Multidrug transporters play key roles for drug resistance, mediating the transport of organic compounds and substrate recognition. The specificity of substrates and functions of multidrug transporters are affected with amino acid substitutions. Hence, it is important to understand the effect of mutations in multidrug resistance proteins on transport function and substrate specificity. In this work, we have analyzed the relationship between amino acid properties and activity of multidrug resistance proteins upon mutations and substrates. We found that the properties for drug activity and kinetic factors depend on amino acid substitutions and specific to substrates. The inclusion of information from neighboring residues from the mutants enhanced our understanding to the activity of multidrug resistant proteins. Further, we have combined amino acid properties using multiple regression technique, which showed a correlation of up to 0.99 between amino acid properties and activity. In addition, we have utilized Naïve Bayes classifier for distinguishing between decrease and increase in IC50 upon mutations using wild type, mutant and neighboring residues, which showed a 10-fold cross-validation accuracy of 86%. Further, we have developed multiple regression models for predicting IC50 upon mutations with a maximum correlation of 0.92. The present method could be used for identifying the mutants in multidrug resistant proteins with enhanced specificity.
Keywords: Amino acid properties; Membrane protein; Multidrug efflux transporter; Multiple regression technique.
Copyright © 2015 Elsevier B.V. All rights reserved.