The SCL-LMO2 Interaction Nucleates Hematopoietic Transcription Factor Complexes: Coupling Protein Stability and Complex Assembly.

The differentiation of hematopoietic cells is tightly controlled by transcription factor complexes, composed of hemato-specific and ubiquitous proteins. The bHLH factor SCL and the LIM-only protein LMO2 are central components of transcription factor complexes and are essential for hematopoiesis. However, the mechanism regulating the assembly of SCL-complexes is unknown. Here we show that SCL, in contrast to LMO, GATA and E proteins, cannot be replaced by other members of its family in hematopoietic gene transactivation and in gel shift assays. Furthermore, we show by GST pull-down assays and by co-immunoprecipitation that interaction with LMO2 is a unique property of SCL, as the neurogenic bHLH NSCL1 related to SCL cannot bind LMO2. By generating SCL-NSCL1 chimeras, and by phylogenetic alignment, we identified the SCL interface that confers transcriptional specificity to the complex. Strikingly, this interface is also necessary for the interaction with LMO2. In contrast with the wild type SCL, the mutant without this interface is not able to enhance erythroid differentiation when overexpressed in hematopoietic cells, as assessed by glycophorin A gene activation, benzidine staining and methylcellulose cultures. Interestingly, we also demonstrate in vivo and in vitro that LMO2 protein levels are greatly increased in the presence of SCL, while mRNA levels remain constant. When the SCL interface described above is mutated, LMO2 protein level is no longer increased, suggesting that the accumulation of LMO2 is mediated by a direct interaction with SCL. In primary hematopoietic cells, when SCL protein levels are genetically reduced by LacZ insertion into one allele in the SCL locus, we observe a dramatic decrease in LMO2 protein levels. In addition, in the TF-1 hematopoietic cell line, most of the LMO2 protein (90%) is found associated with SCL and/or Ldb1, suggesting that free LMO2 is rapidly degraded. Thus, SCL levels determine LMO2 levels in hematopoietic cells. Next, we provide direct evidence that LMO2 is a target for proteasomal degradation. First, we show that LMO2 is ubiquitinated in vivo, by GST purification. Second, by using the ts20 cell line expressing a temperature-sensitive ubiquitin-activating E1 enzyme, we show that LMO2 degradation requires a functional ubiquitin conjugation system, since LMO2 is not degraded when E1 is inactive. Third, we show that the half-life of LMO2 is very short, and it can be increased with the proteasome inhibitor MG132. Finally, a similar increase in LMO2 half-life can be observed when SCL is co-expressed with LMO2. These data indicate that SCL stabilizes LMO2, which is otherwise rapidly degraded by the ubiquitin-proteasome pathway in absence of its interacting partners. Taken together, our results strongly suggest that SCL, by binding and stabilizing LMO2, is a critical determinant of the hematopoietic transcriptional specificity. The interaction between SCL and LMO2 is an essential nucleation step for the assembly of SCL-complexes on DNA, where the regulation of LMO2 levels appears to be the rate-limiting step. We propose that protein stability is a new mechanism of regulation in the formation of SCL complexes, required for proper gene activation during eryhtroid differentiation.