Allosteric Ligand Specificity Determining Regions In The Dihydrofolate Reductase Family

Dihydrofolate reductase (DHFR) is a known target of antibacterial and antiparasitic drugs; repurposing known DHFR inhibitors may represent a cost-effective approach to treat emerging infections in the developing world. Mapping inhibitor specificity determinants of DHFR in terms of amino acid sequence and structure supports design and repurposing of compounds that selectively target pathogenic DHFRs over the human homolog. Previously, a novel, homology-based, computational approach that mines ligand inhibition data was developed to predict residues involved in ligand discrimination in the DHFR family. Eighteen amino acid sequence alignment positions were predicted to be DHFR family-wide ligand specificity determinants, three in the active site, four proximal to the active site, and remaining residues in allosteric regions. Here the validity of the predictions was explored by introducing amino acid mutations at the predicted positions to DHFR from Bacillus stearothermophilus. KI values were determined for 14 single mutants against four DHFR inhibitors, methotrexate, trimethoprim, pyrimethamine, and raltitrexed. Mutations at predicted residues resulted in changes for all single-mutants; each amino acid replacement caused a greater than 11-fold change in KI for at least one of the four inhibitors studied. Many mutations were at residues distal to the active site and located in clusters. Interestingly, specificity determining residues clustered together were found to influence ligand specificity in a more similar manner to each other compared to residues in different parts of the structure. The effects of mutations on KI values were found to be ligand specific; one replacement increased KI for one inhibitor while decreasing the KI for another.