ABSTRACT
We previously demonstrated that the two paralogous RNA-binding proteins PCBP1 and PCBP2 are individually essential for mouse development: Pcbp1-null embryos are peri-implantation lethal, while Pcbp2-null embryos lose viability at midgestation. Midgestation Pcbp2−/− embryos revealed a complex phenotype that included loss of certain hematopoietic determinants. Whether PCBP2 directly contributes to erythropoietic differentiation and whether PCBP1 has a role in this process remained undetermined. Here, we selectively inactivated the genes encoding these two RNA-binding proteins during differentiation of the erythroid lineage in the developing mouse embryo. Individual inactivation of either locus failed to impact viability or blood formation. However, combined inactivation of the two loci resulted in midgestational repression of erythroid/hematopoietic gene expression, loss of blood formation, and fetal demise. Orthogonal ex vivo analyses of primary erythroid progenitors selectively depleted of these two RNA-binding proteins revealed that they mediate a combination of overlapping and isoform-specific impacts on hematopoietic lineage transcriptome, impacting both mRNA representation and exon splicing. These data lead us to conclude that PCBP1 and PCBP2 mediate functions critical to differentiation of the erythroid lineage.
KEYWORDS: murine erythropoiesis, posttranscriptional controls, PCBP1, PCBP2, RNA-binding proteins, conditional gene inactivation
INTRODUCTION
RNA-binding proteins participate in a complex array of posttranscriptional controls essential to cell type specification and somatic development. The critical roles of such posttranscriptional controls are most clearly apparent in settings of global silencing of transcription. Posttranscriptional controls of terminal erythroid differentiation constitute one such highly informative experimental model (1–5).
The poly(C) binding proteins (PCBPs) comprise a widely expressed and multifunctional family of KH domain RNA-binding proteins associated with numerous erythroid and nonerythroid mRNAs (6, 7). Studies focused on globin gene expression in cell culture models have revealed that the PCBPs can integrate nuclear controls over splicing and 3′ processing with cytoplasmic controls over mRNA stabilization and translation (8, 9). These controls are mediated to a large extent via direct interactions of PCBPs with pyrimidine-pure cytosine-rich motifs located at critical sites within nuclear transcripts and cytoplasmic mRNAs (8–13). Transcriptome-wide analyses in a number of experimental settings have further revealed that PCBPs control the structures, abundance, and functions of a wide array of mRNAs encoded by multiple gene subsets throughout the mammalian transcriptome (11, 12).
The PCBPs are encoded by genes representing four dispersed loci: Pcbp1 to Pcbp4. The ubiquitous expression profiles of the major isoforms Pcbp1 and Pcbp2 contrast with the lower and more restricted expression of Pcbp3 and Pcbp4 (14–17). In contrast to the multiexonic gene of Pcbp2, Pcbp1 is carried by an intronless gene that evolved by retrotransposition of a fully processed Pcbp2 mRNA isoform (14, 16). The remarkably high level of structural conservation of the Pcbp1 and Pcbp2 homologs across eukaryotic taxa suggests strong evolutionary constraints in their functions (17).
The PCBP1 and PCBP2 proteins appear to have shared as well as isoform-specific functions (18–20). Recently we have shown that Pcbp1 and Pcbp2 constitute essential genes in mice (21); homozygous germ line inactivation of Pcbp1 renders embryos nonviable in the peri-implantation stage (between E4.5 and E8.5), while Pcbp2-null mouse embryos undergo normal development through E12.5/13.5, at which point they undergo a loss in viability and develop a complex array of phenotypic defects. These defects prominently include alterations in the hematopoietic system (21). The very early (preimplantation) loss of Pcbp1-null embryos obviates detection of an in vivo impact of PCBP1 on hematopoiesis.
In the current study, we specifically focus on the impact of Pcbp1 and Pcbp2 on hematopoietic differentiation. The conditional, erythroid-specific inactivation of the Pcbp1 and/or Pcbp2 locus in the developing mouse embryo in conjunction with targeted ex vivo depletions of these proteins in primary hematopoietic cells reveals that the PCBPs play a critical role in mammalian erythropoiesis.
RESULTS
Selective loss of Pcbp1 or Pcbp2 in the erythroid lineage is compatible with embryonic viability and adult function.
To explore the individual role(s) of Pcbp1 and Pcbp2 in mouse erythropoiesis, we selectively inactivated the corresponding genes in the erythroid lineage of the developing mouse embryo. Mice homozygous for a floxed Pcbp2 allele (Pcbp2fl/fl) were crossed with an EpoR-cre line. The EpoR-cre locus is composed of a Cre recombinase open reading frame (ORF) inserted within the endogenous erythropoietin receptor (EpoR) locus (22). EPOR is specifically activated at the erythroid burst-forming unit (BFU-E) stage erythroid differentiation and is maximally expressed in proerythroblasts within the erythroid CFU (CFU-E) (22, 23). EpoR-cre-mediated recombination within the floxed Pcbp2 locus deletes the first two exons of Pcbp2, thus inactivating gene expression in the erythroid lineage (21).
Cre-mediated inactivation of the PCBP2 locus was confirmed by DNA, RNA, and protein analyses of erythroid cells isolated from Pcbp2fl/fl/EpoR-Cre fetal livers at E14.5 (Fig. 1A to C). Pcbp2fl/fl/EpoR-cre embryos at E14.5 were developmentally intact (Fig. 1D), and corresponding adult mice appeared entirely normal, with the sole exception of a minor decrease in platelet levels (data not shown). Thus, in contrast to the embryonic lethality and hematopoietic abnormalities seen in germ line Pcbp2-null mice (21), the selective homozygous inactivation of the Pcbp2 loci during erythroid lineage development failed to have an impact on hematopoiesis or overall development.
FIG 1.
Embryos lacking Pcbp2 in the erythroid lineage (Pcbp2fl/fl/EpoR-cre) appear normal at E14.5. (A) Genotyping. E14.5 embryos generated from a cross between Pcbp2fl/fl and Pcbp2fl/fl/EpoR-cre mice were genotyped with primers specific for the floxed Pcbp2 and EpoR-cre loci. The genotypes of each of four embryos in a single litter are summarized to the right of the gel. (B) RNA analysis. The Ter119+-enriched cells isolated from the livers of embryos (in panel A) were analyzed for Pcbp2 mRNA. (C) Protein analysis. The Ter119+-enriched cells (as in panel B) were assessed for PCBP2 protein content by Western analysis using a monospecific PCBP2 antibody (see Materials and Methods). Ribosomal protein L7a served as a loading control. (D) Embryo appearance. A Pcbp2fl/fl/EpoR-cre E14.5 embryo (right) demonstrated normal appearance compared with a Pcbp2fl/fl littermate control (left).
We next intercrossed EpoR-cre mice with mice carrying floxed Pcbp1 loci. Pcbp1 inactivation was confirmed by analyses of Pcbp1fl/fl/EpoR-cre embryonic liver erythroid cells (Fig. 2A to C). Pcbp1fl/fl/EpoR-cre mice were born at the expected Mendelian ratio, and the corresponding adults were within normal weight range compared to littermate controls. However, when assessed at E14.5, the Pcbp1fl/fl/EpoR-cre embryos were observed to be pale compared to littermate controls (Fig. 2D). This pallor, along with alterations in a subset of red cell indices (RDW-CV and RDW-SD) and an elevated reticulocyte count suggested a mild level of erythroid stress in the Pcbp1fl/fl/EpoR-cre adults (data not shown). Thus, the selective homozygous inactivation of Pcbp1 loci by EpoR-cre resulted in nonlethal and compensated alterations in the hematopoietic pathway.
FIG 2.
Embryos lacking Pcbp1 in the erythroid lineage (Pcbp1fl/fl/EpoR-cre) demonstrate mild pallor at E14.5. (A) Genotyping. E14.5 embryos generated from a cross between Pcbp1fl/fl and Pcbp1fl/fl/EpoR-cre mice were genotyped by PCR for the floxed Pcbp1 and EpoR-cre loci. The genotypes of seven embryos in a single litter are summarized to the right of the gel. (B) RNA analysis. Ter119+-enriched cell populations from each of the indicated embryos (as in panel A) were analyzed for Pcbp1 mRNA. (C) Protein analysis. The Ter119+-enriched cell populations (as in panel B) were assessed by Western analysis with a monospecific PCBP1 antibody. Ribosomal protein L7a served as a loading control. (D) Embryo appearance. The Pcbp1fl/fl/EpoR-cre E14.5 embryo (right) demonstrated mild pallor compared with a Pcbp1fl/fl littermate control (left).
Combined inactivation of the Pcbp1 and Pcbp2 loci impacts erythroid lineage formation and embryonic viability.
To further explore the in vivo functional relationships of the two Pcbp isoforms in the context of hematopoiesis, we generated intercrosses of mice carrying the floxed Pcbp1 and Pcbp2 alleles and the EpoR-cre locus. Mice compound heterozygous for inactivation of the two Pcbp loci (Pcbp1fl/wt/Pcbp2fl/wt/EpoR-cre) were born at normal Mendelian ratios and appeared healthy at 2 months old, with normal complete blood counts (CBCs). The same intact functional state was observed for embryos that were Pcbp2 null in the erythroblast lineage combined with Pcbp1 haploidy (Pcbp1fl/wt/Pcbp2fl/fl/EpoR-cre) (Fig. 3A). In addition, mice with this genotype were viable at birth and as adults. In contrast, homozygosity for Pcbp1 inactivation in combination with Pcbp2 haploidy (Pcbp1fl/fl/Pcbp2fl/wt/EpoR-cre) was embryonic lethal. Analysis of E14.5 embryos with this genotype demonstrated marked overall pallor with faint and atrophic livers (Fig. 3B). In addition, mice with this genotype were embryonic lethal. These observations led us to conclude that there was a level of redundancy in Pcbp1 and Pcbp2 functions in the erythroid developmental pathway as well as a predominant role for PCBP1 versus PCBP2 in this process.
FIG 3.
Compound conditional inactivations of the Pcbp1 and Pcbp2 loci in the erythroid lineage reveal their essential roles in development of erythroid activity and in sustaining embryonic viability. (A) The mouse embryo with an erythroid lineage lacking PCBP2 combined with haploidy for PCBP1 appears normal at E13.5. A Pcbp1fl/wt/Pcbp2fl/fl/EpoR-cre E13.5 embryo is shown next to corresponding littermate control (floxed alleles without EpoR-cre). Embryos with the Pcbp1fl/wt/Pcbp2fl/fl/EpoR-cre genotype retain viability through term and into adult life. (B) The mouse embryo with an erythroid lineage lacking PCBP1 combined with haploidy for PCBP2 is small and pale at E14.5. A Pcbp1fl/fl/Pcbp2fl/wt/EpoR-cre embryo is shown next to the corresponding littermate control (floxed alleles without EpoR-cre). Embryos with the Pcbp1fl/fl/Pcbp2fl/wt/EpoR-cre genotype are embryonic lethal. (C) Combined homozygous inactivation of PCBP1 and PCBP2 loci results in marked loss of erythroid activity at E11.5. The yolk sacs of two E11.5 littermate embryos are shown along with their indicated genotypes. (D) E12.5 embryos with combined homozygous inactivation of the Pcbp1 and Pcbp2 loci are markedly pale. The embryos shown are from a single litter; corresponding genotypes are shown above each embryo. (E) Compound homozygous inactivation of Pcbp1 and Pcbp2 in the erythroid lineage is embryonic lethal. The numbers in the table reflect a total of 42 embryos from a total of 7 litters.
The most severe hematopoietic phenotype was observed for embryos compound homozygous for loss of the Pcbp1 and Pcbp2 loci (Fig. 3). These embryos (Pcbp1fl/fl/Pcbp2fl/fl/EpoR-cre) had markedly pale yolk sacs at E11.5 (Fig. 3C), extreme overall pallor and absence of liver erythropoiesis at E12.5 (Fig. 3D), and loss of viability by E13.5 (Fig. 3E). Thus, the complete loss of Pcbp1 in the erythroid lineage is well tolerated when both Pcbp2 loci are functioning, is embryonic lethal in the context of Pcbp2 haploidy, and is even more devastating in terms of erythroid development and embryonic viability in embryos that are compound homozygous for inactivation of Pcbp1 and Pcbp2 (Pcbp1fl/fl/Pcbp2fl/fl/EpoR-cre). The foregoing observations support a dose dependency in the combined activities of Pcbp1 and Pcbp2 on erythropoiesis, with a predominant role played by Pcbp1.
Orthogonal evidence for a predominant role of PCBP1 in erythroid terminal differentiation.
An orthogonal assessment of Pcbp1 and Pcbp2 functions in terminal erythroid differentiation was carried out in the mouse erythroleukemia (MEL) cell culture model (24). Depletion of Pcbp2 by short hairpin RNA (shRNA) treatment had no appreciable impact on the induction of terminal erythroid differentiation, while a parallel depletion of Pcbp1 from the MEL cells resulted in a dramatic blockade of hemoglobin synthesis (data not shown). These data further support a predominant role for Pcbp1 compared to Pcbp2 in murine erythroid differentiation.
Distinct contributions of Pcbp1 and Pcbp2 to the erythroid lineage as revealed by transcriptome analysis.
While PCBP1 and PCBP2 protein structures are highly conserved and share a C-rich binding motif (14–17, 25), these two proteins have been shown to function both redundantly as well as with isoform specificity in various contexts (21, 26–33). A combination of redundant and isoform-specific functions in hematopoiesis would be consistent with the results of the preceding conditional gene inactivation studies (Fig. 1 and 3). To further explore the relative roles of PCBP1 and PCBP2 in hematopoiesis, we analyzed the impacts of selective depletion of each protein on the transcriptomes of primary hematopoietic progenitors induced to differentiate ex vivo under controlled conditions (34, 35). This ex vivo culture system supports terminal erythroblast proliferation and differentiation in a manner that closely mimics the in vivo terminal proliferation and maturation of erythroid cells (35). This experimental model, in which differentiation is monitored in a quantitative manner by flow cytometry, allows for erythroid cells at particular developmental stages to be identified and analyzed (35).
Sets of shRNAs targeting PCBP1, PCBP2, and nontargeting control sequences were used to assess the impact of PCBP depletion on ex vivo erythroid differentiation. Hematopoietic progenitor cells were affinity purified from wild-type E14.5 fetal livers, expanded, and selectively depleted of Pcbp1 or Pcbp2 using 4 distinct shRNAs for each isoform. A set of nontargeting (control) shRNAs were used in parallel to control for nonspecific shRNA-induced perturbations of the transcriptome (34, 36).
RNA from each population was isolated at days 0 and 2 of induced differentiation. Sequencing of these RNA populations yielded approximately 35 million reads per sample. Analyses of the transcriptomes at day 0 confirmed an efficient depletion of PCBP1 or PCBP2 (>90%) for both PCBP1 and PCBP2 mRNAs (Fig. 4A) and proteins (Fig. 4B). The differential gene expression (DGE) analysis of the transcriptome sequencing (RNA-seq) data sets revealed that Pcbp1 had a greater overall impact on gene expression than did Pcbp2 (Fig. 5). This difference was observed both prior to and subsequent to erythroid induction (days 0 and 2, respectively) (see Table S2 in the supplemental material). It was additionally noted that the overall number of impacted genes was substantially greater at day 2 versus day 0, a trend that parallels the major shift in transcriptome composition during erythroid differentiation (37). Pathway analysis revealed significant enrichment for pathways relevant to chromatin assembly, cell adhesion, cell differentiation, regulation of signal transduction, cellular response to stress, cell proliferation, regulation of mitotic cell cycle, and erythroid differentiation after erythroid development induction (see Table S3 in the supplemental material).
FIG 4.
Selective depletion of PCBP1 and PCBP2 in hematopoietic stem cell (HSC)/progenitor cells. (A) Depletions of Pcbp1 and Pcbp2 mRNAs in primary HSC/progenitor cells. MSCV-PIG (puromycin-IRES-GFP)-based shRNAs targeting either PCBP1 or PCBP2 were transduced into HSC/progenitor cells purified from wild-type E14.5 mouse livers. The impact of the targeted depletions was quantified 2 days after transfection by qRT-PCR (with values normalized to corresponding glyceraldehyde-3-phosphate dehydrogenase [GAPDH] mRNA levels). P values were calculated by comparing the impact of Pcbp2 and Pcbp1 depletion with the result from parallel controls. ∗∗∗, P < 0.001. (B) Western analysis. The protein analysis confirms efficient depletion of PCBP1 and PCBP2 proteins in shRNA-transduced cells. L7a, large ribosomal subunit 7a. The shRNAs are abbreviated as follows: Scr, Scrambled; Luc, luciferase.
FIG 5.
Comparative impacts of Pcbp1 and Pcbp2 depletions on hematopoietic transcriptomes during ex vivo erythroid differentiation. (A) Genes upregulated by Pcbp1 or Pcbp2 depletions. (Top) Venn diagram displays the numbers of genes in primary hematopoietic progenitors that are upregulated upon Pcbp1 or Pcbp2 depletion prior to erythroid induction (day 0). Pcbp1 depletion versus control is in light blue, Pcbp2 depletion versus control is in light brown, and genes upregulated by both Pcbp1 and Pcbp2 depletions are in white (overlap). (Bottom) A Venn diagram displays the numbers of genes upregulated by depletion of Pcbp1 or Pcbp2 after 2 days of induced erythroid differentiation (day 2). The color code is as in the top diagrams. (B) Genes downregulated by Pcbp1 or Pcbp2 depletions. (Top) A Venn diagram displays the numbers of genes in primary hematopoietic progenitors that are downregulated upon Pcbp1 or Pcbp2 depletion prior to erythroid induction (day 0). The color code is as in panel A. (Bottom) A Venn diagram displays the numbers of genes that are downregulated by depletion of Pcbp1 or Pcbp2 after 2 days of induced erythroid differentiation (day 2). The color code is as in panel A. (C) Pyrimidine-pure/C-rich motifs are enriched in the 3′ UTRs of genes downregulated upon Pcbp1 and Pcbp2 depletions. (Top) Motif maps showing the frequency and positioning of Pcbp1 motifs (defined in the CISBP-RNA database [38]) in the 3′ UTRs of genes that are impacted (upregulated in red and downregulated in blue) upon Pcbp1 depletion in primary embryonic erythroid cells (combined day 0 and day 2 data). The Pcbp1 motifs are significantly enriched in the downregulated genes relative to the control as background (FDR corrected P = 3.56e−06, one-sided Fisher’s exact test). Motif significance was computed with sequences up to 1,000 nucleotides (nt) into the 3′ UTRs of these genes. Frequencies were smoothed using a running mean of 50 nt. (Bottom) Motif maps showing the frequency and positioning of Pcbp2 motifs (CISBP-RNA database [38]) in the 3′ UTRs of genes that are upregulated (red) and downregulated (blue) upon Pcbp2 depletion in primary embryonic erythroid cells (combined day 0 and day 2 data). The color code is as in the top panel. The Pcbp2 motifs are significantly enriched in the downregulated genes, with the control as the background (FDR corrected P = 2.09e−15, one-sided Fisher’s exact test). Motif significance was computed with sequences up to 1,000 nt into the 3′ UTRs of these genes. Frequencies were smoothed using a running mean of 50 nt. (D) The Rbm38 motif (GU-rich) is significantly enriched in the 3′ UTRs of mRNAs that are upregulated upon Pcbp2 depletion. (Top) Motif maps showing the frequency and positioning of Rbm38 motifs in the 3′ UTRs of genes that are upregulated (red) and downregulated (blue) upon Pcbp1 depletion (combined day 0 and day 2 data). The Pcbp1 motif is not significantly enriched in either upregulated genes or downregulated genes relative to control depletion as the background (black). Motif significance was computed with sequences up to 1,000 nt into the 3′ UTRs of these genes. Frequencies were smoothed using a running mean of 50 nt. (Bottom) Motif maps showing the frequency and positioning of the Rbm38 motif in the 3′ UTRs of genes that are upregulated (red) and downregulated (blue) upon Pcbp2 depletion (combined day 0 and day 2 data). The Rbm38 motif is significantly enriched in the genes upregulated upon Pcbp2 depletion, with control depletion as the background (FDR corrected P = 0.003, one-sided Fisher’s exact test). Motif significance was computed with sequences up to 1,000 nt into the 3′ UTRs of these genes. Frequencies were smoothed using a running mean of 50 nt. (E) Impact of Pcbp1 and Pcbp2 depletions on genes linked to hematopoietic/erythropoietic differentiation and function. RT-PCR analyses were carried out on RNA isolated from HSC/progenitor cells post-erythroid induction (day 2 cells). The graph reflects the average and standard deviation of results from 10 separate RNA samples. The shRNAs targeting Pcbp1 are shown as 1-1, 1-3, and 1-4, the shRNAs targeting Pcbp2 are shown as 2-1, 2-2, 2-3, and 2-4, and the control shRNAs are shown as scr1, scr2, Luc1, and Luc2. Each mRNA value was normalized to the corresponding Gapdh mRNA level prior to calculation of the averages and significance values. P values were calculated by comparing the impact of Pcbp1 (top row) and Pcbp2 (low row) depletion with values from parallel controls. ∗, P < 0.05; ∗∗, P < 0.01; n.s., not significant.
The degree of overlap between the sets of Pcbp1- and Pcbp2-impacted genes was evaluated next. This overlap was minimal at day 0 but was substantially increased at day 2, in parallel with the overall increase in the number of impacted genes in the differentiating erythroblasts (Fig. 5A and B). The differential impacts resulting from Pcbp1 versus Pcbp2 depletions during erythroid induction noted in this study were consistent with the distinct patterns and timing of embryonic lethality observed secondary to somatic loss of these two isoforms (21). Thus, these DGE studies supported a combination of redundancy and isoform specificity in Pcbp isoform activities in the differentiating erythroid lineage.
Impacts of Pcbp1 and Pcbp2 on mRNA expression appear to reflect both direct and indirect mechanisms.
To explore the mechanisms of Pcbp impact on the erythroid transcriptome, we searched for C-rich binding site motifs in target mRNAs. This search focused on 3′ untranslated regions (UTRs) of mRNAs that were impacted by depletion of PCBP1 or PCBP2 during the process of the ex vivo erythropoietic differentiation process. Past studies have demonstrated that canonical C-rich binding sites in 3′ UTRs within target mRNAs are linked to major pathways of PCBP-mediated posttranscriptional control, including mRNA stability and 3′ processing (8, 11, 12). The current analysis revealed a significant enrichment in cognate Pcbp1 and Pcbp2 binding sites (CISBP-RNA database [38]) in 3′ UTRs of genes downregulated in the differentiating hematopoietic precursors upon depletion of Pcbp1 (Fig. 5C, top) or Pcbp2 (Fig. 5C, bottom). Pathway analysis of these genes revealed significant enrichment for pathways relevant to chromatin assembly, nucleosome organization, RNA metabolism, chromatin silencing, and negative regulation of gene expression (see Table S4 in the supplemental material). This finding supports a model in which direct PCBP binding plays a positive role in the expression of a subset of mRNAs during the differentiation process.
Remarkably, the motif searches revealed additional motifs, distinct from the poly(C)-rich sequences, that were also enriched in 3′ UTRs of PCBP-impacted mRNAs. Of particular interest was the identification of the consensus motif for RBM38 (GU rich) for many genes (Fig. 5D; see Table S5 in the supplemental material). RBM38 is of particular interest as this RNA-binding protein is downregulated when either Pcbp1 or Pcbp2 was depleted (data not shown), likely due to the presence of several C-rich motifs in its 3′ UTR, and has been reported to have a critical role in erythroid differentiation (39–42). These data indicated that the Rbm38 motif was not significantly enriched in the 3′ UTRs of genes impacted by Pcbp1 depletion. On the other hand, the Rbm38 motif was enriched in the 3′ UTRs of upregulated genes after Pcbp2 was depleted from the primary cells (Fig. 5D). Pathway analysis of these genes containing RBM38 binding sites in the encoded 3′ UTRs revealed significant enrichment for pathways relevant to cell death, organ development, and cell signaling (Table S5). In sum, these data support roles for PCBPs in modulating the developing erythroid transcriptome via direct recognition of the cognate poly(C)-rich motif as well as via indirect actions mediated by alterations in expression of downstream RNA-binding proteins such as RBM38.
Targeted analyses confirm redundant and unique impacts of Pcbp1 and Pcbp2 on erythroid lineage transcripts.
The global transcriptome analyses of primary hematopoietic cells depleted of Pcbp1 and Pcbp2 (described above) pointed to a combination of redundant and isoform-specific actions of the PCBPs on erythroblast gene expression (described above). To further explore this combination of mechanisms, we carried out targeted analyses of 10 genes whose encoded proteins have well-documented roles in hematopoietic lineage commitment and/or erythroid differentiation (Fig. 5E). These data revealed Pcbp1 specifically impacts 2 genes (Ldb1 and EpoR), Pcbp2 specifically impacts 2 genes (Pu.1 and Gata1), and the impacts of Pcbp1 and Pcbp2 depletions on the remaining 6 genes (Alas2, Gfi1b, Tal1, Lmo2, Runx1, and Gata2) overlap. These data support the conclusion that the two Pcbp isoforms mediate redundant as well as isoform-specific roles on erythroid lineage development.
Pcbp1 and Pcbp2 demonstrate different impacts on exon splicing in the erythroid transcriptome.
Pcbps have well-described impacts on alternative splicing pathways in multiple settings (9, 11, 26, 43, 44). With this in mind, we next analyzed our RNA-seq data for cassette exon inclusion/exclusion rates in the setting of Pcbp1 and Pcbp2 depletion in the primary hematopoietic cell culture model (Fig. 6A and B). This alternative splicing (AS) analysis revealed a greater impact by Pcbp1 compared to Pcbp2 depletions at both days 0 and 2 of erythroid induction (see Table S6 in the supplemental material). This difference was observed for enhanced as well as repressed cassette exons. Pathway analysis of these genes with AS revealed significant enrichment for pathways relevant to RNA processing, RNA splicing, erythrocyte homeostasis, and erythrocyte differentiation (see Table S7 in the supplemental material). Thus, the AS and DGE analyses were consistent in revealing that a large subset of the impacts of the two Pcbp isoforms on the transcriptome were isoform specific, that the impact of Pcbp1 had a greater impact than Pcbp2, and that the number of events impacted by Pcbp depletions increased during erythroid lineage induction.
FIG 6.
Pcbp1 and Pcbp2 have distinct impacts on alternative exon splicing in primary hematopoietic cells. (A) Enhanced exon inclusion events in cells depleted of Pcbp1 or Pcbp2 cells at days 0 and 2 of erythroid induction highlight isoform-specific functions. (Top) A Venn diagram displays the numbers of exons whose inclusion is enhanced subsequent to Pcbp1 (blue) and Pcbp2 (yellow) depletion prior to induction (day 0). Each binary comparison is noted below the corresponding circle. Exons whose splicing is impacted by both Pcbp1 and Pcbp2 depletions are shown in the overlap (white). (Bottom) A Venn diagram displays the numbers of exons whose inclusion is enhanced subsequent to Pcbp1 (blue) and Pcbp2 (yellow) depletion after 2 days of erythroid induction (day 2). Each binary comparison is noted below the corresponding circle. Exons whose splicing is impacted by both Pcbp1 and Pcbp2 depletions are shown in the overlap (white). (B) Repressed exon inclusion in cells depleted of Pcbp1 or Pcbp2 at days 0 and 2 of erythroid induction. (Top) A Venn diagram displays the numbers of exons whose inclusion is repressed subsequent to Pcbp1 (blue) and Pcbp2 (yellow) depletion prior to induction (day 0). Each binary comparison is noted below the corresponding circle. Exons whose splicing is impacted by both Pcbp1 and Pcbp2 depletions are shown in the overlap (white). (Bottom) A Venn diagram displays the numbers of exons whose inclusion is repressed subsequent to Pcbp1 (blue) and Pcbp2 (yellow) depletion after 2 days of erythroid induction (day 2). Each binary comparison is noted below the corresponding circle. Exons whose splicing is impacted by both Pcbp1 and Pcbp2 depletions are shown in the overlap (white). (C) Distribution of consensus binding sites for Pcbp1 and Pcbp2 in proximity to alternatively spliced exons impacted by Pcbp1 and Pcbp2 depletion. (Top) Frequency and positions of Pcbp1 motifs (CISBP-RNA database [38]) around the 3′ and 5′ splice sites of cassette exons whose inclusion is enhanced (red) and repressed (blue) by Pcbp1 depletion on days 0 and 2 of erythroid inductions. Frequencies were smoothed using a running mean of 20 nucleotides (nt). The Pcbp1 motif is significantly enriched in the cassette exons with repressed inclusion (blue) in the region of the splice acceptor site (nt −300 to +50; FDR corrected P = 3.23e−06) and the splice donor site (nt −50 to +300; FDR corrected P = 0.01). P values were computed using one-sided Fisher’s exact test. (Bottom) Frequency and positions of Pcbp2 motifs (CISBP-RNA database [38]) in proximity to cassette exons that are enhanced (red) and repressed (blue) upon Pcbp2 depletion at days 0 and 2 of induced erythroid differentiation. Frequencies were smoothed using a running mean of 20 nt. The apparent enrichment of C-rich motifs near the splice acceptor site of repressed exons could not be firmly established as statistically significant due to low number of events in the enhanced set. (D) RT-PCR validation of alternative splicing events. RT-PCR was performed on day 2 RNAs for alternatively spliced exons in 4 genes that have clearly defined roles in hematopoietic lineage commitment: Runx1, exon 6; EpoR, exon 5; Epb41, exon 16; Epb49, exon 3. RNAs were purified from the E14.5 mouse liver and treated with the indicated shRNAs. Control shRNAs are shown as luc1, scr1, and scr2, Pcbp1 shRNAs are shown as 1-1, 1-3, and 1-4, and Pcbp2 shRNAs are shown as 2-1, 2-3, and 2-4. Alternatively spliced exons are indicated by green boxes. The percentage of exon inclusion is calculated and shown at the bottom of the gel (average ± standard deviation). P values were calculated by comparing the impact of Pcbp2 and Pcbp1 depletion with values from parallel controls. ∗, P < 0.05; ∗∗, P < 0.01, n.s., not significant.
We next searched for evidence of direct impacts of PCBPs on splicing in the erythroid transcriptome. MEME analysis of mRNAs whose alternative exon splicing was impacted by Pcbp1 and Pcbp2 depletion revealed enrichment for C-rich motifs in the splice-proximal regions of the alternatively spliced exons (Fig. 6C). In the setting of both Pcbp1 and Pcbp2 depletions, a C-rich motif is significantly enriched in the region of the splice acceptor site of repressed cassette exons; Pcbp1 depletion was linked to significantly enriched peaks in the cassette exons with repressed inclusion (Fig. 6C, blue) adjacent to the splice acceptor sites (nucleotides [nt] −300 to +50; false-discovery rate [FDR] corrected P = 3.23e−06) and splice donor sites (nt −50 to +300; FDR corrected P = 0.01). Enriched peaks in the cassette exons with enhanced inclusion (Fig. 6C, red) in the intronic region were also visually apparent in proximity to the splice donor site (nt +150 to +300), although this did not reach statistical significance (P = not significant [NS]).
Upon Pcbp2 depletion, there were visually apparent peaks of enrichment of the C-rich motif in the cassette exons with repressed inclusion (Fig. 6C, blue) in the intronic region in proximity to the splice donor site (nt +100 to +175 and +250 to +300) and in the intronic region in proximity to the splice acceptor site (nt −50 to 0). There were also visually apparent enriched peaks in the cassette exons with enhanced inclusion (Fig. 6C, red) in the intronic region near the splice acceptor site (nt −300 to −225, −150 to −100, and −50 to 0). These enrichments of C-rich motifs associated with Pcbp2 depletion could not be established as statistically significant due to low number of events in the enhanced and repressed set.
Analysis of genes with Pcbp2-impacted cassette exon splicing revealed significant enrichment for pathways relevant to hematopoietic development, mRNA processing, cellular localization, eryhthrocyte homeostasis, and erythroid differentiation (see Table S8 in the supplemental material).
Targeted validations of alternative exon splicing were carried out on transcripts selectively sensitive to either Pcbp1 or Pcbp2. Runx1 is a critical transcription factor for blood development (45–48). The data confirmed our previous finding that AS control of exon 6 splicing is specific to Pcbp2 (26), and not responsive to depletion of Pcbp1 (Fig. 6D, Runx1). Erythropoietin receptor (EPOR) is critical for terminal erythropoiesis (23). We confirmed that Pcbp1, but not Pcbp2, specifically impacts alternative splicing of exon 5 of the EpoR gene. This event is of functional significance as loss of exon 5 is predicted to trigger an NMD response, which would be consistent with the observed downregulation of EpoR mRNA by Pcbp1 depletion (Fig. 5E). EPB41 and EPB49 are two red blood cell membrane proteins that are important/essential for terminal erythroid development (49–51). We found that Pcbp2 specifically represses inclusion of exon 16 (knockdown of Pcbp2 increases the inclusion of exon 16) of Epb41 (Fig. 6D, Epb41), an exon shown by others to be important in late stage of erythropoiesis (49, 52–55). For Epb49, the data revealed that Pcbp1 enhances splicing of exon 3, while Pcbp2 lacks an effect (Fig. 6D, Epb49). These studies further substantiate the conclusion that the PCBP1 and PCBP2 isoforms have specific impacts on gene expression that are of direct relevance to erythroid development and function.
DISCUSSION
We have previously demonstrated that the paralogous RNA-binding proteins, PCBP1 and PCBP2, are independently essential for mouse embryonic viability (21). Here, we explore their tissue-specific requirement in erythroid lineage. The EpoR-cre driver used in these studies selectively inactivates floxed loci in the erythroid lineage beginning at E12.5 (22). Conditional knockouts revealed that individual loss of Pcbp1 or Pcbp2 failed to significantly impact the developing erythroid lineage (Fig. 1 and 2), while their combined loss resulted in a substantial loss of hematopoiesis: homozygous loss of Pcbp1 in conjunction with haploidy for Pcbp2 inactivation repressed erythroid development and resulted in embryonic lethality at E14.5, and compound homozygous inactivation of Pcbp1 and Pcbp2 loci resulted in an even more severe phenotype, with dramatic loss of fetal liver erythropoiesis and embryonic demise prior to E13.5 (Fig. 3). These findings led us to conclude that formation of the erythroid lineage and embryonic development can be supported individually by either Pcbp1 or Pcbp2, while the combined loss of these two paralogs is incompatible with these critical processes.
Assessing the impact of individual depletions of Pcbp1 and Pcbp2 on the transcriptomes of primary hematopoietic progenitor cells induced toward erythroid differentiation revealed that subsets of genes were uniquely impacted by one or the other of the Pcbp isoforms, while others constitute an overlapping gene set (Fig. 5A and B). A similar combination of isoform-specific and shared (overlapping) impacts was observed in the analysis of alternative exon splicing (Fig. 6A and B). Thus, the transcriptome analyses of primary hematopoietic cells demonstrated a complex set of impacts of Pcbp1 and Pcbp2 on mRNA representation (DGE analysis) and structure (AS analysis), reflecting both isoform-specific and redundant impacts.
The motif analyses of our RNA seq data sets in primary hematopoietic precursors depleted of the individual Pcbp isoforms revealed significant enrichment of C-rich Pcbp binding sites within 3′ UTR sequences of genes that were downregulated upon Pcbp1 or Pcbp2 depletion (Fig. 5C). These data point to a direct positive role in enhancing the expression of a defined gene subset. The motif analysis of mRNAs whose exon splicing reactions were impacted by Pcbp depletion also had significant enrichment for C-rich motifs in proximity to the alternatively splice exons (Fig. 6C). This result is consistent with our prior transcriptome-wide analysis of Pcbp-controlled exon splicing in a human hematopoietic cell line (K562 cells) which identified a significant enrichment of C-rich motifs adjacent to splice sites controlled by PCBPs (11). In the current study, this impact on exonic splicing is most prominently seen in the context of depletion of Pcbp1 compared to Pcbp2 (Fig. 6C). Targeted analysis of exon splicing further linked the roles of Pcbp1- and Pcbp2-mediated AS controls to erythroid development and highlighted isoform-specific functions (Fig. 6D).
While the foregoing studies supported a role for direct Pcbp-mRNA interactions in erythroid development, the fact that these linkages were only defined in a subset of the impacted genes suggested that additional RNAs may be regulated by Pcbp-dependent indirect pathways. Of particular interest was the observation that depletion of Pcbp1 or Pcbp2 caused a significant decrease in the expression of RBM38. This RNA-binding protein has been shown by others to be involved in posttranscriptional controls linked to erythroid differentiation (39–41). The impact of Pcbp depletion on Rbm38 gene expression suggested that one or more of the actions of Pcbp genes may be indirect and mediated via RBM38 actions.
A final observation from the transcriptome analyses that merits mention is that the number of genes impacted by Pcbp1 and Pcbp2 depletions increased markedly subsequent to the 2 days of EPO-induced erythroid differentiation (Fig. 5B). These results suggest that some genes critical/important to erythroid differentiation may not be affected by Pcbp1/2 prior to differentiation (day 0), but rather are affected during the differentiation process. The impact on the alternative exon splicing of the Epb41 gene transcript is a good example of such differentiation-linked posttranscriptional regulation (Fig. 6D).
In summary, we demonstrate in the current report, using both in vivo and ex vivo models, that PCBPs work cooperatively to support pathways of gene expression that are essential to mouse erythroid development.
MATERIALS AND METHODS
Animals.
All experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee at the Perelman School of Medicine (University of Pennsylvania).
The floxed Pcbp2 mouse line has been previously described (21), and the EpoR-cre line was a kind gift from Ursula Klingmüller (22).
Generation of mouse line carrying a floxed Pcbp1 allele.
CRISPR-Cas9-mediated gene editing (56) was utilized to flank the entire intronless Pcbp1 locus with LoxP sites. Single-guide RNAs (sgRNAs) targeting the 5′ and 3′ flanking regions of Pcbp1 (sequences available on request) were selected based on principles that minimize off-target effects (57). sgRNAs and tracrRNA (Alt-R CRISPR-Cas9 tracrRNA no. 1072532) were obtained from Integrated DNA Technologies (CA). The capped and polyadenylated Cas9 mRNA (L-7206; TriLink BioTechnologies) and the homology-directed repair (HDR) templates were microinjected into the cytoplasm of single-cell C57BL/6J embryos (Jackson Laboratory, no. 000664) at the Transgenic Mouse Facility of the University of Pennsylvania. The resultant pups were screened for LoxP sequence insertion by PCR (primer sequences available upon request) and confirmed by Sanger sequencing.
Conditional inactivation of Pcbp1 and/or Pcbp2 in the erythroid lineage.
Pcbp1 and Pcbp2 loci were inactivated in the erythroid lineage by crossing the corresponding floxed Pcbp1 or Pcbp2 mouse lines with the EpoR-cre line (22). Offspring carrying the floxed Pcbp and EpoR-cre alleles were identified by PCR (21, 22).
In vitro expansion and differentiation of murine erythroblasts selectively depleted of Pcbp1 and Pcbp2.
E14.5 fetal liver cell suspension was enriched for hematopoietic progenitors using the EasySep Hematopoietic Progenitor Cell Enrichment kit (StemCell Technologies, no. 19756) supplemented with biotin-conjugated CD71 antibody (BioLegend) (34, 36). The purified hematopoietic progenitor cells were cultured in expansion medium consisting of StemPro34 medium (Invitrogen) supplemented with 2 mM l-glutamine, 1% penicillin-streptomycin, 10 μM 1-thioglycerol, 1 μM dexamethasone, 0.5 U/ml erythropoietin, and 1% murine stem cell factor (SCF) (34) and infected immediately with MSCV-PIG (murine stem cell virus-puromycin-internal ribosome entry site [IRES]-green fluorescent protein [GFP])-based retroviruses encoding shRNAs that target Pcbp1 and Pcbp2, as well as two sets of control shRNAs: luciferase and scrambled sequence (Table 1). The transduced cells were grown in expansion medium for 48 h. These cells were collected (day 0 cells) or washed and resuspended in differentiation medium (Iscove’s modified Dulbecco’s medium [IMDM] with 10% fetal calf serum [FCS], 10% potassium dodecyl sulfate [PDS], 2 mM l-glutamine, 10 μM 1-thioglycerol, 1% penicillin-streptomycin, 5% PFHM-II medium, and 5 U/ml erythropoietin [Amgen]) (34) and incubated for an additional 2 days (day 2 cells).
TABLE 1.
Primer and shRNA sequences
| Gene or shRNA | Primer sequence or shRNA designationa |
|---|---|
| Genes | |
| qRT-PCR | |
| Tal1 | F: TTAGCCAGCCGCTCGCCTCA |
| R: TGGTGAAGATGCGCCGCACT | |
| Lmo2 | F: GAAAGGAAGAGCCTGGACCC |
| R: CACCACATGTCAGCAGGGAT | |
| Ldb1 | F: TGCTGAAGTGCCACGTCTTT |
| R: CACGCTACTTCCGAAGCATTT | |
| Gfi1b | F: CAGGGACAGTGTGGAGGTTC |
| R: CTAGAAAGGACCGTGGCATT | |
| Pu.1 | F: GCTTCCCTTATCAAACCTTGTCCC |
| R: CAGGCGAATCTTTTTCTTGCTGC | |
| Alas2 | F: ATTTGGGCATAAGCAGACAC |
| R: CAGCTCCATGATTCTTCAGG | |
| Beta actin | F: AAGGAGATTACTGCTCTGGCTCCTA |
| R: ACTCATCGTACTCCTGCTTGCTGAT | |
| Runx1 | F: TCAGAAGTGTAAGCCCAGCA |
| R: CTTTCGAAAACGCACCTCTC | |
| Pcbp1 | F: AATCAATGCCAGGCTTTCCTC |
| R: TTAAAACCTGGAATTACCGACCAG | |
| Pcbp2 | F: CATGGGGAGCAGCTAGAACAGA |
| R: TTAAAACCTGGAATCGCTGACTG | |
| Gata1 | F: GCCCAAGAAGCGAATGATTG |
| R: GTGGTCGTTTGACAGTTAGTGCAT | |
| Gata2 | F: CACCCCTAAGCAGAGAAGCAA |
| R: TGGCACCACAGTTGACACACT | |
| Gapdh | F: TGTCAAGCTCATTTCCTGGTATGA |
| R: TCTTACTCCTTGGAGGCCATGT | |
| Epor | F: AGGTCCTGGAAGGCCGCACT |
| R: AGGGTCCAGGTCGCTAGCGG | |
| Alternative splicing RT-PCR | |
| mEpoR | F: CCTATGACCACCCACATCCG |
| R: GCGTCAAGATGAGAGGGTCC | |
| mDmtn | F: GGATGAGGACACCACAGAGC |
| R: GATGGCAGCCAGATCCTTGT | |
| mEpb41 | F: AGCCATTGCTCAGAGTCAGGTCACAG |
| R: CAGTTCACTGATGCTGGCATGGTGC | |
| mRunx1 | F: AATCCGCCACAAGTTGCCA |
| R: ATGGATCCCAGGTACTGGTAGG | |
| shRNAs | |
| Pcbp1 and Pcbp2 shRNAs for fetal liver cellsb | |
| shPcbp1-1 | CTCCGCTAAGAATTTAAAGAAA |
| shPcbp1-2 | GCACCGAGTGTGTGAAGCAGAT |
| shPcbp1-3 | TGCCTACCAATGCCATCTTTAA |
| shPcbp1-4 | CATGTAAGAGTGGAATGTTAAT |
| shPcbp2-1 | CAAGGAGAATCTGTTAAGAAGA |
| shPcbp2-2 | AACCGGATTCAGTGGCATTGAA |
| shPcbp2-3 | CCCGACTAATGCCATCTTCAAA |
| shPcbp2-4 | TCTGCACCAGTTGGCAATGCAA |
| shLuciferase | RHS1705 |
| shScramble | RHS4346 |
| Pcbp1 and Pcbp2 shRNAs for MEL cellsc | |
| shPcbp1 | CGGCGTGCCGCAGTCCGTCACCGAGTGTG |
| shPcbp2 | GGCCTATACCATTCAAGGACAGTATGCCA |
| shScramble | TR30013 |
F, forward; R, reverse.
The sequences shown are on the MSCV-PIG vector (Open Biosystems).
The sequences shown are on the pGFP-V-RS or p-RFP-V-RS vector (Origene).
Transcriptome analysis (RNA-seq).
One microgram of total RNA isolated at day 0 and day 2 cells from each Pcbp1- or Pcbp2-depleted sample (all RNA integrity number [RIN] values of >8.6) was used for library construction after poly(A) selection (Next Generation Sequencing Core, University of Pennsylvania). A total of 9 samples were generated from each of the two time points: three controls, Pcbp1 knockdowns with three distinctly targeting shRNAs (shPcbp1-1, -3, and -4), and Pcbp2 knockdowns with three distinctly targeting shRNAs (shPcbp2-1, -3, and -4). Sequencing was carried out using a 150-nt paired-end sequencing protocol. RNA-seq reads were aligned to GRCm38.p6 annotation using STAR after a quality check and trimming with FastQC and Trim Galore, respectively. Aligned bam files were sorted and indexed with Samtools.
Gene expression analysis.
Transcriptome indices were prepared for Salmon (https://salmon.readthedocs.io/en/latest/salmon.html) with the GRCm38 mouse annotation. Transcripts per million RNA molecules (TPMs), length-scaled TPMs, and effective length quantifications were generated using as input trimmed RNA-seq reads and the Salmon transcriptome indices. Transcript-level TPMs were collapsed into gene-level TPMs with the Bioconductor package tximport. Reads mapping to rRNA, tRNA, and mitochondrial DNA were removed. EdgeR was used to compute log2 fold change and statistical significance of differential expression. Genes with a |fold change| of >1.5 and false-discovery rate of <0.1 were considered differentially expressed.
Alternative splicing analysis.
Splicing events were quantified with MAJIQ and VOILA (58). Differential splicing was computed with the DeltaPSI module of MAJIQ. DeltaPSI quantifies differential splicing (inclusion or exclusion) of a junction between two conditions (e.g., control replicates versus Pcbp1 or Pcbp2 depletion replicates). Cassette splicing events with a |DeltaPSI| value of >0.10 were considered differentially spliced.
Motif enrichment analysis.
One thousand nucleotides in the 3′ UTR sequences were extracted from differentially upregulated, downregulated, and background (nonregulated) genes and were searched by MEME (59) for enriched motifs as defined in the CISBP-RNA database (38). Motif maps of Pcbp1, Pcbp2, and Rbm38 were constructed from the calculated frequencies of occurrence of corresponding motifs at each position in the three data sets. For plotting purposes, the frequencies were smoothed using a running mean of 50 nt.
Sequences were extracted for intronic (300-nt) and exonic (50-nt) regions flanking the 5′ and 3′ splice sites of alternatively spliced cassette exons. In order to construct motif maps, the frequencies of the CISBP-RNA motif (38) for the target RNA-binding protein (RBP) (e.g., PCBP1) in each set of cassette exons were computed at each position (nt −300 to +50 bracketing the 3′ splice sites and nt −50 to 300 bracketing 5′ splice sites). For plotting purposes, frequencies were smoothed over a running mean of 20 nt.
One-sided Fisher’s exact tests with FDR correction were performed, and motifs with an adjusted P value of <0.05 were considered significantly enriched.
mRNA quantification and RT-PCR.
Fetal liver RNA from E12.5 and E14.5 embryos was purified using TRIzol (15596026; Invitrogen). Quantitative reverse transcription-PCR (qRT-PCR) and alternative splicing RT-PCR were performed as described previously (43, 44). Alternative splicing RT-PCR primers and qRT-PCR primers are listed in Table 1.
Western blot analysis.
Total cellular extracts prepared from the TRIzol fraction of the tissue/cell preparations were analyzed as described previously (44). The PCBP1 and PCBP2 antibodies were described previously (13).
Purification of Ter119+ red blood cells.
Mature erythrocytes and erythroid precursor cells were purified from fetal livers by using Anti-Ter-119 microbeads according to the manufacturer’s instructions (catalog no. 130-049-901, magnetically activated cell sorting [MACS]; Miltenyi Biotec, Inc.).
Statistics.
Statistical significance (P values) between paired data sets was determined using two-tailed, unpaired Student's t test.
Software availability.
Detailed information about the availability and versions of software used for the bioinformatics analysis is provided in Table S1 in the supplemental material.
Accession number(s).
All sequencing data generated in this study have been deposited in the Gene Expression Omnibus (GEO) database under accession no. GSE178247 (available at http://www.ncbi.nlm.nih.gov/geo/).
ACKNOWLEDGMENTS
The authors appreciate the generosity of Liebhaber laboratory members for sharing various reagents and thoughts. We thank Daphne Yang, and Gayatri M. Schur for their help with several of the experiments.
This work was supported by a National Institutes of Health grant (MERIT grant R01HL065449 to S.A.L.). The work of Y.B. and A.J. was supported by grant R01GM128096 to Y.B.
X.J. and S.A.L. conceptualized the study, designed the experiments, and supervised the study. X.J., J.H., L.R.G., A.K., and C.D.-L. performed the experiments. E.T. aided in isolation of hematopoietic progenitors under the supervision of M.J.W. A.J. performed the bioinformatics analysis under the supervision of Y.B. X.J., A.J., and S.A.L. wrote the paper.
We declare that we have no conflicts of interest.
Footnotes
Supplemental material is available online only.
REFERENCES
- 1.Liebhaber SA. 1997. mRNA stability and the control of gene expression. Nucleic Acids Symp Ser 1997:29–32. [PubMed] [Google Scholar]
- 2.Waggoner SA, Liebhaber SA. 2003. Regulation of alpha-globin mRNA stability. Exp Biol Med (Maywood) 228:387–395. 10.1177/153537020322800409. [DOI] [PubMed] [Google Scholar]
- 3.Russell JE, Morales J, Liebhaber SA. 1997. The role of mRNA stability in the control of globin gene expression. Prog Nucleic Acids Res Mol Biol 57:249–287. 10.1016/s0079-6603(08)60283-4. [DOI] [PubMed] [Google Scholar]
- 4.Weiss IM, Liebhaber SA. 1994. Erythroid cell-specific determinants of alpha-globin mRNA stability. Mol Cell Biol 14:8123–8132. 10.1128/MCB.14.12.8123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Weiss IM, Liebhaber SA. 1995. Erythroid cell-specific mRNA stability elements in the alpha 2-globin 3′ nontranslated region. Mol Cell Biol 15:2457–2465. 10.1128/MCB.15.5.2457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Waggoner SA, Liebhaber SA. 2003. Identification of mRNAs associated with alphaCP2-containing RNP complexes. Mol Cell Biol 23:7055–7067. 10.1128/MCB.23.19.7055-7067.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Holcik M, Liebhaber SA. 1997. Four highly stable eukaryotic mRNAs assemble 3′ untranslated region RNA-protein complexes sharing cis and trans components. Proc Natl Acad Sci U S A 94:2410–2414. 10.1073/pnas.94.6.2410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ji X, Kong J, Liebhaber SA. 2011. An RNA-protein complex links enhanced nuclear 3′ processing with cytoplasmic mRNA stabilization. EMBO J 30:2622–2633. 10.1038/emboj.2011.171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ji X, Kong J, Carstens RP, Liebhaber SA. 2007. The 3′ untranslated region complex involved in stabilization of human alpha-globin mRNA assembles in the nucleus and serves an independent role as a splice enhancer. Mol Cell Biol 27:3290–3302. 10.1128/MCB.02289-05. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ji X, Kong J, Liebhaber SA. 2003. In vivo association of the stability control protein alphaCP with actively translating mRNAs. Mol Cell Biol 23:899–907. 10.1128/MCB.23.3.899-907.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ji X, Park JW, Bahrami-Samani E, Lin L, Duncan-Lewis C, Pherribo G, Xing Y, Liebhaber SA. 2016. alphaCP binding to a cytosine-rich subset of polypyrimidine tracts drives a novel pathway of cassette exon splicing in the mammalian transcriptome. Nucleic Acids Res 44:2283–2297. 10.1093/nar/gkw088. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ji X, Wan J, Vishnu M, Xing Y, Liebhaber SA. 2013. alphaCP poly(C) binding proteins act as global regulators of alternative polyadenylation. Mol Cell Biol 33:2560–2573. 10.1128/MCB.01380-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chkheidze AN, Lyakhov DL, Makeyev AV, Morales J, Kong J, Liebhaber SA. 1999. Assembly of the alpha-globin mRNA stability complex reflects binary interaction between the pyrimidine-rich 3′ untranslated region determinant and poly(C) binding protein alphaCP. Mol Cell Biol 19:4572–4581. 10.1128/MCB.19.7.4572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Makeyev AV, Chkheidze AN, Liebhaber SA. 1999. A set of highly conserved RNA-binding proteins, alphaCP-1 and alphaCP-2, implicated in mRNA stabilization, are coexpressed from an intronless gene and its intron-containing paralog. J Biol Chem 274:24849–24857. 10.1074/jbc.274.35.24849. [DOI] [PubMed] [Google Scholar]
- 15.Makeyev AV, Eastmond DL, Liebhaber SA. 2002. Targeting a KH-domain protein with RNA decoys. RNA 8:1160–1173. 10.1017/s135583820202808x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Makeyev AV, Liebhaber SA. 2000. Identification of two novel mammalian genes establishes a subfamily of KH-domain RNA-binding proteins. Genomics 67:301–316. 10.1006/geno.2000.6244. [DOI] [PubMed] [Google Scholar]
- 17.Makeyev AV, Liebhaber SA. 2002. The poly(C)-binding proteins: a multiplicity of functions and a search for mechanisms. RNA 8:265–278. 10.1017/s1355838202024627. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Charroux B, Angelats C, Fasano L, Kerridge S, Vola C. 1999. The levels of the bancal product, a Drosophila homologue of vertebrate hnRNP K protein, affect cell proliferation and apoptosis in imaginal disc cells. Mol Cell Biol 19:7846–7856. 10.1128/MCB.19.11.7846. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Colaiacovo MP, Stanfield GM, Reddy KC, Reinke V, Kim SK, Villeneuve AM. 2002. A targeted RNAi screen for genes involved in chromosome morphogenesis and nuclear organization in the Caenorhabditis elegans germline. Genetics 162:113–128. 10.1093/genetics/162.1.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Geng C, Macdonald PM. 2007. Identification of genes that influence gurken expression. Fly (Austin) 1:259–267. 10.4161/fly.5246. [DOI] [PubMed] [Google Scholar]
- 21.Ghanem LR, Kromer A, Silverman IM, Chatterji P, Traxler E, Penzo-Mendez A, Weiss MJ, Stanger BZ, Liebhaber SA. 2016. The poly(C) binding protein Pcbp2, and its retrotransposed derivative Pcbp1, are independently essential to mouse development. Mol Cell Biol 36:304–319. 10.1128/MCB.00936-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Heinrich AC, Pelanda R, Klingmuller U. 2004. A mouse model for visualization and conditional mutations in the erythroid lineage. Blood 104:659–666. 10.1182/blood-2003-05-1442. [DOI] [PubMed] [Google Scholar]
- 23.Wu H, Liu X, Jaenisch R, Lodish HF. 1995. Generation of committed erythroid BFU-E and CFU-E progenitors does not require erythropoietin or the erythropoietin receptor. Cell 83:59–67. 10.1016/0092-8674(95)90234-1. [DOI] [PubMed] [Google Scholar]
- 24.Friend C, Scher W, Holland JG, Sato T. 1971. Hemoglobin synthesis in murine virus-induced leukemic cells in vitro: stimulation of erythroid differentiation by dimethyl sulfoxide. Proc Natl Acad Sci U S A 68:378–382. 10.1073/pnas.68.2.378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kiledjian M, Wang X, Liebhaber SA. 1995. Identification of two KH domain proteins in the alpha-globin mRNP stability complex. EMBO J 14:4357–4364. 10.1002/j.1460-2075.1995.tb00110.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ghanem LR, Kromer A, Silverman IM, Ji X, Gazzara M, Nguyen N, Aguilar G, Martinelli M, Barash Y, Liebhaber SA. 2018. Poly(C)-binding protein Pcbp2 enables differentiation of definitive erythropoiesis by directing functional splicing of the Runx1 transcript. Mol Cell Biol 38:e00175-18. 10.1128/MCB.00175-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Waggoner SA, Johannes GJ, Liebhaber SA. 2009. Depletion of the poly(C)-binding proteins alphaCP1 and alphaCP2 from K562 cells leads to p53-independent induction of cyclin-dependent kinase inhibitor (CDKN1A) and G1 arrest. J Biol Chem 284:9039–9049. 10.1074/jbc.M806986200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gamarnik AV, Andino R. 1997. Two functional complexes formed by KH domain containing proteins with the 5′ noncoding region of poliovirus RNA. RNA 3:882–892. [PMC free article] [PubMed] [Google Scholar]
- 29.Tang YS, Khan RA, Zhang Y, Xiao S, Wang M, Hansen DK, Jayaram HN, Antony A. 2011. Incrimination of heterogeneous nuclear ribonucleoprotein E1 (hnRNP-E1) as a candidate sensor of physiological folate deficiency. J Biol Chem 286:39100–39115. 10.1074/jbc.M111.230938. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Ryu MS, Zhang D, Protchenko O, Shakoury-Elizeh M, Philpott CC. 2017. PCBP1 and NCOA4 regulate erythroid iron storage and heme biosynthesis. J Clin Invest 127:1786–1797. 10.1172/JCI90519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Perrotti D, Calabretta B. 2002. Post-transcriptional mechanisms in BCR/ABL leukemogenesis: role of shuttling RNA-binding proteins. Oncogene 21:8577–8583. 10.1038/sj.onc.1206085. [DOI] [PubMed] [Google Scholar]
- 32.Woolaway K, Asai K, Emili A, Cochrane A. 2007. hnRNP E1 and E2 have distinct roles in modulating HIV-1 gene expression. Retrovirology 4:28. 10.1186/1742-4690-4-28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Chaudhury A, Hussey GS, Ray PS, Jin G, Fox PL, Howe PH. 2010. TGF-beta-mediated phosphorylation of hnRNP E1 induces EMT via transcript-selective translational induction of Dab2 and ILEI. Nat Cell Biol 12:286–293. 10.1038/ncb2029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Khandros E, Thom CS, D'Souza J, Weiss MJ. 2012. Integrated protein quality-control pathways regulate free α-globin in murine β-thalassemia. Blood 119:5265–5275. 10.1182/blood-2011-12-397729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Zhang J, Socolovsky M, Gross AW, Lodish HF. 2003. Role of Ras signaling in erythroid differentiation of mouse fetal liver cells: functional analysis by a flow cytometry-based novel culture system. Blood 102:3938–3946. 10.1182/blood-2003-05-1479. [DOI] [PubMed] [Google Scholar]
- 36.Pimkin M, Kossenkov AV, Mishra T, Morrissey CS, Wu W, Keller CA, Blobel GA, Lee D, Beer MA, Hardison RC, Weiss MJ. 2014. Divergent functions of hematopoietic transcription factors in lineage priming and differentiation during erythro-megakaryopoiesis. Genome Res 24:1932–1944. 10.1101/gr.164178.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.An X, Schulz VP, Li J, Wu K, Liu J, Xue F, Hu J, Mohandas N, Gallagher PG. 2014. Global transcriptome analyses of human and murine terminal erythroid differentiation. Blood 123:3466–3477. 10.1182/blood-2014-01-548305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Ray D, Kazan H, Cook KB, Weirauch MT, Najafabadi HS, Li X, Gueroussov S, Albu M, Zheng H, Yang A, Na H, Irimia M, Matzat LH, Dale RK, Smith SA, Yarosh CA, Kelly SM, Nabet B, Mecenas D, Li W, Laishram RS, Qiao M, Lipshitz HD, Piano F, Corbett AH, Carstens RP, Frey BJ, Anderson RA, Lynch KW, Penalva LOF, Lei EP, Fraser AG, Blencowe BJ, Morris QD, Hughes TR. 2013. A compendium of RNA-binding motifs for decoding gene regulation. Nature 499:172–177. 10.1038/nature12311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Alvarez-Dominguez JR, Zhang X, Hu W. 2017. Widespread and dynamic translational control of red blood cell development. Blood 129:619–629. 10.1182/blood-2016-09-741835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Heinicke LA, Nabet B, Shen S, Jiang P, van Zalen S, Cieply B, Russell JE, Xing Y, Carstens RP. 2013. The RNA binding protein RBM38 (RNPC1) regulates splicing during late erythroid differentiation. PLoS One 8:e78031. 10.1371/journal.pone.0078031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Ulirsch JC, Nandakumar SK, Wang L, Giani FC, Zhang X, Rogov P, Melnikov A, McDonel P, Do R, Mikkelsen TS, Sankaran VG. 2016. Systematic functional dissection of common genetic variation affecting red blood cell traits. Cell 165:1530–1545. 10.1016/j.cell.2016.04.048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Zhang J, Xu E, Ren C, Yan W, Zhang M, Chen M, Cardiff RD, Imai DM, Wisner E, Chen X. 2014. Mice deficient in Rbm38, a target of the p53 family, are susceptible to accelerated aging and spontaneous tumors. Proc Natl Acad Sci U S A 111:18637–18642. 10.1073/pnas.1415607112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Ji X, Humenik J, Liebhaber SA. 2019. A cytosine-rich splice regulatory determinant enforces functional processing of the human α-globin gene transcript. Blood 133:2338–2347. 10.1182/blood-2018-12-891408. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Ji X, Humenik J, Yang D, Liebhaber SA. 2018. PolyC-binding proteins enhance expression of the CDK2 cell cycle regulatory protein via alternative splicing. Nucleic Acids Res 46:2030–2044. 10.1093/nar/gkx1255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Chen MJ, Yokomizo T, Zeigler BM, Dzierzak E, Speck NA. 2009. Runx1 is required for the endothelial to haematopoietic cell transition but not thereafter. Nature 457:887–891. 10.1038/nature07619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Okuda T, van Deursen J, Hiebert SW, Grosveld G, Downing JR. 1996. AML1, the target of multiple chromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis. Cell 84:321–330. 10.1016/S0092-8674(00)80986-1. [DOI] [PubMed] [Google Scholar]
- 47.Tober J, Maijenburg MW, Speck NA. 2016. Taking the leap: Runx1 in the formation of blood from endothelium. Curr Top Dev Biol 118:113–162. 10.1016/bs.ctdb.2016.01.008. [DOI] [PubMed] [Google Scholar]
- 48.Wang Q, Stacy T, Binder M, Marin-Padilla M, Sharpe AH, Speck NA. 1996. Disruption of the Cbfa2 gene causes necrosis and hemorrhaging in the central nervous system and blocks definitive hematopoiesis. Proc Natl Acad Sci U S A 93:3444–3449. 10.1073/pnas.93.8.3444. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Conboy J, Kan YW, Shohet SB, Mohandas N. 1986. Molecular cloning of protein 4.1, a major structural element of the human erythrocyte membrane skeleton. Proc Natl Acad Sci U S A 83:9512–9516. 10.1073/pnas.83.24.9512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Rana AP, Ruff P, Maalouf GJ, Speicher DW, Chishti AH. 1993. Cloning of human erythroid dematin reveals another member of the villin family. Proc Natl Acad Sci U S A 90:6651–6655. 10.1073/pnas.90.14.6651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Siegel DL, Branton D. 1985. Partial purification and characterization of an actin-bundling protein, band 4.9, from human erythrocytes. J Cell Biol 100:775–785. 10.1083/jcb.100.3.775. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Conboy J, Mohandas N, Tchernia G, Kan YW. 1986. Molecular basis of hereditary elliptocytosis due to protein 4.1 deficiency. N Engl J Med 315:680–685. 10.1056/NEJM198609113151105. [DOI] [PubMed] [Google Scholar]
- 53.Conboy JG, Chan J, Mohandas N, Kan YW. 1988. Multiple protein 4.1 isoforms produced by alternative splicing in human erythroid cells. Proc Natl Acad Sci U S A 85:9062–9065. 10.1073/pnas.85.23.9062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Conboy JG, Chan JY, Chasis JA, Kan YW, Mohandas N. 1991. Tissue- and development-specific alternative RNA splicing regulates expression of multiple isoforms of erythroid membrane protein 4.1. J Biol Chem 266:8273–8280. 10.1016/S0021-9258(18)92973-X. [DOI] [PubMed] [Google Scholar]
- 55.Huang SC, Zhang HS, Yu B, McMahon E, Nguyen DT, Yu FH, Ou AC, Ou JP, BenzEJ, Jr.. 2017. Protein 4.1R exon 16 3′ splice site activation requires coordination among TIA1, Pcbp1, and RBM39 during terminal erythropoiesis. Mol Cell Biol 37:e00446-16. 10.1128/MCB.00446-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56.Ran FA, Hsu PD, Wright J, Agarwala V, Scott DA, Zhang F. 2013. Genome engineering using the CRISPR-Cas9 system. Nat Protoc 8:2281–2308. 10.1038/nprot.2013.143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Doench JG, Fusi N, Sullender M, Hegde M, Vaimberg EW, Donovan KF, Smith I, Tothova Z, Wilen C, Orchard R, Virgin HW, Listgarten J, Root DE. 2016. Optimized sgRNA design to maximize activity and minimize off-target effects of CRISPR-Cas9. Nat Biotechnol 34:184–191. 10.1038/nbt.3437. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.Vaquero-Garcia J, Barrera A, Gazzara MR, González-Vallinas J, Lahens NF, Hogenesch JB, Lynch KW, Barash Y. 2016. A new view of transcriptome complexity and regulation through the lens of local splicing variations. eLife 5:e11752. 10.7554/eLife.11752. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Bailey TL, Boden M, Buske FA, Frith M, Grant CE, Clementi L, Ren J, Li WW, Noble WS. 2009. MEME SUITE: tools for motif discovery and searching. Nucleic Acids Res 37:W202–W208. 10.1093/nar/gkp335. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Download MCB.00668-20-s0001.xlsx, XLSX file, 0.01 MB (9.7KB, xlsx)
Table S2. Download MCB.00668-20-s0002.xlsx, XLSX file, 0.6 MB (627.7KB, xlsx)
Table S3. Download MCB.00668-20-s0003.xlsx, XLSX file, 0.3 MB (345.8KB, xlsx)
Table S4. Download MCB.00668-20-s0004.xlsx, XLSX file, 0.1 MB (147.8KB, xlsx)
Table S5. Download MCB.00668-20-s0005.xlsx, XLSX file, 0.06 MB (66.7KB, xlsx)
Table S6. Download MCB.00668-20-s0006.xlsx, XLSX file, 0.2 MB (182.4KB, xlsx)
Table S7. Download MCB.00668-20-s0007.xlsx, XLSX file, 0.10 MB (98.7KB, xlsx)
Table S8. Download MCB.00668-20-s0008.xlsx, XLSX file, 0.08 MB (85.1KB, xlsx)