Theoretical Studies Of The Mechanisms Of Short Polypeptides Aggregation

The proteins with correct3-dimensional folded structure is the basis of living activities. Whereas, the aggregation of misfolded porteins will cause many fatal "misfolding disease", including Prion disorder, Huntington's, Alzhimer's and Pakinson's disease. Recently, a variety of experimental techniques have been applied to explore the detailed architectures of the amyloid aggregates, and providing a number of clearer static picture of the fibrils. However, there is increasing experimental evidence suggests that the oligomeric forms of amyloidogenic peptide instead of mature fibril are responsible for neurotoxity under physiologic conditions. On the other hand, the oligomers also act as nucleating seeds for the growth of amyloid plaque. The ability to form fibrils from misfolded structures may be an intrinsic property of proteins, as various peptides with no clear homology form amyloid structures with similar morphology. Therefore, a systematic study of the aggregation process of misfloded proteins if very important. However, various characters in the process of amyloid foramtion make directly study using conventional experimental techniques became very difficult.On the other hand, all-atom molecular dynamics (MD) simulation can provide complementary information for a more complete characterization of structure and the dynamical aggregation process, as well as fill the gaps between the experimental results. In this thesis, we use MD simmulations to investigate the aggregation process of an amyloidogenic hexapeptide NFGAIL (residues22-27of human islet amyloid polypeptide) in water. In the early stage, the hydrophobic interactions drive the dimerization of peptides and subsequently formation of the disordered tetramer. In the following, through structural fluctuation more inter-mainchain hydrogen bonds formed, and the secondary sturcture of peptide largerly converted from random-coil to (3-extended structure, as a result this disordered species will became partially ordered oligmer. This partially ordered oligmer is still unstable, however the dynamic properties in this process could reflect the characters of formation of the "nucleus" in the early-stage aggregation process of NFGAIL. In the growth phase after formation of "nucleus", two relatively stable, flat, antiparallel β-sheets, with each sheet composed of three antiparallel NFGAIL peptides serving as nucleus. We find that monomer either along the fibrillar axis longitudinally patching to the nucleus, or two monomers form a antiparallel dimer in advance and then perpendicular to the fibrillar axis laterally stacking to the nucleus. Nearly all peptides patching to the nucleus in antiparallel, but one. These results suggest that there should be multiple pathways for fibril growth. During the simulations, the structure of "nucleus" is stable, which implied that the smallest "nucleus" for the aggregation of NFGAIL should be six peptides.As a model system to study the growth of fibril by laterally stacking, we use MD simulations to investigate the water-mediated self-assembly of two amyloidogenic β-sheets of hIAPP22-27peptides (NFGAIL). The initial configurations of β-sheet pairs are packed with two different modes, forming a tube-like nanoscale channel and a slab-like two-dimensional confinement, respectively. For both packing modes, we observe strong water drying transitions occurring in the intersheet region with high occurrence possibilities, suggesting that the "dewetting transition"-induced collapse may play an important role in promoting the amyloid fibrils formation. However, contrary to general dewetting theory prediction, the slab-like confinement (2-D) shows stronger dewetting phenomenon than the tube-like channel (1-D). This unexpected observation is attributed to the different surface roughness caused by different packing modes.The peptide dimer, the smallest oligomer, is an important intermediate even when it is not the critical nucleus. Therefore, the detailed studies of dimerization processes of amyloid peptides are important for an understanding of general principles governing protein folding and aggregation. In this thesis, an enhanced sampling algorithm, the replica exchange method, is employed to explore the dimerization process of a seven-residue fragment of prion Sup-35, GNNQQNY. An orientation-centric dimerization mechanism is proposed:the non-native contacts between the central residues of two monomers initiate the dimerization process; then the two strands reorient themselves into parallel alignment before an increasing number of native contacts build up and a native-like dimer forms. The results also show that the central residues (Asn-3, Gln-4, Gln-5, Asn-6) are crucial to the stability of the dimers. The residue Asn-3, in particular, plays a pivotal role in the early stage of peptide recognition.Experimental evidence suggests that in5M urea solution, formation of fibrilar aggregates by β-lactoglobulin is most rapid. On the other hand, MD evidence suggests that the stability of oligomers of Aβ16-22(KLVFFAE) peptides in8M aqueous urea solution is determined by two opposing effects, namely, by the increased propensity of monomers to form β-strands and the rapid disruption of the oligomers. Herein a question arised:if there is an optimal urea concentration does exists, at this concentration, urea promotes peptide became more extended, at the same time the influence of urea is not so stronger to fiercely disrupt the associations between peptides, and on the whole to facilitate fomation of amyloid fibril. In this thesis, we observed that urea accumulates around the peptide and expels the water molecules from the FSS of the peptide, and promotes formtaion of β-extended structure. The preferential binding of urea over water to the surface of peptide can be attributed to the more favorable van der Waals interaction of urea with the peptide than water. However, we also find that even at a fairly low urea concentration (at1M), urea stiall can significantly slow down the peptide association and the following formation of aggregate. This unexpected phenomena is mainly induced by the accumulation of urea on the surface of peptide, especially deposition around of the mainchain and even forming number of intimate hydrogen bonds with mainchain. It is noteworthy that the influence of urea on the association of sidechains is negligible. In a previous study, we explore the mechanism of urea-induced denaturation of proteins by using molecular dynamics (MD) simulations of hen lysozyme in8M urea, and found the "direct interaction mechanism" whereby urea has a stronger dispersion interaction with protein than water. Here we perform large scale MD simulations of five different peptide/protein systems in aqueous urea to investigate if the above molecular mechanism is unique to the specific proteins. In all cases, accumulations of urea around peptide/proteins are observed, suggesting that urea denature proteins by directly attacks, rather than disrupts water structure as a "water breaker". Consistent with our previous study, our current analyses with interaction energies reveal that the urea's preferential binding to proteins is due to the stronger dispersion interaction of urea with protein than water. Furthermore, the simulations of peptide system at different urea concentration (4.5M and8M) and with different force fields (CHARMM and OPLSAA) suggest that the above mechanism is fairly robust, independent of urea concentrations and the force fields used. These findings suggest that the "dispersion-interaction-driven" mechanism should be a general mechanism for urea-induced denaturation of proteins.Understanding the interactions between carbon nanotube (CNT) and protein/peptide are essential to the CNT-based nanotechnology and biotechnology. Recent experiments have revealed that the π-π stacking interaction between aromatic residues and CNT play a key role in the binding process. On the other hand, there is an increasing interest in modeling protein/peptide-CNT interactions, including π-π stacking interactions, by employing molecular dynamics (MD) simulations in recent years. However, the validation of such approach, as well as a systematic, reliable description of π-π interactions between aromatic residues and CNT are still lacking. Here we study the binding features of three aromatic amino acid (phenylalanine, tyrosine, and tryptophan) analogues and benzene molecule to the outer surface of single-walled CNT in gas phase via π-π interactions, and compare the molecular mechanical (MM) calculations with three popular fixed-charge force fields (OPLSAA, AMBER, and CHARMM) against quantum mechanical (QM) calculations (using DFTB-D method). We consider two energetically favorable configurations of the aromatic molecules relative to CNT, a "flat" one and an "edge" one, with the aromatic rings parallel and perpendicular to the CNT surface, respectively (convenient for π-π stacking and electrostatic interactions of hydrogen atoms on aromatic ring with π electron cloud of CNT, respectively). For the interaction energies of binding, the MM data for both configurations as well as their energy difference are in general good agreement with the corresponding QM data. These results indicate that the classical approaches can appropriately predict the strength of π-π (stacking) interaction between protein/peptide and CNT. Our study provides a comprehensive understanding of the π-π (stacking) interactions between aromatic residues of protein/peptide and CNT, and suggests that MD simulations with fixed-charge force fields are suited for describing such interactions.