All-atom molecular dynamics analysis of multi-peptide systems reproduces peptide solubility in line with experimental observations

Yutaka Kuroda,Satoshi Kosuda,Yuji Sato,Atsushi Suenaga,Makoto Taiji

doi:10.1038/srep19479

Abstract

In order to investigate the contribution of individual amino acids to protein and peptide solubility, we carried out 100 ns molecular dynamics (MD) simulations of 106 Å3 cubic boxes containing ~3 × 104 water molecules and 27 tetra-peptides regularly positioned at 23 Å from each other and composed of a single amino acid type for all natural amino acids but cysteine and glycine. The calculations were performed using Amber with a standard force field on a special purpose MDGRAPE-3 computer, without introducing any “artificial” hydrophobic interactions. Tetra-peptides composed of I, V, L, M, N, Q, F, W, Y, and H formed large amorphous clusters, and those containing A, P, S, and T formed smaller ones. Tetra-peptides made of D, E, K, and R did not cluster at all. These observations correlated well with experimental solubility tendencies as well as hydrophobicity scales with correlation coefficients of 0.5 to > 0.9. Repulsive Coulomb interactions were dominant in ensuring high solubility, whereas both Coulomb and van der Waals (vdW) energies contributed to the aggregations of low solubility amino acids. Overall, this very first all-atom molecular dynamics simulation of a multi-peptide system appears to reproduce the basic properties of peptide solubility, essentially in line with experimental observations.

Highlights

From the 100 ns Molecular dynamics (MD) simulations, we found that the tetra-peptides made of I, V, L, N, Q, F, M, H, W, and Y formed large size amorphous clusters, which consisted of 23 ~ 27 peptides (85 ~ 100% of all peptides in the system), whereas those containing E, D, R, and K did not cluster at all (Fig. 1 and Table 1, Fig. S2)
Similar conclusions were reached using the fraction of monomers (Fig. 1B) or monomers and dimers (Fig. S2), which was expected since the monomer fraction correlates well with the mean cluster size (MCS)
The analysis indicated that the results of a full atom molecular dynamics simulation of a multiple peptide system is be related to experimentally determined “solubility” values

Summary

Introduction

5-residue peptide tags, composed of a single amino acid type, were fused to a model protein, a simplified BPTI variant[12], and the amino acid’s contribution to protein’s solubility was determined by measuring the corresponding BPTI variant’s solubility[13,14]. Overall, these two studies suggest that systematic mutational analysis measuring relative solubility changes, could yield a solubility propensity scale, which might provide an estimate for the relative solubility of a poly-peptide from its amino acid sequence alone. This study is the first of its kind, and it shows that simulations based on solely physico-chemical first principles without introducing artificial parameters can reproduce peptide’s solubility roughly in line with experimental observations

Objectives

Methods

Results

Conclusion