Patterns in the sequence context of protein disulfide bonds

Abstract

Disulfide bonds play an important role in the structural stability of the proteins that contain them. Yet, little is known about the specificity with which they are formed. To address this, a representative set of disulfide bonds from nonhomologous eukaryotic polypeptides was created. The amino acid sequences flanking these disulfide bonds were searched for conserved patterns that may reflect recognition sites by the disulfide bond forming enzyme protein disulfide isomerase (PDI). Several methods of classifying disulfide bonds were explored, and each class was analyzed for conserved sequence patterns. To maximize the chances of finding a conserved recognition site, a simulated annealing algorithm was implemented to divide a set of disulfide-bonded cysteines into two sets of cysteines with an average sequence environment that is as far from randomly-distributed as possible. No significant conserved patterns were found in the set of disulfide bonds or within any of the classification schemes introduced. Additionally, several methods for predicting disulfide bond connectivity were explored. The most successful methods predicted connectivity based on the sequential distance between cysteines.