E2F6 is unique in that it retains the conserved E2F DNA binding and dimerization domains, but lacks the C-terminal transactivation and pocket protein binding domains characteristic of other members. Therefore, E2F6 can act as a competitive inhibitor of DNA binding by other E2F proteins, and when overexpressed can oppose the function of both the oncogenic E2F2 and 3a proteins and the anti-oncogenic E2F4 and 5 family members. This behavior may explain why E2F6 is neutral in our transformation assays. E2F6 may also repress pro-mitogenic E2F-responsive genes, as the C-terminal portion of E2F6 encompassing the marked-box domain has been shown to inhibit gene transcription through the recruitment of co-repressor complexes. This scenario is supported by our in vitro data, in which forced expression of E2F6 delayed serum-induced cell growth. In our experiments, forced expression of E2F1 could support serum-independent growth, which is consistent with previous studies. Dysregulated E2F1 expression can promote hepatocellular adenoma, spontaneous epithelial tumors, or in combination with activated ras or p53 deficiency, accelerate skin tumorigenesis. However, this factor was significantly less efficient in promoting in vitro growth than E2F2 or E2F3, and in our soft agar culture system E2F1 exhibited very weak colony forming activity over control-transduced 3T3 fibroblasts. This weak oncogenic activity could result from posttranslational destabilization of E2F1 through ubiquitination, and indeed we found that E2F1 protein was expressed less efficiently from the same vector as compared to E2F3. These results contrast two previous studies, which showed that stable over-expression of E2F1 in fibroblasts could induce measurable contact-independent cell growth. One of these studies generated stably-transfected rat embryonic fibroblast lines through drug selection, and achieved very high levels of E2F1 expression.The other study utilized a MoMuLV-based vector, which may contribute less background transforming activity than our MSCV-based vector in these studies, and therefore may allow detection of weaker oncogenes.