Tenofovir‐Diphosphate Formation by Creatine Kinase Brain‐Type is Diminished by Naturally Occurring Mutations In Vitro

Abstract

Tenofovir (TFV) is a nucleotide reverse transcriptase inhibitor that is administered as a prodrug for use in both HIV treatment and prevention. TFV requires two sequential phosphorylation steps to form the pharmacologically active metabolite, tenofovir‐diphosphate (TFV‐DP). Previous studies have shown that creatine kinase muscle‐type (CKM) can catalyze the final phosphorylation step in colon tissue. Thus, to discover other TFV activating enzymes, we used CKM as a starting point and identified creatine kinase brain‐type (CKB) as a candidate TFV activating enzyme, as it is 80% homologous to CKM. To test the hypothesis that CKB can phosphorylate TFV in a manner similar to CKM, we performed in vitro activity assays by incubating recombinantly expressed CKB or CKM with reaction substrates, phosphocreatine and tenofovir‐monophosphate. TFV metabolites were detected using ultra‐high performance liquid chromatography tandem mass spectrometry (uHPLC/MS). Analysis of TFV‐DP levels revealed no significant difference between CKB and CKM catalyzed reactions. These data indicate that CKB may contribute to TFV activation in tissues where it is expressed, thus genetic variation in CKB may play a role in inter‐individual variability observed in TFV efficacy. To investigate the potential impact of genetic variation, fifteen naturally occurring missense mutations were chosen and mutant CKB enzymes were recombinantly expressed in E. coli. The purified enzymes were used in our in vitro activity assay and TFV metabolites were subsequently detected by uHPLC/MS. Eight mutations (C74S, R96P, S128R, R132H, R172P, R236Q, R292Q, and H296R) demonstrated a statistically significant reduction in the formation of TFV‐DP compared to that of wild‐type CKB. To gain insight into how these mutations disrupt enzymatic function, we exploited the reverse canonical reaction (ATP dephosphorylation) in an enzyme coupled system, from which Michaelis‐Menten curves were constructed and resulting kinetic parameters were used to calculate catalytic efficiencies. Five mutations (R96P, R132H, R236Q, R292Q, and H296R) resulted in catalytic efficiencies less than 20% of wild‐type, which can be attributed to increases in Km. Lastly, thermal stability was examined using differential scanning fluorimetry and showed four mutations (C74S, R96P, R172P, and H296P) had significantly lower melting temperatures than wild‐type. Additionally, seven mutations (C74S, R96P, R132H, R172G, R172P, R236Q, and H296P) displayed melting curves with double peaks or shouldering, suggesting local domain melting or dimer instability induced by the mutation. Together these data support our hypothesis that CKB contributes to TFV activation and indicates that naturally occurring mutations can diminish TFV‐DP formation, catalytic efficiency, and thermal stability in vitro.