Emergence of antiviral drug resistance is a major challenge to human immunodeficiency virus (HIV) therapy. The archetypal example of this problem is loss of antiviral activity of the nucleoside analogue 3'-azido-3'-deoxythymidine (AZT), caused by mutations in reverse transcriptase (RT), the viral polymerase. AZT resistance results from an imbalance between rates of AZT-induced proviral DNA chain termination and RT-induced excision of the chain-terminating nucleotide. Conversion of the AZT prodrug from its monophosphorylated to diphosphorylated form by human thymidylate kinase (TMPK) is inefficient, resulting in accumulation of the monophosphorylated AZT metabolite (AZT-MP) and a low concentration of the active triphosphorylated metabolite (AZT-TP). We reasoned that introduction of an engineered, highly active TMPK into T cells would overcome this functional bottleneck in AZT activation and thereby shift the balance of AZT activity sufficiently to block replication of formerly AZT-resistant HIV. Molecular engineering was used to link highly active, engineered TMPKs to the protein transduction domain of Tat for direct cell delivery. Combined treatment of HIV-infected T cells with AZT and these cell-permeable, engineered TMPKs restored AZT-induced repression of viral production. These results provide an experimental basis for the development of new strategies to therapeutically increase the intracellular concentrations of active nucleoside analogue metabolites as a means to overcome emerging drug resistance.