Investigation of stability in self-assembled peptides by computational modelling

Amyloid aggregates have been linked to diseases, including neurodegenerative ones such as Amyloid aggregates have been linked to diseases, including neurodegenerative ones such as Alzheimer’s, Parkinson’s, Huntington’s and Prion diseases. Their roles in these as of yet untreatable diseases influenced this field of study on the disease aspect of amyloids. In recent years, with the discovery of non-disease amyloid assemblies and the advances in rational material design had scientists become interested in using self-assembled peptides as possible bio inspired materials. Studied materials are usually known aggregation prone peptides or fragments of these peptides. On the other hand, designing a new material involves creating lots of different sequences and testing the dataset to find the sequence with highest performance. This thesis investigates self-assembling peptides at the molecular level by scanning peptides of various lengths to shed light on the molecular properties contributing to the stability using molecular dynamics simulations. First, we modelled hexa- and heptapeptides from sequences of known aggregation prone proteins and artificially engineered Zinc binding catalytic peptides. Among the 15 molecules (including polymorphs), our simulation results showed that peptides with a strong hydrogen bond network and the peptides which did not have hydrophobic residues exposed to the solvent were more stable. We then reduced the sequence further and modelled tripeptide and amino acid crystals. 6 of 7 tripeptides we modelled were point mutations or functional group modifications and one of them was a cyclic dipeptide. For amino acids, we modelled F, Y and L-DOPA to observe the effect of hydroxylation. We found out that for tripeptides, increasing the hydrophobicity of the peptide and doing that while adding an extra hydroxyl group to increase the number of hydrogen bonds increased stability the most. In case of amino acids, each added hydroxyl group stabilised the assembly. Finally, we have modelled microtubule binding domain of tau protein along with its two familial mutations, P301L and K280Δ as a first step toward studying their effects on fibrillar level. Our results indicate that secondary structure content in P301L is slightly higher than in the other models, which may stabilise its assembly Parkinson’s, Huntington’s and Prion diseases. Their roles in these as of yet untreatable diseases influenced this field of study on the disease aspect of amyloids. In recent years, with the discovery of non-disease amyloid assemblies and the advances in rational material design had scientists become interested in using self-assembled peptides as possible bio inspired materials. Studied materials are usually known aggregation prone peptides or fragments of these peptides. On the other hand, designing a new material involves creating lots of different sequences and testing the dataset to find the sequence with highest performance. This thesis investigates self-assembling peptides at the molecular level by scanning peptides of various lengths to shed light on the molecular properties contributing to the stability using molecular dynamics simulations. First, we modelled hexa- and heptapeptides from sequences of known aggregation prone proteins and artificially engineered Zinc binding catalytic peptides. Among the 15 molecules (including polymorphs), our simulation results showed that peptides with a strong hydrogen bond network and the peptides which did not have hydrophobic residues exposed to the solvent were more stable. We then reduced the sequence further and modelled tripeptide and amino acid crystals. 6 of 7 tripeptides we modelled were point mutations or functional group modifications and one of them was a cyclic dipeptide. For amino acids, we modelled F, Y and L-DOPA to observe the effect of hydroxylation. We found out that for tripeptides, increasing the hydrophobicity of the peptide and doing that while adding an extra hydroxyl group to increase the number of hydrogen bonds increased stability the most. In case of amino acids, each added hydroxyl group stabilised the assembly. Finally, we have modelled microtubule binding domain of tau protein along with its two familial mutations, P301L and K280Δ as a first step toward studying their effects on fibrillar level. Our results indicate that secondary structure content in P301L is slightly higher than in the other models, which may stabilise its assembly