An RBPJ-Drosophila Model Reveals Dependence of RBPJ Protein Stability on the Formation of Transcription-Regulator Complexes

Genetic and Molecular Interactions between HΔCT, a Novel Allele of the Notch Antagonist Hairless, and the Histone Chaperone Asf1 in Drosophila melanogaster

Cellular differentiation relies on the highly conserved Notch signaling pathway. Notch activity induces gene expression changes that are highly sensitive to chromatin landscape. We address Notch gene regulation using Drosophila as a model, focusing on the genetic and molecular interactions between the Notch antagonist Hairless and the histone chaperone Asf1. Earlier work implied that Asf1 promotes the silencing of Notch target genes via Hairless (H). Here, we generate a novel HΔCT allele by genome engineering. Phenotypically, HΔCT behaves as a Hairless gain of function allele in several developmental contexts, indicating that the conserved CT domain of H has an attenuator role under native biological contexts. Using several independent methods to assay protein-protein interactions, we define the sequences of the CT domain that are involved in Hairless-Asf1 binding. Based on previous models, where Asf1 promotes Notch repression via Hairless, a loss of Asf1 binding should reduce Hairless repressive activity. However, tissue-specific Asf1 overexpression phenotypes are increased, not rescued, in the HΔCT background. Counterintuitively, Hairless protein binding mitigates the repressive activity of Asf1 in the context of eye development. These findings highlight the complex connections of Notch repressors and chromatin modulators during Notch target-gene regulation and open the avenue for further investigations.

The Binding of CSL Proteins to Either Co-Activators or Co-Repressors Protects from Proteasomal Degradation Induced by MAPK-Dependent Phosphorylation

The primary role of Notch is to specify cellular identities, whereby the cells respond to amazingly small changes in Notch signalling activity. Hence, dosage of Notch components is crucial to regulation. Central to Notch signal transduction are CSL proteins: together with respective cofactors, they mediate the activation or the silencing of Notch target genes. CSL proteins are extremely similar amongst species regarding sequence and structure. We noticed that the fly homologue suppressor of hairless (Su(H)) is stabilised in transcription complexes. Using specific transgenic fly lines and HeLa RBPJ^KO cells we provide evidence that Su(H) is subjected to proteasomal degradation with a half-life of about two hours if not protected by binding to co-repressor hairless or co-activator Notch. Moreover, Su(H) stability is controlled by MAPK-dependent phosphorylation, matching earlier data for RBPJ in human cells. The homologous murine and human RBPJ proteins, however, are largely resistant to degradation in our system. Mutating presumptive protein contact sites, however, sensitised RBPJ for proteolysis. Overall, our data highlight the similarities in the regulation of CSL protein stability across species and imply that turnover of CSL proteins may be a conserved means of regulating Notch signalling output directly at the level of transcription.

Phospho-Site Mutations in Transcription Factor Suppressor of Hairless Impact Notch Signaling Activity During Hematopoiesis in Drosophila

The highly conserved Notch signaling pathway controls a multitude of developmental processes including hematopoiesis. Here, we provide evidence for a novel mechanism of tissue-specific Notch regulation involving phosphorylation of CSL transcription factors within the DNA-binding domain. Earlier we found that a phospho-mimetic mutation of the Drosophila CSL ortholog Suppressor of Hairless [Su(H)] at Ser269 impedes DNA-binding. By genome-engineering, we now introduced phospho-specific Su(H) mutants at the endogenous Su(H) locus, encoding either a phospho-deficient [Su(H) S269A ] or a phospho-mimetic [Su(H) S269D ] isoform. Su(H) S269D mutants were defective of Notch activity in all analyzed tissues, consistent with impaired DNA-binding. In contrast, the phospho-deficient Su(H) S269A mutant did not generally augment Notch activity, but rather specifically in several aspects of blood cell development. Unexpectedly, this process was independent of the corepressor Hairless acting otherwise as a general Notch antagonist in Drosophila. This finding is in agreement with a novel mode of Notch regulation by posttranslational modification of Su(H) in the context of hematopoiesis. Importantly, our studies of the mammalian CSL ortholog (RBPJ/CBF1) emphasize a potential conservation of this regulatory mechanism: phospho-mimetic RBPJ S221D was dysfunctional in both the fly as well as two human cell culture models, whereas phospho-deficient RBPJ S221A rather gained activity during fly hematopoiesis. Thus, dynamic phosphorylation of CSL-proteins within the DNA-binding domain provides a novel means to fine-tune Notch signal transduction in a context-dependent manner.

Nucleo-cytoplasmic shuttling of murine RBPJ by Hairless protein matches that of Su(H) protein in the model system Drosophila melanogaster

Abstract

CSL transcription factors are central to signal transduction in the highly conserved Notch signaling pathway. CSL acts as a molecular switch: depending on the cofactors recruited, CSL induces either activation or repression of Notch target genes. Unexpectedly, CSL depends on its cofactors for nuclear entry, despite its role as gene regulator. In Drosophila, the CSL homologue Suppressor of Hairless (Su(H)), recruits Hairless (H) for repressor complex assembly, and eventually for nuclear import. We recently found that Su(H) is subjected to a dynamic nucleo-cytoplasmic shuttling, thereby strictly following H subcellular distribution. Hence, regulation of nuclear availability of Su(H) by H may represent a new layer of control of Notch signaling activity. Here we extended this work on the murine CSL homologue RBPJ. Using a ‘murinized’ fly model bearing RBPJ^wt in place of Su(H) at the endogenous locus we demonstrate that RBPJ protein likewise follows H subcellular distribution. For example, overexpression of a H^*NLS3 protein variant defective of nuclear import resulted in a cytosolic localization of RBPJ protein, whereas the overexpression of a H^*NES protein variant defective in the nuclear export signal caused the accumulation of RBPJ protein in the nucleus. Evidently, RBPJ is exported from the nucleus as well. Overall these data demonstrate that in our fly model, RBPJ is subjected to H-mediated nucleo-cytoplasmic shuttling as is Su(H). These data raise the possibility that nuclear availability of mammalian CSL proteins is likewise restricted by cofactors, and may hence present a more general mode of regulating Notch signaling activity.

Graphical abstract

Membrane-Anchored Hairless Protein Restrains Notch Signaling Activity

The Notch signaling pathway governs cell-to-cell communication in higher eukaryotes. In Drosophila, after cleavage of the transmembrane receptor Notch, the intracellular domain of Notch (ICN) binds to the transducer Suppressor of Hairless (Su(H)) and shuttles into the nucleus to activate Notch target genes. Similarly, the Notch antagonist Hairless transfers Su(H) into the nucleus to repress Notch target genes. With the aim to prevent Su(H) nuclear translocation, Hairless was fused to a transmembrane domain to anchor the protein at membranes. Indeed, endogenous Su(H) co-localized with membrane-anchored Hairless, demonstrating their binding in the cytoplasm. Moreover, adult phenotypes uncovered a loss of Notch activity, in support of membrane-anchored Hairless sequestering Su(H) in the cytosol. A combined overexpression of membrane-anchored Hairless with Su(H) lead to tissue proliferation, which is in contrast to the observed apoptosis after ectopic co-overexpression of the wild-type genes, indicating a shift to a gain of Notch activity. A mixed response, general de-repression of Notch signaling output, plus inhibition at places of highest Notch activity, perhaps reflects Su(H)’s role as activator and repressor, supported by results obtained with the Hairless-binding deficient Su(H)^LLL mutant, inducing activation only. Overall, the results strengthen the idea of Su(H) and Hairless complex formation within the cytosolic compartment.