Reduced cilium length is caused by enhanced γ-secretase activity and diminished Foxj1a levels. (a) 2.5-fold increased γ-secretase activity in Rer1p-depleted CL4 cells as indicated by increased NICD production (n = 7). ctrl, control. (b) Overexpression of NICD but not of Cadherin ICD (CICD) reduces cilium length in CL4 cells. (c) Nanomolar concentrations of γ-secretase inhibition (InhX) rescue cilia length in Rer1p-depleted CL4 cells without affecting cilia length in control cells. Acetylated tubulin staining (n = 150; three experiments). NSC, nonspecific control. (d) Reduced deltaA but unaffected notch1 and notch3 levels in 24-hpf SMO. (e) Enhanced Notch signaling in the ear sensory patch, but not in the tail, indicated by increased fluorescence of the Notch reporter line Tp1:hmgb1-mCherry. Quantification is shown of highlighted regions (n > 15). UI, uninjected. (f) FM1-43 dye labeling shows a twofold reduction in the number of neuromasts in 3-dpf SMO. (g) Twofold reduction in the number of hair cells per neuromast (PLL) in 2-dpf SMO-injected ET4 transgenic embryos. Increased Notch signaling (100 pg NICD mRNA) results in less hair cells per neuromast (71% of control), whereas inhibiting Notch signaling (25 µM compound E [cpdE]) increases hair cell number (150% of control) and rescues the effect of SMO from 46 to 93.6% (n > 12). (h) WISH shows normal differentiation of the DFC/KV lineage (lrdr1) but reduced foxj1a levels in DFC (80% epiboly, arrows), KV (bud, arrows), pronephros (arrowheads), and notochord (asterisks). ss, somite stage. (i) Quantification of foxj1a levels by quantitative PCR in SMO. NICD does not affect foxj1a levels in the bud. NI, not injected. (j–l) Partial rescue of pronephros (j and k) and KV (l) cilia length upon coinjection of Foxj1a (n ≥ 11 embryos; three experiments). Ac.tub, acetylated tubulin. Means ± SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.001, Wilcoxon rank sum test (b and c) or Student’s t test (a, e, g, i, k, and l).