List of human NOTCH1 cancer mutations and their Drosophila equivalents used in this study.

Summary of 22 cancer-associated mutations analysed in this study. Drosophila Notch mutations were designed based on reported human NOTCH1 cancer mutations. The table lists each Drosophila mutation alongside its corresponding human mutation, structural location, associated cancer types, and relevant references.

Ligand-independent activation of Notch signal by gain-of-function mutations in the Drosophila Notch NRR

(A) Front and side views of the human NOTCH1 NRR crystal structure, highlighting residues corresponding to four Drosophila gain-of-function mutations: l1N-B (G1572V), 414 (D1577G), CC-SS (C1693S and C1696S), and LGI>AAA (LGI1514–1516AAA). All mutations in this study are uniquely colour-coded for consistency across structural representations and quantitative analyses, including NRE-luciferase assays and protein-level measurements. (B) Enhanced ligand-independent signalling by Notch gain-of-function mutants. NRE-luciferase assay in S2 cells show that all four gain-of-function Notch mutants exhibit significantly increased basal signalling activity compared to wild-type. (C-E) Ligand (Dl) and Dx-induced Drosophila Notch signal. Wild-type Notch signal can be further activated by Dl and Dx (C). Dl-(D) and Dx-(E) induced Notch signal intensity of the four gain-of-function mutants were normalised by the signal from wild-type Notch.

Functional analysis of T-ALL–derived heterodimerisation domain mutations in Drosophila Notch

(A) Positions of residues corresponding to seven Drosophila mutations (F1617P, R1626Q, H1630P, L1632P, R1633P, L1686P, and E1705P) in HD domain shown on human NOTCH1 NRR crystal structure. (B) Basal Notch activity of the seven HD domain mutants, demonstrating that the only two mutants R1626Q and E1705P exhibit accelerated signal intensity in ligand-independent manner, like T-ALL mutants. (C,D) Notch signal activation of HD domain mutants by ligand (C) and Dx (D), normalised by wild-type Notch activated by ligand and Dx, respectively. (E) Impaired secretory trafficking of Notch HD domain mutants due to ER/Golgi-retention. Upper panels: EGFP-tagged wild-type Notch and two HD domain mutants (L1632P and L1686P) were expressed in S2 cells and Notch protein localisation (EGFP, green) was analysed by immunofluorescence. The cells were stained with anti-PDI (for ER, red) and anti-GM130 (for Golgi, blue) antibodies. There was no clear difference between WT and G1515K localisation despite the intense basal signal of G1515K (Figure3B). Lower panels: higher magnification of the box shown in L1632P image. Scale bar: 5µm (upper panels) and 1µm (lower panels). (F) Effect of PEST domain truncation to ligand-independent Notch signal of HD domain mutants. PEST domain in the C-terminal region of Notch was removed from four representative mutants (R1626Q, L1632P, L1686P, and E1705P), and the basal signal was analysed by NRE-luciferase assay. Only R1626Q and E1705P show T-ALL mutants-like synergistic increase of the signal.

Structure-function analysis of the Drosophila Notch LNR domain via cancer-derived mutational mapping.

(A) Distribution of cancer-associated LNR mutations on the human NOTCH1 NRR Structure. Fourteen mutations (E1489V, D1497N, A1504V, G1515K, T1523M, N1529E, E1538R, D1548N, E1553P, R1554C, L1562P, H1570P, D1573V, and Y1578R) were introduced into Drosophila Notch, based on cancer-associated variants reported in the human NOTCH1 LNR domain. The D1548N mutation, which contributes to a calcium binding site, is buried within the LNR and is not visible in the figure. (B) Basal Notch activity of the LNR domain mutants, demonstrating that the only two mutants G1515K and E1553P exhibit apparently high signal intensity in ligand-independent manner. (C,D) Notch signal activation of the LNR mutants by ligand (C) and Dx (D), normalised by wild-type Notch activated by ligand and Dx, respectively. (E) Synergistic enhancement of ligand-independent Notch signalling by combined LNR mutations and PEST domain truncation. The effect of PEST domain truncation on the basal activity of two gain-of-function LNR mutants (G1515K and E1553P) was assessed using the NRE-luciferase assay. The combination led to a synergistic increase in ligand-independent signalling, resembling the behaviour of T-ALL–associated NOTCH1 mutations.

Identification of a novel regulatory hotspot in LNR-C driving ligand-independent Notch activation.

(A) Human NOTCH1 NRR crystal structure highlighting the positions of three surface-exposed LNR-C mutations (F1563A, Y1566A, and H1570A). (B) Strong increase in basal (ligand-independent) Notch signalling induced by the three LNR-C mutations. (C,D) Ligand-dependent (C) and Dx-induced (D) activation of the LNR-C mutants, normalised to wild-type Notch activation under the same conditions. (E) Effect of PEST domain truncation on ligand-independent signalling of the LNR-C mutants.

Intact NRR–NRR homodimerisation despite LNR-C interface mutations

(A,B) Co-immunoprecipitation–based assay for NRR–NRR homodimerisation. (A) Schematic presentation of the assay. EGFP-tagged full-length Notch was immunoprecipitated from S2 cells co-expressing Notch–EGFP and luciferase-tagged Notch LNR domain using an anti-GFP antibody (GFP-trap). Homodimerisation was assessed by measuring co-precipitated luciferase activity. (B) Luciferase signals were normalised to total lysate luciferase activity to quantify binding efficiency. Robust NRR–NRR homodimerisation was observed, and none of the tested interface mutations significantly affected this interaction. (C,D) Split-luciferase assay to assess the effect of LNR-C mutations on NRR–NRR homodimerisation. (C) Schematic representation of the assay. N-terminal and C-terminal fragments of luciferase were fused to Notch LNR constructs and co-expressed in S2 cells. Homodimerisation was detected as reconstituted luciferase activity. (D) Firefly luciferase signals were normalised to Renilla luciferase activity to quantify dimerization efficiency. Strong luciferase activity indicated NRR–NRR homodimerisation, and none of the tested LNR-C interface mutations significantly altered this interaction. All constructs include N-terminal signal peptides to ensure proper membrane topology.

A novel negative regulatory function of LNR-C in Notch protein stability

(A,B) Protein expression levels of Notch cancer mutants in S2 cells. Expression profiles of all cancer-derived Notch mutants used in this study were examined by western blotting (A) and the band intensities of full-length Notch were quantified to compare relative protein levels (B). (C,D) Elevated Notch Protein Levels by LNR-C Interface Mutations. Expression profiles of LNR-C interface mutants and two Drosophila alleles, l1N-B (G1572V) and 414 (D1577G), were examined by western blotting (C) and the band intensities of full-length Notch were quantified to compare relative protein levels (D). (E,F) Enhanced stability of Notch LNR-C interface mutants revealed by cycloheximide (CHX) chase assay. (E) S2 cells expressing LNR-C interface mutants (F1563A or H1570A) were treated with 10 µM cycloheximide for the indicated time and assessed by western blotting. (F) Band intensities of full-length Notch were quantified to evaluate protein stability over time.

Inducibility of NRR mutants by Dl and Dx compared to WT.

A,B, Inducibility of NRR mutants compared to WT for Dl-induced (A,D,F) and Dx-induced (B, E,G) expressed as a % of WT. (A-C) Inducibility of Drosophila NRR mutant alleles to Dl (A) and Dx (B). (C) Fold-decrease of 414 allele after Dx expression compared to 414 basal shows significant decrease. (D,E) Inducibility of cancer-associated HD and LNR domain mutants for Dl (D) and Dx induction. (F,G) Inducibility of LNR-C dimer interface mutants in response to Dl (F), and Dx (G). To calculate relative inducibility first all basal signals were normalised to 1. The “fold difference” was calculated as induced signal -1 for mutant and for WT and mutant value expressed as a percentage of the WT fold difference. Hence 100% indicates fold inducibility of mutant is same as WT, 0% indicates mutant signal does not change in induced conditions compared to its own basal signal, and negative numbers are cases where the mutant signal is reduced in the presence of Dl or Dx, compared its own basal condition with the fold difference expressed as a % of the WT fold difference. * indicates P<0.05 compared to WT (A,B,D-G) or compared to 414 basal (C), by t-test, error bars are SEM.