Abstract The bHLH transcription factor SCL/Tal-1 is required for specification of hematopoietic stem cells and maturation of definitive erythroid cells and megakaryocytes. Additionally, ectopic SCL expression is the most common pathogenetic event in T-cell acute lymphoblastic leukaemia (T-ALL) (up to 60%). However, the mechanism of SCL action is not understood. Therefore, to begin to address this, we have isolated SCL-containing protein complexes in definitive erythroid cells and megakaryocytes. We have used a single step protein purification strategy where SCL is tagged to a sequence that can be biotinylated in vivo (biotag-SCL), and SCL-containing complexes purified by streptavidin beads and identified by mass spectrometry. First, to ensure that biotinylated SCL was functional, we were able to fully rescue both primitive and definitive haematopoietic development from SCL−/− ES cells in vitro with biotag SCL. Then, we expressed biotag-SCL in mouse erythroleukemic (MEL) and megakaryoblastic (L8057) cell lines and pulled-down SCL-containing protein complexes. Western Blot analysis of precipated proteins showed that known partners of SCL (including the transcription factors E2A, GATA1, LMO2 and Ldb1) were present in the pulled-down fractions, thereby validating the technique. Mass spectrometry analysis then led to the identification of several new candidate partners in MEL and L8057 cells. Two such proteins include ETO-2 (human homologue: MTG16), a member of the ETO family of co-repressor proteins involved in the pathogenesis of acute myeloid leukaemia, and SSDP-2, previously identified as a partner of Lbd1. Several approaches were then used to characterise the SCL/ETO-2 interaction. First, the interaction between the endogenous proteins was validated by co-immunoprecipitation and reverse co-immunoprecipitation in both MEL and L8057 nuclear extracts. Second, we showed that endogenous ETO-2 and SCL co-purify in the same high molecular weight fractions in gel-filtration profiles. Then, immunodepletion experiments demonstrated that the composition of the SCL/ETO-2 complexes is different in MEL and L8057 cells. In MEL cells, ETO-2 is part of at least two different complexes comprising other critical haematopoietic regulators including the oncoprotein Gfi-1b, whereas in L8057 cells these interactions are not detected. These results were then confirmed in primary erythroid cells and megakaryocytes. Finally, consistent with the repressor role of ETO-2, transactivation experiments in heterologous cells revealed that ETO-2 expression represses the activator function of a pentameric complex consisting of GATA1/SCL/E2A/LMO2/Ldb-1 on GATA-1 regulatory elements. Taken together, these data establish that SCL interacts with ETO-2 in erythroid cells and megakaryocytes, that the composition of SCL/ETO-2-containing complexes differ in these two lineages and that the SCL/ETO-2 complex may have a repressor activity on genes critical for haematopoietic differentiation. In this context, we also detect SCL binding to ETO-2 in an SCL-expressing T-ALL cell line. Though the exact role of the SCL-ETO-2 complex remains to be elucidated, it is plausible that ectopic T-cell SCL contributes to leukaemia by repressing genes indispensable for normal T cell differentiation. If confirmed, this would shed new light on the mechanism and possible therapeutics for SCL expressing T-ALL.