Supramolecular structural constraints on Alzheimer's beta-amyloid fibrils from electron microscopy and solid-state nuclear magnetic resonance

We describe electron microscopy (EM), scanning transmission electron microscopy (STEM), and solid-state nuclear magnetic resonance (NMR) measurements on amyloid fibrils formed by the 42-residue beta-amyloid peptide associated with Alzheimer's disease (Abeta(1)(-)(42)) and by residues 10-35 of the full-length peptide (Abeta(10)(-)(35)). These measurements place constraints on the supramolecular structure of the amyloid fibrils, especially the type of beta-sheets present in the characteristic amyloid cross-beta structural motif and the assembly of these beta-sheets into a fibril. EM images of negatively stained Abeta(10)(-)(35) fibrils and measurements of fibril mass per length (MPL) by STEM show a strong dependence of fibril morphology and MPL on pH. Abeta(10)(-)(35) fibrils formed at pH 3.7 are single "protofilaments" with MPL equal to twice the value expected for a single cross-beta layer. Abeta(10)(-)(35) fibrils formed at pH 7.4 are apparently pairs of protofilaments or higher order bundles. EM and STEM data for Abeta(1)(-)(42) fibrils indicate that protofilaments with MPL equal to twice the value expected for a single cross-beta layer are also formed by Abeta(1)(-)(42) and that these protofilaments exist singly and in pairs at pH 7.4. Solid-state NMR measurements of intermolecular distances in Abeta(10)(-)(35) fibrils, using multiple-quantum (13)C NMR, (13)C-(13)C dipolar recoupling, and (15)N-(13)C dipolar recoupling techniques, support the in-register parallel beta-sheet organization previously established by Lynn, Meredith, Botto, and co-workers [Benzinger et al. (1998) Proc. Natl. Acad. Sci. U.S.A. 95, 13407-13412; Benzinger et al. (2000) Biochemistry 39, 3491-3499] and show that this beta-sheet organization is present at pH 3.7 as well as pH 7.4 despite the differences in fibril morphology and MPL. Solid-state NMR measurements of intermolecular distances in Abeta(1)(-)(42) fibrils, which represent the first NMR data on Abeta(1)(-)(42) fibrils, also indicate an in-register parallel beta-sheet organization. These results, along with previously reported data on Abeta(1)(-)(40) fibrils, suggest that the supramolecular structures of Abeta(10)(-)(35), Abeta(1)(-)(40), and Abeta(1)(-)(42) fibrils are quite similar. A schematic structural model of these fibrils, consistent with known experimental EM, STEM, and solid-state NMR data, is presented.     

Molecular alignment within beta-sheets in Abeta(14-23) fibrils: solid-state NMR experiments and theoretical predictions

We report investigations of the molecular structure of amyloid fibrils formed by residues 14-23 of the beta-amyloid peptide associated with Alzheimer's disease (Abeta(14-23)), using solid-state nuclear magnetic resonance (NMR) techniques in conjunction with electron microscopy and atomic force microscopy. The NMR measurements, which include two-dimensional proton-mediated (13)C-(13)C exchange and two-dimensional relayed proton-mediated (13)C-(13)C exchange spectra, show that Abeta(14-23) fibrils contain antiparallel beta-sheets with a registry of backbone hydrogen bonds that aligns residue 17+k of each peptide molecule with residue 22-k of neighboring molecules in the same beta-sheet. We compare these results, as well as previously reported experimental results for fibrils formed by other beta-amyloid fragments, with theoretical predictions of molecular alignment based on databases of residue-specific alignments in antiparallel beta-sheets in known protein structures. While the theoretical predictions are not in exact agreement with the experimental results, they facilitate the design of experiments by suggesting a small number of plausible alignments that are readily distinguished by solid-state NMR.     

Amyloid fibril formation by A beta 16-22, a seven-residue fragment of the Alzheimer's beta-amyloid peptide, and structural characterization by solid state NMR

The seven-residue peptide N-acetyl-Lys-Leu-Val-Phe-Phe-Ala-Glu-NH(2), called A beta(16-22) and representing residues 16-22 of the full-length beta-amyloid peptide associated with Alzheimer's disease, is shown by electron microscopy to form highly ordered fibrils upon incubation of aqueous solutions. X-ray powder diffraction and optical birefringence measurements confirm that these are amyloid fibrils. The peptide conformation and supramolecular organization in A beta(16-22) fibrils are investigated by solid state (13)C NMR measurements. Two-dimensional magic-angle spinning (2D MAS) exchange and constant-time double-quantum-filtered dipolar recoupling (CTDQFD) measurements indicate a beta-strand conformation of the peptide backbone at the central phenylalanine. One-dimensional and two-dimensional spectra of selectively and uniformly labeled samples exhibit (13)C NMR line widths of <2 ppm, demonstrating that the peptide, including amino acid side chains, has a well-ordered conformation in the fibrils. Two-dimensional (13)C-(13)C chemical shift correlation spectroscopy permits a nearly complete assignment of backbone and side chain (13)C NMR signals and indicates that the beta-strand conformation extends across the entire hydrophobic segment from Leu17 through Ala21. (13)C multiple-quantum (MQ) NMR and (13)C/(15)N rotational echo double-resonance (REDOR) measurements indicate an antiparallel organization of beta-sheets in the A beta(16-22) fibrils. These results suggest that the degree of structural order at the molecular level in amyloid fibrils can approach that in peptide or protein crystals, suggest how the supramolecular organization of beta-sheets in amyloid fibrils can be dependent on the peptide sequence, and illustrate the utility of solid state NMR measurements as probes of the molecular structure of amyloid fibrils. A beta(16-22) is among the shortest fibril-forming fragments of full-length beta-amyloid reported to date, and hence serves as a useful model system for physical studies of amyloid fibril formation.     

Pathway complexity of Alzheimer's beta-amyloid Abeta16-22 peptide assembly

Recent studies suggest that both soluble oligomers and insoluble fibrils have toxic effects in cell cultures, raising the interest in determining the first steps of the assembly process. We have determined the aggregation mechanisms of Abeta(16-22) dimer using the activation-relaxation technique and an approximate free energy model. Consistent with the NMR solid-state analysis, the dimer is predicted to prefer an antiparallel beta sheet structure with the expected registry of intermolecular hydrogen bonds. The simulations, however, locate three other antiparallel minima with nonnative beta sheet registries and one parallel beta sheet structure, slightly destabilized with respect to the ground state. This result is significant because it can explain the observed dependency of beta sheet registry on pH conditions. We also find that assembly of Abeta(16-22) into dimers follows multiple routes, but alpha-helical intermediates are not obligatory. This indicates that destabilization of alpha-helical intermediates is unlikely to abolish oligomerization of Abeta peptides.     

Observation of multi-step conformation switching in beta-amyloid peptide aggregation by fluorescence resonance energy transfer

We have observed the conformation switching of Abeta(11-25) in the course of amyloid aggregation by employing time-resolved fluorescence resonance energy transfer (FRET). The amyloid peptides undergo multi-step conformational changes during self-assembling such as random coil (monomers), collapsed coil (multimers), micellar structure, and extended beta-sheet in fibrils. We first identified the critical micelle concentration of Abeta(11-25) that occurs at ca. 3 microM for pH 5.0 and ca. 70 microM for pH 7.4. Our experimental results show clearly that the end-to-end distance of micellar Abeta(11-25) becomes much shorter than that of the collapsed coil or fibril structure.     

Probing the role of backbone hydrogen bonding in beta-amyloid fibrils with inhibitor peptides containing ester bonds at alternate positions

Protein-protein interactions are frequently mediated by stable, intermolecular beta-sheets. A number of cytokines and the HIV Protease, for example, dimerize through beta-sheet motifs. Evidence also suggests that the macromolecular assemblies of peptides and proteins in amyloid fibrils are stabilized by intermolecular beta-sheets. In this paper, we report that interfering with the backbone hydrogen bonding of an amyloidgenic peptide (Abeta16-20) by replacing amide bonds with ester bonds prevents the aggregation of the peptide. The ester bonds were incorporated in an alternating fashion so that the peptide presents two unique hydrogen bonding faces when arrayed in an extended, beta-strand conformation; one face of the peptide has normal hydrogen bonding capabilities, but the other face is missing amide protons and its ability to hydrogen bond is severely limited. Analytical ultracentrifugation experiments demonstrate that this ester peptide, Abeta16-20e, is predominantly monomeric under solution conditions, unlike the fibril-forming Abeta16-20 peptide. Abeta16-20e also inhibits the aggregation of the Abeta1-40 peptide and disassembles preformed Abeta1-40 fibrils. These results suggest that backbone hydrogen bonding is critical for the assembly of amyloid fibrils.     

Solid-state NMR as a probe of amyloid structure

Solid state nuclear magnetic resonance (NMR) has developed into one of the most informative and direct experimental approaches to the characterization of the molecular structures of amyloid fibrils, including those associated with Alzheimer's disease. In this article, essential aspects of solid state NMR methods are described briefly and results obtained to date regarding the supramolecular organization of amyloid fibrils and the conformations of peptides within amyloid fibrils are reviewed.     

Reversible mechanical unzipping of amyloid beta-fibrils

Amyloid fibrils are self-associating filamentous structures, the deposition of which is considered to be one of the most important factors in the pathogenesis of Alzheimer's disease and various other disorders. Here we used single molecule manipulation methods to explore the mechanics and structural dynamics of amyloid fibrils. In mechanically manipulated amyloid fibrils, formed from either amyloid beta (Abeta) peptides 1-40 or 25-35, beta-sheets behave as elastic structures that can be "unzipped" from the fibril with constant forces. The unzipping forces were different for Abeta1-40 and Abeta25-35. Unzipping was fully reversible across a wide range of stretch rates provided that coupling, via the beta-sheet, between bound and dissociated states was maintained. The rapid, cooperative zipping together of beta-sheets could be an important mechanism behind the self-assembly of amyloid fibrils. The repetitive force patterns contribute to a mechanical fingerprint that could be utilized in the characterization of different amyloid fibrils.     

Solid-state NMR as a probe of amyloid fibril structure

Amyloid fibrils are intrinsically noncrystalline, insoluble, high-molecular-weight aggregates of peptides and proteins, with considerable biomedical and biophysical significance. Solid-state NMR techniques are uniquely capable of providing high-resolution, site-specific structural constraints for amyloid fibrils, at the level of specific interatomic distances and torsion angles. So far, a relatively small number of solid-state NMR studies of amyloid fibrils have been reported. These have addressed issues about the supramolecular organization of beta-sheets in the fibrils and the peptide conformation in the fibrils, and have concentrated on the beta-amyloid peptide of Alzheimer's disease. Many additional applications of solid-state NMR to amyloid fibrils from a variety of sources are anticipated in the near future, as these systems are ideally suited for the technique and are of widespread current interest.     

An atomic model for the pleated beta-sheet structure of Abeta amyloid protofilaments.

Synchrotron x-ray studies on amyloid fibrils have suggested that the stacked pleated beta-sheets are twisted so that a repeating unit of 24 beta-strands forms a helical turn around the fibril axis (. J. Mol. Biol. 273:729-739). Based on this morphological study, we have constructed an atomic model for the twisted pleated beta-sheet of human Abeta amyloid protofilament. In the model, 48 monomers of Abeta 12-42 stack (four per layer) to form a helical turn of beta-sheet. Each monomer is in an antiparallel beta-sheet conformation with a turn located at residues 25-28. Residues 17-21 and 31-36 form a hydrophobic core along the fibril axis. The hydrophobic core should play a critical role in initializing Abeta aggregation and in stabilizing the aggregates. The model was tested using molecular dynamics simulations in explicit aqueous solution, with the particle mesh Ewald (PME) method employed to accommodate long-range electrostatic forces. Based on the molecular dynamics simulations, we hypothesize that an isolated protofilament, if it exists, may not be twisted, as it appears to be when in the fibril environment. The twisted nature of the protofilaments in amyloid fibrils is likely the result of stabilizing packing interactions of the protofilaments. The model also provides a binding mode for Congo red on Abeta amyloid fibrils. The model may be useful for the design of Abeta aggregation inhibitors.     

An intersheet packing interaction in A beta fibrils mapped by disulfide cross-linking

Most models for the central cross-beta folding unit in amyloid fibrils of the Alzheimer's plaque protein Abeta align the peptides in register in H-bonded, parallel beta-sheet structure. Some models require the Abeta peptide to undergo a chain reversal when folding into the amyloid core, while other models feature very long extended chains, or zigzag chains, traversing the protofilament. In this paper we introduce the use of disulfide bond cross-linking to probe the fold within the core and the packing interactions between beta-sheets. In one approach, amyloid fibrils grown under reducing conditions from each of three double cysteine mutants (17/34, 17/35, and 17/36) of the Abeta(1-40) sequence were subjected to oxidizing conditions. Of these three mutants, only the Leu17Cys/Leu34Cys peptide could be cross-linked efficiently while resident in fibrils. In another approach, double Cys mutants were cross-linked as monomers before aggregation, and the resulting fibrils were assessed for stability, antibody binding, dye binding, and cross-seeding efficiency. Here too, fibrils from the 17/34 double Cys mutant most closely resemble wild-type Abeta(1-40) fibrils. These data support models of the Abeta fibril in which the Leu17 and Leu34 side chains of the same peptide pack against each other at the beta-sheet interface within the amyloid core. Related cross-linking strategies may reveal longer range spatial relationships. The ability of the cross-linked 17/35 double Cys mutant Abeta to also make amyloid fibrils illustrates a remarkable plasticity of the amyloid structure and suggests a structural mechanism for the generation of conformational variants of amyloid.     

Amphotericin B binds to amyloid fibrils and delays their formation: a therapeutic mechanism?

The membrane-active antifungal agent amphotericin B (AmB) is one of the few agents shown to slow the course of prion diseases in animals. Congo Red and other small molecules have been reported to directly inhibit amyloidogenesis in both prion and Alzheimer peptide model systems via specific binding. We propose that it is possible that AmB may act similarly to physically prevent conversion of the largely alpha-helical prion protein (PrP) to the pathological beta-sheet aggregate protease-resistant isoform (PrP(res)) in prion disease and by analogy prevent fibrillization in amyloid diseases. To assess whether AmB is capable of binding specifically to amyloid fibrils as does Congo Red, we have used the insulin fibril and Abeta 25-35 amyloid model fibril system. We find that AmB does bind strongly to both insulin (K(d) = 1.1 microM) and Abeta 25-35 amyloid (K(d) = 6.4 microM) fibrils but not to native insulin. Binding is characterized by a red-shifted AmB spectrum indicative of a more hydrophobic environment. Thus AmB seems to have a complementary face for amyloid fibrils but not the native protein. In addition, AmB interacts specifically with Congo Red, a known fibril-binding agent. In kinetic fibril formation studies, AmB was able to significantly kinetically delay the formation of Abeta 25-35 fibrils at pH 7.4 but not insulin fibrils at pH 2.     

Microscopic Factors that Control β-Sheet Registry in Amyloid Fibrils Formed by Fragment 11-25 of Amyloid β Peptide: Insights from Computer Simulations

Short fragments of amyloidogenic proteins are widely used as model systems in studies of amyloid formation. Fragment 11-25 of the amyloid β protein involved in Alzheimer's disease (Aβ11-25) was recently shown to form amyloid fibrils composed of anti-parallel β-sheets. Interestingly, fibrils grown under neutral and acidic conditions were seen to possess different registries of their inter-β-strand hydrogen bonds. In an effort to explain the microscopic origin of this pH dependence, we studied Aβ11-25 fibrils using methods of theoretical modeling. Several structural models were built for fibrils at low and neutral pH levels and these were examined in short molecular dynamics simulations in explicit water. The models that displayed the lowest free energy, as estimated using an implicit solvent model, were selected as representative of the true fibrillar structure. It was shown that the registry of these models agrees well with the experimental results. At neutral pH, the main contribution to the free energy difference between the two registries comes from the electrostatic interactions. The charge group of the carboxy terminus makes a large contribution to these interactions and thus appears to have a critical role in determining the registry.     

Photoaffinity cross-linking of Alzheimer's disease amyloid fibrils reveals interstrand contact regions between assembled beta-amyloid peptide subunits

The assembly of the beta-amyloid peptide (Abeta) into amyloid fibrils is essential to the pathogenesis of Alzheimer's disease. Detailed structural information about fibrillogenesis has remained elusive due to the highly insoluble, noncrystalline nature of the assembled peptide. X-ray fiber diffraction, infrared spectroscopy, and solid-state NMR studies performed on fibrils composed of Abeta peptides have led to conflicting models of the intermolecular alignment of beta-strands. We demonstrate here the use of photoaffinity cross-linking to determine high-resolution structural constraints on Abeta monomers within amyloid fibrils. A photoreactive Abeta(1-40) ligand was synthesized by substituting L-p-benzoylphenylalanine (Bpa) for phenylalanine at position 4 (Abeta(1-40) F4Bpa). This peptide was incorporated into synthetic amyloid fibrils and irradiated with near-UV light. SDS-PAGE of dissolved fibrils revealed the light-dependent formation of a covalent Abeta dimer. Enzymatic cleavage followed by mass spectrometric analysis demonstrated the presence of a dimer-specific ion at MH(+) = 1825.9, the predicted mass of a fragment composed of the N-terminal Abeta(1-5) F4Bpa tryptic peptide covalently attached to the C-terminal Abeta(29-40) tryptic peptide. MS/MS experiments and further chemical modifications of the cross-linked dimer led to the localization of the photo-cross-link between the ketone of the Bpa4 side chain and the delta-methyl group of the Met35 side chain. The Bpa4-Met35 intermolecular cross-link is consistent with an antiparallel alignment of Abeta peptides within amyloid fibrils.     

Structure and texture of fibrous crystals formed by Alzheimer's abeta(11-25) peptide fragment

Amyloid fibril deposition is central to the pathology of Alzheimer's disease. X-ray diffraction from amyloid fibrils formed from full-length Abeta(1-40) and from a shorter fragment, Abeta(11-25), have revealed cross-beta diffraction fingerprints. Magnetic alignment of Abeta(11-25) amyloid fibrils gave a distinctive X-ray diffraction texture, allowing interpretation of the diffraction data and a model of the arrangement of the peptides within the amyloid fiber specimen to be constructed. An intriguing feature of the structure of fibrillar Abeta(11-25) is that the beta sheets, of width 5.2 nm, stack by slipping relative to each other by the length of two amino acid units (0.70 nm) to form beta ribbons 4.42 nm in thickness. Abeta(1-40) amyloid fibrils likely consist of once-folded hairpins, consistent with the size of the fibers obtained using electron microscopy and X-ray diffraction.     

Aggregating the amyloid Abeta(11-25) peptide into a four-stranded beta-sheet structure

We present a detailed analysis of the structural properties of one monomer of Abeta(11-25) as well as of the aggregation mechanisms for four chains of Abeta(11-25) using the activation-relaxation technique coupled with a generic energy potential. Starting from a random distribution of these four chains, we find that the system assembles rapidly into a random globular state that evolves into three- and four-stranded antiparallel beta-sheets. The aggregation process is considerably accelerated by the presence of preformed dimers. We also find that the reptation mechanism already identified in shorter peptides plays a significant role here in allowing the structure to reorganize without having to fully dissociate.     

Expression and purification of a recombinant peptide from the Alzheimer's beta-amyloid protein for solid-state NMR

Fibrillar protein aggregates contribute to the pathology of a number of disease states. To facilitate structural studies of these amyloid fibrils by solid-state NMR, efficient methods for the production of milligram quantities of isotopically labeled peptide are necessary. Bacterial expression of recombinant amyloid proteins and peptides allows uniform isotopic labeling, as well as other patterns of isotope incorporation. However, large-scale production of recombinant amyloidogenic peptides has proven particularly difficult, due to their inherent propensity for aggregation and the associated toxicity of fibrillar material. Yields of recombinant protein are further reduced by the small molecular weights of short amyloidogenic fragments. Here, we report high-yield expression and purification of a peptide comprising residues 11-26 of the Alzheimer's beta-amyloid protein (Abeta(11-26)), with homoserine lactone replacing serine at residue 26. Expression in inclusion bodies as a ketosteroid isomerase fusion protein and subsequent purification under denaturing conditions allows production of milligram quantities of uniformly labeled (13)C- and (15)N-labeled peptide, which forms amyloid fibrils suitable for solid-state NMR spectroscopy. Initial structural data obtained by atomic force microscopy, electron microscopy, and solid-state NMR measurements of Abeta(11-26) fibrils are also presented.     

Self-assembly in aqueous solution of a modified amyloid beta peptide fragment

The self-assembly in films dried from aqueous solutions of a modified amyloid beta peptide fragment is studied. We focus on sequence Abeta(16-20), KLVFF, extended by two alanines at the N-terminus to give AAKLVFF. Self-assembly into twisted ribbon fibrils is observed, as confirmed by transmission electron microscopy (TEM). Dynamic light scattering reveals the semi-flexible nature of the AAKLVFF fibrils, while polarized optical microscopy shows that the peptide fibrils crystallize after an aqueous solution of AAKLVFF is matured over 5 days. The secondary structure of the fibrils is studied by FT-IR, circular dichroism and X-ray diffraction (XRD), which provide evidence for beta-sheet structure in the fibril. From high resolution TEM it is concluded that the average width of an AAKLVFF fibril is (63+/-18) nm, indicating that these fibrils comprise beta-sheets with multiple repeats of the unit cell, determined by XRD to have b and c dimensions 1.9 and 4.4 nm with an a axis 0.96 nm, corresponding to twice the peptide backbone spacing in the antiparallel beta-sheet.     

Two-dimensional structure of beta-amyloid(10-35) fibrils

Beta-amyloid (Abeta) peptides are the main protein component of the pathognomonic plaques found in the brains of patients with Alzheimer's disease. These heterogeneous peptides adopt a highly organized fibril structure both in vivo and in vitro. Here we use solid-state NMR on stable, homogeneous fibrils of Abeta(10-35). Specific interpeptide distance constraints are determined with dipolar recoupling NMR on fibrils prepared from a series of singly labeled peptides containing (13)C-carbonyl-enriched amino acids, and skipping no more that three residues in the sequence. From these studies, we demonstrate that the peptide adopts the structure of an extended parallel beta-sheet in-register at pH 7.4. Analysis of DRAWS data indicates interstrand distances of 5.3 +/- 0.3 A (mean +/- standard deviation) throughout the entire length of the peptide, which is compatible only with a parallel beta-strand in-register. Intrastrand NMR constraints, obtained from peptides containing labels at two adjacent amino acids, confirm the secondary structural findings obtained using DRAWS. Using peptides with (13)C incorporated at the carbonyl position of adjacent amino acids, structural transitions from alpha-helix to beta-sheet were observed at residues 19 and 20, but using similar techniques, no evidence for a turn could be found in the putative turn region comprising residues 25-29. Implications of this extended parallel organization for Abeta(10-35) for overall fibril formation, stability, and morphology based upon specific amino acid contacts are discussed.     

Abeta40-Lactam(D23/K28) models a conformation highly favorable for nucleation of amyloid

Recent solid-state NMR data (1) demonstrate that Abeta(1)(-)(40) adopts a conformation in amyloid fibrils with two in-register, parallel beta-sheets, connected by a bend structure encompassing residues D(23)VGSNKG(29), with a close contact between the side chains of Asp23 and Lys28. We hypothesized that forming this bend structure might be rate-limiting in fibril formation, as indicated by the lag period typically observed in the kinetics of Abeta(1)(-)(40) fibrillogenesis. We synthesized Abeta(1)(-)(40)-Lactam(D23/K28), a congener Abeta(1)(-)(40) peptide that contains a lactam bridge between the side chains of Asp23 and Lys28. Abeta(1)(-)(40)-Lactam(D23/K28) forms fibrils similar to those formed by Abeta(1)(-)(40). The kinetics of fibrillogenesis, however, occur without the typical lag period, and at a rate approximately 1000-fold greater than is seen with Abeta(1)(-)(40) fibrillogenesis. The strong tendency toward self-association is also shown by size exclusion chromatography in which Abeta(1)(-)(40)-Lactam(D23/K28) forms oligomers even at concentrations of approximately 1-5 microM. Under the same conditions, Abeta(1)(-)(40) shows no detectable oligomers by size exclusion chromatography. Our data suggest that Abeta(1)(-)(40)-Lactam(D23/K28) could bypass an unfavorable folding step in fibrillogenesis, because the lactam linkage "preforms" a bendlike structure in the peptide. Consistent with this view Abeta(1)(-)(40) growth is efficiently nucleated by Abeta(1)(-)(40)-Lactam(D23/K28) fibril seeds.