We have previously shown that the adenovirus E1A oncogene can reactivate the cell cycle in terminally differentiated cells. Current models imply that much or all of this E1A activity is mediated by the release of the E2F transcription factors from pocket-protein control. In contrast, we show here that overexpression of E2F-1, E2F-2 and E2F-4, or a chimeric E2F-4 tethered to a nuclear localization signal cannot reactivate postmitotic skeletal muscle cells (myotubes). This is not due to lack of transcriptional activity, as demonstrated on both a reporter construct and a number of endogenous target genes. Although cyclin E was strongly overexpressed in E2F-transduced myotubes, it lacked associated kinase activity, possibly explaining the inability of the myotubes to enter S phase and accumulate cyclin A. Although E2F is not sufficient to trigger DNA synthesis in myotubes, its activity is necessary even in the presence of E1A, as dominant-negative DP-1 mutants inhibit E1A-mediated cell cycle reentry. Our data show that, to reactivate myotubes, E1A must exert other functions, in addition to releasing E2F. They also establish mouse myotubes as an experimental system uniquely suited to study the most direct E2F functions in the absence of downstream cell cycle effects.