Transcripts for the retinoic acid receptors (RARs) alpha1, alpha2, gamma1, and gamma2 were found in the granulocytic lineage (Gr-1+ cells) through semiquantitative polymerase chain reaction (PCR) analysis. The screening of single cell cDNA libraries derived from hematopoietic progenitors also showed the presence of RARalpha and, to a lesser extent, RARgamma transcripts in committed granulocyte (colony-forming unit-granulocyte [CFU-G]) or granulocyte-macrophage (CFU-GM) colony-forming cells. The contribution of RARalpha1 and gamma to hematopoietic cell differentiation was therefore investigated in mice bearing targeted disruption of either one or both of these loci. Because RARgamma and RARalpha1gamma compound null mutants die shortly after birth, bone marrow cells were collected from fetuses at 18.5 days postcoitum (dpc) and evaluated for growth and differentiation in culture in the presence of Steel factor (SF), interleukin-3 (IL-3), and erythropoietin (Epo). The frequency of colony-forming cells from bone marrow populations derived from RARalpha1/gamma double null mice was not significantly different from that of RARgamma or RARalpha1 single nulls or from wild-type controls. In addition, the distribution of erythroid, granulocyte, and macrophage colonies was comparable between hematopoietic cells from all groups, suggesting that lineage commitment was not affected by the lack of RARalpha1 and/or RARgamma. Colony cells were then harvested individually and evaluated by morphologic criteria. While terminal granulocyte differentiation was evident in wild-type cells and colonies from either single null mutant, colonies derived from RARalpha1-/-gamma-/- bone marrow populations were blocked at the myelocyte and, to a lesser extent, at the metamyelocyte stages, whereas erythroid and macrophage differentiation was not affected. Together, these results indicate that both RARalpha1 and gamma are required for terminal maturation in the granulocytic lineage in vitro, but appear to be dispensable for the early stages of hematopoietic cell development. Our results raise the possibility that in acute promyelocytic leukemia (APL), the RARalpha fusion proteins cause differentiation arrest at a stage when further maturation requires not only RARalpha, but also RARgamma. Finally, bone marrow cells appear to differentiate normally in vivo, suggesting an effective compensation mechanism in the RARalpha1/gamma double null mice.