Combinations of drugs targeting viral proteins have been used to limit or control drug resistance, which is the most important cause of treatment failure in HIV-1-infected individuals. We suggest an alternative approach, namely to target cellular proteins, which are less prone to mutations than viral proteins. Here we show that simultaneous inhibition of a cellular protein (by hydroxyurea) and a viral protein (by ddI) produces a consistent and sustained suppression of HIV-1 for as long as 40 weeks in the absence of virus rebound. We identified the mechanism to explain this lack of rebound: although the combination of the two drugs did not prevent the emergence of mutant viral strains resistant to didanosine (ddI) in these patients, the mutants were still sensitive to standard doses of ddI in the presence of hydroxyurea. These in vivo results were consistent with our in vitro observations: HIV-1 molecular clones resistant to ddI were rendered sensitive to this drug (at concentrations routinely achievable in vivo) after addition of hydroxyurea. This phenomenon can be explained by the observation that hydroxyurea decreases the level of dATP, the cellular competitor of ddI. A low level of dATP favors the incorporation of ddI, even if the viral reverse transcriptase is resistant to this nucleoside analog. This is a novel mechanism of control of resistance and it explains the efficacy of a treatment that is well tolerated, simple, and inexpensive.