CNCC-specific deletion of Prmt1 elevates intron retention in the embryonic mandibular process.

(A) Expression levels of PRMT1-9 mRNAs in primary isolated cranial neural crest cells (CNCC) from Wnt1-Cre; R26RtdTomato mouse embryo heads at E13.5 and E15.5. TPM, transcript per million. (B) Diagram illustrating the isolation of CNCCs from embryonic mandibles, followed by poly(A)+ mRNA isolation and sequencing (n=4 in control and Prmt1 CKO group). (C) Prmt1 deletion in CNCC caused changes in alternative splicing (AS). Changes in AS events were analyzed by rMATS using RNA sequencing data and significant changes in each type of AS were shown in a stacked bar chart. SE, skipped exon. IR, intron retention. MXE, mutually exclusive exons. A5SS, alternative 5’ splice site. A3SS, alternative 3’ splice site. (D & E) Intron retention was prevalent in CNCCs and altered by Prmt1 deletion. Track view of genes demonstrating intronic and exonic expression with blue boxes indicating exons and blue lines indicating introns. Intron retention was elevated by Prmt1 deletion in Pex12, Mmp23, and Ecm1 (D) and reduced in Tbx1 (E). (F) Quantification of intron expression in Pex12, Mmp23, Ecm1 and Tbx1 by the percentage of intron-retaining mRNAs in each gene, calculated from IRI analysis based on RNA-seq data. * p<0.05 Prmt1 CKO vs Control. Control: Wnt1-Cre; R26RtdTomato. Prmt1 CKO: Wnt1-Cre; Prmt1fl/fl; R26RtdTomato.

CNCC-specific Prmt1 deletion reduced matrix gene expression in the developing mandibles.

(A) Volcano plot illustrating upregulation of 160 and downregulation of 303 genes in the mandibular primordium of Prmt1-deficient embryos at E13.5, compared to control mandibles. (B) Heatmap showing differential gene expression between control and Prmt1-deficient mandibles. (C) GO analysis of pathway enrichment in downregulated genes demonstrating glycosaminoglycan (GAG) degradation and extracellular matrix (ECM) organization as the top pathways. (D) Ingenuity Pathway Analysis suggesting connective tissue, bone, and cartilage development as affected biological processes based on downregulated genes. (E) GO analysis of pathway enrichment in upregulated genes demonstrating adult behavior and cytokine-mediated signaling and p53 signal transduction as top pathways. Control: Wnt1-Cre; R26RtdTomato (n=4). Prmt1 CKO: Wnt1-Cre; Prmt1fl/fl; R26RtdTomato (n=4).

PRMT1 regulates intron retention in ECM and GAG degradation genes.

(A-C) Intron retention increased in the majority of ECM gene transcripts that were downregulated in Prmt1 deficient embryos, as illustrated by a scatter plot based on intron retention index (IRI). The red line delineates unchanged levels of intron retention. Genes with the top differential IR were represented by red dots and labeled. ECM gene transcripts Adamts16 and Cthrc1 demonstrating higher intron retention in Prmt1- deleted embryos were illustrated by track view in B and quantified for intronic (left) and exonic (right) expression as shown in C. (D) Higher intron retention and lower mRNA abundance of ECM genes were validated in additional embryo samples by RT-PCR. Primers that span the intronic or intron-exon junction region were used to assess intronic expression. Primers that span the exonic region were used to examine exonic expression that indicated mRNA abundance. (E-H) Reduced expression of LOXL1 and FBLN5 was examined at the protein level by immunostaining (E, G) and quantified (F, H). Ee and Ef illustrated the plane of section and the region of analysis for E and G. (I-K) Intron retention increased in the majority of GAG degradation gene transcripts that were downregulated in Prmt1 deficiency, as indicated by a scatter plot based on IRI. The red line defines where intron retention is unchanged. Genes were represented by black dots and labeled in red. GAG degradation genes St6galnac3 and Galnt11 demonstrating higher intron retention in Prmt1-deleted embryos were illustrated by track view (J) and quantified for intronic (left) and exonic (right) expression (K). * p<0.05 Prmt1 CKO vs Control. Control: Wnt1-Cre; R26RtdTomato. Prmt1 CKO: Wnt1-Cre; Prmt1fl/fl; R26RtdTomato. Scale bar: (E,G), 100µm.

IR-triggered NMD functions as a basal and stress-responsive mechanism for mRNA decay in CNCCs.

(Aa, Ab) Treatment with the NMD inhibitor NMDI14 (NMDI) led to accumulation of intron-retaining (intron) transcripts of Gpx1, Adamts2, and Alpl in CNCCs from Control (Prmt1 Het) and Prmt1 CKO embryos, analyzed by RT-PCR. Gpx1 serves as a positive control to validate NMD inhibition by NMDI14. (Ba–Bf) NMDI14 caused accumulation of intron-retaining (intron) and total mRNAs (exon) of Adamts2, Alpl, Eln, Matn2, Loxl1 and Bgn in CNCCs. CNCCs were isolated from E13.5 Wnt1-Cre; R26RtdTomato and analyzed by RT-PCR. Primers that span the intronic or intron-exon junction region were used to assess intronic expression. Primers that span the exonic region were used to examine exonic expression that indicated mRNA abundance. (Ca, Cd) Intron retention of matrix transcripts Adamts2 and Fbln5 was detected in four independent control embryos by RNAseq, as illustrated by track views. (Cb-Cf) NMDI14 treatment caused accumulation of intron-retaining Adamts2 and Fbln5 transcripts. CNCCs isolated from E13.5 Wnt1-Cre; R26RtdTomato embryos were treated by DMSO or NMDI14 (NMDI), followed by mRNA extraction and assessment with semi-quantitative PCR. Primers (indicated by the red arrows) were designed to span regions (red line) of intron 19 of Adamts2 or intron 7 of Fbln5. * p<0.05 NMDI vs. DMSO of the same group.

PRMT1 methylates SFPQ, EWSR1 and TRA2B in CNCCs.

(A) Expression levels of splicing factors SFPQ, SRSF1, EWSR1, TAF15, TRA2B, HnRNPA1, WDR70 and G3BP1 in primary isolated CNCCs from Wnt1-Cre; R26RtdTomato mouse embryonic heads at E13.5 and E15.5. TPM, transcript per million. (B-C) Subcellular localization of splicing factors in CNCCs by immunostaining using sagittal sections of the mandibular process from wild-type mouse embryos, which revealed nuclear expression of SFPQ (a-d), EWSR1 (e-h), and TAF15 (i-l); even distribution of TRA2B between nucleus and cytosolic compartment (m-p), and cytoplasmic expression of SRSF1 (q-t) and G3BP1 (u-x). The subcellular distribution was quantified in C and presented as nuclear to cytosolic signal ratio (Nuc/Cyto Ratio). (D-G) SFPQ methylation diminished in the mandibular (D&E) and maxillary (F&G) processes of Prmt1 deficient embryos at E13.5. (H-K) Reduction of EWSR1 and TRA2B methylation was observed in the mandibular processes of Prmt1 deficient embryos at E13.5. Methylation was detected by proximity ligation assay (PLA). Green puncta indicated methyl-SFPQ, TRA2B or EWSR1. Nuclei were counterstained with DAPI (blue). Representative images are shown for Control (Da-Dc; Fa-Fc; Ha-Hc; Ja-Jc) and Prmt1 CKO (Dd-Df; Fd-Ff; Hd-Hf; Jd-Jf). Higher magnification views in (Dc, Df, Fc, Ff, Hc, Hf, Jc, Jf) illustrate methyl-SFPQ, TRA2B and EWSR1 (green puncta) in the nuclei, as indicated by white arrows. E, G, I and K showed quantification of PLA puncta normalized to cell number in four biological replicates, presented as mean ± SEM. * p<0.05 Prmt1 CKO vs Control. Control: Wnt1-Cre; R26RtdTomato. Prmt1 CKO: Wnt1-Cre; Prmt1fl/fl; R26RtdTomato. Scale bar = 100µm in B, D, F, H and J except in enlarged panels, where scale bar = 25µm (Bd, Bh, Bl, Bp, Bt, Bx, Dc, Df, Fc, Ff, Hc, Hf, Jc, Jf).

PRMT1 depletion reduces SFPQ protein levels via proteasomal degradation.

(A-D) SFPQ protein level was significantly reduced in the mandibular (A&B) and maxillary (C&D) processes of Prmt1 deficient embryos. SFPQ protein was detected by immunostaining and quantified in B and D. (E & F) SFPQ protein levels declined dramatically in the Prmt1 deficient embryonic head, as detected by Western blotting (E) and quantified with ImageJ (F). (G) PRMT1 deletion did not alter SFPQ mRNA levels in CNCC. (H-K) SFPQ protein accumulated in Prmt1-deficient CNCCs upon MG132 treatment, and the accumulated SFPQ protein remained un-methylated. CNCCs isolated from Prmt1 CKO embryos and littermate controls (Prmt1 heterozygous) were treated with MG132. SFPQ protein was detected by immunostaining in H and quantified in I. SFPQ methylation was detected by PLA in J and quantified in K.* p<0.05 Prmt1 CKO vs Control. Control, Wnt1-Cre; R26RtdTomato. Prmt1 CKO, Wnt1-Cre; Prmt1fl/fl; R26RtdTomato. Scale bar =100µm in A & C. Scale bar =25µm in H & J.

PRMT1-SFPQ pathway regulates matrix genes in CNCCs.

(A) SFPQ knockdown caused around 50% reduction in Sfpq expression in CNCCs. CNCCs were transfected with two independent siRNAs targeting SFPQ, or control siRNA, followed by poly(A)+ mRNA extraction and RNA sequencing. * p<0.01 siSFPQ vs siControl. (B) The two independent siRNAs targeting SFPQ caused similar transcriptomic and intronic changes, shown by scatter plot of DEG (Ba) and IRI (Bb). DEG, differentially expressed genes. IRI, intron retention index. (C) SFPQ depletion altered intron retention in CNCCs. Changes of IR events in genes showing increased (Up, yellow color) or decreased (Down, orange color) IR were illustrated by stacked bar graphs. Pie chart demonstrated differentially regulated genes, with shaded areas among downregulated genes (orange) highlighting their overlap with IR elevated genes (red). (D) GO analysis of genes with elevated IR following SFPQ depletion. (E&F) Heatmap and GO analysis of SFPQ-regulated genes in CNCC. (G&H) Pol II CUT&Tag analysis in ST2 cells transfected with control or SFPQ siRNAs showing Pol II recruitment in downregulated and upregulated genes (G), promoter regions (Ha) and gene body (Hb). P-value was indicated at the top. siControl, control siRNA. siSFPQ#1 and siSFPQ#2, two independent SFPQ siRNAs.

SFPQ regulates intron retention of Wnt signaling and neuronal genes in CNCCs.

SFPQ depletion in CNCCs elevated intron retention and decreased mRNA abundance of Wnt signaling components (A) and neuronal genes (B). The levels of mRNA abundance (left) and intron retention (right) were illustrated by a two-sided bar graph. TPM, transcripts per million. * p<0.01 siSFPQ vs siControl. siControl, control siRNA. siSFPQ#1 and siSFPQ#2, two independent SFPQ siRNAs.

SFPQ depletion reduces long gene expression via intron retention triggered NMD.

(A) SFPQ depletion promoted intron retention and reduced mRNA abundance of Col4a2, St6galnac3, and Ptk7 in ST2 cells. Bar chart showing RT-PCR analysis of intronic and exonic expression. (B&C) NMD inhibitor NMDI14 caused the accumulation of retained introns and total mRNAs of Col4a2, St6galnac3, and Ptk7 in ST2 cells. Bar chart showing RT-PCR analysis of intronic (B) and exonic expression (C) in DMSO or NMDI-treated cells. NMDL14-mediated inhibition of NMD was validated using Gpx1 as a positive control. *p<0.05 siSFPQ vs. siControl. # p<0.05 NMDI vs. DMSO treatment of the same group. (D) SFPQ binding peaks were mapped to retained intron 1 but not spliced intron 6 of Ptk7. Peak distribution from published Sfpq CLIP-seq data using E13.5 brain (top) and track view of RNA-seq data using siControl or siSFPQ-transfected CNCCs (bottom) for Ptk7, with the retained Ptk7 intron 1 (red box) and spliced intron 6 (green box) highlighted. (E) SFPQ binding peaks were significantly enriched in retained intron regions within CNCCs. Violin plot displaying the density of SFPQ binding peaks in introns with elevated retention compared to introns with reduced retention or no change. P-value was calculated using Mann-Whitney U test and indicated at the top. (F) SFPQ binding peaks were preferentially enriched in genes with higher intron retention when compared to genes with no IR change. P = 0.07 using Fisher’s exact test. (G) SFPQ-regulated genes were significantly longer than average. Violin plot displaying the distribution of length for genes showing increased intron retention (IR Up), decreased intron retention (IR Down), or unchanged intron retention (No Change). Median length of the IR Up group was highlighted with solid blue line. Median length of the IR Down and No Change groups was highlighted with dotted blue line. P-value was calculated using Mann-Whitney U test and indicated at the top.

SFPQ, EWSR1, TAF15, and TRA2B regulate distinct transcriptional and splicing programs.

(A) Genes with increased IR events in SFPQ-depleted CNCCs demonstrated 8.28% (64 out of 773) overlap with Prmt1 deficient CNCCs. (B-E) Matrix genes Col4a2, Adam12, Ntn1, App, St6Galnac3, Galnt10 and Asph were regulated by both Prmt1 deletion and SFPQ depletion. Bar graph showing elevated IR (B, C) and reduced mRNA abundance (D, E) in CNCCs. TPM, transcripts per million. * p<0.05 siSFPQ vs siControl. * p<0.05 Prmt1 CKO vs Control. Control: Wnt1-Cre; R26RtdTomato. Prmt1 CKO: Wnt1-Cre; Prmt1fl/fl; R26RtdTomato. (F-S) ST2 cells were transfected with control or SFPQ, EWSR1, TAF15, TRA2B siRNAs and mature mRNA was extracted for sequencing. GSEA demonstrated enrichment of ECM genes among downregulated genes upon depletion of SFPQ, EWSR1, TRA2B, and TAF15 (F-I). Pie charts showed the percentage of genes with increased (dark grey) or decreased IRI (light grey) upon SFPQ, EWSR1, TAF15 and TRA2B deletion (J-M). rMATS analysis showed widespread and significant splicing changes following depletion of SFPQ, EWSR1, TAF15, and TRA2B (N). Postn mRNA expression was decreased in all four depletion groups (O). Exon skipping events of Postn were noted in all four depletion groups, illustrated by Sashimi plots (P) and quantified by bar graphs based on rMATS analysis (Q-S). * p<0.05 siRNAs vs siControl. siControl, control siRNA. siSFPQ#1 and siSFPQ#2, two independent SFPQ siRNAs.