Recently, we demonstrated that the oxygen dependence of the formation of DNA-protein crosslinks (DPCs) in irradiated mammalian cells measured by the alkaline elution technique is the mirror image of the oxygen dependence of radiation-induced cell killing. Consequently, these radiation-induced DPCs could be used to detect hypoxic cells or estimate the hypoxic fraction of cells in solid tumors. Although several techniques, including alkaline elution, gas chromatography/mass spectrometry (GC/MS) and nitrocellulose filter binding, have been used to measure radiation-induced DPCs, the published reports suggest that the characteristics of these DPCs may depend on both the type of sample irradiated (cellular compared to model systems, oxygenated compared to hypoxic, etc.) and the technique used to measure these radiation-induced DPCs. In the present study, the radiation-induced DPCs measured by our alkaline elution technique with and without proteinase K in the lysis solution were characterized by studying the dependence of their formation on temperature in hypoxic rat 9L brain tumor cells. Exponentially growing 9L rat brain tumor cells were rendered hypoxic at 4 degrees C or at 37 degrees C and then irradiated with either 7.5 Gy or 15 Gy. The cells were trypsinized at 4 degrees C, either immediately after the irradiation or after one half-time of strand break repair at 37 degrees C. The results demonstrated that the radiation-induced DPCs produced in 9L cells under hypoxic conditions, measured by our alkaline elution technique after low to moderate radiation doses, required metabolism for their formation, unlike the radiation-induced DPCs reported by others using the GC/MS or nitrocellulose filter binding technique.