PMC

(A–B) SFRS2 is required to support self-renewal. Lentiviral shRNA-mediated depletion of SFRS2 disrupted pluripotency in H1 ESC as monitored by (A) colony morphology and (B) expression of OCT4 and NANOG. (C) Depletion of OCT4 in H1 ESC for 2 and 5 days led to a coordinate decrease in the expression of SFRS2. (D) OCT4 selectively binds the proximal promoter region of SFRS2 in H1 ESC. (E) OCT4 depletion in human clone 9 iPSC disrupts luciferase expression downstream of the native SFRS2 promoter. (F) Mutation or deletion of the predicted OCT4 binding motif (ATGCCAAT) in the proximal SFRS2 promoter region decreased downstream luciferase expression in clone 9 iPSC. See also . Error bars represent the mean ± SEM.

(A–D) Lentiviral-mediated expression of MBD2a but not MBD2c disrupts pluripotency in H1 ESC. Expression of MBD2 isoforms as monitored by (A) qRT-PCR and (B) western blot. Pluripotency in H1 ESC was assessed by (C) colony morphology and (D) expression of OCT4, NANOG, and SOX2. (E–G) Exogenous expression of MBD2c but not MBD2a enhances reprogramming efficiency in BJ DF. (E) Expression of MBD2 isoforms in infected BJ DF as monitored by qRT-PCR. Reprogramming efficiency was assessed by the (F) number and (G) fold change of TRA-1-60 colonies measured across biological replicates. (H) Exogenous MBD2a and MBD2c independently bind to OCT4 and NANOG promoter regions in (top) H1 ESC and (bottom) 293T cells. (I–J) MBD2 interacts with the NuRD complex in an isoform-specific manner. (I) Members of the NuRD transcription repressor complex (HDAC1, HDAC2, MTA2, Mi-2, and RbAp46) co-immunoprecipitate with exogenous FLAG-HA-MBD2a but not FLAG-HA-MBD2c in 293T cells. Neither MBD2 isoform interacts with the SIN3A-Histone deacetylase complex. (J) Co-immunoprecipitation of MTA2 in 293T cells overexpressing FLAG-HA-tagged MBD2 isoforms confirmed the preferential interaction between MBD2a and MTA2, a core NuRD complex member. (K) Proposed model illustrating a putative positive feedback loop, in which the splicing factor SFRS2 along with microRNAs, controlled by the core pluripotency genes, regulate the expression of MBD2 isoforms that either support (MBD2c) or oppose (MBD2a) expression of OCT4, NANOG, and SOX2 through recruitment of the NuRD complex. Error bars represent the mean ± SEM.

(A) Exon and protein graph for the methyl-CpG binding protein MBD2. Dashed lines indicate splice sites. Protein segments corresponding to each exon are annotated with predicted functional domains and primer locations. (B–C) Verification of the distinct MBD2 isoforms in H1 ESC and BJ DF by (B) qRT-PCR and (C) western blot. (D) Lentiviral shRNA-mediated depletion of OCT4 and SFRS2 independently led to a significant increase in MBD2a expression along with reduced levels of MBD2c after 5 days in H1 ESC. (E) Exogenously expressed SFRS2-FLAG-HA preferentially binds to MBD2 pre-mRNA at intron 2 (primer pairs II and III, each spanning into exon 2 and exon 3, respectively) but not inside exon 2 (primer pair I) in H1 ESC. (F) miR-301b and miR-130b suppress luciferase expression in the context of wild-type but not mutated sequences corresponding to the 3′-UTR of SFRS2 in HeLa cells. (G–I) miR-302 cluster members target the 3′-UTR of MBD2 in an isoform-specific manner. (G) miR-302 cluster members specifically suppressed luciferase expression upstream of the wild-type but not mutated MBD2a 3′-UTR sequence in HeLa cells. (H) miR-302 cluster members did not affect luciferase expression upstream of the MBD2c 3′-UTR sequence in HeLa cells. (I) Overexpression of miR-302 cluster members in 293T cells reduced expression of endogenous MBD2a. Error bars represent the mean ± SEM.

(A) Independent hierarchical clustering of microarray and proteomic data demonstrated that hPSC are molecularly distinct compared to DF at the level of (left) gene expression, (middle) protein expression, and (right) protein phosphorylation. (B, left) Analysis of pluripotency signature genes according their membership in Gene Ontology (GO) Biological Processes (BP) revealed enrichment of multiple pathways. (B, right) Further analysis of genes within each GO-BP pathway and measurement class (gene, protein, and phosphoprotein) based on physical interactions with three positive reference sets (PRS) of pluripotent factors (see , , and ) suggested that the RNA splicing pathway is strongly associated with pluripotency. Abbreviations: Main., maintenance; Org., organization; Reg., regulation. (C) Splicing factors within the pluripotent molecular signature were individually ranked (see and ), with SFRS2 as the top candidate. (D) Alternatively spliced genes associated with hESC are enriched for physical interactions with positive reference sets (PRS) of pluripotent factors. Null distributions (grey bars) were created by random selection (10,000 iterations) of identical size gene sets from the background of all genes detected by exon-junction microarray. (E) The methyl-DNA binding protein MBD2 displays the strongest alternative splicing pattern between hESC and DF based on the linear regression analysis (see ).