Cellular resistance to the cytotoxic purine analogues 8-azaguanine (AG) and 6-thioguanine (TG) is usually mediated by a mutation leading to the loss or reduction in hypoxanthine phosphoribosyltransferase (HPRT) activity. However, stable AG-resistant variants have often been shown to contain wild-type levels of HPRT, while cellular resistance to TG is always accompanied by a profound deficiency in HPRT activity. Such AG-resistant, HPRT-positive cells are still sensitive to TG. To investigate the basis of this differential sensitivity, we examined the inhibition of the HPRT activity by AG and TG in whole cells, in cell-free extracts, and with purified mouse HPRT. In addition, the relative incorporation and utilization of AG and TG by L929 cells were determined under a variety of culture conditions. Results show that, compared to TG, AG is generally a very poor substrate for HPRT. Incorporation of radioactive AG by HPRT-positive cells was extremely sensitive to the free purine concentrations in the medium, so that under the usual culture conditions employing undialyzed serum, cellular uptake and utilization was minimal even when relatively high levels of AG were present. In contrast, the incorporation of radioactive TG was comparable to that of a natural substrate, hypoxanthine. The results indicate that the differential cellular sensitivity to AG and TG is due to the difference between these two guanine analogues as substrates of HPRT. Additional data indicate also that cellular resistance to TG is mediated exclusively by HPRT deficiency, but resistance to very high levels of AG may result through at least two other mechanisms not involving HPRT deficiency. These observations may help resolve some of the conflicting data in the literature, and demonstrate that TG is a better selective agent for the HPRT-deficient phenotype.