Fig 2: Deletion of embigin disrupts sebaceous gland homeostasis(A) Hematoxylin and eosin staining of the tail skin of adult WT and Emb-/- mice.(B) Adult tail wholemounts from WT and Emb-/- as well as from Embflox/flox and K14CreEmbflox/flox mice labeled with antibodies against ITGA6 (green) and DAPI (blue).(C) Quantification of area covered by SGs per HF. The average of =9 HFs per mouse quantified from wholemounts is shown (Cre-, n = 4 mice; K14CreEmbflox/flox, n = 4 mice; WT, n = 4 mice; Emb-/-, n = 4 mice).(D) Quantification of size of suprabasal sebocytes (WT, n = 72, Emb-/-, n = 72) from Herovici stained sections. Cells were pooled from 3 WT and 3 Emb-/- mice. Red circles delineate examples of individual suprabasal cells.(E) Quantification of suprabasal sebocyte number per sebaceous gland from cryosections. Average ± SD of mice is shown (2–3-month WT, n = 5; 2–3-month Emb-/-, n = 4; 7-month WT, n = 3; 7-month Emb-/-, n = 3). At least 7 SGs per mouse were quantified.(F) Tail wholemounts from WT and Emb-/- mice labeled with antibodies against Ki67 (red) and ITGA6 (green) with DAPI nuclear counterstain (blue).(G) Quantification of Ki67+ cells in WT and Emb-/- SGs (per) SG or normalized to SG size (WT, n = 37; Emb-/-, n = 28). SGs were pooled from 3 WT and 3 Emb-/- mice.(H) Cryosections of Cre- and K14CreEmbflox/flox mouse tail SG on the same day (0d) or 11 days after (11d) EdU injection stained with antibodies to EdU (green) and basement membrane marker laminin (red) with DAPI nuclear counterstain (blue).(I) Quantification of the percentage of suprabasal EdU+ cells per SG (0-day Cre-, n = 30; 0-day K14 K14CreEmbflox/flox, n = 32; 11-day Cre-, n = 48; 11-day K14CreEmbflox/flox, n = 45). SGs were pooled from 3 mice per group.(J) Cryosections of neonatal (P1–P2) WT and Emb-/- tail HFs stained for stearoyl-CoA desaturase-1 (Scd1), which is involved in lipid synthesis (green), and basement membrane marker laminin (red) with DAPI nuclear counterstain (blue).(K) Quantification of SCD1+ cell number per stage V–VI HF (WT, n = 48; Emb-/-, n = 35). HFs were pooled from 4 WT and 3 Emb-/- mice. In stage V–VI HF, sebocytes can be detected and start forming a sebaceous gland but the SG is not yet localized on the HF posterior wall (Paus et al., 1999).(L) Cryosections of WT and Emb-/- tail skin stained with a lipid marker LipidTOX (green) and ITGA6 (red) with DAPI nuclear counterstain (blue).(M) Quantification of SG lipids (LipidTOX pixels) in WT and Emb-/- SGs (WT, n = 30; Emb-/-, n = 18). SGs were pooled from 6 WT and 4 Emb-/- mice.(N) Quantification of total lipids on mouse skin surface by hexane extraction and TLC (WT, n = 3 mice; Emb-/-, n = 3 mice).Two-tailed t test (C, D, G, I, K, M, and N) or one-tailed t test (E) for independent means was used to determine statistical significance. ***p < 0.0005, **p < 0.005, *p < 0.05. The length of scale bars is 50 µm.

Fig 3: SunCatcher enables functional and phenotypic analysis of individual clones and clonal dynamics.a Functional analysis of Met1 BCs in monoculture and within the BC Pool. Column 1: proliferation rate (doubling time) of each BC over 9 days in culture, ranked from fastest to slowest. Each row represents data for the BC indicated in column 1. Column 2: Met1 BC Pool clonal dynamics in vitro, represented as fold-change (FC) in BC composition after 7 days (d) relative to initial admixture (data values shown in Table 1). Column 3: Met1 tumor clonal dynamics in vivo, represented as fold-change in BC composition at experimental endpoint relative to composition at time of injection (data values shown in Table 1). Column 4: average contribution of each BC as a percent of total BCs in each resulting tumor (n = 10). b Growth of Met1 BC Pool tumors (n = 4) and indicated individual BCs tumors (n = 6 per cohort). c Growth of indicated BC tumors (n = 6 per cohort). d Immunofluorescence images of indicated BCs stained for cytokeratin 8 (CK8, green), cytokeratin 14 (CK14, magenta), f-actin (red) (top), and epithelial cell adhesion molecule (EpCAM, green), zinc finger E-box-binding homeobox 1 (Zeb1) transcription factor (magenta), f-actin (white) (bottom). Nuclei were counterstained with DAPI (blue); Scale bars = 50 µm. Images represent 1 of 4 images per BC from 2 independent observations. e Pearson correlation matrix for individual Met1 BCs analyzed for: proliferation rate, CK8, CK14, Zeb1, EpCAM, programmed death ligand 1 (PD-L1), major histocompatibility complex type I (MHC-I) (associated with Supplementary Fig. 4 and Supplementary Tables 6–8). f Pearson correlation matrix for the 7 phenotypes assessed in vitro: proliferation rate, CK8, CK14, Zeb1, EpCAM, PD-L1, MHC-I (associated with Supplementary Tables 10, 11). Pearson correlation coefficients: CK8 v CK14 *p < 0.0215, 1-tailed, 90% CI; PD-L1 v EpCAM **p = 0.0096, 2-tailed, 90% CI; Zeb1 v EpCAM ***p = 0.0009, 1-tailed, 90% CI; MHC-I v CK14 ****p < 0.0001, 2-tailed, 90% CI. g Principal component (PC) analysis of BCs based on 9 parameters (proliferation, in vitro dynamics, in vivo dynamics, CK8, CK14, Zeb1, EpCAM, PD-L1, MHC-I) stratified by indicated percent composition of Met1 BC Pool tumors at experimental endpoint (associated with Supplementary Tables 9, 10). h Loadings plot from PCA shown in g. Source data are provided as a Source Data file.

Fig 4: ECE20 is a developmental enhancer active in basal ectoderm.(A) Genomic locations of En1 candidate enhancers (ECEs) ECE18 and ECE20 (boxed) in the mouse genome (mm10). En1 Topological associating domain (TAD). (B) Strategy used to generate skin-specific transgenic mice to assess spatial and temporal activity of mouse ECE20 homolog (mECE20) in vivo. In situ hybridization for En1 mRNA (purple) in mouse volar limb skin at P2.5. Representative images of volar skin from the interfootpad (IFP) medial volar skin and footpad (FP) volar skin are shown. Nascent eccrine gland (black arrow). (C) Representative images of control and mECE20 transgenic P2.5 volar and dorsal limb autopods stained with anti-GFP antibody. eGFP is visualized using HRP-DAB-coupled immunohistochemistry so signal is brown. Nascent volar eccrine gland (black arrow). Dorsal hair follicle (head arrow). (D) Strategy to reconstitute transgenic human-derived developing ectoderm in hybrid human-mouse skin patches to test spatial and temporal activity of human ECE20 (hECE20) in vivo. (E) Representative images of a section from human developing skin infected with hECE20 or control virus and stained with antibody against GFP (brown) and also stained with the basal keratinocyte specific marker Keratin 14 (KRT14, Red). eGFP (black double head arrow) is visualized using HRP-DAB coupled immunohistochemistry so positive clones are brown. Primary human basal keratinocyte cells double positive for GFP and KRT14 (white double head arrow). DAPI nuclear stain (blue). See also S1 Fig.

Fig 5: Wnt5a’s role in skin morphogenesis is ancestral in mammals.(A) DAPI-stained transverse sections showing dermal cell density in transgenic mice and controls. Yellow box denotes zoomed-in region. (B) Quantification of cell density in transgenic mice and controls. (C) H&E and KRT14 stained mouse skin from transgenic mice and controls. (D) Quantification of epidermal thickness in transgenic mice and controls. (E) E14.5 mouse showing the early ear pinna. Dotted line shows the rostro-caudal axis along which sections were taken. (F) ROIs used for quantifying dermal cell density and epidermal thickness. (G) Sections stained with DAPI and KRT14. (H and I) Quantification of cell density (H) and epidermal thickness (I) in E14.5 mouse pinna. (J) In situ hybridization for Wnt5a in E14.5 mouse pinna. Statistical significance in (B) and (D) (n = 4; ****P < 0.001) and in (H) and (I) (****P < 0.0001, **P < 0.01, and *P < 0.05; n = 3) was assessed using a general mixed effects model ANOVA test. Post hoc pairwise comparisons indicated that region 4 was significantly different from other regions. All other comparisons were statistically not significant and are not displayed. Dotted lines in (C) and (J) delineate the dermis-epidermis boundary. Scale bars, 200 µm (zoomed out) and 50 µm (zoomed in) (A), 100 µm (C), 200 µm (F), 50 µm (G), and 100 µm (zoomed out) and 50 µm (zoomed in) (J).