Identification of a C-Terminal Negative Regulatory Sequence in NOTCH1 That Influences Signal Strength and Leukemogenic Activity.

NOTCH1 is a large type I transmembrane receptor that regulates normal T cell development via a signaling pathway that relies on regulated proteolysis. Ligand-binding induces sequential proteolytic cleavages in NOTCH1 that release the NOTCH1 intracellular domain (ICN1), allowing it to translocate to the nucleus and activate target genes. It does so by forming a short-lived nuclear complex with two other proteins, the DNA-binding factor CSL and transcriptional co-activators of the Mastermind family. The half-life of this complex appears to be regulated by C-terminal sequences, including a PEST domain, but the specific residues that mediate this activity are unknown. Recent work has shown that human T-ALL is frequently associated with C-terminal truncations, which uniformly remove sequences lying between residues 2520 and 2556. This region includes the highly conserved sequence WSSSSP, which lies immediately C-terminal of residue 2520 and represents a possible site for regulatory phosphorylation events. To test this, we introduced S to A substitutions, singly and in combination. The mutation of WSSSSP (S4) to WAAAAP (A4) stimulated a NOTCH-sensitive reporter gene to an extent equivalent to a deletion spanning residues 2520-2556. Single, double, or triple S to A substitutions had additive effects, implying that all four S residues contribute to negative regulation. Cells expressing the A4 mutant also accumulated higher levels of ICN1, consistent with an effect on ICN1 stability. These effects correlated with an observed hypophosphorylation of the A4 mutant relative to normal ICN1. One kinase implicated in ICN1 phosphorylation is CDK8, which is recruited through the C-terminal tail of Mastermind polypeptides. However, the A4 mutant is hypophosphorylated even when complexed with CSL and a truncated form of Mastermind-1 that cannot recruit CDK8, implying that S4 phosphorylation is not CDK8-dependent. To determine whether the S4 site influences leukemogenesis, we asked if the A4 mutant would convert a weak, non-leukemogenic NOTCH1 allele, delta(EGF-LNR), to a leukemogenic allele. Both delta(EGF-LNR)-S4 and delta(EGF-LNR)-A4 viruses induced the appearance of an abnormal CD4+CD8+GFP+ cell population in the peripheral blood of mice reconstituted with transduced bone marrow progenitors, but only delta(EGF-LNR)-A4 mice progressed to T-ALL. These leukemias arose in all delta(EGF-LNR)-A4 mice by 150 days post-transplant, involved hematopoietic tissues and viscera, were clonal or oligoclonal based on proviral integration site analysis, and were transplantable to secondary recipients. Proviruses appeared to be intact in the T-ALLs, and Western blot analysis showed that the tumor cells expressed NOTCH1 polypeptides of the expected size. In contrast, the CD4+CD8+GFP+ cell population in delta(EGF-LNR)-S4 mice slowly disappeared, and these mice have remained leukemia-free. We conclude that the S4 sequence is likely to be the target of an unknown kinase (or kinases) that negatively regulates NOTCH1, and that the loss of this sequence may contribute to the gains in NOTCH1 function that commonly occur human T-ALL.