Intrinsic Local Balancing of Hydrophobic and Hydrophilic Residues in Folded Protein Sequences

Abstract

Protein sequences may evolve to avoid highly hydrophobic local regions of sequence, in part because such sequences promote nonnative aggregation. Hydrophobic local sequences are avoided in proteins even in buried regions, where native structure requirements tend to favor them. In this dissertation, I describe three explorations of this hydrophobic suppression. In Chapter 2, I examine the occurrence of hydrophobic and polar residues in completely buried β-strand elements, and find evidence for hydrophobic suppression that decreases as a β-strand becomes more exposed. In Chapter 3, I present a generalized study of the tendency of local sequences to deviate from the hydropathy (hydrophobicity) expected based on their solvent exposure. First, I examined the hydropathy of local and nonlocal sequence groups over a large range of solvent exposures, within folded protein domains in the ASTRAL Compendium database; second, I calculated the tendency of residues within 10 positions of a nonpolar or polar reference residue to deviate from the hydropathy expected based on their structural environment. Both analyses suggested that protein sequences exhibit 'local hydropathic balance' across a range of 6-7 residues, meaning that polar and nonpolar residues are more dispersed in the sequence than expected based on solvent exposure patterns. This balance occurs in all major fold classes, domain sizes and protein functions. An unexpected finding was that it partly arises from a tendency of buried or exposed residues to be flanked by polar or nonpolar residues, respectively. This relationship may result from evolutionary selection for folding efficiency, which might be enhanced by reduced local competition for buried or exposed sites during folding. Finally, in Chapter 4, I present several exploratory analyses, including a decision-tree approach, to visualize the influence of a large number of sequence-structure properties on residue hydrophobicity. Overall, the work in this dissertation confirms that hydrophobic suppression and local hydropathic balance in general are intrinsic properties of folded proteins. I speculate that local hydropathic balance results from selection for reduced aggregation propensity, increased folding efficiency and increased native state specificity. The concept of local hydropathic balance might be used to improve the properties of designed and engineered proteins.