The divergent DSL ligand Dll3 does not activate Notch signaling but cell autonomously attenuates signaling induced by other DSL ligands

Abstract

Mutations in the DSL (Delta, Serrate, Lag2) Notch (N) ligand Delta-like (Dll) 3 cause skeletal abnormalities in spondylocostal dysostosis, which is consistent with a critical role for N signaling during somitogenesis. Understanding how Dll3 functions is complicated by reports that DSL ligands both activate and inhibit N signaling. In contrast to other DSL ligands, we show that Dll3 does not activate N signaling in multiple assays. Consistent with these findings, Dll3 does not bind to cells expressing any of the four N receptors, and N1 does not bind Dll3-expressing cells. However, in a cell-autonomous manner, Dll3 suppressed N signaling, as was found for other DSL ligands. Therefore, Dll3 functions not as an activator as previously reported but rather as a dedicated inhibitor of N signaling. As an N antagonist, Dll3 promoted Xenopus laevis neurogenesis and inhibited glial differentiation of mouse neural progenitors. Finally, together with the modulator lunatic fringe, Dll3 altered N signaling levels that were induced by other DSL ligands.

Figure 1.

Dll3 does not activate N signaling in trans. (A) Cell surface expression of HA-tagged Dll1, Dll3, and Dll1 + Dll3 L cell lines was determined by biotinylation, streptavidin (SAV) pull-down, and blotting with anti-HA mAb (12CA5). (B) NIH 3T3 cells transfected with N and CSL reporter were cocultured with J1, Dll1, Dll3, mDll3, or parental (L) cell lines and were assayed for luciferase activity. Error bars reflect the SD of the mean from three experiments. RLU, relative luciferase units. (C) Parental C2C12 and cell lines expressing N1 were cocultured with L, J1, Dll1, or Dll3 cells, and myogenesis was monitored by myosin light chain 2 (MLC2) mRNA expression. Loading and transfer of RNA was monitored by methylene blue staining of 18S rRNA. (D) NIH 3T3 cells transfected with N1, CSL reporter, and control or LFng were cocultured with Dll1, J1, rDll3, mDll3, or L cells and were assayed for luciferase activity (n = 3). (E) Parental L, Dll1, or Dll1 + Dll3 cells were cocultured with NIH 3T3 cells transfected with CSL reporter and N or vector. There was no statistically significant difference between Dll1 and Dll1 + Dll3 (n = 3).

Figure 2.

Dll3 does not bind to N. (A) Western Blot (WB) analysis using anti-Fc quantitated Dll3Fc, D1Fc, and D1^NTD3Fc for binding assays in B. Numbers represent the dilution of condition media used in the binding assays. (B) 293T cells transfected with vector or N1 were assayed for binding of Fc, Dll3Fc, D1^NTD3Fc, and D1Fc by flow cytometry. Fold binding over Fc control is plotted against serial dilutions. (C) Binding of N1Fc to 293T cells transfected with vector or HA-tagged Dll1, Dll3, or D1^NTD3 is shown as fold binding over vector control (n = 3). Error bars represent SD. (D) Alignment of the DSL domains of rat and mDll3 with other Dll family members identifies Dll3 as highly divergent. Amino acid sequences conserved in Dll3 are red, whereas amino acid sequences conserved amongst the other Dll-related proteins are blue.

Figure 3.

Dll3 cell autonomously inhibits N signaling. (A) NIH 3T3 cells cotransfected with N1 or N2 and either HA-tagged Dll1, rat Dll3 (rDll3), mDll3, or vector along with a CSL reporter cocultured with Dll1, J1, or L cells. Error bars reflect the SD of the mean from four experiments (*, P < 0.01; **, P < 0.001; ***, P < 0.0001). (B) Western blot of HA-tagged Dll1 and Dll3 from NIH 3T3 cells indicates expression in CSL reporter assays in A. (C) 293T cells cotransfected with N1Δmyc and either vector, Dll1, or Dll3 plasmids and cell surface expression of N1 were analyzed after biotinylation/SAV pull-down and anti-myc (9E10) Western blotting. (D) 293T cells transfected with vector or HA-N1 and either vector, Dll1, or Dll3 plasmids were stained live, and mean fluorescence intensity was determined by FACS. Error bars represent SD. (E) 293T cells cotransfected with Dll1 or Dll3 and N1Δmyc or vector plasmids and cell surface expression of Dll1 or Dll3 were determined by biotinylation/SAV pull-down and anti-HA antibody (12CA5) Western blotting. (F) Lysates from 293T cells transfected with N1 and either HA-tagged Dll1, Dll3, or vector were incubated with N1 antibodies and HA antibody Western blot to detect Dll1 or Dll3 (12CA5; top) or with N1 antibody (93–4; bottom). Middle panel is a HA Western blot of Dll1 and Dll3 from WCL.

Figure 4.

Injection of mDll3 mRNA promotes neurogenesis in X. laevis . Representatives of embryos injected with either synthetic lacZ mRNA alone (A) or in combination with 250 pg Dll1 mRNA (B), 250 pg mDll3 mRNA (C), or 1,000 pg mDll3 mRNA (D) and stained for X-gal (light blue) and β-tubulin (purple) expression. Arrows indicate injection site.

Figure 5.

Dll3 suppresses D1Fc-induced astrocytic differentiation. Embryonic day 11.5 mouse cortical NSCs cotransfected with reporters for GFAP (A) or S100β (B) and either vector or Dll3 were treated with control Fc or D1Fc to induce astrogliogenesis. (C) NSCs cotransfected with GFP and either vector or Dll3 were cultured with D1Fc for 4 d and stained for GFAP. (D) Quantitation of transfected cells expressing GFAP from six independent experiments. (E) GFAP expression in NSCs infected with control (lane 1) or Dll3 adenovirus (lane 2) as determined by Western blotting. (F) Neurogenic stage mouse cortical NSCs cotransfected with NeuroD promoter luciferase construct and vector or Dll3. Promoter activation is plotted as relative luciferase units for six experiments. P < 0.05. *, significant differences between Dll3 and vector; **, significant increase between Fc and D1Fc. Error bars represent SD. AU, arbitrary units. (G) Tubulin (TuJ1) expression in NSCs infected with control (lane 1) or Dll3 adenovirus (lane 2) as determined by Western blotting.

Figure 6.

LFng and Dll3 dynamically modulate N signaling. (A) NIH 3T3 cells cotransfected with N1 and CSL reporter as well as vector or Dll3 with increasing amounts of LFng or control DNA were cocultured with L or Dll1 cells and assayed for luciferase activity. (B) NIH 3T3 cells cotransfected with N1 and CSL reporter as well as control or LFng, and either vector or increasing amounts of Dll3 DNA were cocultured with L or Dll1 cells. P < 0.01. *, a significant difference between vector and Dll3 (low); **, a significant difference between Dll3 (low) and Dll3 (high). n = 5. (C) Lysates from 293T cells cotransfected with N1Δmyc, HA-tagged Dll3, and LFng or control DNA incubated with anti-HA antibody to capture Dll3 immunoprecipitates, which were identified by anti-myc Western blotting to detect N1 (right). Lysate from N1 cells was mixed with equal amounts of D3 lysate and was analyzed alongside other samples (Mix Lys). WCLs were analyzed by anti-myc (top left) and anti-HA Western blotting (bottom left). (D) Parental L or J1 cells cocultured with NIH 3T3 cells that were transfected with N1, CSL reporter, vector, or Dll3 and control or LFng. (*, P < 0.005; **, P < 0.001; n = 7). Error bars represent SD.