A Comparative Study of Amyloid Fibril Formation by Residues 15–19 of the Human Calcitonin Hormone: A Single β-Sheet Model with a Small Hydrophobic Core
Nurit Haspel,David Zanuy,Buyong Ma,Haim Wolfson,Ruth Nussinov
DOI: https://doi.org/10.1016/j.jmb.2004.11.002
IF: 6.151
2005-01-01
Journal of Molecular Biology
Abstract:Experimentally, the human calcitonin hormone (hCT) can form highly stable amyloid protofibrils. Further, a peptide consisting of hCT residues 15-19, DFNKF, was shown to create highly ordered fibrils, similar to those formed by the entire hormone sequence. However, there are limited experimental data regarding the detailed 3D arrangement of either of these fibrils. We have modeled the DFNKF protofibril, using molecular dynamics simulations. We tested the stabilities of single sheet and of various multi sheet models. Remarkably, our most ordered and stable model consists of a parallel-stranded, single beta-sheet with a relatively insignificant hydrophobic core. We investigate the chemical and physical interactions responsible for the high structural organization of this single beta-sheet amyloid fibril. We observe that the most important chemical interactions contributing to the stability of the DFNKF organization are electrostatic, specifically between the Lys and the C terminus, between the Asp and N terminus, and a hydrogen bond network between the Asn side-chains of adjacent strands. Additionally, we observe hydrophobic and aromatic pi stacking interactions. We further simulated truncated filaments, FNKF and DFNK. Our tetra-peptide mutant simulations assume models similar to the penta-peptide. Experimentally, the FNKF does not create fibrils while DFNK does, albeit short and less ordered than DFNKF. In the simulations, the FNKF system was less stable than the DFNK and DFNKF. DFNK also lost many of its original interactions becoming less organized, however, many contacts were maintained. Thus, our results emphasize the role played by specific amino acid interactions. To further study specific interactions, we have mutated the penta-peptide, simulating DANKF, DFNKA and EFNKF. Here we describe the model, its relationship to experiment and its implications to amyloid organization.