Zfp206 Is a Transcription Factor That Controls Pluripotency of Embryonic Stem Cells

Authors


Abstract

Zfp206 (ZNF206 in human) encodes a zinc finger- and SCAN domain-containing protein that is highly expressed in pluripotent ESC. Upon differentiation of human and mouse ESC, Zfp206 expression is quickly repressed. Zfp206 was found to be expressed throughout embryogenesis but absent in adult tissues except testis. We have identified a role for Zfp206 in controlling ESC differentiation. ESC engineered to overexpress Zfp206 were found to be resistant to differentiation induced by retinoic acid. In addition, ESC with knocked-down expression of Zfp206 were more sensitive to differentiation by retinoic acid treatment. We found that Zfp206 was able to enhance expression from its own promoter and also activate transcription of the Oct4 and Nanog promoters. Our results show that Zfp206 is an embryonic transcription factor that plays a role in regulating pluripotency of embryonic stem cells.

Disclosure of potential conflicts of interest is found at the end of this article.

Introduction

Embryonic stem cells (ESC) are in vitro equivalents of epiblast cells from the inner cell mass (ICM) of mammalian blastocysts [1]. ESC have the defining properties that they are pluripotent, thus giving rise to all cell types of the organism, and the capacity for unlimited self-renewal [2]. ESC provide an opportunity to characterize the molecular pathways that regulate the maintenance of the undifferentiated, pluripotent state and the earliest steps in commitment toward the various tissue lineages [3, 4].

Significant progress has been made recently in elucidating the transcriptional networks operative in ESC. Technological advances in gene expression profiling have made it possible to comprehensively identify the genes expressed in ESC and their differentiated derivatives [5, [6]–7]. Results from these genome-wide studies have identified a number of transcription factors that are uniquely expressed in the pluripotent state and downregulated upon differentiation, raising the possibility that they play a role in regulating differentiation. Among these are three transcription factors, Oct4, Sox2, and Nanog, that are essential for maintaining pluripotency of ESC and early embryos [8, [9], [10], [11]–12]. Hundreds of direct target genes that are regulated by Oct4, Sox2, and Nanog have been identified by chromatin immunoprecipitation experiments [13, 14]. However, it remains unclear which of these target genes are essential for regulating differentiation.

We have analyzed the ESC-derived data from the comprehensive expression and transcription factor binding site mapping studies described above to identify additional regulators of pluripotency. Our search focused on genes that encode transcription factors that were (a) expressed in undifferentiated human and mouse ESC, (b) repressed in differentiated ESC, and (c) direct targets of Oct4, Sox2, and/or Nanog. Described here is a functional characterization of one transcription factor, Zfp206, that fits these selection criteria. Our functional analysis of this transcription factor indicates that it acts to maintain ESC in a pluripotent state.

Materials and Methods

Cell Culture

E14 mouse ES cells were cultured as described previously [20]. H1 human ESC line (WiCell Research Institute, Madison, WI, http://www.wicell.org) was passaged as described [20]. Culture of other cell lines, differentiation protocol, alkaline phosphatase (AP) staining procedures, and colony forming assay procedures are described in the supplemental online Materials and Methods.

Gene Expression Studies

Northern blot was carried out with a 60-base pair (bp) oligonucleotide (TCGGAAGGATCGAGAGCTTGGGTGAGGTTGCAAGTTGGGGT-GAGATTGTAACTGGTGGGC). In situ hybridization was carried out with a probe targeting the last exon of Zfp206. Real-time polymerase chain reaction (RT-PCR) was carried out using TaqMan MGB probe (Applied Biosystems). Full protocols are described in the supplemental online Materials and Methods.

Immunostaining

Staining of blastocysts was carried out with an antibody targeting a region between the SCAN and C2H2 domains of Zfp206. Synthesis of antibodies and the full protocol for immunostaining are described in the supplemental online Materials and Methods.

Construction of Plasmids

The construction of plasmids that overexpress or knock down Zfp206 expression can be found in the supplemental online Materials and Methods. Essentially, a 19-bp sequence TGCAGAATGCCTGACAATGA was used for Zfp206 knockdown.

RNA Interference

The sequence of RNA interference (RNAi) oligonucleotides for GFP, Pou5f1, and Sox2 was described previously [15].

Luciferase Reporter Assays

Dual Luciferase System (Promega, Madison, WI, http://www.promega.com) was used. The different constructs used and other details are described in the supplemental online Materials and Methods.

Statistical Analyses

Data are shown as averages ± SD. Student's nonpaired t test was used to determine the statistical significance, where indicated. All student statistical analyses were done with Excel 2003 (Microsoft, Redmond, WA, http://www.microsoft.com) with the GraphPad online software (GraphPad Software, Inc., San Diego, http://www.graphpad.com).

Results

Zfp206 Is Expressed in Pluripotent Embryonic Stem Cells

Our comprehensive gene expression profiling of human and mouse ESC lines has led to identification of hundreds of genes that are differentially expressed between undifferentiated and differentiated states [5, 6]. The expression data led us to one particularly interesting and uncharacterized gene called Zfp206 (ZNF206 in human), which was highly expressed in ESC and quickly repressed upon differentiation induced by several independent methods (Fig. 1). Deep sequencing of cDNA libraries from human ESC revealed 670 tags per million (TPM) transcripts for ZNF206 in the undifferentiated cells, but none were identified from the three libraries made from differentiated cells (Fig. 1A). A query of expression data derived by massively parallel signature sequencing (MPSS) revealed 92 TPM in human ESC for ZNF206, and none were detected in EB generated from these ESC (Fig. 1A). Likewise, MPSS data of mouse ESC and three differentiated derivatives of these cells indicated that Zfp206 was repressed upon differentiation (Fig. 1A). It was found that Zfp206 is expressed in mouse EB but not in EB derived from human ESC. Although there is a possibility that this is due to evolutionary differences in Zfp206 expression in mouse and human EB, we do not think this is likely since the results presented here were obtained from 4 days and 12 days aggregated mouse and human EB, respectively. Such a difference in the stage of EB may account for the difference of Zfp206 expression. The differentiation-dependent expression pattern of Zfp206 was confirmed in human ESC, mouse ESC, and a human embryonal carcinoma cell line, N-TERA2, by quantitative RT-PCR (Q-RT-PCR). Each cell line was treated with retinoic acid (RA) for up to 8 days to induce differentiation. Zfp206 levels were repressed in all cell lines by RA (Fig. 1B). The decrease in Zfp206 expression was similar to that observed for the pluripotency genes Oct4 and Nanog (Fig. 1B). Northern blots and Q-RT-PCR demonstrated that Zfp206 expression was also repressed in ESC differentiated by hexamethylene bis-acetamide, by dimethyl sulfoxide, and by aggregation into EB (Fig. 1C; supplemental online Fig. 1).

Figure Figure 1..

Expression of Zfp206 in undifferentiated ESC. (A): The number of tags representing transcripts from ZNF206 (human) and Zfp206 (mouse) per million total transcripts expressed in undifferentiated ESC and variously differentiated ESC as determined by MPSS [6] and EST sequencing [5]. Differentiated cells were EB, DMSO- or RA-treated EPL, and NS. (B): Gene expression levels of Zfp206, Oct4 (Pou5f1), and Nanog were determined by quantitative real-time polymerase chain reaction in a human ESC line (H1), mouse ESC, and a human embryonal carcinoma line (NTERA2) treated with RA for the times indicated. Data (representative of five separate experiments) are expressed as average expression levels ± SD, relative to untreated cells on day 0. (C): Northern blot of Zfp206 RNA in mouse ESC, undifferentiated and differentiated for 8 days as indicated. (D): Western blot of Zfp206 in mouse ESC induced to differentiate by addition of RA in the presence (+) and absence (−) of leukemia inhibitory factor for the number of days indicated. Blots were first probed with a Zfp206 antibody (Ab-NT1) and reprobed with a tubulin antibody. Abbreviations: DMSO, dimethyl sulfoxide; E14, embryonic day 14; EPL, epiblast-like cells; ES, embryonic stem; EST, expressed sequence tag; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; MPSS, massively parallel signature sequencing; NS, neurospheres; RA, retinoic acid.

To evaluate Zfp206 protein in ESC, rabbit antisera were generated against the 134-amino acid (aa) N-terminal sequence of mouse Zfp206 (Ab-NT1) and a 14-aa peptide (Ab-P1). These antisera detected a doublet of ∼100 kDa by Western blot of mouse ES cell extracts (Fig. 1D). The specificities of these antisera were confirmed, as they detect exogenous Zfp206 that is similar in size to the endogenous protein (supplemental online Fig. 2). The vector-expressed protein appeared as a doublet of ∼100 kDa when blotted with the Zfp206 antibodies or a myc antibody that recognizes an epitope tag placed on the C-terminus of Zfp206. Immunoblotting with these sera showed that levels of Zfp206 protein decreased in mouse ESC differentiated by RA in the presence or absence of leukemia inhibitory factor (Fig. 1D). Thus, expressions of both Zfp206 mRNA and protein are similarly altered during ESC differentiation, with the highest levels observed in the undifferentiated state.

The expression of Zfp206 was highly restricted to embryonic development. Northern blot analysis of 13 adult tissues indicated that Zfp206 was expressed only in testis (Fig. 2A). These results were confirmed by the more sensitive Q-RT-PCR, although just barely detectable expression was observed in the adult brain and liver, which was not detected by Northern (data not shown). We next determined the expression of Zfp206 during embryonic development at daily intervals from E4.5 to E18.5 by Q-RT-PCR (not shown) and Northern (Fig. 2B). Zfp206 was expressed at each stage, with somewhat higher levels seen at E10.5–E13.5.

Figure Figure 2..

Expression of Zfp206 in mouse embryos and adult tissues. (A): Zfp206 mRNA detection in adult mouse tissues. Northern blot analysis was performed by sequential hybridization for Zfp206 and Gapdh. Total RNA was extracted from the following: 1, brain; 2, heart; 3, lung; 4, liver; 5, spleen; 6, kidney; 7, stomach; 8, small intestine; 9, skeletal muscle; 10, thymus; 11, testis; 12, uterus; and 13, placenta. (B): Zfp206 mRNA detection in mouse embryos by Northern blot. Analysis was by sequential hybridization with an RNA probe for mouse Zfp206 and Gapdh control. RNA was extracted from pooled mouse embryos collected at the indicated day post coitum. (C–R): Sectioned E7.5 embryos (C–F) and whole mount E9.5–E12.5 embryos (G–R) were analyzed for Zfp206 expression by RNA in situ hybridization. Fgf8 was used as a positive control for a regionally restricted gene expression pattern, and a no-probe hybridization was performed as Neg. Sectioned embryos were analyzed in both sagittal (C, D, E) and transverse (F, F′) planes. Abbreviations: al, allantoic bud; E, embryonic day; ee, embryonic ectoderm; een, embryonic endoderm; epc, ectoplacental cone; exo, extraembryonic ectoderm; fl, forelimb bud; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; hl, hind limb bud; iem, intraembryonic mesoderm; Neg, negative control.

The restricted expression of Zfp206 in the embryo prompted us to perform in situ hybridizations to assess the spatial distribution of this gene's expression in embryos between days E7.5 and E12.5. At E7.5, expression of Zfp206 was clearly visible in embryonic tissue but absent from both extraembryonic and maternal tissues (Fig. 2C, 2F). It was visible in embryonic ectoderm and adjacent intraembryonic mesoderm at lower levels. There was no staining in extraembryonic endoderm and ectoderm, posterior amniotic fold, nascent allantois, or ectoplacental cone. At later stages (E9.5) Zfp206 showed low expression levels throughout most of the embryo, regardless of germ layer (Fig. 2G). This pattern was maintained through E12.5, after which expression of Zfp206 dropped to near background levels (Fig. 2J, 2M, 2P). Exceptions were mesenchymal cells of the forelimb and hind limb buds, which showed elevated expression from E10.5 to E12.5. It has been shown previously that Zfp206 is expressed in the epiblast of preimplantation embryos but not in the trophectoderm (TE) [16]. The temporal and spatial distribution of Zfp206 expression during postimplantation development indicates that the function of this gene is primarily restricted to the embryo, although not to any particular germ lineage.

In the preimplantation embryo, it has been previously reported that Zfp206 mRNA is most abundant at the two-cell stage [17] and preferentially localized to the ICM [16]. To locate Zfp206 protein within the blastocyst, we used our Zfp206 antisera (Ab-P1) and indirect immunofluorescence confocal microscopy (Fig. 3A–3C). In fully expanded blastocysts, Zfp206 was present in the nucleus of all cells, both TE and ICM. This pattern was consistent in all eight blastocysts studied. The discrepancy between this and the preferential localization of Zfp206 mRNA to the ICM suggests that ZFP206 is relatively stable and remains abundant in trophectoderm despite reduced mRNA levels. In morulas, all nuclei were positive for Zfp206 (data not shown). This panblastocyst expression suggests that Zfp206 function is not restricted to the epiblast lineage in the preimplantation embryo.

Figure Figure 3..

Localization of Zfp206 protein in the blastocyst and ESC. (A–C): Indirect immunofluorescence and confocal microscopy produced an optical section through the middle of a single fully expanded blastocyst stained with Hoechst (A), Zfp206 antibody Ab-P1 (B), and an overlay of these two (C). Note that the inner cell mass is located at the top of the image and that the cells within the blastocoel are a result of a mural trophectoderm in-fold occurring during the mounting process. (D): Schematic diagram of the protein and gene structure of Zfp206. (E–L): Subcellular localization of Zfp206 protein in mouse ESC. Cells were analyzed by immunostaining with a Zfp206 antibody, Ab-P1 (green), which showed exclusive nuclear staining (F, J). 4,6-Diamidino-2-phenylindole (blue) was used as a specific nuclear label (E, I), and phalloidin (red) served as specific stain for the cytoskeleton of the cytoplasm (G, K). The overlay of all three images shows the localized expression of Zfp206 in the nucleus (H, L). Results were similar between embryonic day 14 (E14) cells (E–H) and E14 cells that were transiently transfected to overexpress Zfp206 (I–L). Abbreviations: ORF, open reading frame; ZnF, zinc finger.

Zfp206 Encodes a Zinc Finger- and SCAN Domain-Containing Transcription Factor

We have isolated several full-length mouse cDNAs for Zfp206 from ESC mRNA. Our longest cDNA (isoform 1) encodes a protein of 782 amino acids that contains a SCAN domain and fourteen C2H2-type zinc fingers (Fig. 3D). Zfp206 is highly conserved across mammalian genomes, particularly in the SCAN and zinc finger domains, with overall identity of 72% between mouse and human (supplemental online Fig. 3). We isolated several variant cDNAs of Zfp206 that indicated the expression of multiple mRNA splice forms in mouse ESC (supplemental online Fig. 4). Several of these (isoforms 4–7) were generated by RT-PCR with primers that amplified only the predicted coding regions, so the 5′ and 3′ exonic structures of these remain to be determined. We have confirmed by RT-PCR and sequencing that these alternative splice forms are expressed in mouse ESC (E14). A recent report has identified at least six mRNA isoforms of Zfp206 [17], and we have evidence for at least seven splice variants (supplemental online Fig. 4), which together represent at least nine isoforms. All but one of the isoforms contains exon 2, which encodes the SCAN domain. There are alternative forms of exon 6, which encodes the zinc finger domains, resulting in proteins with 5–14 zinc fingers.

SCAN-containing proteins that have multiple zinc finger domains represent a subgroup of eukaryotic transcription factors implicated to play roles in cell survival and differentiation [18]. Thus, the presence of these structural domains in Zfp206 strongly suggests that it acts as a transcription factor. Immunostaining of ESC supports this, as endogenous Zfp206 staining was specific to the nucleus (Fig. 3E–3H). Exogenously expressed Zfp206 also localized to the nucleus (Fig. 3I–3L). The nuclear localization of Zfp206 is consistent with a role as a transcriptional regulator.

Zfp206 Is a Regulator of ESC Differentiation

Expression of Zfp206 correlated tightly with the differentiation state of ESC. Its expression in undifferentiated ESC and repression upon differentiation suggested that it may play a role in maintaining pluripotency. We investigated this possibility by examining the effect of Zfp206 overexpression (OE) and expression knockdown (KD) in ESC. To this end, mouse ESC were infected with a lentiviral construct to generate a pool of cells that stably express Zfp206 (Fig. 4A). To achieve the highest possible expression, infected cells were sorted three times by fluorescence-activated cell sorting (FACS) for the highest levels of vector-expressed Venus protein (details given in the supplemental online Materials and Methods). It is likely that the selected pool of stably transfected cells is highly polyclonal, as we never sorted fewer than 20% of cells in any round of sorting and the total number of cells collected was at least 5 × 104 for each round. From two independent sorting experiments, we achieved a Zfp206 RNA expression level that was 13-fold and 8-fold higher than that of wild-type ESC or cells infected and sorted with empty vector (EV control) (Fig. 4C). Western blots confirmed that Zfp206 protein levels were elevated in the pool of 206-OE cells (Fig. 4E). The 206-OE cells appeared morphologically normal, with no evidence of differentiation (Fig. 5A). FACS analysis of 206-OE cells showed normal expression of pluripotency markers SSEA1 and Oct4 (data not shown). We looked more carefully by Q-RT-PCR for changes in expression of genes that are indicative of ESC differentiation. No altered expression was observed in 206-OE cells for several genes associated with pluripotency, including Oct4, Sox2, Nanog, Esrrb, and Tdgf1 (Fig. 5B). In addition, none of the lineage markers of differentiation tested were induced in 206-OE cells (Fig. 5B). These data indicated that Zfp206 overexpression did not induce the ESC to differentiate. Interestingly, for a few differentiation markers, a slight repression was consistently observed in 206-OE cells, including the lineage markers for endoderm (Gata4, Gata6, Mixl1, and Ipf1), trophectoderm (Cdx2), and primitive ectoderm (Fgf5) (Fig. 5B). These genes are typically expressed at low levels in ESC, and it is not clear what the significance of further reduced levels is. Possible explanations for these expression differences are that the 206-OE cells have fewer spontaneously differentiated cells in the culture or that Zfp206 suppresses expression of these genes.

Figure Figure 4..

OE and KD expression of Zfp206 in mouse ESC. (A): A full-length cDNA for Zfp206 was inserted into the lentiviral vector CS-II for sustained OE in stably infected embryonic day 14 cells. Transgene expression was driven from the EF1a promoter, and the Venus fluorescent protein was expressed from an IRES for sorting of green fluorescent protein (GFP)-expressing cells. (B): The lentiviral vector pLL3.7 was constructed to express a short hairpin RNA (shRNA) that targets Zfp206 for expression KD. The shRNA was expressed from the U6 promoter, and a CMV promoter drove expression of the GFP for selection of GFP-positive cells. (C, D): Expression of Zfp206 was assessed by quantitative real-time polymerase chain reaction in pools of cells that were enriched for high-level OE in lentiviral infected ESC, Zfp206-OE (C), and knocked-down expression by RNA interference, Zfp206-KD (D). The results are expressed as the average ± SD of three separate experiments (normalized to cells expressing empty vector) and shown to be statistically significant by Student's t test (∗, p < .05; ∗∗, p < .01). (E, F): Levels of Zfp206 protein in the pool of OE and KD cells was assessed by Western blot analysis with Zfp206 antibody, Ab-NT1. Triplicate preparations of cell extracts were analyzed in the Zfp206-OE cells (E) and the Zfp206-KD cells (F). Cells infected with an empty vector control were analyzed in parallel. The blots were subsequently probed with anti-tubulin to establish equivalency of sample loading. Abbreviations: bp, base pairs; IRES, internal ribosome entry site; KD, knockdown; OE, overexpression.

Figure Figure 5..

Characterization of ESC with Zfp206-OE and expression KD. (A): Photographs of cultured Zfp206-OE and Zfp206-KD cells. The cell morphology is indistinguishable from control cells that contain empty vectors only. (B): Expression of lineage markers was assessed by quantitative real-time polymerase chain reaction in the Zfp206-OE and Zfp206-KD cells. The marker genes tested are indicated and grouped together by specificity of lineage. Data are presented as expression level relative to the corresponding empty vector cells (the average of three biological replicates). (C): Colony-plating assay of Zfp206-OE and Zfp206-KD cells. The numbers of cells indicated were plated at low density to generate individual colonies that were scored for differentiation by alkaline phosphatase staining. Data are expressed as the average ± SD of three separate experiments and shown statistically significant according to Student's t test (∗, p < .05; ∗∗, p < .01) between empty vector and Zfp206-OE or -KD cells. Abbreviations: KD, knockdown; OE, overexpression.

To test whether sustained overexpression of Zfp206 had an impact on ESC, we performed a colony-plating assay. This assay scores for the ability of ESC to seed colonies of fully undifferentiated cells as assessed by alkaline phosphatase staining (AP+). 206-OE and EV control cells were grown as undifferentiated colonies on feeder cells and then reseeded at low densities without feeders. Threefold more undifferentiated colonies emerged from the 206-OE cells than the EV control cells (p < .01) (Fig. 5C). These results indicate that Zfp206 contributes to maintenance of ESC pluripotency.

We next determined whether sustained expression of Zfp206 could repress differentiation. The 206-OE cells were cultured in the presence of RA. The EV control cells undergo typical differentiation in response to RA, as evidenced by a robust morphological change (Fig. 6A) and repression of pluripotency markers (data not shown). Cultures of 206-OE cells were treated with RA for 6 days, and it was readily apparent by morphology and AP staining that differentiation was inhibited, although not completely blocked (Fig. 6A). Because the cultures were passaged on alternate days, it was not possible to score individual colonies for differentiation. Therefore, to quantify differentiation, we scanned wells to measure the AP-stained area in each of the cultures using fluorescence imaging microscopy. Substantially more AP+ cells (approximately twofold) were observed in 206-OE cultures relative to EV control cultures (Fig. 6C). We also performed FACS analysis on these cells to assess the changes in pluripotency marker Oct4. As expected, there were very few (∼20%) Oct4-expressing cells for the EV control, indicating a robust differentiation after 6 days of exposure to RA. In contrast, after RA treatment, the numbers of Oct4+ cells was higher (∼60%) in 206-OE cells, indicating that differentiation was impeded (Fig. 6E). To confirm the altered differentiation state of the 206-OE cells, expression of additional pluripotency markers was assessed by Q-RT-PCR (Fig. 6G). Over the 6 days, Zfp206 levels were 15–50-fold higher in the 206-OE cells relative to EV-OE cells. All the pluripotency markers tested—Oct4, Nanog, Sox2, Otx2, Ebaf, and Esrrb—were expressed at substantially higher levels throughout the time course of differentiation in the 206-OE cultures relative to EV control. The observed changes in RNA expression were similarly reflected in the changes in protein levels determined by Western blots of Oct4, Sox2, and Nanog over the same time course of differentiation in these cells (supplemental online Fig. 5). In our differentiation experiments (Figs. 5, 6), no significant differences were observed between EV control cells and the wild-type ESC (data not shown). The OE data presented here were from cells that overexpressed Zfp206 RNA at 13-fold higher levels than control cells. However, similar results were obtained from independently sorted cells that overexpressed Zfp206 RNA at eightfold higher levels (data not shown) It is not clear whether high-level overexpression is required or whether simply sustaining endogenous levels would be sufficient. These data indicate that sustained overexpression of Zfp206 renders ESC less responsive to differentiation and further strengthens our hypothesis that Zfp206 plays a role in maintaining pluripotency.

Figure Figure 6..

Impact of Zfp206 on ESC differentiation. (A, B): Cultures of Zfp206-OE, Zfp206-KD cells, and empty vector controls were treated with retinoic acid (RA) for 6 days. Cultures were stained for AP+ to identify undifferentiated cells. Representative photomicrographs are shown for all four cell lines. (C, D): Twenty-five individual fields for each cell line from one of three replicate experiments were evaluated by fluorescence microscope imaging to assess the percentage of cells that were AP+. The data are expressed as the average percentage ± SD of the AP+ areas for each line and shown statistically significant by Student's t test (∗, p < .05; ∗∗, p < .01) between empty vector and Zfp206-OE or -KD cells. (E, F): The extent of differentiation of Zfp206-OE and Zfp206-KD cells after 6 days of RA treatment was assessed by fluorescence-activated cell sorting analysis for Oct4 expression. One representative of three replicate experiments is shown. (G, H): Expression of pluripotency markers was assessed by quantitative real-time polymerase chain reaction in the Zfp206-OE (G) and Zfp206-KD cells (H). Gene expressions of Zfp206-OE and Zfp206-KD cells were normalized to cells transfected with empty vector at the corresponding time point and represented as average ± SD of four separate experiments. Student's t test was carried out to show significant differences at different time points compared with day 0 (∗, p < .05; ∗∗, p < .01). Abbreviations: AP, alkaline phosphatase; KD, knockdown; OE, overexpression.

Given our results that Zfp206 expression contributed to maintenance of the pluripotent state, we then knocked down expression by RNA interference and evaluated the differentiation response. A lentiviral vector was constructed that expressed short hairpin RNA (shRNA) that specifically targets Zfp206 transcripts (Fig. 4B). A pool of ESC was selected that had high-level expression of enhanced green fluorescent protein, which was driven from a linked promoter. The pool of 206-KD cells showed a 10–1,010-fold reduction in expression of Zfp206 mRNA (Fig. 4D) and a comparable repression of protein (Fig. 4F). The repression of Zfp206 had no apparent morphological effect on the ESC (Fig. 5A). FACS analyses were performed on these cells to assess the expression of differentiation markers SSEA1 and Oct4, which were not significantly different in the 206-KD cells relative to controls (data not shown). We performed Q-RT-PCR on the 206-KD cells to assess the expression of pluripotency and differentiation markers (Fig. 5B). There were no differences in the expression of the pluripotency markers Oct4, Sox2, Nanog, Esrrb, and Tdgf1. A few markers of differentiation showed small increases in expression in the knockdown cells, including Gata4, Fgf5, and Cdx2. Interestingly, these same markers were slightly repressed in the Zfp206 overexpression cells. We also observed a fourfold increase in Ipf1, a marker of early pancreatic lineage, in the 206-KD cells. The biological significance, if any, of the induction of expression of these endoderm markers is not clear. Overall, the morphology and marker expression data indicated that knockdown of Zfp206 expression was not sufficient to induce differentiation. This is in contrast to pluripotency regulating transcription factors Oct4, Sox2, and Nanog, which, when repressed, result in ESC differentiation.

It was surprising to us that knockdown of Zfp206 expression did not induce differentiation, given our results that enforced expression of the gene contributed to maintenance of the pluripotent state. To further examine the role of Zfp206 as a regulator of ESC differentiation, we tested our 206-KD cells to see whether suppression of Zfp206 rendered cells more responsive to differentiation induction. The 206-KD cells were treated with RA to induce differentiation. Over a 6-day time course, 206-KD cells were observed to be morphologically differentiated to greater extent than control cells (Fig. 6B). We performed the same analysis on the knockdown cultures as was done for the overexpressers to quantify the degree of differentiation. A twofold reduction in the number of AP+ cells was measured in the 206-KD cultures relative to EV control cultures at day 6 post-RA treatment (Fig. 6D). FACS analysis of these cultures showed a twofold reduction in Oct4+ cells for 206-KD relative to EV control after 6 days in RA (Fig. 6F). We also looked by Q-RT-PCR for changes in expression of pluripotency markers in the 206-KD cells in response to RA. For all the markers tested (Oct4, Nanog, Sox2, Otx2, Ebaf, and Esrrb), the expression was significantly reduced in 206-KD cells relative to control cells (Fig. 6H). These data establish that ESC with knocked down expression of Zfp206 have enhanced responsiveness to differentiate, which is opposite to the effect we observed when Zfp206 was overexpressed. Together, the knockdown and overexpression data indicate that Zfp206 plays a role in maintenance of ESC pluripotency.

Zfp206 Is a Transcriptional Activator

It is predicted that Zfp206 functions as a transcriptional regulator because of its nuclear localization and the presence of zinc fingers and a SCAN domain. To determine whether Zfp206 acts as a regulator of transcription, we tested whether it could enhance activity of three promoters expressed in ESC. Constructs containing the promoter and 5′ flanking regions of Zfp206, Nanog, and Oct4, each linked to a luciferase reporter, were transfected into HEK293, a cell line that does not express Zfp206. The promoter for Cdx2, which is not expressed in ESC, was tested as well. A basal level of transcriptional activity from these promoters was measured as a function of the luciferase activity in the transfected cells. We tested the activity of Zfp206 on these promoters by cotransfection of a vector that expresses Zfp206. Expression of Zfp206 increased luciferase activity from the promoters of Nanog and Oct4 by 2.5-fold (Fig. 7A). Zfp206 activated its own promoter by more than fourfold. The Cdx2 promoter was not activated by Zfp206, indicating that there is sequence specificity for activation of gene expression. To test for specificity of our reporter assays, the same promoter constructs were cotransfected with vectors that express two other zinc-finger transcription factors, Zfp263 and Zfp331. These factors did not enhance expression from the four promoters. Collectively, these results indicate that Zfp206 encodes a transcriptional activator capable of enhancing expression of two ESC pluripotency genes, Oct4 and Nanog, and from its own promoter.

Figure Figure 7..

Zfp206 regulates promoters from Zfp206, Oct4, and Nanog. (A): The promoter regions of Oct4, Nanog, Zfp206, and Cdx2 were linked to a luciferase reporter gene. A Zfp206 expression vector and each luciferase reporter were cotransfected into HEK293 cells. The action of Zfp206 on these four promoters was assessed by measuring the luciferase activity relative to an empty vector control. The activities of two other transcription factors, Zfp263 and Zfp331, were measured against the same promoters. (B): The expression of the promoters linked to the luciferase reporter was evaluated in mouse ESC. RNA interference with shRNA vector was used to repress the expression of Zfp206. An irrelevant shRNA to green fluorescent protein and an empty vector served as negative controls. The change in reporter activity resulting from each RNA interference is given relative the empty vector. Experiments were repeated three times, with six technical replicates in each repetition. Significance values relative to empty vector control are from Student's t test (∗, p < .05; ∗∗, p < .01).

To confirm the above results in ESC, the luciferase reporters were cotransfected with shRNA directed against Zfp206 (Fig. 7B). Knockdown of Zfp206 resulted in a significant reduction (40%–60%) in the expression from the Oct4, Nanog, and Zfp206 promoters. There was little change in activity from the Cdx2 promoter. An shRNA targeting Gfp did not alter the expression of any of these reporter constructs. These results are consistent with the previous data showing that Zfp206 regulates the expression of Oct4, Nanog, and Zfp206, although additional experiments are required to establish whether Zfp206 is acting directly or indirectly on these promoters.

Discussion

Considerable work has established that three transcription factors—Oct4, Sox2, and Nanog—play an essential role in both epiblast and ESC in regulating pluripotency. Knockdown or knockout of these genes leads to differentiation of ESC and epiblast defects. The dominant phenotypes exhibited by loss of function for these genes have placed them in a prominent position, perhaps even at the top of a hierarchy, in the transcriptional network that operates to control differentiation. Recently, hundreds of target genes, directly regulated by Oct4, Sox2, and Nanog, have been identified by genome-wide mapping of their binding sites [13, 14]. These data provide an opportunity to map in more detail the transcriptional network that regulates ESC differentiation.

We have selected several genes as candidate regulators of ESC pluripotency by searching expression data sets for genes that are selectively expressed in undifferentiated ESC and that have binding sites for Oct4, Sox2, and/or Nanog in relatively close proximity to their transcription start sites. Among the genes on our selected short list was Zfp206, a previously uncharacterized gene that encodes a protein containing 14 zinc fingers and a SCAN domain. A detailed analysis indicates that Zfp206 is expressed in ESC and in certain areas of the developing embryo, but not in differentiated ESC or adult tissues (except testis) (Figs. 1, 2; supplemental online Fig. 1). The predominant expression of Zfp206 in ESC and ICM suggests that this gene plays a role in the maintenance of pluripotency. The expression of Zfp206 in other regions of the developing embryo and in adult testis (Fig. 2) indicates that it may have additional roles in regulating cellular differentiation. The expression data presented here are consistent with those of Zhang et al. [17], which showed that Zfp206 mRNA is expressed predominantly in ES cells and preimplantation embryos. To confirm that Zfp206 protein is expressed similarly to the RNA, we went on to generate an antibody specific to Zfp206. Our data now establish definitively that Zfp206 is expressed in the mouse blastocysts (Fig. 3) and not simply an untranslated transcript. Our antibody was useful in establishing that Zfp206 expressed not only in the ICM as previously described but also in the trophectoderm, indicating that this transcription factor may also regulate extraembryonic development.

Two strategies were used to test our hypothesis that Zfp206 regulates pluripotency. The first was to see whether sustained overexpression of the gene rendered ESC resistant to differentiation. The second test was to determine whether knockdown of expression by RNA interference induced differentiation. Positive results were obtained in both tests, indicating that Zfp206 is a regulator of pluripotency. In separate overexpression studies, we noted that ESC engineered to express Zfp206 at 8-fold (data not shown) or 13-fold (Fig. 4) elevated levels were less likely to undergo spontaneous differentiation when plated at low densities (Fig. 5A). There is always a concern with overexpression studies that the high-level expression leads to nonphysiological effects. We observed no apparent differences in morphology, growth rates, or marker gene expression for the overexpressers of Zfp206 relative to the vector-only cells or the parent E14 line. In addition, the overexpression results were consistent with the inverse result that we observed in the knockdown studies. In the knockdown experiments, we achieved a 10-fold repression of Zfp206, but this was not sufficient to induce ESC differentiation. However, in our colony-plating assay, fewer undifferentiated colonies were observed, indicating that a reduction in Zfp206 led to increased spontaneous differentiation (Fig. 5). These results are consistent with those published by Zhang et al. [17] and support the hypothesis that Zfp206 is important to maintain pluripotency. We provide additional support for this role as the Zfp206-OE and -KD cells showed an altered response to RA-induced differentiation. We found that Zfp206 overexpressers were impaired in their ability to differentiate in response to RA treatment (Fig. 6). Moreover, ESC with reduced levels of Zfp206 were significantly more inclined toward differentiation upon RA treatment (Fig. 6). It is interesting that Zfp206 overexpression does not completely block RA-induced differentiation. Likewise, knockdown of Zfp206 is not sufficient to induce ESC differentiation. Other pluripotency genes, such as Nanog, are sufficient to prevent differentiation by overexpression and to induce differentiation by knockdown [9, 12]. It is possible that we have not achieved the levels of overexpression and knockdown expression required to prevent or induce differentiation. It is also possible that Zfp206 plays more of a supporting role than a dominant role in maintaining pluripotency. We are attempting to knock down expression of other genes in concert with Zfp206 to test this idea. It will also be important to generate ESC that are null for Zfp206 to see whether they lose pluripotency.

We initiated our studies on Zfp206 under the assumption that it acts as a transcription factor. This is likely a reasonable assumption given that the gene encodes a protein that contains a SCAN domain and 14 zinc fingers (C2H2 type), structural motifs that are associated with a large family of transcription factors found only in vertebrates [18]. Typically, transcription factors with C2H2 zinc fingers recognize and interact with DNA in a sequence-specific manner, although in some cases they bind to single- and double-stranded RNA. Transcription factors with many zinc finger domains are thought to have multiple ligand specificities [19]. Our own experiments showed that Zfp206 could drive transcription selectively from the promoter regions of Nanog, Oct4, and Zfp206, but not from the Cdx2 (Fig. 7). Thus, it appears that Zfp206 activates transcription from a select set of target genes in ESC. Additional experiments are required to identify the binding motif recognized by Zfp206 and to map its target genes by chromatin immunoprecipitation. During preparation of this report, another group published results about the role of Zfp206 as a transcriptional regulator of ESC differentiation [17]. In that study, they were able to identify potential target genes of Zfp206 by performing gene expression microarrays on ESC that over- and underexpress Zfp206. Several targets were identified, including family members Zscan4, Thoc4, Tcstv1, and eIF-1A, which were shown to be coexpressed with Zfp206 in preimplantation embryos. It will be interesting to determine whether Zfp206 is able to directly regulate the transcription of these genes.

Expression of alternative splice versions of Zfp206 with various numbers of the zinc fingers may contribute to an expanded repertoire of binding targets [19]. The SCAN domains mediate interactions among other proteins, either as homodimers or as heterodimers [18]. Various interactions are thus possible between Zfp206 and other transcription factors expressed in ESC, which would also have an impact on selection of genes that are targets for regulation by Zfp206.

It is of interest that Zfp206 activated transcription from the Oct4 and Nanog promoters in our cotransfection assays (Fig. 7). The presence of binding sites for Oct4 and Nanog in Zfp206 had indicated that it is a direct target of regulation by these established pluripotency transcription factors [13, 14]. In fact, we have recently shown that Zfp206 is a direct target of Oct4 and Sox2 [20]. Collectively, these data indicate that Zfp206 acts upstream and downstream of Oct4 and Nanog to make up a regulatory loop that controls pluripotency. Similar self-reinforcing loops have been demonstrated between Oct4, Sox2, and Nanog [15] [21, 22].

The results presented here place Zfp206 within the transcriptional network that contributes to pluripotency of stem cells of the mammalian embryo. It will be useful to identify target genes that are regulated by Zfp206 and to map these targets to the entire transcriptional network that regulates differentiation. A better understanding of such a transcriptional network and how it intersects with signaling pathways that react to external stimuli will lead to improved methods for controlling ESC differentiation in vitro.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Acknowledgements

We extend our gratitude to Yi-Ling Koh, Tng-Jie Lun, Hian-Cheong Ng, Phin Peng, Seow Peng, Shi-Feng Xue, Boon Tat, Ru-Bing Liu, Renee Tan, Wei-Jia Wang, Tan-Siok Lay, Guo-Qing Tong, Galih Kunarso, and Yi-Xun Li for technical assistance; Andrew Hutchins, Jonathan Loh, David Rodda, Wai-Leong Tam, and Jin-Qiu Zhang for reagents; and Manjiri Bakre, Huck-Hui Ng, and Bing Lim for intellectual contributions to this project.

Ancillary