Generation of betaPix conditional trap (betaPixct) allele by an homologous recombination (HDR)-mediated knock-in method.

(A) Schematic diagram illustrates the HDR-mediated Zwitch strategy for generating zebrafish knock-in allele at the betaPix locus. (B) Genomic PCR analysis of the F1 embryos confirming the right Zwitch insertions. (C) Sanger sequencing confirming the junction of betaPixct (after HDR-mediated insertion) or betaPixm (after Cre-mediated inversion) that are highlighted by RED lines in (A). (D) Representative stereomicroscopy images of erythrocytes stained with o-dianisidine in betaPixct/ct and betaPixm/m embryos at 48 hpf. Brain hemorrhages, indicated with arrow, in Cre mRNA-injected embryos (betaPixm/m). Lateral views, anterior to the left. (E) Quantification of hemorrhagic parameters in (D). Left panel showing hemorrhage percentages, with independent experiments as dots. Right panel showing hemorrhage areas, with each dot representing one embryo, # represents the numbers of embryos scored for each analysis, and three or more individual experiments conducted. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (F) qRT-PCR analysis showing the expression of betaPix, RFP, and alphaPix in betaPixct/ct, betaPixct/m, and betaPixm/m embryos at 48 hpf. Each dot represents one embryo. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test, individual P values mentioned in the figure. cryaa, α A-crystallin; PA, polyadenylation signal; SA, splice acceptor; T2A, T2A self-cleaving peptide. Individual scale bars indicated in the figure.

betaPixm/m mutant have brain hemorrhages, central artery defects and abnormal glial structure that was partially rescued by Pak1 inhibitor IPA-3 treatment.

(A) Left panel showing the maximum intensity projection of the glial structures in the hindbrain of betaPixct/ct and betaPixm/m embryos at 48 hpf. Lateral view, anterior to left. Middle panel showing representative optical sections and right panel showing the higher magnifications of boxed area, presenting atypical glial structures with disoriented arrangements (yellow arrows) in betaPixm/m embryos. (B) Quantification of glial parameters in (A). Left panel showing the average glia length, and right panel showing glia length index normalized to individual head length, which each dot represents one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (C) Whole-mount RNA in situ hybridization revealed nestin and pax2a expression pattern in betaPixct/ct and betaPixmlm embryos at 48 hpf. Dorsal view, anterior to the left. (D) Optical sections of glial structure (green) and blood vessels (magenta) in the heads of siblings and CRISPR-mediated betaPix F0 knockout embryos. Arteries in the hindbrain of betaPix KO mutants had developmental defects (white arrowheads), showing shorter distance between basilar artery and glial cell bodies. (E) 3D reconstruction of the sox2-positive precursors (green) and vasculatures (magenta) in the heads of siblings and CRISPR-mediated betaPix F0 knockout embryos at 48 hpf. Box areas are shown in higher magnifications at the middle panels, with optical sections shown in the right panels. Arrows indicate CtA with enlarged perivascular space. (F) Representative stereomicroscopy images of o-dianisidine staining of betaPixct/ct and betaPixm/m embryos at 48 hpfthat were treated with DMSO or PAK inhibitor IPA3. Brain hemorrhages indicated with arrows. (G) Quantification of brain hemorrhagic parameters in (F). Left panel showing hemorrhage percentages, with independent experiments as dots. Right panel showing hemorrhage areas, with each dot representing one embryo, # represents the numbers of embryos scored for each analysis, and three or more individual experiments conducted. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test, individual P values mentioned in the figure. (H) Left panels showing 3D reconstruction of the glial structure (green) and vasculature (magenta) in the heads of betaPixct/ct and betaPixm/m embryos at 48 hpftreated with DMSO or IPA3. Lateral view, anterior to left. Box areas are shown in higher magnifications at the middle panels. Defects in hindbrain central arteries indicated in white arrows, while defects in radial glia indicated in yellow arrows. Right panels showing schematic diagrams. Glia (grey) and CtAs (pink) develop normally in DMSO or IPA3-treated betaPixct/ct embryos, with fine radial glial processes and characteristic arch vasculture. Yet in betaPixm/m embryos, abnormal development of the glia and central artery presented. In IPA-3-treated betaPixm/m embryos, central arterial defects were partially rescued with relatively complete arch architecture, and glial processes defects significantly rescued. (I) Quantification of CtA parameters in (H). Left panel showing the average CtA length, and right panel showing the CtA length index normalized to individual head length, with each dot representing one embryo. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test, individual P values mentioned in the figure. (J) Quantification of glia parameters in (H). Left panel showing the average glia length, and right panel showing glia length index normalized to individual head length, which each dot represents one embryo. Individual scale bars indicated in the figure. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test, individual P values mentioned in the figure. Individual scale bars indicated in the figure. BA, basilar artery.

Glial-specific betaPix knockouts recapture global betaPix mutant phenotypes.

(A) Representative stereomicroscopy images of erythrocytes stained with o-dianisidine in betaPixct/ct siblings and gfap:GFP-Cre; betaPixct/ct mutant embryos treated with DMSO or IPA3 at 48 hpf. Brain hemorrhages indicated with arrows in glial-specific betaPix knockouts. Lateral view with anterior to the left. (B) Left panels showing 3D reconstruction of the vasculature (magenta) in the heads at 48 hpf, lateral view with anterior to the left. Box areas are shown in higher magnifications of brain vasculatures at the right panels. CtA defects indicated in yellow arrows in gfap:GFP-Cre; betaPixct/ct mutant embryos. (C) Quantification of brain hemorrhages in (A) and CtA parameters in (B). Left panel showing hemorrhage percentages, with independent experiments as dots. Middle panel showing hemorrhage areas with each dot representing one embryo. # represents the numbers of embryos scored for each analysis, three or more individual experiments conducted. Right panel showing CtA length index normalized to individual head length, with each dot representing one embryo. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test, individual P values mentioned in the figure. Individual scale bars indicated in the figure.

Single cell transcriptome reveals that a subcluster of glial progenitor and stathmin family members are associated with betaPix mutation.

(A) Experimental strategy for single cell RNA sequencing of embryonic heads from wild-type siblings and betaPix CRISPR mutants at 1 dpf and 2 dpf. (B) UMAP visualization and clustering of cells labeled by cell type. Four samples were aggregated and analyzed together. (C) Proportions of 24 cell clusters were differentially distributed among four sample groups. ctrl_ld, PBS-injected siblings at 1 dpf; ko_ld, betaPix CRISPR mutants at 1 dpf; ctrl_2d, PBS-injected siblings at 2 dpf; ko_2d, betaPix CRISPR mutants at 2 dpf. (D) Enriched GO terms for differentially expressed genes for progenitor sub-cluster comparing ko_2d to ctrl_2d groups. (E) Violin plots showing the expression of the stathmin family genes by all cells among four sample groups (left panel) or by progenitor sub-cluster among four sample groups (right panel). (F) qRT-PCR analysis showing expression of betaPix, stmn1a, stmn1b, stmn4l and ppfia3 in glia of FACS-sorted betaPixct/ct siblings and betaPixm/m mutants at 48 hpf. Each dot represents cells sorted from one embryo. Data are presented in mean ± SEM. ***P <0.005; ****P<0.001; unpaired Student’s t test. (G) qRT-PCR analysis showing expression of betaPix, stmn1a, stmn1b, stmn4l and ppfia3 in glia of FACS-sorted betaPixct/ct siblings and gfap:GFP-Cre; betaPixct/ct mutants at 48 hpf. Each dot represents cells sorted from one embryo. Data are presented in mean ± SEM. ***P <0.005; ****P<0.001; unpaired Student’s t test. (H) Whole-mount RNA in situ hybridization revealing down-regulation of stmn1a, stmn1b and stmn4l in betaPixct/ct siblings and gfap:GFP-Cre; betaPixct/ct mutant embryos at 48 hpf. Individual scale bars indicated in the figure.

Stathmin acts downstream of betaPix in glial migration via regulating tubulin polymerization.

(A) Whole-mount RNA in situ hybridization showing that stmn1a, stmn1b and stmn4l expression in gfap:GFP-Cre; betaPixct/ct embryos were partially rescued by Pak1 inhibitor IPA3 treatment at 48 hpf. Dorsal views, anterior to the left. (B) qRT-PCR analysis showing that stmn1a and stmn4l expression were rescued in gfap:GFP-Cre; betaPixct/ct mutants by IPA3 treatment at 48 hpf. Each dot represents one embryo. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test, individual P values mentioned in the figure. (C) Representative stereomicroscopy images of erythrocytes stained with o-dianisidine in siblings and stmn1a/1b/4l CRISPR mutants at 48 hpf. Brain hemorrhages, indicated with arrows, appeared in stmn1a/1b/4l mutants. Lateral views with anterior to the left. (D) Quantification of hemorrhagic parameters in (C). Left panel showing hemorrhage percentages, with independent experiment as dot. Right panel showing hemorrhage areas with each dot representing one embryo. # represents the numbers of embryos scored for each analysis, three or more individual experiments conducted. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (E) 3D reconstruction of glial structure (green) and vasculature (magenta) in the heads at 48 hpf. Lateral view with anterior to the left. CtA defects (white arrows) indicated in stmn1a/1b/4l mutants. (F) Quantification of CtA and glia parameters in (E). Length index normalized to individual head length, with each dot representing one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (G) Representative stereomicroscopy images of erythrocytes stained with o-dianisidine in bbhfn40a and bbhfn40a; Tg(gfap:GFP-stmn1b) embryos at 48 hpf. Brain hemorrhages, indicated with arrows, decreased in bbhfn40a mutants with glia-specific overexpression of stmn1b, compared with bbhfn40a mutant siblings. Lateral views with anterior to the left. (H) Quantification of hemorrhagic parameters of (G). Left panel showing hemorrhage percentages, with independent experiment as dot. Right panel showing hemorrhage areas with each dot representing one embryo. # represents the numbers of embryos scored for each analysis, three or more individual experiments conducted. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (I) 3D reconstruction of the gfap:GFP-stmn1b overexpression (green) and vasculature (magenta) in the heads of bbhfn40a siblings and bbhfn40a; Tg(gfap:GFP-stmn1b) mutants at 48 hpf. Lateral view with anterior to left. White arrows indicate CtA defects. (J) Quantification of CtA and glia parameters in (I). Length index normalized to individual head length, with each dot representing one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (K) Representative stereomicroscopy images of U251 cells at 0 and 24 hours after wounding. U251 cells were transfected with negative control siRNA or betaPIX siRNA separately, in combination with pcDNA3.1 vector, betaPIX overexpression plasmid, and STMN1 overexpression plasmid. The wound edges are highlighted by dashed lines, with arrow lines indicating the wound width. (L) Quantification of wound closure in (K), showing *P<0.05, ***P<0.005 compared to negative control siRNA with empty vector transfection. ##P<0.01, ###P<0.005 compared to betaPix knockdown with empty vector transfection. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test. (M) Representative immunofluorescence image of alpha-tubulin and DAPI signals in U251 cells. U251 cells were transfected with negative control siRNA or betaPIX siRNA separately, in combination with pcDNA3.1 vector, betaPIX overexpression plasmid, and STMN1 overexpression plasmid. Box areas are shown in higher magnifications. Arrows indicate protrusions at the cell periphery. (N) Quantification of cell percentages with protrusions in (M). ***P<0.005 compared to negative control siRNA with empty vector transfection. #P<0.05, ##P<0.01 compared to betaPix knockdown with empty vector transfection. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test. Individual scale bars indicated in the figure.

Zfhx3/4 acts downstream of betaPix to regulate vascular integrity development

(A) Whole-mount RNA in situ hybridization revealing that Vegfaa decreased in bbhfn40a mutants compared with siblings at 36 hpf. Lateral views with anterior to the left. Arrows indicate vegfaa expression in the CtAs that shown depletion in bbhfn40a mutants. (B) qRT-PCR analysis revealing that ZFHX3, ZFHX4 and VEGFA decreased in U251 cells transfected with betaPIX siRNA, which were rescued by betaPIX overexpression. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test, individual P values mentioned in the figure. (C) Whole-mount RNA in situ hybridization showing that vegfaa decreased in CRISPR-mediated zfhx3/4 F0 knockout embryos at 36 hpf. Lateral views with anterior to the left. Arrows indicate vegfaa expression in the CtAs that shown reduction in zfhx3/4 knockouts. (D) Representative stereomicroscopy images of erythrocytes stained with o-dianisidine in siblings and CRISPR-mediated zfhx3/4 F0 knockout embryos at 48 hpf. Arrows indicated brain hemorrhages in the knockout brains. Lateral views with anterior to the left. (E) Quantification of hemorrhagic parameters in (D). Left panel showing hemorrhage percentages, with independent experiment as dot. Right panel showing hemorrhage areas with each dot representing one embryo. # represents the numbers of embryos scored for each analysis, three or more individual experiments conducted. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (F) Left panel showing 3D reconstruction of the glial structure (green) and vasculature (magenta) in the hindbrain of siblings and CRISPR-mediated zfhx3/4 F0 knockouts at 48 hpf. Lateral view with anterior to the left. White arrows indicate CtA defects. Right panels showing schematic diagrams. Glia (grey) developed normally in both treatment, while defects in central artery (pink) presented in zfhx3/4 knockout embryos. (G) Quantification of CtA and glia length parameters of (F). Length index normalized to individual head length, with each dot representing one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (H) Representative stereomicroscopy images of erythrocytes stained with o-dianisidine in CRISPR-mediated betaPix FO knockout embryos with or without zfhx4 mRNA injection at 48 hpf. Arrows indicate brain hemorrhages. Lateral views with anterior to the left. (I) Quantification of hemorrhagic parameters in (H). Left panel showing hemorrhage percentages, with independent experiment as dot. Right panel showing hemorrhage areas with each dot representing one embryo. # represents the numbers of embryos scored for each analysis, three or more individual experiments conducted. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (J) Left panel showing 3D reconstruction of the glial structure (green) and vasculature (magenta) in CRISPR-mediated betaPix F0 knockout embryos with or without zfhx4 mRNA injection at 48 hpf. White arrows indicate CtAs and yellow arrows indicate glia. Right panels showing schematic diagrams. Glia (grey) and CtA (pink) developmental defects rescued in zfhx4 treatment. (K) Quantification of CtA and glia length parameters in (J). Length index normalized to individual head length, with each dot representing one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (L) Whole-mount RNA in situ hybridization revealed that Zfhx4 and Vegfaa decreased in CRISPR-mediated betaPix F0 knockout embryos at 36 hpf, which were rescued by Zfhx4 mRNA injection. Lateral views with anterior to the left. Arrows indicate hindbrain regions. (M) Dot plots of several angiogenesis-associated genes expression in endothelial cell cluster. Dot size indicates the percentage of cells with gene expression, and dot color represents the average gene expression level. Individual scale bars indicated in the figure.

Schematic diagram on the function of glial betaPix in zebrafish vascular integrity development of the hindbrain.

betaPix is enriched in glia and regulates the PAK1-Stathmin axis on microtublin stabilization, thus fine-tuning glial cell migration and their interactions with vanscular endothelial cells; and in paralelle, betaPix may also regulate Zfhx3/4-Vegfaa signaling in glia, which then modulates angiogenesis during cerebral vessel development and maturation. Deletion of betaPix affects glial cell migration and interaction with cerebral endothelial cells. MT, Microtubule.

Generation of betaPix conditional trap (betaPixct) allele by an HDR-mediated knock-in method.

(A) Sanger sequencing results of the original pZwitch (+3) and the modified vector with glycine-serine-glycine spacer (GSG) in front of the 2A peptide. (B) Sanger sequencing results of the genomic PCR products at both junction sites from the betaPixct F1 offspring showing the correct knock-in in the betaPix locus. (C) The screening process for betaPixct knock-in zebrafish showing about 1% efficiency with correct homologous recombinations. (D) Representative stereomicroscopy images of o-dianisidine staining in bbhfn40a embryos at 48 hpf. Lateral view, anterior to the left in left panel; Dorsal view, anterior to the top in right panel. (E) Quantification of hemorrhagic parameters in (D). Left panel showing hemorrhage percentages, with independent experiment as dot. Right panel showing hemorrhage areas with each dot representing one embryo. # represents the numbers of embryos scored for each analysis, three or more individual experiments conducted. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (F) 3D reconstruction of the betaPix expression (magenta) in betaPixct/+ brains of embryos microinjected with PBS or Cre mRNA at 24 hpf, 48 hpf, 72 hpf and 144 hpf. Cre-induced RFP expression indicates successful inversion of gene trap system (arrows). Dorsal view, anterior to the top. (G) 3D reconstruction of the neurod1 transgenic expression (green) and betaPix expression (magenta) in betaPixcl/+ brains of embryos microinjected with PBS or Cre mRNA at 24 hpf, 48 hpf, 72 hpf and 144 hpf. Cre-induced RFP expression indicates successful inversion of gene trap system (arrows), marking strong betaPix expression at the midbrain hindbrain boundary (MHB), and milder expression at hindbrain at early stages. At 144 hpf, betaPix expresses ubiquitously in the brain, outlining the cerebellum (arrowheads). Dorsal view, anterior to the top. Individual scale bars indicated in the figure.

BetaPixm/m mutant had brain hemorrhages, central arteries defects, and abnormal glial structure.

(A) 3D reconstruction of the vasculature (green) in the heads with lateral view, anterior to left. Endogenous betaPix expression (magenta) only shown in Cre mRNA-injected mutant embryos. Box areas are shown in higher magnifications of vasculatures at the right panels. Central arteries (CtA) defects (white arrows) were evident in betaPixm/m embryos at 48 hpf. Arrow lines indicating the CtA length. (B) Quantification of CtA parameters in (A). Left panel showing the average CtA length, and right panel showing the CtA length index normalized to individual head length, with each dot representing one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (C) 3D reconstruction of the glia structure. Box areas showing higher magnifications of glia at the right panels. D, dorsal; V, ventral. Arrow lines indicating the glia length. Individual scale bars indicated in the figure.

CRISPR-mediated betaPix F0 knockouts had similar phenotypes as betaPixm/m mutants.

(A) Schematic diagram illustrates guide RNA sites for CRISPR-mediated betaPix F0 knockout. E, exon. (B) qRT-PCR revealing that betaPix decreased in CRJSPR-mediated betaPix F0 knockout embryos compared with wild-type siblings at 48 hpf. Each dot represents one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (C) Maximal intensity projection of the glial structure of the hindbrains in siblings and CRISPR-mediated betaPix F0 knockouts at 48 hpf, lateral view with anterior to left. (D) Whole-mount RNA in situ hybridization revealing up-regulated nestin and down-regulated pax2a expression in CRISPR-mediated betaPix F0 knockout embryos at 48 hpf. Dorsal view with the anterior to the left. (E) Quantification of hemorrhagic parameters of siblings and CRISPR-mediated betaPix F0 knockout embryos at 48 hpf. Left panel showing hemorrhage percentages, with independent experiment as dot. Right panel showing hemorrhage areas with each dot representing one embryo. # represents the numbers of embryos scored for each analysis, three or more individual experiments conducted. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (F) Quantification of CtA parameters of siblings and CRISPR-mediated betaPix Fo knockout embryos at 48 hpf. Left panel showing average CtA length, andright panel showing CtA length index normalized to individual head length, with each dot representing one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (G) Quantification of glial parameters of siblings and CRISPR-mediated betaPix Fo knockout embryos at 48 hpf. Left panel showing average glia length, and right panel showing glia length index normalized to individual head length, with each dot representing one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. Individual scale bars indicated in the figure.

Generating glial-specific betaPix knockout zebrafish.

(A) Schematic diagram illustrates establishment of glial-specific betaPix knockout zebrafish. (B) 3D reconstruction of the heads at 48 hpf with lateral view, anterior to left. Glial-specific Cre expression (green) was only shown in gfap:GFP-Cre; betaPixct/ct mutant embryos, which overlaps with betaPix:RFP expression (magenta). (C) Genomic PCR analysis of the betaPixct or betaPixm-unique sequences in betaPixct/ct siblings, gfap:GFP-Cre; betaPixct/+heterozygous mutants, and gfap:GFP-Cre; betaPixct/ct homozygous mutants. (D) qRT-PCR analysis revealing Cre expression in gfap:GFP-positive glial population and gfap:GFP-negative non-glial population that were sorted from gfap:GFP-Cre; betaPixct/ct embryos. Each dot represents one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (E) qRT-PCR analysis showing betaPix, betaPix-RFP and alphaPix expression in betaPixct/ct control siblings and gfap:GFP-Cre; betaPixct/ct mutant embryos at 48 hpf. Each dot represents one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (F) Whole-mount RNA in situ hybridization revealing that alphaPix was comparable but betaPixct/ct decreased in gfap:GFP-Cre; betaPixct/ct mutants compared with betaPixct/ct siblings at 48 hpf. Lateral view with anterior to the left. Individual scale bars indicated in the figure.

Neither endothelial-, neuronal- nor mural-specific deletion of betaPix caused brain hemorrhages and abnormal CtA phenotypes.

(A) Representative stereomicroscopy images of o-dianisidine staining in betaPixct/ct, kdrl:Cre; betaPixct/ct, huC:GFP-Cre; betaPixct/ct, and acta2:GFP-Cre; betaPixct/ct embryos at 48 hpf. (B) Quantification of hemorrhage parameters in (A). Up panel showing hemorrhage percentages, with independent experiment as dot. Down panel showing hemorrhage areas with each dot representing one embryo. # represents the numbers of embryos scored for each analysis, three or more individual experiments conducted. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (C) 3D reconstruction of the vasculature of betaPixct/ct, kdrl:Cre; betaPixct/ct, huC:GFP-Cre; betaPixct/ct, and acta2:GFP-Cre; betaPixct/ctembryos at 48 hpf. Lateral view with anterior to left. Boxed areas of the hindbrains are shown in higher magnifications at the right panels. (D) Quantification of CtA parameters in (C). Up panel showing average CtA length, and down panel showing CtA length index normalized to individual head length, with each dot representing one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. Individual scale bars indicated in the figure.

Glial progenitor sub-cluster of betaPix knockouts has down-regulation of stathmin family members and up-regulation of pak1 gene.

(A) Dot plot of marker genes expression in each cell type. Dot size indicates the percentage of cells expressed, and dot color represents the average expression level. Box area and arrow highlighting gfap expressions. Clusters with high gfap levels including neuronal and glial progenitors, hindbrain, ventral diencephalon, ventral midbrain and floor plate. (B) UMAP visualization and clustering of cells labeled by cell type across four sample groups. (C) UMAP feature plots displaying relative expression levels of selected transcripts among four sample groups. Cells are colored by expression level.

CRISPR-mediated stmn1a/1b/4l F0 knockouts have similar but milder phenotypes as betaPix knockouts.

(A) qRT-PCR analysis showing that IPA3 treatment rescued stmn1a and stmn4l expression in CRISPR-mediated betaPix F0 knockouts at 48 hpf. Each dot represents one embryo. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test, individual P values mentioned in the figure. (B) Schematic diagram illustrates 4-guide RNAs mediated F0 knockout strategy in the loci of stmn1a, stmn1b and stmn4l. (C) Whole-mount RNA in situ hybridization confirming that stmn1a, stmn1b and stmn4l decreased in CRISPR-mediated stmn1a/lb/41 F0 knockouts at 48 hpf. Dorsal views with anterior to the left. (D) qRT-PCR analysis confirtning that stmn1a, stmn1b and stmn4l decreased in mutant embryos of CRISPR-mediated stmn1a/lb/4l F0 knockouts at 48 hpf. Each dot represents one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (E) qRT-PCR analysis confirming that stmn1a, stmn1b and stmn4l decreased in the FACS-sorted glia of CRISPR-mediated stmn1a/lb/4l F0 knockouts at 48 hpf. Each dot represents sorted cells from one embryo. Data are presented inmean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (F) Optical sections of glial structure in the hindbrains of siblings and CRISPR-mediated stmn1a/1b/4l F0 knockouts at 48 hpf. Lateral view with anterior to the left. Abnormal glial structures with disoriented arrangements (yellow arrows) in stmn1a/1b/4l mutant embryos. (G) Quantification of average CtA and glial length in Figure 5 (E). Each dot represents one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (H) Whole-mount RNA in situ hybridization revealed that nestin increased while pax2a decreased in CRISPR-mediated stmn1a/1b/4l F0 knockout embryos at 48 hpf. Dorsal views with anterior to the left. (I) 3D reconstruction and whole-mount RNA in situ hybridization showing that gfap:GFP-stmn1b transgenic overexpression rescued glial defects in bbhfn40a; Tg(gfap:GFP-stmn1b) embryos at 48 hpf. Lateral views with anterior to the left. (J) Whole-mount RNA in situ hybridization showing that nestin and pax2a were normally expressed in bbhfn40a; Tg(gfap:GFP-stmn1b) embryos at 48 hpf. Dorsal views with anterior to the left. (K) Quantification of average CtA and glia length in Figure 5 (I). Each dot represents one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (L) qRT-PCR analysis confirming the efficacy of betaPix transgenic over-expression and betaPix siRNA knockdown in U251 cells. Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test, individual P values mentioned in the figure. (M) Representative stereomicroscopy images of A172 cells at 0 and 18 hours after wounding. betaPIX siRNA treatment decreased A172 cell migration, which were rescued by either betaPIX or STMNI overexpression. The wound edges are highlighted by dashed lines, with arrow lines indicating the wound width. (N) Quantification of wound closures in (M). Data are presented in mean ± SEM; one-way ANOVA analysis with Dunnett’s test. **P<0.01 compared to negative control siRNA and empty vector transfection. ##P<0.01, ###P<0.005 compared to betaPix knockdown and empty vector transfecti on. Individual scale bars indicated in the figure.

Zfhx3/4 acts downstream of betaPix to regulate vascular integrity development

(A) Violin plots showing that Zfhx3 and Zfhx4 decreased in progenitor sub-cluster of betaPix knockouts. (B) qRT-PCR analysis revealing that Zfhx3 and Zfhx4 decreased in FACS-sorted glia of gfap:GFP-Cre; betaPixct/ct mutants compared with betaPixct/ct siblings at 48 hpf. Each dot represents the cells sorted from one embryo. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (C) Schematic diagram illustrates 4-guide RNAs mediated F0 knockout strategy in the Zfhx3 and Zfhx4 loci. (D) Whole-mount RNA in situ hybridization confirming Zfhx3/4 efficiency in CRISPR-mediated Zfhx3/4 F0 knockout embryos at 48 hpf. Arrows indicate significant reduction of Zfhx3 and Zfhx4 levels in hindbrain by Zfhx3/4 double knockout. (E) qRT-PCR analysis revealing that Zfhx3, Zfhx4, and Vegfaa decreased in CRISPR-mediated Zfhx3/4 F0 knockout embryos at 36 hpf. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (F) Whole-mount RNA in situ hybridization revealing that nestin increased in CRISPR-mediated Zfhx3/4 F0 knockout embryos at 36 hpf. Arrows indicate the hindbrain regions. (G) qRT-PCR analysis confirming overexpression of Zfhx4 in CRISPR-mediated betaPix F0 knockout embryos with Zfhx4 mRNA injection at 70% epiboly. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. (H) Representative stereomicroscopy images of erythrocytes stained with a-dianisidine showing that brain hemorrhages (arrows) decreased in bbhfn40a mutants with Zfhx4 mRNA injection at 48 hpf. Lateral views with anterior to the left. (I) Quantification of hemorrhagic parameters in (H). Left panel showing hemorrhage percentages, with independent experiment as dot. Right panel showing hemorrhage areas with each dot representing one embryo. # represents the numbers of embryos scored for each analysis, three or more individual experiments conducted. Data are presented in mean ± SEM; unpaired Student’s t test with individual P values mentioned in the figure. Individual scale bars indicated in the figure.

Primers used for siRNA

Primers used for guide RNA synthesis

Primers used for WISH probe synthesis

Primers used for qRT-PCR