Abstract

NOTCH1 is a type I transmembrane receptor that regulates T cell development via a signaling pathway that relies on regulated proteolysis. During its maturation, most NOTCH1 is cleaved at a position 70 amino acids external to the transmembrane domain by a furin-like protease, creating extracellular (NEC) and transmembrane (NTM) subunits that are held together non-covalently by a juxtamembranous heterodimerization (HD) domain. Ligand-binding to NEC promotes cleavage by i) metalloproteases at a site in the ectodomain of NTM, followed by ii) gamma-secretase within the transmembrane domain. This releases the NOTCH1 intracellular domain (ICN1), allowing it to translocate to the nucleus and activate target genes. Normally, proteolysis is constrained prior to ligand-binding by an extracellular negative regulatory region consisting of 3 iterated LNR repeats and the N- and C-terminal portions of the HD domain, which flank the furin cleavage site. Recent work has shown that human T-ALL is frequently associated with gain of function mutations that map to the HD domain of NOTCH1. These mutations are distributed in both parts of the HD domain and include point mutations, short insertions, and deletions, suggesting that there might be variation in their relative strength and the mechanisms by which they act. To investigate these issues, we introduced 16 of the HD domain mutations found in primary T-ALLs or T-ALL cell lines into a full-length NOTCH1 cDNA, and tested their ability to activate a NOTCH sensitive luciferase reporter gene. Except for the "mutation" R1609S, which was found in only one primary T-ALL sample, all of the mutations stimulated NOTCH1 signaling. These increases in signaling were abolished by a gamma-secretase inhibitor and were associated with increased rates of metalloprotease-mediated cleavage, indicating that activation proceeds through the normal series of proteolytic events. The mutations also caused gains in function when introduced into NOTCH1 polypeptides lacking the ligand-binding region of NEC, indicating that the HD domain mutations can cause ligand-independent receptor activation. Since NEC dissociation can lead to activation of NOTCH signaling (and is a proposed mechanism for normal ligand-mediated NOTCH activation), one simple way for HD domain mutations to act is through the destabilization of NOTCH1 heterodimers. To test this model, each mutation was introduced into soluble NOTCH1 mini-receptors bearing N-terminal FLAG and C-terminal HA tags. When expressed transiently, the normal NOTCH1 mini-receptor was secreted into conditioned media as a furin-processed heterodimer. Certain activating HD domain mutations, such as L1601P, resulted in complete dissociation of the furin-processed mini-receptor subunits under native conditions, and all other HD domain mutations save one were more sensitive to urea-induced dissociation than normal NOTCH1. The exception was an unusual insertional mutation (identified in the P12-Ichikawa cell line) that introduces a 14 amino acid direct repeat sequence at a position immediately N-terminal of the metalloprotease cleavage site. We hypothesize that this mutation, which was associated with the greatest increases in signaling in NOTCH1 reporter gene assays, displaces protective HD and LNR domain residues and thereby unveils the metalloprotease cleavage site. We conclude that most T-ALL-associated HD domain mutations confer ligand-independent gain-of-function on NOTCH1 receptors, but vary in strength and are likely to act through several distinct mechanisms.

Author notes

Corresponding author