Phylogenetic analysis and embryonic expression of Stag paralogues. (A) Phylogenetic analysis of predicted protein sequences using the maximum likelihood approach. The accession numbers for the protein sequences used in this analysis are listed in Supplementary Table 1. (B) Whole-mount in situ hybridization of stag1a, stag1b, stag2a, and stag2b during early embryogenesis. Lateral views are shown, anterior to the left. Scale bars are 50 μm for embryos at 12 and 15 hpf and 100 μm for embryos at 24 hpf. (C,D) mRNA expression of stag paralogues at the indicated time points during (C) early embryogenesis and (D) late embryogenesis. Each data point represents mRNA isolated from a pool of 30 embryos. Graphs are mean +/- one standard deviation. Expression was normalized to the reference genes, b-actin and rpl13a (Supplementary Figure 1A).

Generation of zebrafish stag germline mutants. CRISPR-Cas9 genome editing was used to generate germline mutations in stag1a, stag1b, and stag2b. (A) Exon diagrams of the respective paralogues showing details of the editing strategy. sgRNA binding sites are marked by red arrowheads with the type of mutation generated indicated above. (B) mRNA levels of the four paralogues in each of the mutant lines indicated on the x-axis. Each data point represents mRNA isolated from a pool of 30 embryos. All graphs are mean +/- one standard deviation. *P ≤ 0.05, **P ≤ 0.01, and ****P ≤ 0.0001; one-way ANOVA. Expression was normalized to the reference gene, b-actin (Supplementary Figure 1C) (C)stag2b^nz207 mutants have displaced pigment cells in the tail fin, zoom-ins are shown in insets. Mutants also show mild developmental delay with late swim bladder inflation as indicated by the black arrow. Scale bars are 200 μm.

Stag mutations alter the number of gata1- and spi1-positive cells in 24 hpf zebrafish embryos. (A) Lateral views of gata1 expression in whole-mount embryos at 24 hpf; anterior to the left. gata1 expression is reduced in stag1a^nz204+/- and stag1a^nz204-/- embryos and is rescued upon injection of functional stag1a mRNA. gata1 expression is reduced in stag2b^nz207+/- and stag2b^nz207-/- embryos and is rescued upon injection of functional stag2b mRNA. Reduced expression is indicated by arrows. (B)spi1 expression in whole-mount embryos at 24 hpf. Left panels show lateral views and right panels show ventral views; anterior to the left. The number of spi1-positive cells is increased in stag1a^nz204+/- and stag1a^nz204-/- embryos (yellow arrows). spi1 expression in stag2b^nz207 heterozygous embryos and stag2b^nz207 homozygous mutant embryos is comparable to wildtype. Scale bars are 100 μm. The number of embryos is indicated in lower-right-hand corners. (C) mRNA levels of haematopoietic stem cell marker cmyb and erythroid or myeloid lineage markers in stag1a^nz204 and stag2b^nz207 homozygous mutant embryos at 48 hpf. The bar graph shows the mean +/- one standard deviation. *P ≤ 0.05, **P ≤ 0.01; one-way ANOVA. Expression was normalized to the reference gene, b-actin.

stag1a and stag2b mutations alter the number of scl-positive cells in the posterior lateral mesoderm at 15 hpf. (A)runx1 expression in whole-mount embryos at 15 hpf. Posterior views of the PLM are shown; dorsal to the top. In stag1a^nz204 homozygous mutant embryos, runx1 expression is slightly increased. In stag2b^nz207 homozygous mutant embryos, runx1 expression is slightly reduced. Changes in expression are marked by arrows and the number of embryos is indicated below each panel. (B)scl expression in whole-mount embryos at 15 hpf. Posterior views of the PLM are shown; dorsal to the top. In stag1a^nz204 homozygous mutant embryos, expanded expression of scl laterally into the PLM is dampened upon injection of functional stag1a mRNA. In stag2b^nz207 homozygous mutant embryos, scl expression is reduced in the anterior PLM and is rescued upon injection of functional stag2b mRNA. Changes in expression are marked by arrows and the number of embryos is indicated below each panel. (C) mRNA levels of mesoderm-derived haematopoietic and endothelial markers at 15 hpf. The bar graph shows the mean +/- one standard deviation. *P ≤ 0.05, **P ≤ 0.01; one-way ANOVA. Expression was normalized to the reference genes, b-actin and rpl13a (Supplementary Figure 1B).

stag1a and stag2b mutations affect cell identity in the posterior lateral mesoderm at 12 hpf. (A)scl expression in whole-mount embryos at 12 hpf, posterior views of the PLM; dorsal to the top. In stag1a^nz204 homozygous mutant embryos scl expression is expanded. In stag2b^nz207 homozygous mutant embryos, scl expression is reduced. Changes in expression are marked by arrows and the number of embryos is indicated below each panel. (B)pax2 expression in whole-mount embryos at 12 hpf, posterior views of the PLM; dorsal to the top. In stag1a^nz204 homozygous mutant embryos, pax2 expression in the PLM is markedly reduced. In stag2b^nz207 homozygous mutant embryos, pax2 expression is comparable to wild type. Changes in expression are marked by arrows and the number of embryos is indicated below each panel. (C)pax2 expression in whole-mount embryos at 12 hpf, lateral views of the head region; anterior to the left. Anterior pax2 expression is specifically reduced in the optic stalk of stag2b^nz207 homozygous mutant embryos. (D) Multiplexed in situ HCR of scl (Alexa Fluor 488, false color yellow), gata1 (Alexa Fluor 594, false color red), and fli1 (Alexa Fluor 647, false color blue) expression at 15 hpf. High magnification maximum intensity projections of a single PLM stripe; posterior views with anterior to the left. Expression domains of scl broadly overlap gata1 and fli1 in all embryos. Ectopic scl expression, indicated by white arrow, in stag1a^nz204 homozygous mutant embryos does not overlap gata1 or fli1 expression domains. In stag2b^nz207 homozygous mutant embryos, expression of all three markers is reduced. Scale bars are 10 μm. The number of embryos analyzed is indicated below the respective panels.

stag1a and stag2b mutations differentially alter the production of primitive myeloid cells in the anterior lateral mesoderm at 12 hpf. (A)scl expression in whole-mount embryos at 12 hpf. Ventral views of ALM are shown; dorsal to the top. scl expression is comparable to wildtype in stag1a^nz204 homozygous mutant embryos and reduced in stag2b^nz207 homozygous mutant embryos. (B)scl expression in whole-mount embryos at 15 hpf. Top panels show lateral views and bottom panels show ventral views of the ALM. Expanded scl expression in the ALM of stag1a^nz204 mutants is rescued upon injection of functional stag1a mRNA. The reduced scl expression in the ALM of stag2b^nz207 mutants is rescued upon injection of functional stag2b mRNA. Changes in expression are marked by arrows and the number of embryos is indicated below each panel. (C) Multiplexed in situ HCR of runx1 and spi1 at 15 hpf. Maximum intensity projections of a single ALM stripe; lateral views with dorsal to the top. Expression domains of runx1 (Alexa Fluor 647) and spi1 (Alexa Fluor 514, false color blue) broadly overlap in all embryos. Both runx1 and spi1 are expanded in stag1a^nz204 embryos but reduced in stag2b^nz207 embryos. Changes in expression are indicated by white arrows. Scale bars are 10 μm. The number of embryos analyzed is indicated below the respective panels.

Hypothetical model explaining the effects of Stag1a and Stag2b loss on primitive erythropoiesis. In stag1a^nz204 mutants, an expansion of early haematopoietic progenitors driven by increased scl expression may lead to precocious differentiation that exhausts the progenitor pool. In stag2b^nz207 mutants, a limited pool of haematopoietic progenitors resulting from reduced scl expression leads to a reduction in primitive erythropoiesis.

Unillustrated author statements