|Home | About | Journals | Submit | Contact Us | Français|
Notch signaling affects many developmental and cellular processes and has been implicated in congenital disorders, stroke, and numerous cancers. The Notch receptor binds its ligands Delta and Serrate and is able to discriminate between them in different contexts. However, the specific domains in Notch responsible for this selectivity are poorly defined. Through genetic screens in Drosophila, we isolated a mutation, Notchjigsaw, that affects Serrate- but not Delta-dependent signaling. Notchjigsaw carries a missense mutation in epidermal growth factor repeat-8 (EGFr-8) and is defective in Serrate binding. A homologous point mutation in mammalian Notch2 also exhibits defects in signaling of a mammalian Serrate homolog, Jagged1. Hence, an evolutionarily conserved valine in EGFr-8 is essential for ligand selectivity and provides a molecular handle to study numerous Notch-dependent signaling events.
The evolutionarily conserved Notch (N) signaling pathway affects numerous cell fate and differentiation events as well as proliferation and cell death (1). Signal activation is initiated by the binding of N receptor to ligands, Delta (Dl) or Serrate (Ser) (2). The majority of the extracellular domain of N receptor is composed of epidermal growth factor repeats (EGFrs) (Fig. 1A). EGFr-11 and EGFr-12 are necessary for ligand-receptor interactions with both Dl and Ser (3), whereas EGFr-24 to EGFr-29 (Abruptex domain) negatively regulate these interactions (4). Although the in vivo role of most EGFrs is unknown, mutations in these repeats are associated with numerous human diseases, including cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) (5), Alagille syndrome (ALGS) (6), aortic valve diseases (AVDs) (7–9), and squamous cell carcinoma (SCC) (10–12). Mammals have four N paralogs (NOTCH1 to NOTCH4) that share similar structural organizations, but only one N gene exists in Drosophila, simplifying structure-function analysis in vivo (table S1).
To obtain N mutations, we performed an F3 mosaic genetic screen on the X chromosome (fig. S1) and isolated 42 additional alleles of N. Twenty-one alleles carry different missense mutations and were grouped into eight distinct classes on the basis of molecular and phenotypic features (tables S2 and S3). All mutated residues are conserved in most human N paralogs.
One mutation, Valine361-to-Methionine (V361M) in EGFr-8, Njigsaw, exhibits defects in the wing margin without affecting venation or bristle development in mutant clones (Fig. 1, B to F, and table S3). Hemizygous mutants of Njigsaw are pupal lethal, and Njigsaw/+ flies do not display wing notching. Njigsaw fails to complement the lethality of null alleles of N and is rescued by a genomic rescue construct. Homozygous mutant clones in the wing exhibit strong notching and occasional ectopic wing margin formation (Fig. 1, D and E). We did not observe any bristle density or cell fate defects (Fig. 1F and figs. S2 and S3). Hence, Njigsaw displays a specific phenotype for a lethal N allele as inductive signaling is impaired, whereas lateral inhibition and lineage decisions remain unaffected.
Because the phenotypes associated with Njigsaw are similar to Ser loss of function (13, 14), we determined whether N-Ser signaling is compromised. In the early wing primordium, N is activated at the boundary between the dorsal and ventral compartments (Fig. 1G). The dorsal domain expresses both Ser and Fringe (Fng) (15), whereas Dl is mainly expressed ventrally (16). Fng, a β3-N-acetylglucosaminyltransferase that adds N-acetylglucosamine (GlcNAc) to O-fucosylated EGFrs, modifies N so that it can be activated by Dl but not Ser (17, 18). As a result, Dl activates N in the dorsal domain, and Ser activates N in the ventral domain in cells flanking the dorsal-ventral boundary (Fig. 1, G and H). In Njigsaw hemizygous discs, N activation is severely reduced or lost (Fig. 1I). Furthermore, cells that activate N signaling are present in the dorsal but not ventral compartment (Fig. 1I′). These data indicate that Njigsaw is defective in N-Ser but not N-Dl signaling. This was confirmed in mosaic tissues (fig. S4). Last, Njigsaw mutant clones can be ectopically activated by overexpression of Dl but not by Ser (fig. S5).
The defects in N-Ser signaling in Njigsaw mutants are not confined to the wing. Ser is required during hematopoiesis for crystal cell specification in the lymph glands (19). In Njigsaw mutant animals, the crystal cell marker Lozenge is not expressed (fig. S6, A and B), resulting in a loss of circulating crystal cells (fig. S6, C to E). Furthermore, similar to Njigsaw (fig. S6, F and G), loss of Ser causes a loss of salivary imaginal ring cells (20). Hence, multiple processes requiring N-Ser signaling are severely impaired in Njigsaw mutants.
To determine the cause of ectopic wing margin formation in Njigsaw mutant clones (Fig. 1E), we monitored N activity. We observed ectopic N signal activation in wild-type (WT) cells adjacent to Njigsaw mutant cells in the ventral compartment near the dorsal-ventral boundary (Fig. 2, A and B). Binding of N to Ser in cis mediates the removal of Ser from the plasma membrane in WT cells (ligand-cis-inhibition) (Fig. 2D) (21), whereas in mutant cells, Ser is up-regulated (Fig. 2C and fig. S7A). The ectopic activation of N in the neighboring WT cells can be suppressed by reducing the levels of Ser specifically in Njigsaw mutant cells (fig. S7B), suggesting that Njigsaw mutant cells are defective in ligand-cis-inhibition (Fig. 2E). Hence, Njigsaw is defective in Ser-mediated signaling, both in trans and in cis, whereas N-Dl signaling appears unaffected.
Changes in the glycosylation status of EGFrs by Fng are known to modulate the ligand selectively of Notch (17, 18). When one copy of fng is removed, notching in Njigsaw clones is enhanced, whereas overexpression of Fng in Njigsaw mutant clones partially suppresses notching (figs. S8 and S9). These data suggest that Njigsaw and Fng affect the same process independently.
To test whether Njigsaw affects the glycosylation status of N, we analyzed glycopeptides using mass spectrometry (22, 23). The Njigsaw mutation does not cause a significant change in the level of O-fucosylation and Fng-mediated elongation of O-fucose on EGFr-8 (Fig. 3, A to C, and fig. S10). In addition, other EGFrs do not display alterations in O-fucosylation and Fng-mediated elongation (fig. S11), indicating that these EGFrs are properly folded, which is a requirement for glycosylation. Hence, the Njigsaw mutation has little or no effect on the sugar modification status of N, again suggesting that the defects in N-Ser signaling are independent of Fng.
Because both trans- and cis-signaling of N-Ser are defective in Njigsaw mutant cells, and the expression, exocytosis, and endocytosis of Njigsaw are indistinguishable from NWT (figs. S12 and 13), we tested whether Njigsaw affects Ser binding (17, 24). In contrast to NWT, Njigsaw-Ser binding is significantly diminished, and this interaction is unaffected by the presence of Fng (Fig. 3D). Moreover, overexpression of O-fut1, a sugar-modifying enzyme and molecular chaperone that can strengthen NWT-Ser binding (24), cannot improve Njigsaw-Ser binding (Fig. 3D). In contrast, Njigsaw interacts with Dl in the presence of Fng (Fig. 3E). These data indicate that Njigsaw directly impairs N-Ser binding without disrupting N-Dl binding.
To assess whether Njigsaw (V361M) plays a similar role for a mammalian N, we performed a signaling assay (25). We introduced V327M (table S2) into mouse Notch2 and tested its ability to respond to Jagged1 or Delta-like1 (Dll1). Notch2-Jagged1 signaling is severely decreased, whereas Notch2-Dll1 remains unaffected (Fig. 3, F and G). In addition, a V327A mutation causes a further decrease in Notch2-Jagged1 signaling, showing that loss of valine is responsible for the Njigsaw phenotype. In contrast, a mutation in the conserved O-fut1/Fng–modification site within Notch2 EGFr-8 (T314A) does not affect Jagged1- or Dll1-mediated signaling.
Among the identified N mutations (tables S2 and S3) and other reported N alleles (26, 27), Njigsaw separates N-Ser from N-Dl binding and signaling in vivo. Previously, two N alleles have been shown to exhibit wing margin defects only when mutant cells are in the ventral compartment, suggesting defective Ser-mediated signaling (28). However, these mutations do not affect the EGFrs, and the nature of the signaling defects remains uninvestigated. Our data suggest that in addition to EGFr-11 and EGFr-12, which are required for both Dl and Ser binding (3), EGFr-8 plays a critical role in N-Ser interactions. The Njigsaw mutation does not obviously affect the trafficking or glycosylation of N. Therefore, EGFr-8 may directly be involved in ligand-receptor interactions together with EGFr-11 and EGFr-12 or may be crucial for N to adopt a conformation favoring Ser binding.
We thank the Bloomington Drosophila Stock Center, the Vienna Drosophila RNAi Center, and the Developmental Studies Hybridoma Bank for distribution of flies and antibodies. We are grateful to S. Artavanis-Tsakonas, G. Bornkamm, S. Bray, S. Chiba, P. Heitzler, H. Hiral, K. Irvine, H. Jafar-Nejad, A. Martinez-Arias, M. Milan, L. Strobl, G. Weinmaster, and C. Wesley for flies and reagents. We thank K. Schulze for critical reading and editing of the manuscript. We acknowledge S. Gibbs, A. Jawaid, Y. Chen, C. Benitez, B. McNulty, A. Lin-Moore, D. Bei, L. Wang, Y. Kawagishi, and T. Busby for technical assistance. We thank R. Kopan, S. Blacklow, M. Milan, K. Cook, N. Giagtzoglou, K. Venken, H. Tsuda, C. Yao, and H. Takeuchi for helpful comments and discussion. This research was supported in part by the NIH (1RC4GM096355-01 to H.J.B. and 5R01GM061126-12 to R.S.H). Confocal microscopy at Baylor College of Medicine is supported by the Intellectual and Developmental Disabilities Research Center (NIH 5P30HD024064). S.Y. was supported by a fellowship from the Nakajima Foundation. W.-L.C. was supported by Taiwan Merit Scholarships Program sponsored by the National Science Council (NSC-095-SAF-I-564-015-TMS). V.B. received support from the NIH (5T32-HD055200) and the Edward and Josephine Hudson Scholarship Fund. B.X. was supported by the Houston Laboratory and Population Science Training Program in Gene-Environment Interaction from the Burroughs Wellcome Fund (grant 1008200). H.S. received support from a supplement to NIH 5R01GM067858 and the Research Education and Career Horizon Institutional Research and Academic Career Development Award Fellowship 5K12GM084897. H.J.B. is a Howard Hughes Medical Institute Investigator.