Fig 2: scRNA-seq analysis identified Notch1-Dll4 signaling in the VSN dichotomy. (A) UMAP dimension plot of Seurat object3 specifically focusing on the VSN apical-basal split. Letters A and B on the right of the UMAP denote apical and basal VSN branches, respectively. (B,C) Feature plots of known marker genes specific to neuronal progenitors, precursors and VSNs. (D) Feature plot of Bcl11b expression across the VSN dichotomy. The arrows highlight continuous Bcl11b expression in the basal VSN branch. Bcl11b expression appears at a later stage in the apical VSN branch. (Di) Feature plot highlighting Bcl11b+ and Bcl11b- cells at the split point used for differential gene expression analysis. (Dii) Volcano plot highlighting Notch-related genes differentially expressed between Bcl11b+ and Bcl11b- clusters. (E) Feature plots of Dll4 (arrowhead) and Notch1 (arrow) demonstrate their mutually exclusive expression at the VSN dichotomy. (F) Co-expression feature plots of Dll4 and downstream Notch signaling targets like Hes5, Hey1, Nrarp and Ccnd1 (arrows) show that active Notch signaling occurs transiently at the split point and early stages of the basal VSN trajectory. (G) Co-expression feature plot of Dll4 versus Meis2 highlights Dll4 expression at the start of the apical VSN branch. (H) Co-expression feature plot of Hey1 versus Tfap2e shows active Notch signaling in the basal VSN branch. (I) Schematic showing the inferred Notch1-Dll4 signaling required for establishing the early VSN dichotomy. (J) Dot plot of a few selected genes that are differentially expressed between the Dll4+ and Notch1+ clusters at the VSN dichotomy.

Fig 3: Ectopic expression of NICD at the Neurog1+ stage diverts neuronal precursors towards basal VSN fate. (A) Neurog1CreERT2/R26tdTom and Neurog1CreERT2/R26NICD pups were injected with tamoxifen at P1 and perfused at 7 and 14 dpi. (Ai) Expected results show that NICD overexpression at Neurog1+ stage may drive progenitors towards the basal VSN fate. (B-E) Immunofluorescence anti tdTom/Meis2/Tfap2e in Neurog1CreERT2-traced control pups and Neurog1CreERT2/R26NICD mutant pups at 14 dpi. Panels B and C highlight tdTom/Meis2 staining and panels D and E highlight tdTom/Tfap2e (AP-2ε) staining. Arrows show traced Tfap2e+ basal VSNs and arrowheads show Meis2+ apical VSNs in both control and inducible mice. (F,G) Quantification of percentage traced tdTom+ cells or GFP+ cells at 7 dpi and 14 dpi that are Meis2+, Tfap2e+ VSNs or Meis2/Tfap2e double negative cells in control and NICD mice. Statistical analysis based on arcsine-transformed values of the percentage data; unpaired two-tailed t-test; n=3 biological replicates. Data shown as mean±s.e.m. At 7 dpi, the average number of recombined cells in the control group was 368.1±23.8 and in the mutant group it was 128.3±35.2. At 14 dpi, the average number of recombined cells in the control group was 509±46.6 and in the mutant group it was 148.5±20.7.

Fig 4: Canonical Notch signaling establishes basal VSN fate. (A) Ascl1CreERT2/Rbpjfl/fl/R26tdTom and Ascl1CreERT2/R26tdTom pups were injected with tamoxifen at P1 and perfused at 7 dpi. (Ai) Expected results with Rbpj KO driving the progenitors towards the apical VSN fate. (B-Cii) Immunofluorescence anti Meis2, Tfap2e and tdTom in Ascl1CreERT2-induced control pups and Rbpj KO pups at 7 dpi. Arrows highlight traced Meis2+ and arrowheads highlight traced Tfap2e+ basal VSNs. (D) Quantification of the percentage of traced VSNs that are Meis2+, Tfap2e+, and Meis2/Tfap2e double negative cells in control and Rbpj KO mice at 7 dpi. Statistical analysis based on arcsine-transformed values of the percentage data; unpaired two-tailed t-test; n=3,4 biological replicates. Data shown as mean±s.e.m. At 7 dpi, the average number of tdTom traced cells in control group was 217.5±25.7 and in cKO it was 177±13.9.

Fig 5: R26AP-2e mouse line mouse generation and characterization of rescued Tfap2e expression in Tfap2eNull mice.(A) Knock-in strategy through homologous recombination to generate the R26AP-2e mouse line. The CAG-loxP-stop-loxP-mouse-Tfap2e cassette as integrated into the first intron of Rosa26. (B–D’) Immunohistochemistry on P21 wild-type (WT) (B, B’), Tfap2eNull (C,C’), and Tfap2eRescue (D,D’) mice against AP-2e. (B,B’) In WT mice AP-2e expression is in the marginal zones (MZ) and in the basal regions of the VSNs. (C,C’) In Tfap2eNull mice, some AP-2e immunoreactivity is observed in the MZ but lost in the central regions where more mature neurons reside. (D,D’) In the Tfap2eRescue, AP-2e is expressed in the basal region, but with less intensity and density at the MZ and in central regions compared to WT controls. (E–G) Immunostainings against V2R2 (magenta) and Meis2 (green) counterstained with 4’,6’-diamidino-2-phenylindole (DAPI) (white). (H–J) Immunostainings against Gao (magenta) and Gai2 (green) counterstained with DAPI. (K–M) Immunostainings against Krt18 (magenta) counterstained with DAPI (white). (N) Quantifications of the percentage of vomeronasal sensory neurons (VSNs) expressing apical/basal markers in WT, Tfap2eNull, and Tfap2eRescue mice. Tfap2eNull mice show a dramatic reduction in basal VSNs and apical VSNs occupy most of the epithelium. The Tfap2eRescue has an intermediate phenotype between WT and Tfap2eNull mice, where the vomeronasal epithelium (VNE) contains more basal VSNs than in the Tfap2eNull mice but does not reach the equivalency of the WT (p<0.05 = *, p<0.01 = **, p<0.001 = ***, p<0.0001 = ****). One-way ANOVA. Error bars are standard deviation. N=4 for WT in % AP-2e, % meis2, % Gao, % Gai2, and % Krt18. N=6 for WT in % V2R2. N=2 for Tfap2eNull in % AP-2e. N=4 for Tfap2eNull in % Meis2, % V2R2, and % Krt18. N=6 for Tfap2eNull in % Gao and % Gai2. N=4 for Tfap2eRescue in % AP-2e, % Meis2, and % V2R2. N=6 for Tfap2eRescue in % Gao, % Gai2, and % Krt18.