The sequence dependence of fiber organization. A comparative molecular dynamics study of the islet amyloid polypeptide segments 22-27 and 22-29.
ABSTRACT Amyloid fiber formation and the possible polymorphism of molecular arrangements depend on the polypeptide length and composition. Here, we seek the chemical clues underlying these processes. Our starting point is based on the experimental observation that some short peptide segments are able to develop fibers that are very similar to those of their original parent proteins. We focus our study on the NFGAILSS peptide, derived from the human islet amyloid polypeptide (residues 22-29). This peptide turned out to be a perfect example, illustrating the fact that the amyloid microscopic organization is highly complex, rather than simply involving hydrogen bond formation. Furthermore, obtaining a reliable molecular model has allowed us to analyze the differences between the amyloid structure we have obtained for this peptide and that obtained for the previously studied, two residues shorter, segment (residues 22-27, NFGAIL). This comparative study yields some clues about chemical events that govern the aggregation of proteins into oriented fibers, such as molecular packing between sheets and the degree of interaction specificity. We characterize the important role played by the hydrophobic and aromatic residues in the inter-sheet association and present new approaches toward the understanding of the nature of events that are likely to take place during fibril formation. These include analysis of interaction patterns derived from specific sheet-associated packing.
- SourceAvailable from: David Zanuy[Show abstract] [Hide abstract]
ABSTRACT: In this paper we present a detailed atomic model for a protofilament, the most basic organization level, of the amyloid fibre formed by the peptide DFNKF. This pentapeptide is a segment derived from the human calcitonin, a natural amyloidogenic protein. Our model, which represents the outcome of extensive explicit solvent molecular dynamics (MD) simulations of different strand/sheet organizations, is a single beta-sheet filament largely without a hydrophobic core. Nevertheless, this structure is capable of reproducing the main features of the characteristic amyloid fibril organization and provides clues to the molecular basis of its experimental aggregation behaviour. Our results show that the side chains' chemical diversity induces the formation of a complex network of interactions that finally determine the microscopic arrangement of the strands at the protofilament level. This network of interactions, consisting of both side chain-side chain and backbone-side chain interactions, confers on the final single beta-sheet arrangement an unexpected stability, both by enhancing the association of related chemical groups and, at the same time, by shielding the hydrophobic segments from the polar solvent. The chemical physical characterization of this protofilament provides hints to the possible thermodynamical basis of the supra molecular organization that allows the formation of the filaments by lateral association of the preformed protofibrils. Its regular, highly polarized structure shows how other protofilaments can assemble. In terms of structural biology, our results clearly indicate that an amyloid organization implies a degree of complexity far beyond a simple nonspecific association of peptide strands via amide hydrogen bonds.Physical Biology 07/2004; 1(1-2):89-99. DOI:10.1088/1478-3967/1/2/005 · 3.14 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Current data suggest that globular domains may form amyloids via different mechanisms. Nevertheless, there are indications that the initiation of the process takes place invariably in the less stable segments of a protein domain. We have studied the sequence and structural conservation of beta(2)-microglobulin that deposits into fibrils in dialysis-related amyloidosis. The dataset includes 51 high-resolution non-redundant structures of the antibody constant domain-like proteins (C1) and 132 related sequences. We describe a set of 30 conserved residues. Among them, 23 are conserved structurally, 16 are conserved sequentially and nine are conserved both sequentially and structurally. Strands A (12-18), G (91-95) and D (45-55) are the less conserved and stable segments of the domain, while strands B (22-28), C (36-41), E (62-70) and F (78-83) are the conserved and stable segments. We find that the conserved residues form a cluster with a network of interactions. The observed pattern of conservation is consistent with experimental data including H/D exchange, urea denaturation and limited proteolysis that suggest that strands A and G do not participate in the amyloid fibril. Additionally, the low conservation of strand D is consistent with the observation that this strand may acquire different conformations as seen in crystal structures of bound and isolated beta(2)-microglobulin. We used a docking technique to suggest a model for a fibril via stacking of beta(2)-microglobulin monomers. Our analysis suggests that the favored monomer building block for fibril elongation is the conformation of the isolated beta(2)-microglobulin, without the beta-bulge on strand D and without strands A and G participating in the fibril beta-sheet structure. This monomer retains all the conserved residues and their network of interactions, increasing the likelihood of its existence in solution. The inter-strand interaction between the two (monomer) building blocks forms a new continuous beta-sheet such that addition of monomers results in a fibril model that has the characteristic cross-beta structure.Journal of Molecular Biology 07/2003; 330(1):159-74. DOI:10.1016/S0022-2836(03)00557-6 · 3.96 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Publisher’s description: Volume one of this two volume sequence focuses on the basic characterization of known protein structures as well as structure prediction from protein sequence information. The 11 chapters provide an overview of the field, covering key topics in modeling, force fields, classification, computational methods, and struture prediction. Each chapter is a self contained review designed to cover (1) definition of the problem and an historical perspective, (2) mathematical or computational formulation of the problem, (3) computational methods and algorithms, (4) performance results, (5) existing software packages, and (6) strengths, pitfalls, challenges, and future research directions. Table of contents: Modeling protein structures. Empirical force fields. Knowledge-based energy functions for computational studies of proteins.