a computational method for the systematic screening of reaction barriers in enzymes: searching for bacillus circulans xylanase mutants with greater activity towards a synthetic substrate

We present a semi-empirical (PM6-based) computational method for systematically estimating the effect of all possible single mutants, within a certain radius of the active site, on the barrier height of an enzymatic reaction. The intent of this method is not a quantitative prediction of the barrier heights, but rather to identify promising mutants for further computational or experimental study. The method is applied to identify promising single and double mutants of Bacillus circulans xylanase (BCX) with increased hydrolytic activity for the artificial substrate ortho-nitrophenyl β-xylobioside (ONPX2). The estimated reaction barrier for wild-type (WT) BCX is 18.5 kcal/mol, which is in good agreement with the experimental activation free energy value of 17.0 kcal/mol extracted from the observed kcat using transition state theory (Joshi et al., 2001). The PM6 reaction profiles for eight single point mutations are recomputed using FMO-MP2/PCM/6-31G(d) single points. PM6 predicts an increase in barrier height for all eight mutants while FMO predicts an increase for six of the eight mutants. Both methods predict that the largest change in barrier occurs for N35F, where PM6 and FMO predict a 9.0 and 15.8 kcal/mol increase, respectively. We thus conclude that PM6 is sufficiently accurate to identify promising mutants for further study. We prepared a set of all theoretically possible (342) single mutants in which every amino acid of the active site (except for the catalytically active residues E78 and E172) was mutated to every other amino acid. Based on results from the single mutants we construct a set of 111 double mutants consisting of all possible pairs of single mutants with the lowest barrier for a particular position and compute their reaction profile. None of the mutants have, to our knowledge, been prepared experimentally and therefore present experimentally testable predictions.