3'-Azido-3'-deoxythymidine-(5')-tetraphospho-(5')-adenosine, the product of ATP-mediated excision of chain-terminating AZTMP, is a potent chain-terminating substrate for HIV-1 reverse transcriptase

The resistance of HIV-1 to 3'-azido-3'-deoxythymidine (AZT) involves phosphorolytic excision of chain-terminating AZT-5'-monophosphate (AZTMP). Both pyrophosphate (PPi) and ATP act as excision substrates in vitro, but the intracellular substrate used during replication of AZT-resistant HIV is still unknown. PPi-mediated excision produces AZT-5'-triphosphate (AZTTP), which could be immediately re-used as a substrate for viral DNA chain termination. In contrast, ATP-mediated excision produces the novel compound AZT-(5')-tetraphospho-(5')-adenosine (AZTp4A). Since little is known of the interaction of AZTp4A with HIV-1 RT, we carried out kinetic and molecular modeling studies to probe this. AZTp4A was found to be a potent inhibitor of HIV-1 RT-catalyzed DNA synthesis and of both ATP- and PPi-mediated AZTMP excision. AZTp4A is in fact an excellent chain-terminating substrate for AZT-resistant RT-catalyzed DNA synthesis, better than AZTTP (k(pol)/Kd = 6.2 and 11.9 for AZTTP and AZTp4A, respectively). The affinity of AZT-resistant HIV-1 RT for AZTp4A is at least 30,000-fold greater than that for the excision substrate ATP and approximately 10-fold greater than that for AZTTP. Dissociation of newly formed AZTp4A from RT may therefore provide a significant rate-limiting step for continued HIV-1 DNA synthesis. Our studies show that the products of PPi- and ATP-mediated excision of chain-terminating AZTMP (AZTTP and AZTp4A, respectively) are both potent chain-terminating substrates for HIV-1 RT, suggesting that there is no obvious benefit to HIV using ATP instead of PPi as the excision substrate.     

The antiretrovirus drug 3'-azido-3'-deoxythymidine increases the retrovirus mutation rate.

It was previously observed that the nucleoside analog 5-azacytidine increased the spleen necrosis virus (SNV) mutation rate 13-fold in one cycle of retrovirus replication (V. K. Pathak and H. M. Temin, J. Virol. 66:3093-3100, 1992). Based on this observation, we hypothesized that nucleoside analogs used as antiviral drugs may also increase retrovirus mutation rates. We sought to determine if 3'-azido-3'-deoxythymidine (AZT), the primary treatment for human immunodeficiency virus type 1 (HIV-1) infection, increases the retrovirus mutation rate. Two assays were used to determine the effects of AZT on retrovirus mutation rates. The strategy of the first assay involved measuring the in vivo rate of inactivation of the lacZ gene in one replication cycle of SNV- and murine leukemia virus-based retroviral vectors. We observed 7- and 10-fold increases in the SNV mutant frequency following treatment of target cells with 0.1 and 0.5 microM AZT, respectively. The murine leukemia virus mutant frequency increased two- and threefold following treatment of target cells with 0.5 and 1.0 microM AZT, respectively. The second assay used an SNV-based shuttle vector containing the lacZ alpha gene. Proviruses were recovered as plasmids in Escherichia coli, and the rate of inactivation of lacZ alpha was measured. The results indicated that treatment of target cells increased the overall mutation rate two- to threefold. DNA sequence analysis of mutant proviruses indicated that AZT increased both the deletion and substitution rates. These results suggest that AZT treatment of HIV-1 infection may increase the degree of viral variation and alter virus evolution or pathogenesis.