Ma, Quigley, Omran, and Kintner

Supplementary Figure legends

Supplementary Figure 1 (related to Figure 1): Interaction of endogenous Multicilin and E2f4.

Embryos were injected at the 2-4 cell stage with RNA encoding activated Notch, ICD, or with a DNA binding mutant of Su(H) to inhibit Notch signaling in order to increase or repress, respectively, the expression of Multicilin and MMC differentiation (Stubbs et al. 2012). Protein extracts were prepared from 30 animal caps that were isolated at stage 10, cultured to stage 12 when MCC precursors form. Endogenous Multicilin was pulled down using a rabbit anti-Multicilin antibody, resolved by PAGE, Western blotted, and then probed with the same antibody to Multicilin or to an antibody that recognizes Xenopus E2f4.

Supplementary Figure 2 (related to Figure 1): Domain mapping of the EDM complex.

(A) The domain structure of Multicilin and deletion mutants analyzed. CC (coiled-coil domain) and TIRT (45 amino acid, c-terminal domain) are denoted. Extracts were prepared from animal caps isolated from embryos injected at 2-4 cell stage with RNAs as indicated and subjected to Western blot analysis prior (input) or after immunoprecipitation (IP), using the indicated antibody. Nonspecific bands in lanes 1, 3 and 8 are marked with an asterisk. (B) The domain structure for Dp1 and deletion mutations analyzed. DNA binding domain (DBD) and Dimerization Domain (DD) are denoted. Extracts were prepared from animal caps isolated from embryos injected at 2-4 cell stage with RNAs as indicated and subjected to Western blot analysis prior (input) or after immunoprecipitation (IP), using the indicated antibody. (C) Domain structure for E2f4 and deletion mutations used to map regions required for complex formation with Multicilin. DNA binding domain (DBD), dimerization domain (DD) and c-terminal domain (CT) are denoted. Extracts were prepared from animal caps isolated from embryos injected at 2-4 cell stage with RNAs encoding flag-Multicilin, Dp1, and different GFP-tagged forms of E2f4, as indicated. Upper panels show a Western blot analysis of protein extracts, probed with an α-FLAG or an α-GFP antibody, while the lower panels show a Western blot analysis of proteins after pull down using an α-FLAG antibody, and probed with either an α-FLAG or α-GFP antibodies.

Supplementary Figure 3 (Related to Fig. 2): Centriole duplication and PSC differentiation in embryos injected with RNA encoding E2f4 or E2f4ΔCT

(A) Centriole number was scored in outer cells based on doublets, split singlets, or a singlet as illustrated in Fig. 2a. No outer cells were identified with more than two centrioles. The number of cells with only one centriole is likely overestimated since doublets can be mistaken for singlets at certain angles along the Z-axis and since singlets in a split pair can lie deep in the cytoplasm, prior to mitotis, where they are difficult to image. Data is based on 100-150 cells from 6 embryos (B-F) Embryos were injected with RNA encoding mGFP (green) as a control or with RNA encoding E2f4, or E2f4ΔCT, fixed at stage 28 and stained with an antibody directed against Atpb1/2 (red) and ae1 (blue) to identify α-PSCs (apical Atpb1/b2+ (C) and basal ae1+ (E)) or β-PSCs (basal Atpb1/b2+ (D) and ae1- (E). Graph (B) summarizing the average number of different cells types per field based on Atpb1/b2 and ae1 staining to identify α-PSCs and β-PSCs, and cell morphology based on mGFP labeling of cell boundaries and cilia for multiciliate cells (MCC), and outer cells (OCs). mGFP staining labels cells with multiple cilia but is not sufficient to score cells projected a few cilia, as in Fig. 2. Data is based on sixteen fields (196 microns2) taken from eight embryos. Experimental values significantly different from control values based on a two-tailed t-test (p<0.05) are marked with asterisks.

Supplementary Figure 4 (related to Fig. 3) Gmnn misexpression inhibits MCC differentiation.

(A-B) Embryos were injected with RNAs encoding mRFP (red) and Hyls1-GFP (green) to label the plasma membrane and centrioles/basal bodies, respectively, along with RNA encoding Gmnn (B). Shown are representative confocal images of the skin of a control (A) and Gmnn (B) RNA injected embryos. (C) Different cell types in the skin were scored based on large outer cells morphology (OCs), small apical domain cells (INCs) mostly likely PSCs (Supplementary Figure 3), multiciliate cells (MCCs) or cells projecting one or two cilia (panel b, arrow). Data is based on sixteen fields (196 microns2) taken from eight embryos. (D) Average number of basal bodies in MCCs in control and Gmnn injected embryos based on 16 MCCs from eight embryos. (E) Plot summarizing the average surface area of MCCs in control and Gmnn samples based on 16 cells from eight embryos. Experimental values significantly different from control values based on a two-tailed t-test (p<0.05) are marked with asterisks.

Supplementary Figure 5: Validation of E2f4 ChIPseq.

(A) Proportions of the types of genomic features occupied by E2f4 in mouse CH12 cells (melanoma, ENCODE ref), in Xenopus animal caps, and in Xenopus animal caps in the presence of Multicilin. The Xenopus laevis v7.1 genome build is in hundreds of thousands of scaffolds; "NA" denotes peaks mapped to scaffolds, generally small ones, with no annotated features. (B-G) The two most enriched sequence motifs in ChIPseq peaks along with p-value for mouse (B-C), Xenopus (D-E) and Xenopus in the presence of Multicilin (F-G). Histograms measure density of that motif in ChIPseq peaks (motif per bp per peak) from each experiment, with highest motif density in peak centers. (H) E2f4 peaks in promoters of genes in Xenopus animal caps and mouse CH12 cells; overlap shows E2f4 peaks found in the promoters of orthologs of both species. (I) Overlap of E2f4 peaks between Xenopus animal caps, mouse CH12 cells, and KEGG pathway cell cycle genes (Supplementary Table 2). (J) Similar overlap but with centriole genes as in (Supplementary Table 3).

Supplementary Figure 6: (related to Figure 5): Binding of E2f4 to centriolar and cell cycle genes

(A-D) Genome browser screen shots of E2f4 +/- Multicilin CHiPseq and ICD +/- Multicilin RNAseq, here examining the centriole genes (A) plk4 and (B) deup1/ccdc67, with closeups on their respective promoters (A’,B’). Also shown are the cell cycle genes (C) cdc6 and (D) pcna. Reads were normalized to a fixed library size to facilitate comparison.

Supplementary Figure 7 (related to Fig. 6) E2f4 acts as an activator to drive basal body assembly during MCC differentiation.

(A) Diagram showing the different forms of E2f4 used here. (B-E) Animal blastomeres of 2-4 cell stage embryos were injected with RNAs encoding mRFP (red) and Hyls1-GFP (green) to label the plasma membrane and centrioles/basal bodies, respectively, along with RNA encoding E2f4ΔCT (C) or E2f4ΔCT-VP16 (D). Shown are representative confocal images taken of the skin after embryos were fixed, and stained with the acetylated tubulin antibody to label cilia (blue). (E) Data based on twenty fields (98 microns2) from ten embryos are plotted to summarize average number of outer cells (OCs), INCs (small apical domain), multiciliate cells or ciliated cells (CCs). Plot summarizing the average number of basal bodies in MCCs. (F) Plot showing average basal body number that form in control, E2f4 or E2f4ΔCT-VP16 injected embryos, based on twenty MCCs from ten embryos (G) Centriole number was scored in outer cells based on doublets, split singlets, or a singlet as marked with Hyls1-GFP expression, as in Supplementary Figure 3A. Asterisks mark values that differ significantly from controls based on two-tailed t-test (p<0.05).

Supplementary Figure 8 (related to Fig. 6): E2f4CT-VP16 fails to rescue mutants of Multicilin that cannot form an EDM complex.

(A-D) Animal blastomeres of 2-4 cell stage embryos were injected with RNAs encoding mRFP (red) and Chibby-GFP (green) to label the plasma membrane and basal bodies, respectively, along with RNA encoding Multicilin R270H, MulticilinΔTIRT, E2f4, or E2f4ΔCT-VP16, as indicated. Shown are representative confocal images taken of the skin after embryos were fixed, and stained with the acetylated tubulin antibody to label cilia (blue). (E) Percent MCCs in the outer epithelium, based on basal body number and acetylated tubulin staining, summarizing data from ten fields from five embryos. (F) Average outer cell size based on twenty cells from ten embryos. Note that MulticilinR270H increases cell size, indicative of reduced cell division, as shown previously for MulticilinTIRT (Stubbs et al. 2012). (G-I) Animal blastomeres of 2-4 cell stage embryos were injected with RNAs encoding mRFP (red) and Chibby-GFP (green) to label the plasma membrane and basal bodies, respectively, along with RNA encoding Multicilin RG355D, and E2f4ΔCT-VP16, as indicated. Shown are representative confocal images taken of the skin after embryos were fixation, and staining with the acetylated tubulin antibody to label cilia (blue). (J) The percentage of cells in the outer epithelium that are multiciliated, based on basal body number and acetylated tubulin staining, summarizing data from ten fields from five embryos. (K) Basal body number based on twenty cells from five embryos. Note that E2f4ΔCT-VP16 consistently increases basal body formation. (L) Average outer cell size based on twenty cells from ten embryos. Note that Multicilin with a disabled TIRT domain consistently increases cell size, indicative of reduced cell division (Stubbs et al. 2012). (M) Basal body number based on twenty MCCs from five embryos, using embryos shown in Figure 6B-E. Note that E2f4ΔCT-VP16 consistently increases basal body number. All values significantly different from controls (p<0.05) based on a two-tailed t-test are marked with asterisks.

Supplementary Figure 9: Quantitative analysis of RNA expression in animal caps expressing ICD along with Multicilin, MulticilinR370H, and/or E2f4ΔCT-VP16.

Embryos were injected at the two-four cells stage with RNAs encoding ICD, to block MCC differentiation, Multicilin, MulticilinR370H, and/or E2f4ΔCT-VP16. Animal caps were isolated at stage 10, cultured for 9 hours and then extracted for RNA. RNA samples were assayed in triplicate for the expression of the indicated gene using quantitative RT-PCR. RNA levels are shown after normalization to ubiquitously expressed odc RNA, and set relative to a value of 1 for ICD controls.

Supplementary Figure 10: Western blots used in Figs. 1 and 6, and Supplementary Fig. 2

Supplementary Tables

Table 1: RNAseq analysis of skin progenitors expressing Multicilin-HGR alone (B) or in the presence of E2f4CT (A).

Table 2: Significant changes in centriole gene expression (p<0.05) based on RNAseq analysis of skin progenitors expressing Multicilin-HGR alone (B) or in the presence of E2f4CT (A).

Table 3: Significant changes in cell cycle gene expression (p<0.05) based on RNAseq analysis of skin progenitors expressing Multicilin-HGR alone or in the presence of E2f4CT.

Table 4: Position of ChIP seq peaks that map near promoters and increase 1.5 fold (p<0.05) obtained with E2f4-GFP in the presence of Multicilin versus E2f4-GFP alone.

Table 5: Position of ChIP seq peaks that map near promoters and increase 1.5 fold (p<0.05) obtained with E2f4-GFP alone versus in the presence of Multicilin.