RNA polymerase II-specific transcription requires, in addition to auxiliary protein factors, the hydrolysis of the beta-gamma phosphate bond of ATP. The nonhydrolyzable analog of ATP, imidoadenosine triphosphate does not suffice for specific in vitro transcription (Bunick, D., Zandomeni, R., Ackerman, S., and Weinmann, R. (1982) Cell 29, 877-886), although it can be incorporated into RNA. The experiments presented here suggest two energy-dependent steps in RNA polymerase II transcription. One of these steps is required at, or close to, the point of initiation, as determined by 5' end primer extension analysis. In vitro transcription occurs efficiently in vitro when imidoadenosine triphosphate is supplemented with dATP to fulfill the energy requirement. In the presence both of imidoadenosine triphosphate and imidoguanosine triphosphate, the concentration of dATP required for transcription initiation is dramatically increased. This suggests that ATP and GTP are co-substrates in transcription initiation, supporting the role of protein kinase II in this process (Zandomeni, R., Zandomeni, M. C., Shugar, D., and Weinmann, R. (1986) J. Biol. Chem. 261, 3414-3419). The concentration of dATP required for maximal initiation is inadequate for the production of full-length transcripts, suggesting a second energy-dependent step in the RNA elongation process. Since the elongation step is unaffected by the presence of imidoguanosine triphosphate, GTP beta-gamma phosphate bond hydrolysis appears to be required only for initiation.