Fig 2: Splicing factor SRSF2 modulates NUMB exon 9 splicing AThe putative SRSF2 binding motifs on NUMB exon 9 (labeled as red), predicted by RBPmap. M1 refers to motif 1 and M2 refers to motif 2.BWestern blotting showing the overexpression of SRSF2 in 293T cells, which was co‐transfected with NUMB reporters of WT, ΔM1, or ΔM2. ΔM1 or ΔM2 represents plasmid deleting corresponding binding site.CA minigene splicing reporter assay showing SRSF2 promoted the generation of NUMB_L isoform via M1 but not M2. The upper panel indicating the pET01 vector that contains two constitutive NUMB exons. The representative RT–PCR electropherogram showing the expression of NUMB_L and NUMB_S isoforms in WT, ΔM1, or ΔM2 transfected 293T cells with or without SRSF2 overexpression, respectively.DThe bar graph showing the normalized exon 9 inclusion from the RT–PCR electropherogram of (C) to the WT control cells. P‐values were determined by an unpaired two‐tailed Student’s t‐test.EOn day 5 of hematopoietic differentiation, the inclusion or exclusion of NUMB exon 9 were detected by RT–PCR with 1.25 nM PLB supplementation or SRSF2 exogenous expression alone or their combinational treatment from day 2.5. The bar graph showing the signaling intensity of each specific band from the RT–PCR electropherogram.F, GThe top panel depicts the location of specific primers targeting NUMB_S and NUMB_L isoform. The bottom bar graph showing the expression of NUMB_S and NUMB_L in cells on day 5 of differentiation with 1.25 nM PLB supplementation or SRSF2 exogenous expression alone or their combinational treatment from day 2.5 with RT–qPCR. Data information: Results given are mean ± SD. P‐values were calculated by one‐way ANOVA followed by Dunnett’s test in (E), (F), and (G). *P < 0.05, **P < 0.01, ***P < 0.001. All experiments were conducted for at least 3 biological replicates. Source data are available online for this figure.

Fig 3: Disruption of the splicing factor switch impacts EPC specification The schematic depicting experimental strategy for RNA‐Seq. APLNR+ cells on day 2 were FACS‐purified and treated with 1.25 nM or 2.5 nM PLB from day 2.5 to 5 during hematopoietic differentiation. Then, the day 2‐purified APLNR+ cells, day 5‐differentiated cells with or without PLB treatment were harvested for RNA‐Seq. n = 2 technical replicates.The violin plot represents distributions of statistically significant ΔPSI for different types of splicing events by comparison of PLB‐treated cells vs DMSO control. The number of differential splicing events regulated by 2.5 nM PLB is denoted in brackets. we analyzed the differential splicing events with genes FPKM > 1 in at least one differentiation stage. Among them, the cutoffs of PSI > 0.2 and FDR < 0.05 were used to define differential splicing events. The dotted line indicates ΔPSI = 0.The heatmap shows the expression of differentially expressed splicing factors in APLNR+ cells, day 5‐differentiated cells treated with DMSO or 2.5 nM PLB. The differentially expressed splicing factors were defined as fold change of FPKM > 1.5 between APLNR+ cells and DMSO‐treated cells. The heatmap was scaled with Z‐Score using the log2(FPKM) expression of differentially expressed splicing factors. n = 2 technical replicates.The Venn diagram presents the intersection between highly expressed genes in APLNR+ cells (APLNR+ cells vs. DMSO‐treated cells; FC > 1.5) and highly expressed genes in PLB on day 5 (PLB‐treated cells vs DMSO; FC > 1.5). The STRING diagram (middle) shows the protein interaction networks among the 38 candidates. The heatmap on the right illustrates the fold change of 7 splicing regulators between APLNR+ and DMSO‐treated cells.RT–qPCR showing the mRNA level of SRSF2 in APLNR+ cells and day 5‐differentiated cells without or with 2.5 nM PLB treatment. The SRSF2 expression was normalized to that in APLNR+ cells. The ACTB gene was used as a control. P‐values were determined by an unpaired two‐tailed Student’s t‐test.Schematic of experimental design. During hematopoietic differentiation, stable cell lines harboring SRSF2 expression vector and its corresponding empty vector control were treated with DOX alone or in combination with 1.25 nM PLB from day 2.5 to 5. The cells were harvested for analysis on day 5 of differentiation.The Western blotting illustrates the overexpression of SRSF2 in sorted GFP+ cells on day 5 during differentiation, which were induced with DOX from day 2.5 to 5. GAPDH was used as the loading control. The bar graph presents the quantification of SRSF2 protein expression. The P‐value was determined by an unpaired two‐tailed Student’s t‐test.The representative FACS plots show the frequency of CD31+CD34+ EPCs in the GFP+ population at day 5 with 1.25 nM PLB supplementation or SRSF2 exogenous expression alone or their combinational treatment from day 2.5, as described in (G).The bar graph represents the percentage of CD31+CD34+ cells from (H). P‐values were calculated by one‐way ANOVA followed by Tukey’s test.The Western blotting shows the depletion of SRSF2 in the DOX‐inducible knockdown system. GAPDH acts as the loading control.The representative FACS plots show the frequency of CD31+CD34+ EPCs at day 5 with or without SRSF2 depletion by administration of DOX from day 2.5 to 5.The bar graph representing the percentage of CD31+CD34+ cells from (K). P‐values were determined by an unpaired two‐tailed Student’s t‐test. Data information: Results given are mean ± SD. *P < 0.05, **P < 0.01. All experiments were conducted for at least 3 biological replicates unless stated otherwise. Source data are available online for this figure.

Fig 4: Effects of ectopic expression of constitutive splicing factors on human hematopoietic differentiation AVolcano plot shows the differential splicing events regulated by PLB treated from day 2.5 to 5. The dotted line denotes FDR = 0.05.B, CThe scatter plot showing the correlation of splicing factors form Fig 3c (B) and all expressed genes (C) between day 2‐APLNR+ cells and day 5‐PLB‐treated cells. r refers to the correlation coefficient.DThe protein–protein interaction network of 38 intersected splicing factors (Fig 3D) (STRING: https://string‐db.org/).EThe mRNA expression of SF3A3, SNRPD1, and SNRPE in day 2‐APLNR+ cells, day 5‐DMSO–treated, and day 5‐PLB‐treated cells. N = 2 technical replicates.FThe upper panel illustrates the DOX‐inducible overexpression system. Western blotting confirmed the SRSF2 overexpression upon DOX induction with anti‐FLAG antibody. GAPDH was used as the loading control.GRT–qPCR assay of the relative mRNA expression of SF3A3, SNRPD1, and SNRPE after overexpression upon DOX induction. The ACTB gene was used as a control. Data were normalized to the mRNA level of empty vector controls cells.HWestern blotting showing the protein expression of SF3A3, SNRPD1, and SNRPE after overexpression with anti‐FLAG antibody. The GAPDH gene was used as a control.IRepresentative FACS plots of CD31+CD34+ cells in SF3A3, SNRPD1, and SNRPE overexpressed (GFP+) cells.JStatistical analysis of the frequency of CD31+CD34+ in SF3A3, SNRPD1, and SNRPE overexpressed (GFP+) cells. Data information: P‐values were calculated by one‐way ANOVA followed by Dunnett’s test. ns represents no significant difference. Results given are mean ± SD.. All experiments were conducted for at least 3 biological replicates unless stated otherwise Source data are available online for this figure.

Fig 5: NUMB alternative splicing impacts NOTCH activity during EPC generation The top 10 Reactome terms enriched from differential splicing events regulated by PLB. To define the differential splicing events, genes with FPKM > 1 in at least one differentiation stage were selected. Among them, the differential splicing events were identified by the double cutoffs of delta PSI > 0.2 and FDR < 0.05.GSEA analysis of NOTCH signaling in isoform level between DMSO control and 2.5 nM of PLB‐treated cells.The bar graph representing the percentage of CD31+CD34+ cells at day 5 of differentiation after treatment with DMSO and various concentrations of DAPT starting from day 2.5.The heatmap plotting the expression of NUMB_S and the key components in NOTCH signaling in day 2‐APLNR+ cells, day 5‐DMSO–treated, and day 5‐PLB‐treated cells (2.5 nM), respectively. The NOTCH components were divided into groups of upregulated, downregulated, and unchanged by comparison between DMSO‐ and PLB‐treated cells. The heatmap was scaled with Z‐Score using the log2(FPKM) expression of indicated genes. n = 2 technical replicatesRT–qPCR assay showing the expression of HES1 mRNA in the GFP+ cells on day 5 of differentiation with 1.25 nM PLB supplementation alone or in combination with NUMB_S ectopic expression from day 2.5.RT–qPCR analysis confirmed the overexpression of HES1 mRNA after DOX induction. The ACTB gene was used as a control. The data were normalized to the mRNA level in empty vector control cells.Western blotting showing the level of HES1 overexpression after DOX induction. The GAPDH gene was used as a loading control.The representative FACS plots indicating the generation of CD31+CD34+ EPCs on day 5 of differentiation with or without HES1 overexpression. DOX was treated from day 2.5 to 5 to induce HES1 overexpression.Statistical analysis of the percentage of CD31+CD34+ EPCs of (H).A hypothetical model. During hematopoietic differentiation from hESCs, a splicing factor switch occurs during the transition from mesoderm APLNR+ cells to EPCs. Concomitant with the switch of components in the splicing machinery, the splicing regulator SRSF2 is downregulated, leading to the enrichment of an EPC‐induced NUMB_S isoform. Alternative splicing of NUMB controls NOTCH activity, probably by suppressing HES1 expression and, ultimately, regulates EPC specification. The blue and red colors represent the two switching clusters within splicing factors during hematopoietic differentiation. Data information: Results given are mean ± SD. P‐values were determined by an unpaired two‐tailed Student’s t‐test in (C), (E), (F), and (I). **P < 0.01, ***P < 0.001. All experiments were conducted for at least 3 biological replicates. Source data are available online for this figure.