It is now conceivable that leukemogenesis requires two types of mutations, class I and class II mutations. We previously established a mouse bone marrow-derived HF6, an IL-3-dependent cell line, that was immortalized by a class II mutation MLL/SEPT6 and can be fully transformed by class I mutations such as FLT3 mutants. To understand the molecular mechanism of leukemogenesis, particularly progression of myelodysplastic syndrome (MDS) to acute leukemia, we made cDNA libraries from the samples of patients and screened them by expression-cloning to detect class I mutations that render HF6 cells factor-independent. We identified RasGRP4, an activator of Ras, as a candidate for class I mutation from three of six patients (MDS/MPD = 1, MDS-RA = 1, MDS/AML = 2, CMMoL/AML = 1 and AML-M2 = 1). To investigate the potential roles of RasGRP4 in leukemogenesis, we tested its in vivo effect in a mouse bone marrow transplantation (BMT) model. C57BL/6J mice transplanted with RasGRP4-transduced primary bone marrow cells died of T cell leukemia, myeloid leukemia, or myeloid leukemia with T cell leukemia. To further examine if the combination of class I and class II mutations accelerated leukemic transformation, we performed a mouse BMT model in which both AML1 mutant (S291fsX300) and RasGRP4 were transduced into bone marrow cells. The double transduction led to early onset of T cell leukemia but not of AML in the transplanted mice when compared to transduction of RasGRP4 alone. Thus, we have identified RasGRP4 as a gene potentially involved in leukemogenesis and suggest that RasGRP4 cooperates with AML1 mutations in T cell leukemogenesis as a class I mutation.