Structural insights into the polymorphism of amyloid-like fibrils formed by region 20-29 of amylin revealed by solid-state NMR and X-ray fiber diffraction

Many unrelated proteins and peptides can assemble into amyloid or amyloid-like nanostructures, all of which share the cross-beta motif of repeat arrays of beta-strands hydrogen-bonded along the fibril axis. Yet, paradoxically, structurally polymorphic fibrils may derive from the same initial polypeptide sequence. Here, solid-state nuclear magnetic resonance (SSNMR) analysis of amyloid-like fibrils of the peptide hIAPP 20-29, corresponding to the region S (20)NNFGAILSS (29) of the human islet amyloid polypeptide amylin, reveals that the peptide assembles into two amyloid-like forms, (1) and (2), which have distinct structures at the molecular level. Rotational resonance SSNMR measurements of (13)C dipolar couplings between backbone F23 and I26 of hIAPP 20-29 fibrils are consistent with form (1) having parallel beta-strands and form (2) having antiparallel strands within the beta-sheet layers of the protofilament units. Seeding hIAPP 20-29 with structurally homogeneous fibrils from a 30-residue amylin fragment (hIAPP 8-37) produces morphologically homogeneous fibrils with similar NMR properties to form (1). A model for the architecture of the seeded fibrils is presented, based on the analysis of X-ray fiber diffraction data, combined with an extensive range of SSNMR constraints including chemical shifts, torsional angles, and interatomic distances. The model features a cross-beta spine comprising two beta-sheets with an interface defined by residues F23, A25, and L27, which form a hydrophobic zipper. We suggest that the energies of formation for fibril form containing antiparallel and parallel beta-strands are similar when both configurations can be stabilized by a core of hydrophobic contacts, which has implications for the relationship between amino acid sequence and amyloid polymorphism in general.     

Structural regulation of a peptide-conjugated graft copolymer: a simple model for amyloid formation

The self-assembly of peptides and proteins into beta-sheet-rich high-order structures has attracted much attention as a result of the characteristic nanostructure of these assemblies and because of their association with neurodegenerative diseases. Here we report the structural and conformational properties of a peptide-conjugated graft copolymer, poly(gamma-methyl-L-glutamate) grafted polyallylamine (1) in a water-2,2,2-trifluoroethanol solution as a simple model for amyloid formation. Atomic force microscopy revealed that the globular peptide 1 self-assembles into nonbranching fibrils that are about 4 nm in height under certain conditions. These fibrils are rich in beta-sheets and, similar to authentic amyloid fibrils, bind the amyloidophilic dye Congo red. The secondary and quaternary structures of the peptide 1 can be controlled by manipulating the pH, solution composition, and salt concentration; this indicates that the three-dimensional packing arrangement of peptide chains is the key factor for such fibril formation. Furthermore, the addition of carboxylic acid-terminated poly(ethylene glycol), which interacts with both of amino groups of 1 and hydrophobic PMLG chains, was found to obviously inhibit the alpha-to-beta structural transition for non-assembled peptide 1 and to partially cause a beta-to-alpha structural transition against the 1-assembly in the beta-sheet form. These findings demonstrate that the amyloid fibril formation is not restricted to specific protein sequences but rather is a generic property of peptides. The ability to control the assembled structure of the peptide should provide useful information not only for understanding the amyloid fibril formation, but also for developing novel peptide-based material with well-defined nanostructures.     

A designed protein interface that blocks fibril formation

Protein fibril formation is implicated in many diseases, and therefore much effort has been focused toward the development of inhibitors of this process. In a previous project, a monomeric protein was computationally engineered to bind itself and form a heterodimer complex following interfacial redesign. One of the protein monomers, termed monomer-B, was unintentionally destabilized and shown to form macroscopic fibrils. Interestingly, in the presence of the designed binding partner, fibril formation was blocked. Here we describe the complete characterization of the amyloid properties of monomer-B and the inhibition of fiber formation by the designed binding partner, monomer-A. Both proteins are mutants of the betal domain of streptococcal protein-G. The free monomer-B protein forms amyloid-type fibrils, as determined by transmission electron microscopy and the change in fluorescence of Thioflavin T, an amyloid-specific dye. Fibril formation kinetics are influenced by pH, protein concentration, and seeding with preformed fibrils. Under all conditions tested, monomer-A was able to inhibit the formation of monomer-B fibrils. This inhibition is specific to the engineered interaction, as incubation of monomer-B with wild-type protein-G (a structural homologue) did not result in inhibition under the same conditions. Thus, this de novo-designed heterodimeric complex is an excellent model system for the study of protein-based fibril formation and inhibition. This system provides additional insight into the development of pharmaceuticals for amyloid disorders, as well as the potential use of amyloid fibrils for self-assembling nanostructures.     

Atomic force microscopic imaging of seeded fibril formation and fibril branching by the Alzheimer's disease amyloid-β protein

Background: Amyloid plaques composed of the fibrillar form of the amyloid-β protein (Aβ) are the defining neuropathological feature of Alzheimer's disease (AD). A detailed understanding of the time course of amyloid formation could define steps in disease progression and provide targets for therapeutic intervention. Amyloid fibrils, indistinguishable from those derived from an AD brain, can be produced in vitro using a seeded polymerization mechanism. In its simplest form, this mechanism involves a cooperative transition from monomeric Aβ to the amyloid fibril without the buildup of intermediates. Recently, however, a transient species, the Aβ amyloid protofibril, has been identified. Here, we report studies of Aβ amyloid protofibril and its seeded transition into amyloid fibrils using atomic force microscopy.Results: Seeding of the protofibril-to-fibril transition was observed. Preformed fibrils, but not protofibrils, effectively seeded this transition. The assembly state of Aβ influenced the rate of seeded growth, indicating that protofibrils are fibril assembly precursors. The handedness of the helical surface morphology of fibrils depended on the chirality of Aβ. Finally, branched and partially wound fibrils were observed.Conclusions: The temporal evolution of morphologies suggests that the protofibril-to-fibril transition is nucleation-dependent and that protofibril winding is involved in that transition. Fibril unwinding and branching may be essential for the post-nucleation growth process. The protofibrillar assembly intermediate is a potential target for AD therapeutics aimed at inhibiting amyloid formation and AD diagnostics aimed at detecting presymptomatic disease.     

beta-Sheet mediated self-assembly of dipeptides of omega-amino acids and remarkable fibrillation in the solid state

Single crystal X-ray diffraction studies show that the extended structure of dipeptide Boc-beta-Ala-m-ABA-OMe (m-ABA: meta-aminobenzoic acid) self-assembles in the solid state by intermolecular hydrogen bonding to create an infinite parallel beta-sheet structure. In dipeptide Boc-gamma-Abu-m-ABA-OMe (gamma-Abu: gamma-aminobutyric acid), two such parallel beta-sheets are further cross-linked by intermolecular hydrogen bonding through m-aminobenzoic acid moieties. SEM (scanning electron microscopy) studies reveal that both the peptides and form amyloid-like fibrils in the solid state. The fibrils are also found to be stained readily by Congo red, a characteristic feature of the amyloid fiber whose accumulation causes several fatal diseases such as Alzheimer's, prion-protein etc.     

Solid-state NMR study of amyloid nanocrystals and fibrils formed by the peptide GNNQQNY from yeast prion protein Sup35p

Sup35p is a prion protein found in yeast that contains a prion-forming domain characterized by a repetitive sequence rich in Gln, Asn, Tyr, and Gly amino acid residues. The peptide GNNQQNY7-13 is one of the shortest segments of this domain found to form amyloid fibrils, in a fashion similar to the protein itself. Upon dissolution in water, GNNQQNY displays a concentration-dependent polymorphism, forming monoclinic and orthorhombic crystals at low concentrations and amyloid fibrils at higher concentrations. We prepared nanocrystals of both space groups as well as fibril samples that reproducibly contain three (coexisting) structural forms and examined the specimens with magic angle spinning (MAS) solid-state nuclear magnetic resonance. 13C and 15N MAS spectra of both nanocrystals and fibrils reveal narrow resonances indicative of a high level of microscopic sample homogeneity that permitted resonance assignments of all five species. We observed variations in chemical shift among the three dominant forms of the fibrils which were indicated by the presence of three distinct, self-consistent sets of correlated NMR signals. Similarly, the monoclinic and orthorhombic crystals exhibit chemical shifts that differ from one another and from the fibrils. Collectively, the chemical shift data suggest that the peptide assumes five conformations in the crystals and fibrils that differ from one another in subtle but distinct ways. This includes variations in the mobility of the aromatic Tyr ring. The data also suggest that various structures assumed by the peptide may be correlated to the "steric zipper" observed in the monoclinic crystals.     

Cytochrome display on amyloid fibrils

Protein amyloid fibrils can be functionalized by coating the core protofilament with high concentrations of proteins and enzymes. This can be done elegantly by appending a functional domain to an amyloidogenic protein monomer, then assembling the monomers into a fibril. To display an array of biologically functional porphyrins on the surface of protein fibrils, we have fused the sequence of the small, soluble cytochrome b562 to an SH3 dimer sequence that can form classical amyloid fibrils rapidly under well-defined conditions. The resulting fusion protein also forms amyloid fibrils and, in addition, binds metalloporphyrins, at half of the porphyrin binding sites as shown by UV-vis and NMR spectroscopies. Once metalloporphyrins are bound to the fibrils, the resulting holo-cytochrome domains are spectroscopically identical to the wild type cytochrome. The concentration of metalloporphyrins on a saturated fibril is estimated to be of the order of approximately 20 mM, suggesting that they could be interesting systems for applications in nanotechnology.     

Monitoring and inhibition of insulin fibrillation by a small organic fluorogen with aggregation-induced emission characteristics

Amyloid fibrillation of proteins is associated with a great variety of pathologic conditions. Development of new molecules that can monitor amyloidosis kinetics and inhibit fibril formation is of great diagnostic and therapeutic value. In this work, we have developed a biocompatible molecule that functions as an ex situ monitor and an in situ inhibitor for protein fibrillation, using insulin as a model protein. 1,2-Bis[4-(3-sulfonatopropoxyl)phenyl]-1,2-diphenylethene salt (BSPOTPE) is nonemissive when it is dissolved with native insulin in an incubation buffer but starts to fluoresce when it is mixed with preformed insulin fibril, enabling ex situ monitoring of amyloidogenesis kinetics and high-contrast fluorescence imaging of protein fibrils. Premixing BSPOTPE with insulin, on the other hand, inhibits the nucleation process and impedes the protofibril formation. Increasing the dose of BSPOTPE boosts its inhibitory potency. Theoretical modeling using molecular dynamics simulations and docking reveals that BSPOTPE is prone to binding to partially unfolded insulin through hydrophobic interaction of the phenyl rings of BSPOTPE with the exposed hydrophobic residues of insulin. Such binding is assumed to have stabilized the partially unfolded insulin and obstructed the formation of the critical oligomeric species in the protein fibrillogenesis process.     

Absolute structural constraints on amyloid fibrils from solid-state NMR spectroscopy of partially oriented samples

We demonstrate that absolute, molecular-level structural information can be obtained from solid-state NMR measurements on partially oriented amyloid fibrils. Specifically, we show that the direction of the fibril axis relative to a carbonyl 13C chemical shift anisotropy (CSA) tensor can be determined from magic-angle spinning (MAS) sideband patterns in 13C NMR spectra of fibrils deposited on planar substrates. Deposition of fibrils on a planar substrate creates a highly anisotropic distribution of fibril orientations (hence, CSA tensor orientations) with most fibrils lying in the substrate plane. The anisotropic orientational distribution gives rise to distorted spinning sideband patterns in MAS spectra from which the fibril axis direction can be inferred. The experimentally determined fibril axis direction relative to the carbonyl CSA tensor of Val12 in fibrils formed by the 40-residue beta-amyloid peptide associated with Alzheimer's disease (Abeta1-40) agrees well with the predictions of a recent structural model (Petkova et al. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 16742-16747) in which Val12 is contained in a parallel beta-sheet in the cross-beta motif characteristic of amyloid fibrils.     

Self-assembly and morphological characterization of two-component functional amyloid proteins

Functional amyloid has been increasingly applied as self-assembling nanostructures to construct multifunctional biomaterials. However, little has been known how different side domains, varied fusion positions and subunits affect self-assembly and morphologies of amyloid fibrils. Here, we constructed three groups of two-component amyloid proteins based on CsgA, the major protein components of Escherichia coli biofilms, to bridge these gaps. We showed that all fusion proteins have amyloid features, as indicated by Congo red assay. Atomic force microscopy (AFM) indeed reveals that these fusion proteins are able to self-assemble into fibrils, with an average diameter of 0.5-2 nm and length of hundreds of nanometers to several micrometers. The diameter of fibrils increases with the increase of the molecular weight of fusion domains, while the dynamic assembly of recombinant proteins was delayed as a result of the introduction of fusion domains. Moreover, fusion of the same functional domains but at intermediate position seems to cause the most interference on fibril assembly compared with those fused at C or Nterminus, as mainly short and irregular fibrils were detected. This phenomenon appears more pronounced for randomly coiled mussel foot proteins (Mfps) than for rigid chitin-binding domain (CBD). Finally, increase of the molecular weight of tandem repeats in protein monomer seemed to increase the fibril diameter of the resultant fibrils, but either reduction of the tandem repeats of CsgA to one single belta-sheet loop or increase in the number of tandem repeats of CsgAs from one to four produced shorter and intermittent fibrils compared with CsgA control protein. These studies therefore provide insights into self-assembly of two-component amyloid proteins and lay the foundation for rational design of multifunctional molecular biomaterials.     

The organization of aromatic side groups in an amyloid fibril probed by solid-state 2H and 19F NMR spectroscopy

Some 25 diseases are associated with proteins and peptides that assemble into amyloid fibrils composed of beta-strands connected by hydrogen bonds oriented parallel to the fiber long axis. There is mounting evidence that amyloid formation involves specific interactions between amino acid side groups, which bring together beta-sheets to form layers with buried and exposed faces. This work demonstrates how a combination of solid-state 2H and 19F NMR experiments can provide constraints on fibril architecture by probing the environment and spatial organisation of aromatic side groups. It is shown that phenylalanine rings within fibrils formed by a decapeptide fragment of the islet amyloid polypeptide, amylin, are highly motionally restrained and are situated within 6.5 A of one another. Taken together with existing structural constraints for this peptide, these results are consistent with a fibril architecture that comprises layers of two or more beta-sheets, with the aromatic residues facing into the inter-sheet space and possibly engaged in pi-pi interactions. The methods presented will be of general utility in exploring the architecture of fibrils of larger, full-length peptides and proteins, including amylin itself.     

Molecular Dynamics Simulation Studies on the Aggregation of Amyloid-β Peptides and Their Disaggregation by Ultrasonic Wave and Infrared Laser Irradiation

Alzheimer’s disease is understood to be caused by amyloid fibrils and oligomers formed by aggregated amyloid-β (Aβ) peptides. This review article presents molecular dynamics (MD) simulation studies of Aβ peptides and Aβ fragments on their aggregation, aggregation inhibition, amyloid fibril conformations in equilibrium, and disruption of the amyloid fibril by ultrasonic wave and infrared laser irradiation. In the aggregation of Aβ, a β-hairpin structure promotes the formation of intermolecular β-sheet structures. Aβ peptides tend to exist at hydrophilic/hydrophobic interfaces and form more β-hairpin structures than in bulk water. These facts are the reasons why the aggregation is accelerated at the interface. We also explain how polyphenols, which are attracting attention as aggregation inhibitors of Aβ peptides, interact with Aβ. An MD simulation study of the Aβ amyloid fibrils in equilibrium is also presented: the Aβ amyloid fibril has a different structure at one end from that at the other end. The amyloid fibrils can be destroyed by ultrasonic wave and infrared laser irradiation. The molecular mechanisms of these amyloid fibril disruptions are also explained, particularly focusing on the function of water molecules. Finally, we discuss the prospects for developing treatments for Alzheimer’s disease using MD simulations.     

Genetic engineering combined with deep UV resonance Raman spectroscopy for structural characterization of amyloid-like fibrils

Elucidating the structure of the cross-beta core in large amyloid fibrils is a challenging problem in modern structural biology. For the first time, a set of de novo polypeptides was genetically engineered to form amyloid-like fibrils with similar morphology and yet different strand length. Differential ultraviolet Raman spectroscopy allowed for separation of the spectroscopic signatures of the highly ordered beta-sheet strands and turns of the fibril core. The relationship between Raman frequencies and Ramachandran dihedral angles of the polypeptide backbone indicates the nature of the beta-sheet and turn structural elements.     

Self-assembly of amylin(20-29) amide-bond derivatives into helical ribbons and peptide nanotubes rather than fibrils

Uncontrolled aggregation of proteins or polypeptides can be detrimental for normal cellular processes in healthy organisms. Proteins or polypeptides that form these amyloid deposits differ in their primary sequence but share a common structural motif: the (anti)parallel beta sheet. A well-accepted approach for interfering with beta-sheet formation is the design of soluble beta-sheet peptides to disrupt the hydrogen-bonding network; this ultimately leads to the disassembly of the aggregates or fibrils. Here, we describe the synthesis, spectroscopic analysis, and aggregation behavior, imaged by electron microscopy, of several backbone-modified amylin(20-29) derivatives. It was found that these amylin derivatives were not able to form fibrils and to some extent were able to inhibit fibril growth of native amylin(20-29). However, two of the amylin peptides were able to form large supramolecular assemblies, like helical ribbons and peptide nanotubes, in which beta-sheet formation was clearly absent. This was quite unexpected since these peptides have been designed as soluble beta-sheet breakers for disrupting the characteristic hydrogen-bonding network of (anti)parallel beta sheets. The increased hydrophobicity and the presence of essential amino acid side chains in the newly designed amylin(20-29) derivatives were found to be the driving force for self-assembly into helical ribbons and peptide nanotubes. This example of controlled and desired peptide aggregation may be a strong impetus for research on bionanomaterials in which special shapes and assemblies are the focus of interest.     

Aβ-binding molecules: Possible application as imaging probes and as anti-aggregation agents

As amyloid β (Aβ) is at the centre of pathogenesis of Alzheimer's disease (AD), Aβ aggregate-specific probes for in vivo studies of Aβ are potentially important for the early diagnosis and the assessment of new treatment strategies in the AD brain by noninvasive imaging. Several series of compounds derived from Congo red (CR) and Thioflavin T (ThT) have been evaluated as potential probes for the Aβ imaging. They include a diversity of core structures contributing to their affinities to Aβ. Small-molecule inhibi- tors were known to inhibit the formation of Aβ oligomers and fibrils. This inhibition has to be performed in such a way that these inhibitors bind to Aβ in the binding channel where Aβ-binding probes should sit. Therefore, several of them were used as novel core structures to develop Aβ probes, with their de- rivatives exhibiting good Aβ affinities. This approach will facilitate the design of a variety of candidates for Aβ probe molecules and anti-aggregation-therapeutic drugs. Moreover, the finding of Aβ probes with diverse core structures recognized by binding sites on Aβs will likely provide a promising per- spective for the design of 99mTc-labeled probe-derived molecules.     

Design of peptides that form amyloid-like fibrils capturing amyloid beta1-42 peptides

Amyloid beta-peptide (Abeta) plays a critical role in Alzheimer's disease (AD). The monomeric state of Abeta can self-assemble into oligomers, protofibrils, and amyloid fibrils. Since the fibrils and soluble oligomers are believed to be responsible for AD, the construction of molecules capable of capturing these species could prove valuable as a means of detecting these potentially toxic species and of providing information pertinent for designing drugs effective against AD. To this aim, we have designed short peptides with various hydrophobicities based on the sequence of Abeta14-23, which is a critical region for amyloid fibril formation. The binding of the designed peptides to Abeta and the amplification of the formation of peptide amyloid-like fibrils coassembled with Abeta are elucidated. A fluorescence assay utilizing thioflavin T, known to bind specifically to amyloid fibrils, revealed that two designed peptides (LF and VF, with the leucine and valine residues, respectively, in the hydrophobic core region) could form amyloid-like fibrils effectively by using mature Abeta1-42 fibrils as nuclei. Peptide LF also coassembled with soluble Abeta oligomers into peptide fibrils. Various analyses, including immunostaining with gold nanoparticles, enzyme-linked immunosorbent assays, and size-exclusion chromatography, confirmed that the LF and VF peptides formed amyloid-like fibrils by capturing and incorporating Abeta1-42 aggregates into their peptide fibrils. In this system, small amounts of mature Abeta1-42 fibrils or soluble oligomers could be transformed into peptide fibrils and detected by amplifying the amyloid-like fibrils with the designed peptides.     

Observation of metastable Aβ amyloid protofibrils by atomic force microscopy

Background: Brain amyloid plaque, a diagnostic feature of Alzheimer's disease (AD), contains an insoluble fibrillar core that is composed primarily of variants of the β-amyloid protein (Aβ). As Aβ amyloid fibrils may initiate neurodegeneration, the inhibition of fibril formation is a possible therapeutic strategy. Very little is known about the early steps of the process, however.Results: Atomic force microscopy was used to follow amyloid fibril formation in vitro by the Aβ variants Aβ1-40 and Aβ1-42. Both variants first form small ordered aggregates that grow slowly and then rapidly disappear, while prototypical amyloid fibrils of two discrete morphologies appear. Aβ1-42 aggregates much more rapidly than Aβ1-40, which is consistent with its connection to early-onset AD. We propose that the metastable intermediate species be called Aβ amyloid protofibrils.Conclusions: Aβ protofibrils are likely to be intermediates in the in vitro assembly of Aβ amyloid fibrils, but their in vivo role has yet to be determined. Numerous reports of a nonfibrillar form of Aβ aggregate in the brains of individuals who are predisposed to AD suggest the existence of a precursor form, possibly the protofibril. Thus, stabilization of Aβ protofibrils may be a useful therapeutic strategy.     

Common Fibril Structures Imply Systemically Conserved Protein Misfolding Pathways In Vivo

Systemic amyloidosis is caused by the misfolding of a circulating amyloid precursor protein and the deposition of amyloid fibrils in multiple organs. Chemical and biophysical analysis of amyloid fibrils from human AL and murine AA amyloidosis reveal the same fibril morphologies in different tissues or organs of one patient or diseased animal. The observed structural similarities concerned the fibril morphology, the fibril protein primary and secondary structures, the presence of post-translational modifications and, in case of the AL fibrils, the partially folded characteristics of the polypeptide chain within the fibril. Our data imply for both analyzed forms of amyloidosis that the pathways of protein misfolding are systemically conserved; that is, they follow the same rules irrespective of where inside one body fibrils are formed or accumulated.     

Common Fibril Structures Imply Systemically Conserved Protein Misfolding Pathways In Vivo

Systemic amyloidosis is caused by the misfolding of a circulating amyloid precursor protein and the deposition of amyloid fibrils in multiple organs. Chemical and biophysical analysis of amyloid fibrils from human AL and murine AA amyloidosis reveal the same fibril morphologies in different tissues or organs of one patient or diseased animal. The observed structural similarities concerned the fibril morphology, the fibril protein primary and secondary structures, the presence of post‐translational modifications and, in case of the AL fibrils, the partially folded characteristics of the polypeptide chain within the fibril. Our data imply for both analyzed forms of amyloidosis that the pathways of protein misfolding are systemically conserved; that is, they follow the same rules irrespective of where inside one body fibrils are formed or accumulated.     

Structural model of the amyloid fibril formed by beta(2)-microglobulin #21-31 fragment based on vibrational spectroscopy

A structural model of amyloid fibril formed by the #21-31 fragment of beta2-microglobulin is proposed from microscope IR measurements on specifically 13C-labeled peptide fibrils and Raman spectra of the dispersed fibril solution. The 13C-shifted amide frequency indicated the secondary structure of the labeled residues. The IR spectra have demonstrated that the region between F22 and V27 forms the core part with the extended beta-sheet structure. Raman spectra indicated the formation of a dimer with a disulfide bridge between C25 residues.