C/EBPα dimerizes via its leucine zipper domain and then binds DNA via the adjacent basic region. One or both copies of the C/EBPα gene harbors point mutations in 5–8% of human AMLs. One common group of mutations leads to expression of C/EBPαp30, lacking an N-terminal transactivation domain. In addition, insertions and deletions commonly occur in the first α-helix of the leucine zipper, preventing DNA-binding - these alterations are typically in-frame, suggesting that the C/EBPαLZ mutants retain a role in transformation. We have now found that C/EBPα, C/EBPαp30, and four C/EBPαLZ mutants inhibit the apoptosis which occurs when the 32Dcl3 or HF1 myeloid or Ba/F3 or FL5.12 lymphoid cell lines are withdrawn from IL-3; for example, 30% of Ba/F3-αER cells survived at 24 hrs and 0–2% at 48 hrs in the absence of estradiol, and 180%, 80%, and 40% survived at 1, 2, or 3 days in its presence. However, TNF death receptor-mediated apoptosis was not affected. C/EBPα-ER did not alter expression of STAT3, P-STAT3, STAT5, P-STAT5, MAPK, P-MAPK, PI3K, AKT, c-Jun, JunB, bcl-xL, or bim in Ba/F3 cells. However, bcl-2 protein levels, which diminish several-fold after IL-3 withdrawal, were elevated 5-fold at 16 hours by C/EBPα-ER, but not in vector-transfected cells exposed to estradiol. Northern blotting demonstrated induction of transcription from the bcl-2 P2 promoter. Induction of bcl-2 by C/EBPα was previously seen in t(14;18) lymphoma lines, in which the bcl-2 gene is fused to the immunoglobulin heavy chain promoter, and C/EBPα was shown to transactivate P2-luciferase, in part via interaction with a Cdx:A-Myb complex. We have now also identified a consensus C/EBP-binding site, using EMSA, in the P2 promoter. C/EBPα-ER also induced endogenous bcl-2 in NIH 3T3 cells, in transduced normal murine myeloid progenitors, and in the splenocytes of H2K-Eμ-C/EBPα transgenic mice. C/EBPα enhanced survival of splenocytes exposed to low-dose ionizing radiation. The ability of additional C/EBPα mutants to induce bcl-2 and inhibit apoptosis was assessed. The BRM2 mutant, which does not bind E2F, Δ3–8, lacking transactivation domains, GZ, having the GCN4 leucine zipper, and L12V, which cannot dimerize, each retained these activities. In contrast, BR3, with 4 mutant residues in the basic domain, did not induce bcl-2 and only weakly prevented cell death. The BR3 mutations were then studied individually. Only R300G prevented DNA-binding; this mutant did not induce bcl-2 and only weakly inhibited cell death. K298T and K302N each potently induced bcl-2 and prevented apoptosis. Finally, R297G only weakly induced bcl-2 and had an intermediate effect on survival. Perhaps R300G alters the conformation of R297, explaining R300G's inability to induce bcl-2. In summary, C/EBPα stimulates cell survival, in part via direct activation of the bcl-2 P2 promoter. This induction is optimal when C/EBPα retains its ability to bind DNA, but can also occur via protein:protein interaction, perhaps mediated in part by R297 with the Cdx:A-Myb complex or with other proteins which regulate the bcl-2 gene, such as NF-κB. Targeting this interaction with the outer surface of the C/EBPα basic region may be an effective means to identify novel therapeutics for AMLs expressing C/EBPα oncoproteins.