Computer-aided subsite mapping of α-amylases

Abstract

Subsite mapping is a crucial procedure in the characterization of α-amylases (EC 3.2.1.1), which are extensively used in starch-based industries and in diagnosis of pancreatic and salivary glands disorders. A computer-aided method has been developed for subsite mapping of α-amylases, which substitutes the difficult, expensive, and time-consuming experimental determination of action patterns to crystal structures based energy calculations. Interaction energies between enzymes and carbohydrate substrates were calculated after short energy minimization by a molecular mechanics program. A training set of wild type and mutant amylases with known experimental action patterns of 13 enzymes of wide range of origin was used to set up the procedure. Calculations for training set resulted in good correlation in case of subsite binding energies (r(2)=0.827-0.929) and bond cleavage frequencies (r(2)=0.727-0.835). A set of eight novel barley amylase 1 mutants was used to test our model. Subsite binding energies were predicted with r(2)=0.502 correlation coefficient, while bond cleavage frequency prediction resulted in r(2)=0.538. Our computer-aided procedure may supplement the experimental subsite mapping methods to predict and understand characteristic features of α-amylases.