and Anders O ¨ hman 2 1 Department of Medical Biochemistry and Biophysics, Umea ˚ University, Sweden 2 Department of Chemistry, Umea ˚ University, Sweden Introduction Abnormal protein assemblies in form of amyloid are linked to more than 25 different syndromes, of which the neurodegenerative disorder Alzheimer’s disease (AD) is the most well-known example [1]. The main Keywords Alzheimer’s disease; amyloid-b peptide; Cu 2+ ;H⁄ D exchange NMR; Zn 2+ Correspondence A. O ¨ hman, Department of Chemistry, Umea ˚ University, SE-901 87 Umea ˚ , Sweden Fax: +46 90 786 5944 Tel: +46 90 786 5919 E-mail: anders.ohman@chem.umu.se A. Olofsson, Department of Medical Biochemistry and Biophysics, Umea ˚ University, SE-901 87 Umea ˚ , Sweden Fax: +46 90 786 5944 Tel: +46 90 786 5921 E-mail: anders.olofsson@medchem.umu.se (Received 10 March 2009, revised 4 May 2009, accepted 26 May 2009) doi:10.1111/j.1742-4658.2009.07113.x Alzheimer’s disease, a neurodegenerative disorder causing synaptic impair- ment and neuronal cell death, is strongly correlated with aggregation of the amyloid-b peptide (Ab). Divalent metal ions such as Cu 2+ and Zn 2+ are known to significantly affect the rate of aggregation and morphology of Ab assemblies in vitro and are also found at elevated levels within cerebral plaques in vivo. The present investigation characterized the architecture of the aggregated forms of Ab(1–40) and Ab(1–42) in the presence or absence of either Cu 2+ or Zn 2+ using quenched hydrogen ⁄ deuterium exchange combined with solution NMR spectroscopy. The NMR analyses provide a quantitative and residue-specific structural characterization of metal- induced Ab aggregates, showing that both the peptide sequence and the type of metal ion exert an impact on the final architecture. Common features among the metal-complexed peptide aggregates are two solvent- protected regions with an intervening minimum centered at Asn27, and a solvent-accessible N-terminal region, Asp1–Lys16. Our results suggest that Ab in complex with either Cu 2+ or Zn 2+ can attain an aggregation-prone b-strand–turn–b-strand motif, similar to the motif found in fibrils, but where the metal binding to the N-terminal region guides the peptide into an assembly distinctly different from the fibril form. Structured digital abstract l MINT-7102414, MINT-7102427: Abeta (uniprotkb:P05067) and Abeta (uniprotkb:P05067) bind ( MI:0407)byfluorescence technologies (MI:0051) l MINT-7103341, MINT-7103348: Abeta (uniprotkb:P05067) and Abeta (uniprotkb:P05067) bind ( MI:0407)byatomic force microscopy (MI:0872) l MINT-7102371, MINT-7102380: Abeta (uniprotkb:P05067) and Abeta (uniprotkb:P05067) bind ( MI:0407)bycircular dichroism (MI:0016) l MINT-7102390, MINT-7102399: Abeta (uniprotkb:P05067) and Abeta (uniprotkb:P05067) bind ( MI:0407)bybiophysical (MI:0013) l MINT-7103367, MINT-7103374: Abeta (uniprotkb:P05067)andAbeta (uniprotkb:P05067) bind ( MI:0407)bynuclear magnetic resonance (MI:0077) Abbreviations AD, Alzheimer’s disease; AFM, atomic force microscopy; Ab, amyloid-b peptide; H ⁄ D, hydrogen ⁄ deuterium; ThT, thioflavin T. FEBS Journal 276 (2009) 4051–4060 ª 2009 The Authors Journal compilation ª 2009 FEBS 4051 protein component of the plaques found in patients with AD is the amyloid-b peptide (Ab). This is a proteolytic excision product derived from the signifi- cantly larger amyloid precursor protein, representing an ensemble of peptides of various lengths, each of which has distinguishing biophysical properties. Frag- ments of 39–43 residues are all clinically relevant, but the most abundant are Ab(1–40) and Ab(1–42) [2]. The fold and mode of assembly of Abs are deter- mined by their primary sequences in combination with the chemical properties of the solvent. Abs assemble via several different routes into aggregates ranging in size from dimers [3] to smaller oligomers [4–7], protofi- brils [8,9], and fully mature amyloid fibers. Elucidation of amyloid fiber and aggregate structures, as well as their path of formation, is crucial to understanding the pathological process of AD. Although the size, insolu- bility and noncrystalline behavior of fibers and aggre- gates make conventional structural studies difficult, extensive solid-state and solution-state NMR investiga- tions of both Ab(1–40) and Ab(1–42) fibrils have dem- onstrated that the peptides stack perpendicularly along the fibril axis via a parallel in-register assembly of a b-strand–turn–b-strand motif, forming a characteristic cross-b structure with a highly solvent-protected core of two intermolecular b-sheets [10–13]. X-ray micro- crystallography of short Ab fragments has revealed very tight packing of side chains within the cross-b structure [14]. The aggregation properties of Abs are strongly affected by their affinities for certain metal ions [15– 19]. Elevated concentrations of Cu 2+ ,Zn 2+ and Fe 2+ are found within deposited plaques in vivo [20], and both Cu 2+ and Zn 2+ affect the toxicity of Ab [21–25]. The in vivo effect of metals can be complex [26,27], and the presence of Cu 2+ ,Zn 2+ or Ca 2+ significantly accelerates aggregation of Ab in vitro [19,28,29]. The binding of Cu 2+ and Zn 2+ to monomeric Ab has been extensively studied and occurs via three histidines (His6, His13, and His14) and a fourth ligand suggested to be either Tyr10 [30–32], Glu11 [33], or the N-termi- nus [34,35]. A weaker binding site for zinc, involving Asp23–Lys28, has also been identified [35]. Ab assem- blies formed in the presence of metals are not charac- terized as amyloid, because they lack tinctorial properties and fibrillar morphology, and they can be completely dissociated using chelating agents or by a slight increases in pH [19,28]. Interesting parallels can be drawn between these assemblies and the diffuse pla- ques found in AD patients, which also lack fibrillar morphology and which are largely solubilized by che- lating agents [36]. Intriguingly, it appears that metal binding induces a fold that effectively protects against further propagation into fibrillar form [37] and that removal of the bound metals causes dissociation of the peptide aggregates rather than further assembly into amyloid. Identification of structural differences between the fibrillar and the metal-bound aggregate forms is therefore of interest with respect to under- standing the mode of assembly and pinpointing the mechanism by which metals guide the path of aggre- gation away from the fibrillar trail. The present investigation assessed how the growth and architecture of Ab(1–40) and Ab(1–42) aggregates are affected by the presence or absence of Cu 2+ or Zn 2+ , using turbidity measurements, thioflavin T (ThT) fluorescence, CD spectroscopy, atomic force microscopy (AFM), and, in particular, a very powerful quenched hydrogen ⁄ deuterium (H ⁄ D) exchange NMR method. Although this NMR method was developed and utilized for studies of amyloid fibrils [12,38–43], it is equally applicable to the quantitative and residue- specific identification of the core structures in metal- bound Ab aggregates. Significant structural changes with an overall less protected structure were observed in the metal-complexed Ab aggregates as compared with the Ab fibrils. Differences were also identified between Cu 2+ -bound and Zn 2+ -bound aggregates and between Ab(1–40) and Ab(1–42) aggregates, indicating that both the type of bound metal and the peptide sequence dictate the mode of peptide assembly. Results and Discussion Ab aggregation in the presence of Cu 2+ and Zn 2+ Turbidity measurements showed that the aggregation rates of both Ab(1–40) and Ab(1–42) were significantly increased in the presence of either Cu 2+ or Zn 2+ (Fig. 1A), in agreement with previous investigations [19,28,44,45]. Both Ab(1–40) and Ab(1–42) reached a plateau where aggregation was essentially complete after 40 min (Fig. 1A). It is noteworthy that the pla- teau values for both peptides were considerably higher when Zn 2+ was added, suggesting a different mode of assembly within the metal-induced aggregates. Addi- tion of EDTA to our samples efficiently reversed the aggregation and reduced the intensity to background levels (Fig. 1A), which parallels previous findings [19,28]. During the time of the analysis, no detectable turbidity was observed for either Ab(1–40) or Ab(1–42) in the absence of divalent metals, i.e. when EDTA was included in the reaction mixture. SDS ⁄ PAGE and gel filtration analysis of the EDTA-containing samples confirmed that Ab was present in its monomeric form (data not shown). NMR analysis of metal-complexed Ab aggregates A. Olofsson et al. 4052 FEBS Journal 276 (2009) 4051–4060 ª 2009 The Authors Journal compilation ª 2009 FEBS The amount of amyloid or amyloid-like structure within the different Ab assemblies was quantified by ThT fluorescence. To ensure that trace amounts of divalent metals did not affect the formation and analy- sis of the fibril samples, a chelator in a two-fold molar excess of the peptide was present in all experiments on fibrillar samples. With Cu 2+ present, approximately 10% of the normal fibrillar ThT signal for both Ab(1–40) and Ab(1–42) aggregates was detected, whereas Zn 2+ -induced aggregates resulted in a larger ThT response, corresponding to approximately 40% of the fibrillar counterpart (Fig. 1B). A considerable change between the spectra from metal-complexed Ab aggregates and from fibrils was detected using CD spectroscopy (Fig. 1C). The spectra from the metal-complexed peptides were characteristic for aggregated samples, and suggest a lower degree of secondary structure content as compared with the spectra from the fibril samples. The CD measurements were carried out at two different protein concentra- tions (50 and 100 lm) to ensure that the results were not biased by the inherent light scattering of aggre- gates (data shown only for 50 lm). All forms of aggregates were examined using AFM (Fig. 2). The results were in accordance with previous investigations in which aggregates formed in the pres- ence of a molar excess of divalent metal ions failed to attain the classic morphology of amyloid fibrils [37,46]. The aggregates of Ab(1–40) and Ab(1–42) formed in the presence of a chelating agent gave rise to the clas- sic fibrillar morphology, with an average fibril diame- ter of about 7 nm. It is well known that Ab(1–40) and Ab(1–42) fibrils consist of bundles of thinner filaments, each having a cross-section of about 3 nm [47]. This was most clearly seen within the Ab(1–42) variant (Fig. 2F), which also had a twinned ultrastructure. For both Ab(1–40) and Ab(1–42), the presence of divalent metal ions efficiently inhibited fibril formation and, as expected, resulted in more amorphous aggregates. It was not, however, possible to morphologically distin- guish aggregates formed in the presence of Cu 2+ from those formed in the presence of Zn 2+ . Taken together, these results suggest that Ab aggre- gates formed in the presence of Cu 2+ or Zn 2+ are less compact than fibrillar complexes and are assembled from Abs with altered secondary structure. The results of ThT analysis suggest that Zn 2+ -complexed Ab may tolerate assembly into a more fibrillar-like fold, whereas Cu 2+ more efficiently prevents formation of a fibrillar assembly. Furthermore, all metal-bound Ab aggregates could be efficiently converted to monomers through the addition of a chelating agent, suggesting that their mode of assembly is different from that of the fibrillar forms. Quenched H ⁄ D exchange NMR on Ab assemblies The use of quenched H⁄ D exchange in combination with NMR spectroscopy is a highly useful technique with which to efficiently pinpoint the solvent accessibil- ity of individual amide protons within peptide assem- blies in a residue-specific and quantitative manner, thereby providing detailed structural information [12,38–43]. In this study, aggregates of Ab(1–40) and AB C Fig. 1. Abs were characterized by using (A) turbidity measurements, (B) a ThT assay, and (C) CD spectroscopy. (A) Turbidity measurements on samples containing 50 l M Ab(1–40) (solid line) and Ab(1–42) (broken lines) were started immediately after the addition of Cu 2+ [(iii) and (iv)], Zn 2+ [(i) and (ii)] or EDTA [(v) and (vi)] and continued for 164 min. After 100 min, 400 lM EDTA was added, and the absorbance was measured for an additional 64 min. (B) ThT analysis of aggregated Ab samples containing 100 l M Ab(1–40) and Ab(1–42) in the presence of Cu 2+ (light gray bars), Zn 2+ (open bars), or EDTA (dark gray bars). The respective emissions of the fibrillar forms of Ab(1–40) and Ab(1–42) were set to 100% intensity. (C) Far-UV CD spectra from samples containing 50 l M Ab(1–40) (solid lines) or Ab(1–42) (broken lines), either in the form of aggregates formed in the presence of Cu 2+ [(i) and (iii)] or Zn 2+ [(ii) and (iv)], or as fibrils formed in the presence of EDTA [(v) and (vi)]. A. Olofsson et al. NMR analysis of metal-complexed Ab aggregates FEBS Journal 276 (2009) 4051–4060 ª 2009 The Authors Journal compilation ª 2009 FEBS 4053 Ab(1–42) peptides in the presence of Cu 2+ ,Zn 2+ or EDTA were investigated. The chelating agent EDTA was used to remove trace amounts of divalent metals that may affect fibril formation. H ⁄ D exchange was carried out by preincubating samples in D 2 O for 24 h before analysis. The length of incubation ensures detection of only those amide protons that are pro- tected as a result of hydrogen bonding within second- ary structure elements or because they are deeply buried in the core of the fibril [12,13]. Samples verified by AFM and ThT analysis to con- tain the fibrillar forms of Ab(1–40) and Ab(1–42) (fibrils formed in the presence of EDTA) displayed amide protection patterns with two bell-shaped pro- tected regions (Fig. 3C,F). Ab(1–40) fibrils showed partial to full protection for Glu3–Ser26 and Gly29– Val40, separated by exposed residues in the turn region centered at Asn27–Lys28 (Fig. 3C), whereas the pro- tection pattern of the fibrillar form of Ab(1–42) included residues Phe4–Arg5, Tyr10–Ser26, and Gly29–Ala42 (Fig. 3F). The degree of protection in the N-terminal region of Ab(1–42) is notably lower than in Ab(1–40). These results are similar to those of our recently published investigations on Ab(1–40) and Ab(1–42) fibrils [12,13]; the small differences observed are probably due to the use of different growth condi- tions during fibril preparation [150 mm NaCl (pH 7.4) and 1 mm EDTA versus 50 mm NaCl (pH 7.0) in the previous investigations]. The protection patterns observed are in good agreement with current models extrapolated from solid-state NMR data [10,48], dou- ble compensatory mutagenesis combined with H ⁄ D exchange NMR [11], and cysteine scanning mutagene- sis [49], where the Abs (which form two b-strands with a turn region involving Val24–Ala30) stack in an in- register parallel arrangement to form the fibril [10,48] (Fig. 4). The pattern for Ab(1–40) fibrils also fits well with our recent report, in which a new general method to quantitatively determine the exchange rates of amide protons within fibrils is described and applied to fibrils formed by Ab(1–40) [43]. The presence of either Cu 2+ or Zn 2+ during aggre- gate ⁄ fibril growth had a significant impact on the solvent protection pattern for both Ab(1–40) and Ab(1–42). When compared to the protection ratios in the absence of metal (Fig. 3C,F), the overall protection ratio in the presence of metal was significantly reduced (Fig. 3A,B,D,E). In the metal-complexed aggregates, the Asp1–Lys16 region essentially lacks solvent protec- tion. This is reasonable, as metal binding occurs within the first 14 residues, inducing a low degree of second- ary structure [33,50]. However, the remaining parts of both Ab(1–40) and Ab(1–42) still display two fairly well-protected bell-shaped regions with an intervening minimum centered on Asn27. This is strong evidence that metal-complexed Ab aggregates attain a b-strand- turn–b-strand structural arrangement similar to the structural arrangement within mature fibrils (Fig. 4). The metal-complexed Ab(1–40) aggregates also show pronounced differences from the fibrillar form in the C-terminal region as well as in the turn region of the peptide. Assuming a fibril-like structural arrangement in which two protofilaments formed from stacked pep- tides are laterally assembled, the N-terminal metal- binding region of the peptide will be in close proximity to both the C-terminal and turn residues, possibly interfering with these regions. This explains how an altered conformation in the N-terminal region upon metal binding can affect sections of the molecule closer to the C-terminus (see the fibril cross-section in Fig. 4). Furthermore, differences in solvent protection are A B C D E F Fig. 2. Tapping mode AFM images of recombinant Ab(1–40) and Ab(1–42) aggregates. (A–C) Morphology of Ab(1–40) in the pres- ence of Cu 2+ ,Zn 2+ , and EDTA, respectively. (D–F) Morphology of Ab(1–42) in the presence of Cu 2+ ,Zn 2+ , and EDTA, respectively. Scale bar: 1 lm. NMR analysis of metal-complexed Ab aggregates A. Olofsson et al. 4054 FEBS Journal 276 (2009) 4051–4060 ª 2009 The Authors Journal compilation ª 2009 FEBS observed between the Cu 2+ -induced and Zn 2+ -induced aggregates, reflecting the differences in metal-binding properties of Ab [26,30,33–35]. Zn 2+ significantly reduces the overall protection, in particular for residues close to the turn region (Fig. 3B). For Zn 2+ -induced Ab(1–40) aggregates, the solvent-protected residues include Leu17–Glu22 and Ile31–Val36 (Fig. 3B), whereas in Cu 2+ -induced Ab(1–40) aggregates, Leu17– Gly25 and Lys28–Val36 are protected (Fig. 3A). The increased solvent accessibility of the turn region within Zn 2+ -induced Ab(1–40) aggregates correlates well with a proposed second weak binding site for Zn 2+ involving Asp23–Lys28 in the turn region [35]. Comparison of the solvent protection pattern of Cu 2+ -induced and Zn 2+ -induced Ab(1–42) aggregates (Fig. 3D,E) with Ab(1–42) fibrils (Fig. 3F) reveals a reduced overall level of protection and completely exposed residues in the N-terminal region. The solvent-protected residues in the Ab(1–42) metal- complexed aggregates (Leu17–Glu22 ⁄ Asp23 and Ile31– Ile41) form two characteristic bell-shaped regions, suggesting a mode of assembly similar to that of the fibrillar samples. Figure 3D,E demonstrates that there is no significant difference in solvent exposure in the turn regions of A b(1–42) aggregates formed in the presence of either Cu 2+ or Zn 2+ . The C-terminal region in the Ab(1–42) aggregates is affected by metals to a lesser extent than the C-terminal region of the Ab(1–40) aggregates, suggesting they are assembled in a way that places the C-terminal residues in the aggre- gate core. This is similar to the arrangement found in Ab(1–42) fibrils, where the C-terminal residues strengthen the hydrophobic interactions within the core of the fibrillar fold via a shift in the protofilament assembly that positions the C-terminal residues of Ab(1–42) in a more hydrophobic environment [10,48] (Fig. 4). Implications for aggregate ⁄ fibril formation Although the properties of fibrils and metal-bound aggregates of Abs differ significantly, the H ⁄ D protec- tion patterns of metal-bound Ab aggregates still sug- gest a partly preserved b-strand–turn–b-strand fold similar to the fibrillar form (Fig. 3). The metal-induced Ab aggregates have fewer protected residues and a A B C D E F Fig. 3. Solvent protection ratios for the backbone amide protons of Ab(1–40) and Ab(1–42) aggregates. Protection is defined as the ratio of the observed intensity after a 24 h preincubation period in D 2 O over the intensity in a completely protonated sample (defined as 100%). (A–C) Solvent protection ratios for Ab(1–40) aggregates in the presence of Cu 2+ ,Zn 2+ , and EDTA, respectively. (D–F) Ratios for Ab(1–42) aggregates in the presence of Cu 2+ ,Zn 2+ , and EDTA, respectively. Rings correspond to 0% protection, and crosses represent residues that exchange too quickly to be detected. Error bars indicate the experimental uncertainty of the measurements. A. Olofsson et al. NMR analysis of metal-complexed Ab aggregates FEBS Journal 276 (2009) 4051–4060 ª 2009 The Authors Journal compilation ª 2009 FEBS 4055 lower degree of solvent protection than fibrils, suggest- ing that formation of b-strands and subsequently fibrils is hindered, and that fewer intermolecular inter- actions are needed for assembly. This scenario would explain the high aggregation rates observed in the tur- bidity measurements (Fig. 1A). The pronounced lack of H ⁄ D protection close to the known metal-binding site within the N-terminal regions of both Ab(1–40) and Ab(1–42) clearly shows that only the protected C-terminal regions are involved in the assembly. How- ever, assembly in the presence of metals is unlikely to involve the same intermolecular interactions as within fibrils, as removal of the metals causes dissociation into monomers instead of the continued assembly into amyloid fibrils expected from a more fibril-like fold. One speculative explanation of this behavior is that, in the presence of metals, the peptides assemble into an array with b-strand–turn–b-strand motifs (Fig. 4) stacked in an antiparallel fashion, distinctly different from the parallel assembly within fibrils. This sugges- tion is supported by previous observations that peptide fragments of Ab preferably stack in an antiparallel fashion along the fibril axis [14,51–53]. In conclusion, our results support the previous notion that metal binding by the N-terminal residues (Asp1– Lys16) prevents assembly into a fibrillar structure. On the basis of our quenched H ⁄ D exchange NMR data, we propose that both Cu 2+ -induced and Zn 2+ -induced Ab aggregates assemble via a b-strand–turn–b-strand motif, resembling the motif found in fibrils. Experimental procedures Isotope-enriched chemicals were purchased from Cambridge Isotope Laboratories (Andover, MA, USA). All peptides, including uniformly 15 N-labeled Ab(1–40) and Ab(1–42), were obtained in lyophilized form from Alexotech AB (Umea ˚ , Sweden) (http://www.alexotech.com). Reagents and buffers were purchased from Sigma-Aldrich (St Louis, MO, USA), unless otherwise stated. Sample preparation Lyophilized Ab(1–40) and Ab(1–42) were briefly dissolved in 10 mm NaOH, sonicated for 30 s, and centrifuged (10 min at 12 000 g) to remove residual oligomeric species, as previously described [54]. This treatment efficiently monomerizes the peptides and facilitates dilution in 20 mm Tris (pH 7.4) and 150 mm NaCl (for turbidity measure- ments) or 10 mm Tris (pH 7.4) and 150 m m NaCl (for ThT assay and CD spectroscopy) to a peptide concentration of either 50 or 100 lm, with a two-fold molar excess of CuCl 2 , ZnCl 2 , or EDTA. Fibrils ⁄ aggregates were formed by incu- bating the samples for 14 days at 37 °C with agitation. Turbidity measurements Turbidity was measured by recording the absorbance at 405 nm. Measurements on samples containing 50 lm Ab(1–40) and Ab(1–42) were started immediately after the addition of Cu 2+ ,Zn 2+ , or EDTA, and thereafter recorded for 164 min. The samples were kept still at room tempera- ture except from 5 s before each measurement. Each mea- surement was performed in triplicate. After 100 min, A B C D E F Fig. 4. The determined solvent protection ratios for residues within Ab(1–40) and Ab(1–42) aggregates are mapped onto corresponding dimer of two cross-b units taken from cross-sections of the Ab(1–40) and Ab(1–42) fibril (aggregate) models, respectively. The color code is varied between the following extremes: navy blue for complete, and red for no solvent protection. Residues with no pro- tection ratios available are depicted in gray. Cu 2+ is shown in brown, and Zn 2+ in bluish gray. (A–C) Protection ratios of Ab(1–40) in the presence of Cu 2+ ,Zn 2+ , and EDTA, respectively. (D–F) Pro- tection ratios for Ab(1–42) in presence of Cu 2+ ,Zn 2+ , and EDTA, respectively. The image was prepared in MOLMOL [55]. NMR analysis of metal-complexed Ab aggregates A. Olofsson et al. 4056 FEBS Journal 276 (2009) 4051–4060 ª 2009 The Authors Journal compilation ª 2009 FEBS 400 lm EDTA was added to all wells and the absorbance was measured for an additional 64 min. ThT analysis Samples containing 100 lm aggregated Abs were mixed with 10 lm ThT and 50 lm phosphate buffer (pH 6.5) in a 10 mm quartz cuvette. Emission intensities were recorded at 25 °C, using a JASCO spectrofluorometer (Jasco Int. Co., Ltd., Tokyo, Japan), with excitation and emission wavelengths of 450 and 482 nm, respectively, and a 3 nm bandwidth for both emission and excitation. Each measure- ment was performed in triplicate. CD spectroscopy Far-UV CD spectra were collected on samples containing either 50 or 100 lm Ab, using a JASCO J-810 spectropola- rimeter (Jasco) at 25 °C. Spectra were recorded between 200 and 250 nm, and averaged over 10 scans with a band- width of 1 nm, a response time of 1 s, a pitch of 0.5 nm, and a scan rate of 20 nmÆs )1 , in a 2 mm quartz cuvette. AFM A portion of each Ab aggregate sample was diluted in water to approximately 1 lm peptide and applied to freshly cleaved ruby red mica (Goodfellow, Cambridge, UK). The material was allowed to adsorb for 30 s, and then washed with distilled water three times and air dried. AFM analysis was performed using a Nanoscope IIIa multimode atomic force microscope (Digital Instruments, Santa Barbara, USA) in tapping mode in air. A silicon probe was oscillated at abount 280 kHz, and images were collected at an opti- mized scan rate corresponding to 1 Hz. Quenched H ⁄ D exchange NMR Lyophilized 15 N-labelled Ab(1–40) and Ab(1–42) (obtained from Alexotech AB) (http://www.alexotech.com) were monomerized as described above, and this was followed by addition of 10· buffer (20 mm Tris, pH 7.4, 150 mm NaCl) to a final peptide concentration of 500 lm. A two-fold molar excess of either ZnCl 2 , CuCl 2 or EDTA was added to each sample. Fibril ⁄ aggregate solutions were prepared (5–8 days of incubation at 37 °C with agitation at 130 r.p.m.) and recovered by centrifugation (2 min at 13 000 g), and each was then split into two; one half was preincubated in D 2 O for 24 h (by dissolving the pellet 30 times in 20 mm Tris, pD 7.0, 150 mm NaCl and a two-fold molar excess of ZnCl 2 , CuCl 2 , or EDTA), and one served as a fully protonated reference sample. At the end of the incubation period and immediately prior to NMR analysis, the Ab assemblies were recovered by short centrifugation steps (13 000 g) and subsequently dissociated into NMR- detectable monomers (to  2mm monomeric Ab) using an optimized solution of hexafluoroisopropanol as described in [12], with the addition of 2 m m diethylenetriamine penta- acetic acid as chelating agent. NMR data were recorded and analyzed as previously described [12,13], resulting in residue-specific solvent protection ratios for the backbone amide protons within Ab(1–40) and Ab(1–42) aggregates. Observed ratios were mapped onto models of the various aggregates. These models are based on our previous fibril models [12,13], the model of Ab(9–40) by Tycko et al. [10], the solution structure of Ab(1–16) with and without bound Zn 2+ [33], and the proposed filament packing arrangement for Ab(1–42) [10,48]. Modifications and energy minimiza- tion were performed in molmol [55] and swiss-pdbviewer [56], respectively. Acknowledgements We thank R. Tycko for kindly providing us with the coordinates of the Ab(9–40) amyloid model. This work was supported by the Magn. Bergvalls Foundation, Carl Trygger Foundation, Alzheimerfonden, Socialsty- relsen, Hja ¨ rnfonden, A ˚ ke Wibergs Foundation, Bern- hard och Signe Ba ¨ ckstro ¨ ms stiftelse, O. E. och Edla Johanssons vetenskapliga stiftelse, Go ¨ ran Gustafssons Foundation, Swedish Research Science Council, Ernst Schering Foundation, Centre for Biomedical Engineer- ing at Wrocaw University of Technology, Gun och Bertil Stohnes stiftelse, Loo och Hans Ostermans stift- else fo ¨ r geriatrisk forskning and patients’ association FAMY ⁄ AMYL. References 1 Chiti F & Dobson CM (2006) Protein misfolding, func- tional amyloid, and human disease. Annu Rev Biochem 75, 333–366. 2 Iwatsubo T, Odaka A, Suzuki N, Mizusawa H, Nukina N & Ihara Y (1994) Visualization of A beta 42(43) and A beta 40 in senile plaques with end-specific A beta monoclonals: evidence that an initially deposited species is A beta42(43). Neuron 13, 45–53. 3 Roher AE, Chaney MO, Kuo YM, Webster SD, Stine WB, Haverkamp LJ, Woods AS, Cotter RJ, Tuohy JM, Krafft GA et al. (1996) Morphology and toxicity of Abeta-(1-42) dimer derived from neuritic and vascular amyloid deposits of Alzheimer’s disease. J Biol Chem 271, 20631–20635. 4 Lambert MP, Barlow AK, Chromy BA, Edwards C, Freed R, Liosatos M, Morgan TE, Rozovsky I, Trom- mer B, Viola KL et al. (1998) Diffusible, nonfibrillar ligands derived from Abeta1–42 are potent central A. Olofsson et al. NMR analysis of metal-complexed Ab aggregates FEBS Journal 276 (2009) 4051–4060 ª 2009 The Authors Journal compilation ª 2009 FEBS 4057 nervous system neurotoxins. Proc Natl Acad Sci USA 95, 6448–6453. 5 Lacor PN, Buniel MC, Chang L, Fernandez SJ, Gong Y, Viola KL, Lambert MP, Velasco PT, Bigio EH, Finch CE et al. (2004) Synaptic targeting by Alzhei- mer’s-related amyloid beta oligomers. J Neurosci 24, 10191–10200. 6 Klein WL, Stine WB Jr & Teplow DB (2004) Small assemblies of unmodified amyloid beta-protein are the proximate neurotoxin in Alzheimer’s disease. Neurobiol Aging 25, 569–580. 7 Necula M, Kayed R, Milton S & Glabe CG (2007) Small molecule inhibitors of aggregation indicate that amyloid beta oligomerization and fibrillization path- ways are independent and distinct. J Biol Chem 282, 10311–10324. 8 Walsh DM, Lomakin A, Benedek GB, Condron MM & Teplow DB (1997) Amyloid beta-protein fibrillogenesis. Detection of a protofibrillar intermediate. J Biol Chem 272, 22364–22372. 9 Kheterpal I, Lashuel HA, Hartley DM, Walz T, Lansbury PT Jr & Wetzel R (2003) Abeta protofibrils possess a stable core structure resistant to hydrogen exchange. Biochemistry 42, 14092–14098. 10 Petkova AT, Yau WM & Tycko R (2006) Experimental constraints on quaternary structure in Alzheimer’s beta-amyloid fibrils. Biochemistry 45, 498–512. 11 Lu ¨ hrs T, Ritter C, Adrian M, Riek-Loher D, Bohrmann B, Do ¨ beli H, Schubert D & Riek R (2005) 3D structure of Alzheimer’s amyloid-beta(1-42) fibrils. Proc Natl Acad Sci USA 102, 17342–17347. 12 Olofsson A, Sauer-Eriksson AE & O ¨ hman A (2006) The solvent protection of alzheimer amyloid-beta-(1-42) fibrils as determined by solution NMR spectroscopy. J Biol Chem 281, 477–483. 13 Olofsson A, Lindhagen-Persson M, Sauer-Eriksson AE &O ¨ hman A (2007) Amide solvent protection analysis demonstrates that amyloid-beta(1-40) and amyloid-beta(1-42) form different fibrillar structures under identical conditions. Biochem J 404, 63–70. 14 Sawaya MR, Sambashivan S, Nelson R, Ivanova MI, Sievers SA, Apostol MI, Thompson MJ, Balbirnie M, Wiltzius JJ, McFarlane HT et al. (2007) Atomic structures of amyloid cross-beta spines reveal varied steric zippers. Nature 447, 453–457. 15 Bush AI, Pettingell WH, Multhaup G, Paradis M, Vonsattel JP, Gusella JF, Beyreuther K, Masters CL & Tanzi RE (1994) Rapid induction of Alzheimer A beta amyloid formation by zinc. Science 265, 1464–1467. 16 Garzon-Rodriguez W, Yatsimirsky AK & Glabe CG (1999) Binding of Zn(II), Cu(II), and Fe(II) ions to Alzheimer’s A beta peptide studied by fluorescence. Bioorg Med Chem Lett 9, 2243–2248. 17 Tougu V, Karafin A & Palumaa P (2008) Binding of zinc(II) and copper(II) to the full-length Alzheimer’s amyloid-beta peptide. J Neurochem 104, 1249–1259. 18 Noy D, Solomonov I, Sinkevich O, Arad T, Kjaer K & Sagi I (2008) Zinc-amyloid beta interactions on a millisecond time-scale stabilize non-fibrillar Alzheimer- related species. J Am Chem Soc 130, 1376–1383. 19 Atwood CS, Moir RD, Huang X, Scarpa RC, Bacarra NM, Romano DM, Hartshorn MA, Tanzi RE & Bush AI (1998) Dramatic aggregation of Alzheimer abeta by Cu(II) is induced by conditions representing physiologi- cal acidosis. J Biol Chem 273, 12817–12826. 20 Lovell MA, Robertson JD, Teesdale WJ, Campbell JL & Markesbery WR (1998) Copper, iron and zinc in Alzheimer’s disease senile plaques. J Neurol Sci 158, 47–52. 21 Huang X, Cuajungco MP, Atwood CS, Hartshorn MA, Tyndall JD, Hanson GR, Stokes KC, Leopold M, Multhaup G, Goldstein LE et al. (1999) Cu(II) potenti- ation of alzheimer abeta neurotoxicity. Correlation with cell-free hydrogen peroxide production and metal reduc- tion. J Biol Chem 274, 37111–37116. 22 Yoshiike Y, Tanemura K, Murayama O, Akagi T, Murayama M, Sato S, Sun X, Tanaka N & Takashima A (2001) New insights on how metals disrupt amyloid beta-aggregation and their effects on amyloid-beta cytotoxicity. J Biol Chem 276, 32293–32299. 23 Garai K, Sahoo B, Kaushalya SK, Desai R & Maiti S (2007) Zinc lowers amyloid-beta toxicity by selectively precipitating aggregation intermediates. Biochemistry 46, 10655–10663. 24 Lovell MA, Xie C & Markesbery WR (1999) Protection against amyloid beta peptide toxicity by zinc. Brain Res 823, 88–95. 25 Huang X, Cuajungco MP, Atwood CS, Moir RD, Tanzi RE & Bush AI (2000) Alzheimer’s disease, beta- amyloid protein and zinc. J Nutr 130, 1488S–1492S. 26 Bush AI (2003) The metallobiology of Alzheimer’s disease. Trends Neurosci 26, 207–214. 27 Phinney AL, Drisaldi B, Schmidt SD, Lugowski S, Coronado V, Liang Y, Horne P, Yang J, Sekoulidis J, Coomaraswamy J et al. (2003) In vivo reduction of amyloid-beta by a mutant copper transporter. Proc Natl Acad Sci USA 100, 14193–14198. 28 Huang X, Atwood CS, Moir RD, Hartshorn MA, Vonsattel JP, Tanzi RE & Bush AI (1997) Zinc-induced Alzheimer’s Abeta1-40 aggregation is mediated by conformational factors. J Biol Chem 272, 26464– 26470. 29 Isaacs AM, Senn DB, Yuan M, Shine JP & Yankner BA (2006) Acceleration of amyloid beta-peptide aggregation by physiological concentrations of calcium. J Biol Chem 281, 27916–27923. 30 Curtain CC, Ali F, Volitakis I, Cherny RA, Norton RS, Beyreuther K, Barrow CJ, Masters CL, Bush AI NMR analysis of metal-complexed Ab aggregates A. Olofsson et al. 4058 FEBS Journal 276 (2009) 4051–4060 ª 2009 The Authors Journal compilation ª 2009 FEBS & Barnham KJ (2001) Alzheimer’s disease amyloid-beta binds copper and zinc to generate an allosterically ordered membrane-penetrating structure containing superoxide dismutase-like subunits. J Biol Chem 276, 20466–20473. 31 Tickler AK, Smith DG, Ciccotosto GD, Tew DJ, Curtain CC, Carrington D, Masters CL, Bush AI, Cherny RA, Cappai R et al. (2005) Methylation of the imidazole side chains of the Alzheimer disease amyloid-beta peptide results in abolition of superoxide dismutase-like structures and inhibition of neurotoxic- ity. J Biol Chem 280, 13355–13363. 32 Antzutkin ON (2004) Amyloidosis of Alzheimer’s Abeta peptides: solid-state nuclear magnetic resonance, electron paramagnetic resonance, transmission electron microscopy, scanning transmission electron microscopy and atomic force microscopy studies. Magn Reson Chem 42, 231–246. 33 Zirah S, Kozin SA, Mazur AK, Blond A, Cheminant M, Segalas-Milazzo I, Debey P & Rebuffat S (2006) Structural changes of region 1–16 of the Alzheimer disease amyloid beta-peptide upon zinc binding and in vitro aging. J Biol Chem 281, 2151– 2161. 34 Syme CD, Nadal RC, Rigby SE & Viles JH (2004) Copper binding to the amyloid-beta (Abeta) peptide associated with Alzheimer’s disease: folding, coordina- tion geometry, pH dependence, stoichiometry, and affinity of Abeta-(1-28): insights from a range of complementary spectroscopic techniques. J Biol Chem 279, 18169–18177. 35 Danielsson J, Pierattelli R, Banci L & Gra ¨ slund A (2007) High-resolution NMR studies of the zinc-binding site of the Alzheimer’s amyloid beta-peptide. FEBS J 274, 46–59. 36 Cherny RA, Legg JT, McLean CA, Fairlie DP, Huang X, Atwood CS, Beyreuther K, Tanzi RE, Masters CL & Bush AI (1999) Aqueous dissolution of Alzheimer’s disease Abeta amyloid deposits by biometal depletion. J Biol Chem 274, 23223–23228. 37 Smith DP, Ciccotosto GD, Tew DJ, Fodero-Tavoletti MT, Johanssen T, Masters CL, Barnham KJ & Cappai R (2007) Concentration dependent Cu 2+ induced aggregation and dityrosine formation of the Alzheimer’s disease amyloid-beta peptide. Biochemistry 46, 2881– 2891. 38 Alexandrescu AT (2001) An NMR-based quenched hydrogen exchange investigation of model amyloid fibrils formed by cold shock protein A. Pac Symp Biocomput 6, 67–78. 39 Reference withdrawn. 40 Ippel JH, Olofsson A, Schleucher J, Lundgren E & Wijmenga SS (2002) Probing solvent accessibility of amyloid fibrils by solution NMR spectroscopy. Proc Natl Acad Sci USA 99, 8648–8653. 41 Olofsson A, Ippel JH, Wijmenga SS, Lundgren E & O ¨ hman A (2004) Probing solvent accessibility of trans- thyretin amyloid by solution NMR spectroscopy. J Biol Chem 279, 5699–5707. 42 Hoshino M, Katou H, Hagihara Y, Hasegawa K, Naiki H & Goto Y (2002) Mapping the core of the beta(2)- microglobulin amyloid fibril by H ⁄ D exchange. Nat Struct Biol 9, 332–336. 43 Olofsson A, Sauer-Eriksson AE & O ¨ hman A (2009) Amyloid fibril dynamics revealed by combined hydro- gen ⁄ deuterium exchange and nuclear magnetic reso- nance. Anal Biochem 385, 374–376. 44 Huang X, Atwood CS, Moir RD, Hartshorn MA, Tanzi RE & Bush AI (2004) Trace metal contamination initiates the apparent auto-aggregation, amyloidosis, and oligomerization of Alzheimer’s Abeta peptides. J Biol Inorg Chem 9, 954–960. 45 House E, Collingwood J, Khan A, Korchazkina O, Berthon G & Exley C (2004) Aluminium, iron, zinc and copper influence the in vitro formation of amyloid fibrils of Abeta42 in a manner which may have consequences for metal chelation therapy in Alzheimer’s disease. J Alzheimers Dis 6, 291–301. 46 Ha C, Ryu J & Park CB (2007) Metal ions differentially influence the aggregation and deposition of Alzheimer’s beta-amyloid on a solid template. Biochemistry 46, 6118–6125. 47 Malinchik SB, Inouye H, Szumowski KE & Kirschner DA (1998) Structural analysis of Alzheimer’s beta(1–40) amyloid: protofilament assembly of tubular fibrils. Bio- phys J 74, 537–545. 48 Sato T, Kienlen-Campard P, Ahmed M, Liu W, Li H, Elliott JI, Aimoto S, Constantinescu SN, Octave JN & Smith SO (2006) Inhibitors of amyloid toxicity based on beta-sheet packing of Abeta40 and Abeta42. Bio- chemistry 45, 5503–5516. 49 Shivaprasad S & Wetzel R (2006) Scanning cysteine mutagenesis analysis of Abeta-(1-40) amyloid fibrils. J Biol Chem 281, 993–1000. 50 Ma QF, Hu J, Wu WH, Liu HD, Du JT, Fu Y, Wu YW, Lei P, Zhao YF & Li YM (2006) Characterization of copper binding to the peptide amyloid-beta(1-16) associated with Alzheimer’s disease. Biopolymers 83, 20–31. 51 Bu Z, Shi Y, Callaway DJ & Tycko R (2007) Molecular alignment within beta-sheets in Abeta(14-23) fibrils: solid-state NMR experiments and theoretical predic- tions. Biophys J 92, 594–602. 52 Klimov DK & Thirumalai D (2003) Dissecting the assembly of Abeta16-22 amyloid peptides into antipar- allel beta sheets. Structure 11, 295–307. 53 Lansbury PT Jr, Costa PR, Griffiths JM, Simon EJ, Auger M, Halverson KJ, Kocisko DA, Hendsch ZS, Ashburn TT, Spencer RG et al. (1995) Structural model for the beta-amyloid fibril based on interstrand align- A. Olofsson et al. NMR analysis of metal-complexed Ab aggregates FEBS Journal 276 (2009) 4051–4060 ª 2009 The Authors Journal compilation ª 2009 FEBS 4059 ment of an antiparallel-sheet comprising a C-terminal peptide. Nat Struct Biol 2, 990–998. 54 Fezoui Y, Hartley DM, Harper JD, Khurana R, Walsh DM, Condron MM, Selkoe DJ, Lansbury PT Jr, Fink AL & Teplow DB (2000) An improved method of preparing the amyloid beta-protein for fibrillogenesis and neurotoxicity experiments. Amyloid 7, 166–178. 55 Koradi R, Billeter M & Wu ¨ thrich K (1996) MOLMOL: a program for display and analysis of macromolecular structures. J Mol Graph 14, 51–55, 29–32. 56 Guex N & Peitsch MC (1997) SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis 18, 2714–2723. NMR analysis of metal-complexed Ab aggregates A. Olofsson et al. 4060 FEBS Journal 276 (2009) 4051–4060 ª 2009 The Authors Journal compilation ª 2009 FEBS . distin- guish aggregates formed in the presence of Cu 2+ from those formed in the presence of Zn 2+ . Taken together, these results suggest that Ab aggre- gates formed in the presence of Cu 2+ or Zn 2+ are. the form of aggregates formed in the presence of Cu 2+ [(i) and (iii)] or Zn 2+ [(ii) and (iv)], or as fibrils formed in the presence of EDTA [(v) and (vi)]. A. Olofsson et al. NMR analysis of metal-complexed. Quenched hydrogen ⁄ deuterium exchange NMR characterization of amyloid-b peptide aggregates formed in the presence of Cu 2+ or Zn 2+ Anders Olofsson 1 , Malin Lindhagen-Persson 1 ,