A Novel Role for an RNA Polymerase III Subunit POLR3G in Regulating Pluripotency in Human Embryonic Stem Cells

Abstract

The pluripotency of human embryonic stem cells (hESC) could have great potential for the development of cell replacement therapies. Previous studies have converged on the finding that OCT4, SOX2, and NANOG serve as key regulators in the maintenance of hESC. However, other signals that regulate hESC maintenance remain poorly studied. Here we describe a novel role of an RNA polymerase III (Pol III) subunit, POLR3G, in the maintenance of pluripotency in hESC. We demonstrate the presence of POLR3G in undifferentiated hESC, human induced pluripotent stem cells (hiPSC), and early mouse blastocysts. Downregulation of POLR3G is observed on differentiation of hESC and hiPSC, suggesting that POLR3G can be used as a molecular marker to readily identify undifferentiated pluripotent stem cells from their differentiated derivatives. Using an inducible shRNA lentiviral system, we found evidence that decreased levels of POLR3G result in loss of pluripotency and promote differentiation of hESC to all three germ layers but have no effect on cell apoptosis. On the other hand, overexpression of POLR3G has no effect on pluripotency and apoptosis in undifferentiated hESC. Interestingly, hESC expressing elevated levels of POLR3G are more resistant to differentiation. Furthermore, our experimental results show that POLR3G is a downstream target of OCT4 and NANOG, and our pharmacological study indicated that POLR3G expression can be readily regulated by the Erk1/2 signaling pathway. This study is the first to show an important role of POLR3G in the maintenance of hESC, suggesting a potential role of Pol III transcription in regulating hESC pluripotency. STEM CELLS 2011;29:1517–1527

INTRODUCTION

Since the initial derivation of human embryonic stem cells (hESC) [1, 2], many studies have focused on understanding the molecular switches that regulate hESC pluripotency and differentiation. In particular, the transcription factors OCT4, SOX2, and NANOG have been found to form a core regulatory transcriptional network that regulates stem cell pluripotency [3, 4]. However, the role of many downstream effectors of these transcription factors remains unknown. Learning how to control stem cell pluripotency may improve current methods of maintaining hESC for a prolonged period of time. Moreover, the recent development of human induced pluripotent stem cells (hiPSC) provided a novel way to generate pluripotent cells by reprogramming terminally differentiated cells using a cocktail of transcription factors [5, 6]. However, such reprogramming efficiency is extremely low and identification of novel pluripotency genes may prove helpful in developing more efficient methods to generate iPSC.

Using fluorescence-activated cell sorting (FACS) coupled with global gene expression profiling, a previous study identified a novel stem cell gene, POLR3G, to be the most highly upregulated gene in undifferentiated hESC compared with differentiated cells [7]. In this regard, POLR3G has not previously been associated with any aspect of stem cell biology. POLR3G, also known as RPC32 or RPC7, is a subunit specific to RNA polymerase III (Pol III), with no counterpart in RNA polymerase I or II. In humans, POLR3G is known to form a subcomplex with two other Pol III subunits POLR3F (aka RPC39) and POLR3C (aka RPC62) [8]. Human Pol III lacking this POLR3G-3F-3C subcomplex loses the ability to initiate transcription, but transcription elongation and termination remain normal [8]. Indeed, a previous study in mice suggested that this POLR3G-3F-3C subcomplex plays a role in the recruitment of the Pol III holoenzyme to its target genes [9]. In S. cerevisiae, deletion of the yeast POLR3G homolog, C31, is lethal, whereas mutation of the C-terminal tail impairs cell growth and transcription initiation [10, 11]. On the other hand, siRNA-mediated knockdown of POLR3G has been demonstrated to impair anchorage independent growth in HeLa cells, suggesting a role of POLR3G in oncogenic transformation [12]. However, the precise role of POLR3G in pluripotent stem cells remains unknown. This study aims to characterize the expression profile of POLR3G and study its function in hESC.

MATERIALS AND METHODS

Cell Culture and Treatment

H1 and H9 hESC (WiCell, WI, USA) and hiPSC were cultured as previously described [5, 13]. In some experiments, hESC were treated with 25 ng/ml of BMP4 (R&D Systems, MN, USA) to induce differentiation as described previously [14]. Embryoid bodies (EBs) were formed by growing hESC in suspension in a low attachment culture plate in media supplemented with 20% knockout serum replacement. Treatment of hESC with pharmacological inhibitors U0126 (30 μM; Calbiochem, CA, USA), U0124 (30 μM, Calbiochem), LY294002 (10 μM, Sigma, MO, USA), rapamycin (20 nM, Cell Signaling Technology, MA, USA) was carried out as described previously [15].

Quantitative PCR

Total RNA was extracted from H1 or H9 cells using the RNeasy kit (Qiagen, CA, USA). cDNA was synthesized using the high capacity cDNA reverse transcription kit (Applied Biosystems, CA, USA). Taqman mastermix, Taqman probes, and SYBR green mastermix were all purchased from Applied Biosystems. The sequences for SYBR green primers used would be provided on request. Primer sequences for tRNA Leu, 7SL, and 5S are described previously [16, 17]. Quantitative PCR (qPCR) was performed in triplicate in 384-well plates and analyzed in a 7500 or 7900HT qPCR machine (Applied Biosystems). The results were analyzed using the ΔΔC_t method as described previously [18].

Western Blot

Standard Western blot procedures were carried out as described previously [15]. Briefly, H1 and H9 lysates were separated by polyacrylamide gel electrophoresis and transferred to a polyvinylidene fluoride (PVDF) membrane (GE Healthcare, WI, USA). The membrane was blotted with an antibody against POLR3G (Santa Cruz, CA, USA), OCT4 (Santa Cruz), or FLAG (Sigma) followed by a horseradish peroxidase (HRP)-conjugated secondary antibody. Chemiluminescent detection reagent (ECL Plus, GE Healthcare) was used to detect the HRP signal on film or using the Gel-Doc system (Biorad, CA, USA). As a loading control, the membrane was stripped and reblotted with antibody against β-tubulin (Sigma) or actin (Santa Cruz) followed by HRP secondary antibody.

Lentiviral Transduction and Nucleofection

Nucleofection of hESC with siRNA against OCT4 and NANOG were carried out as previously described [19, 20]. For knockdown of POLR3G in hESC, the TRIPZ lentiviral vector was used (Open Biosystems/Thermo Fisher Scientific, AL, USA). The short hairpin RNA (shRNA) sequences used would be provided on request. Lentiviruses were produced following the manufacturer's instructions. Briefly, lentiviruses were produced using HEK293T cells and harvested at 48 and 72 hours post-transfection. Lentiviruses were concentrated by ultracentrifugation or using PEG-it solution (System Biosciences, CA, USA). Trypsinized hESC were transduced with lentiviruses in the presence of Sequabrene (5 μg/ml, Sigma) and ROCK inhibitor Y27632 (10 μM, Calbiochem) overnight. Stable cell lines were established by selection with puromycin (1 μg/ml, Calbiochem). For induction of shRNA expression, lentiviral-transduced hESC were treated with doxycycline (1–2 μg/ml, Clontech, CA, USA). For overexpression of POLR3G in hESC, the coding sequence of POLR3G was first amplified and a FLAG tag was inserted into the N-terminal of POLR3G by polymerase chain reaction (PCR). The FLAG-tagged POLR3G was cloned into a TOPO vector (Invitrogen, CA, USA) and subcloned into the pCDH-EF1-IRES-neo lentiviral vector (System Biosciences). Lentiviruses were produced following the manufacturer's instructions (System Biosciences). Briefly, lentiviruses were produced using HEK293TN cells, harvested at 48 hours post-transfection and concentrated using PEG-it solution (System biosciences). Trypsinized hESCs were transduced with lentiviruses overnight in the presence of Sequabrene (5 μg/ml, Sigma) and ROCK inhibitor Y27632 (10 μM, Calbiochem). Stable cell lines were established by selection with G418 (300 μg/ml, Sigma).

Chromatin Immunoprecipiation

Chromatin immunoprecipitation (ChIP) assays with H9 cells were carried out as described previously [21]. Briefly, H9 cells were crosslinked with 1% formaldehyde, neutralized, and lysed with sodium dodecyl sulfate (SDS) lysis buffer. The samples were sonicated and cleared with preblocked protein A sepharose beads (GE Healthcare). The samples were then incubated with an antibody against POLR3G (Santa Cruz), NANOG (R&D Systems), or OCT4 (R&D Systems) followed by incubation with preblocked protein A sepharose beads (GE Healthcare). The immunoprecipitates were then washed and eluted. Crosslinks were reversed and DNA was purified using a gel cleanup kit (Eppendorf, Hamburg, Germany or MP biomedicals, OH, USA). PCR was performed with the Taq PCR core kit (Qiagen). Amplicons were run on agarose gels and analyzed using the Gel-Doc Imager (Bio-Rad).

Immunocytochemistry

Unfertilized mouse oocytes, E0.5, E1.5, E2.5, and E3.5 mouse embryos were collected by crossing B6C3F1 females to CB6F1 males. In other experiments, H1 hESCs were treated with or without BMP4 (25 ng/ml, R&D) for 3 days. The cells were fixed in 4% paraformaldehyde and permeabilized with 0.5% Triton X-100. Subsequently, the samples were blocked with 10% serum and incubated with antibodies in the following order: anti-POLR3G antibody (Santa Cruz), Alexa fluor 594 (Invitrogen), anti-OCT4 antibody (Santa Cruz), and Alexa fluor 488 (Invitrogen). The samples were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and visualized under a fluorescent microscope (Zeiss, USA).

FACS Analysis

H1 or H9 hESC were trypsinized, blocked with 10% serum, and incubated with TRA181 antibody (Santa Cruz), TRA160 antibody (Santa Cruz), or anti-SSEA4 antibody (Stem Cell Technologies, Vancouver, Canada or Developmental Studies Hybridoma Bank, IA, USA), followed by the appropriate Alexa fluor 488 secondary antibodies (Invitrogen). For nuclear immunostaining, the samples were first fixed in 4% paraformaldehyde and permeabilized with 100% cold ethanol. The samples were then blocked with 10% serum and incubated with reagents in the following order: antibody against turbo RFP (tRFP; Evrogen, Moscow, Russia), anti-rabbit Alexa fluor 568 (Invitrogen), anti-OCT4 antibody (Santa Cruz), and anti-mouse Alexa fluor 488 (Invitrogen). Flow cytometric analysis experiments were carried out on a FACS-Diva flow cytometer (BD Biosciences, CA, USA). For FACS experiments, samples were immunostained using the same protocol and sorted into microfuge tubes using the FACS-Aria II cell sorter (BD Biosciences). For the cell apoptosis assay, floating apoptotic bodies were collected and H9 hESC were harvested by trypsinization. hESC apoptosis was quantified using the in situ cell death detection kit (Roche, USA).

RESULTS

POLR3G Is Enriched in Undifferentiated hESC and hiPSC

Using qPCR, we characterized the mRNA expression of POLR3G in undifferentiated hESC and their differentiated derivatives. Upon spontaneous differentiation by EB formation, POLR3G mRNA was downregulated from 63% in day 5 EB to 10% in day 21 EB compared to undifferentiated hESC (Fig. 1A). Western blot analysis revealed a consistent downregulation of POLR3G protein upon EB formation (Fig. 1B). Similar results were obtained in hESC directed to differentiate by BMP4 treatment (data now shown). Recent studies have revealed the heterogeneous nature of hESC cultures [22, 23]. Given that residual undifferentiated hESC may exist even after prolonged differentiation, this may potentially confound the reliability of our qPCR results. To address this issue, we fractionated the undifferentiated and differentiated hESC population using known stem cell surface markers TRA-160 and SSEA-4 by FACS. Consistent with our previous findings, POLR3G transcripts were downregulated in SSEA4-negative hESC (18%) and TRA-160 negative hESC (2%, Fig. 1C). Moreover, we have consistently observed downregulation of POLR3G upon differentiation in multiple hESC lines including H1, H9, and MEL1 (data not shown). Using immunocytochemistry, POLR3G expression in undifferentiated hESC was detected as foci in the nucleocytoplasm but excluded from the nucleoli (Fig. 1D), a localization pattern previously reported in other cell types [24]. Also, we observed the disappearance of POLR3G nuclear staining in OCT4-negative hESC on BMP4 treatment. Together, our results suggest that POLR3G is highly enriched in undifferentiated hESC.

Figure 1. POLR3G is a molecular marker for undifferentiated hESC and hiPSC. (A): Quantitative PCR (qPCR) analysis of POLR3G expression in undifferentiated hESC (H9) compared to embryoid bodies at various timepoints. Average relative expression levels and standard deviations are shown. (B): Western blot of POLR3G expression in undifferentiated H9 compared to embryoid bodies at various timepoints. β-Tubulin was used as protein loading control. (C) qPCR analysis of POLR3G expression in H9 FACS-sorted with SSEA4 (left panel) or TRA160 (right panel). Average relative expression levels and standard deviations are shown. (D): Immunocytochemistry of POLR3G, OCT4, and DAPI counterstain in undifferentiated hESC (top panel: lower magnification; middle panel: higher magnification) and hESC differentiated by BMP4 treatment (bottom panel). Representative pictures are shown. (E): qPCR analysis of POLR3G expression in undifferentiated hiPSC (IMR90C2) compared to embryoid bodies at various timepoints. Average relative expression levels and standard deviations are shown. (F): qPCR analysis of POLR3G expression in hESC (MEL1) or hiPSC (IMR90C2, IPS-Foreskin-CL1) that are fractionated with GCTM2 and TG30 using FACS. Average relative expression levels and standard deviations are shown. All experiments from this figure were repeated at least three times. Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole; PCR, polymerase chain reaction; EB, embryoid bodies; FACS, fluorescence-activated cell sorting; hESC, human embryonic stem cells; iPSC, induced pluripotent stem cells.

We extended the characterization of POLR3G expression to two different hiPSC lines, a lung fibroblast-derived hiPSC IMR90C2 and a foreskin fibroblast-derived hiPSC IPS-Foreskin-CL1 [5]. We observed a more rapid downregulation of POLR3G in hiPSC compared to hESC. On EB differentiation, the level of POLR3G transcripts in IMR90C2 decreased to 42% in day 4 EB, and only 13% of POLR3G transcripts remained by day 14 (Fig. 1E). Similar results are found in IPS-Foreskin-CL1 (data not shown). Furthermore, we used FACS to separate undifferentiated and differentiated hiPSC using two stem cell markers GCTM-2 and TG30 [25]. As illustrated in Figure 1F, POLR3G transcript levels are decreased in GCTM-2/TG30 double negative cells (2% in IMR90C2 and 19% in IPS-Foreskin-CL1) when compared to GCTM-2/TG30 double-positive hiPSC. In summary, our results demonstrate that POLR3G is highly enriched in multiple cell lines of undifferentiated hESC and hiPSC that were sorted by FACS using different stem cell markers. These results provide strong evidence that POLR3G can be used as a molecular marker to identify undifferentiated hESC or hiPSC.

POLR3G Protein Is First Expressed in the Nucleus of 8–16 Cell Mouse Embryos

As POLR3G showed an enriched expression in undifferentiated hESC, we wondered if POLR3G is also expressed during early mouse development. As illustrated in Figure 2A–2C, POLR3G expression is absent in unfertilized oocytes and first detected in zygotes and the two-cell stage embryos. As POLR3G expression is mainly in the cytoplasm, we anticipated that POLR3G is not participating as part of the Pol III complex during these stages of development. POLR3G expression starts to localize to the nucleus in the 8–16 cell stage embryo and early blastocysts, an expression pattern similar to that observed for the stem cell marker OCT4 (Fig. 2D, 2E). In summary, our results suggested that POLR3G may participate in Pol III transcription starting from 8 to 16 cell stage in mouse embryos.

Figure 2. POLR3G is first expressed in the nucleus from the 8–16 cell stage in preimplantation embryos. Immunocytochemistry of POLR3G, OCT4, and DAPI counterstain in (A) the unfertilized oocyte, (B) a zygote, (C) the two-cell embryo, (D) the 8–16 cell embryo, and (E) a blastocyst. Representative pictures from two independent experiments are shown. Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole.

POLR3G Plays an Important Role to Maintain Pluripotency in Undifferentiated hESC

To determine the functional significance of POLR3G in hESC, we carried out a knockdown study using a Tet-on lentiviral-shRNA system. The lentiviral vector used allows coexpression of tRFP and the corresponding shRNA in the presence of doxycycline (Fig. 3A). Using this system, we successfully established stable hESC cell lines with doxycycline-inducible POLR3G-shRNA expression. No tRFP-positive hESC were detected in the absence of doxycycline, suggesting that the coexpression of tRFP and shRNA is subject to tight regulation by doxycycline (Fig. 3B). Using two different shRNAs directed against POLR3G, we knocked down the level of POLR3G transcripts to 63% and 52%, respectively, in hESC when compared to hESC expressing nonsilencing shRNA (Fig. 3C). Western blot analysis confirmed downregulation of POLR3G protein levels (Fig. 3D). Importantly, doxycycline treatment has no effect on the level of POLR3G transcripts in hESC transduced with nonsilencing shRNA, further supporting the specificity of our observed knockdown of POLR3G in hESC (Fig. 3C).

Figure 3. POLR3G plays an important role to maintain pluripotency in hESC. (A): Schematic diagram of the Tet-On shRNA lentiviral vector used to knockdown expression of POLR3G. (B): Morphology of hESC transduced with lentiviruses carrying nonsilencing shRNA in the absence (top panel) or presence of doxycycline induction for 7 days (second panel), and hESC transduced with lentiviruses carrying shRNA two against POLR3G in the presence of doxycycline induction for 7 days (third and fourth panels). Representative phase contrast and tRFP images are shown. (C) Quantitative PCR (qPCR) analysis of POLR3G expression in hESC transduced with lentiviruses carrying nonsilencing shRNA or shRNA against POLR3G, in the absence or presence of doxycycline induction for 7 days. Average relative expression levels and standard deviations are shown. (D): Western blot analysis of POLR3G protein expression in hESC transduced with lentiviruses carrying shRNA against POLR3G in the absence or presence of doxycycline induction for 7 days. β-Actin was used as protein loading control. (E): qPCR analysis of OCT4, SOX2, and NANOG expression in POLR3G knockdown hESC with and without doxycycline induction for 7 days. Average relative expression levels and standard deviations are shown. (F): qPCR analysis of markers from the endodermal (AFP, SOX17, PDX1, and GATA4), mesodermal (BRACHYURY, MIXL1, and EOMES), ectodermal (PAX6, MAP2, and SOX1) and trophectodermal (CGB7) lineages in POLR3G-knockdown hESC that are fluorescence-activated cell sorted for tRFP. Average relative expression levels and standard deviations are shown. (G): Flow cytometry analysis of OCT4+/tRFP+ cells in hESC transduced with lentiviruses carrying nonsilencing shRNA or shRNA against POLR3G. Average percentage and standard deviations are shown. (H): Flow cytometric analysis of TUNEL+ cells in hESC transduced with lentiviruses carrying nonsilencing shRNA or shRNA against POLR3G. Wild-type hESC treated with 5% DMSO were used as a positive control to induce cell death. Average percentage and standard deviations are shown. All experiments in this figure were repeated at least three times. Abbreviations: cPPT, central polypurine tract; DMSO, dimethyl sulfoxide; hESC, human embryonic stem cell; IRES, internal ribosome entry site; LTR, long terminal repeat; rtTA, reverse tetracycline-controlled transactivator; shRNA, short hairpin RNA; TRE, tetracycline response element; tRFP, turbo RFP; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling; UBC, ubiquitin C promoter; WRE, woodchuck hepatitis virus response element; Zeo, zeocin resistance gene.

Interestingly, knockdown of POLR3G resulted in dramatic changes to the morphology of hESC and caused their differentiation (Fig. 3B). Predominantly two types of differentiated cells were observed, a bulky cell type with multiple granules and a spindle-like cell type (Fig. 3B). As the two shRNAs directed against POLR3G generated a similar phenotype, we decided to carry out subsequent analysis using the shRNA that gave us higher levels of knockdown. Further analysis using qPCR suggested that knockdown of POLR3G resulted in downregulation of OCT4 (57%), SOX2 (53%), NANOG (70%) in hESC (Fig. 3E). Moreover, we quantified cells that were positive for OCT4 and tRFP to assess pluripotency in POLR3G knockdown hESC. The POLR3G knockdown hESC showed a decreased level of OCT4+/tRFP+ cells (39% compared to 82% in the control), indicating loss of hESC pluripotency when POLR3G is knocked down (Fig. 3G). To determine if POLR3G knockdown biased hESC to differentiate into any particular lineage, we interrogated the POLR3G knockdown hESC with a set of differentiation markers from all three germ layers using qPCR. In this experiment, we used FACS to enrich for the POLR3G shRNA transduced cells by separating the tRFP+ and tRFP− population. Knockdown of POLR3G resulted in upregulation of markers from all three lineages, including BRACHYURY (227%), MIXL1 (206%), EOMES (294%), SOX17 (229%), PDX1 (203%), GATA4 (278%), PAX6 (272%), and MAP2 (614%) (Fig. 3F). SOX1 and CGB7 transcripts were not detected in either control or POLR3G knockdown hESC. Using FACS to separate tRFP+ and tRFP− population, we also showed that POLR3G knockdown modulate expression of Pol III transcribed genes in hESC (Supporting Information Fig. 1A), as we observed changes in levels of tRNA Leu (212%), 7SL (58%), and 5S (274%) in POLR3G knockdown hESC compared to control. In addition, we did not detect changes in apoptosis in POLR3G knockdown hESC (5% TUNEL+ cells compared to 5% in control, Fig. 3H). Taken together, our results suggest that POLR3G knockdown results in loss of pluripotency, promoting hESC to differentiate into mesodermal, endodermal, and ectodermal lineages.

POLR3G Overexpression Has No Effect on Maintenance of hESC Pluripotency

Next we sought to overexpress POLR3G to determine its functional significance in hESC. Using the lentiviral system illustrated in Figure 4A, we successfully generated a stable cell line of hESC that constitutively express a Flag-tagged POLR3G. Our qPCR results demonstrate a sevenfold increase of POLR3G transcripts in the POLR3G overexpressing hESC compared to hESC transduced with an empty lentiviral vector (Fig. 4C). Also, the exogenous flag-tagged POLR3G protein can be readily detected by Western blot (Fig. 4D). Using probes that specifically recognize the endogenous POLR3G transcripts, we demonstrated that ectopic expression of POLR3G did not significantly affect the level of endogenous POLR3G transcripts (130%, Fig. 4C), suggesting POLR3G does not autoregulate its own expression at the transcriptional level. However, overexpression of POLR3G did not result in any changes in hESC morphology, as the cells remain undifferentiated (Fig. 4B). Moreover, POLR3G overexpressing hESC possess similar levels of OCT4 (117%), SOX2 (97%), and NANOG transcripts (116%) compared to hESC expressing an empty vector control (Fig. 4E). We further assessed markers of pluripotency in POLR3G overexpressing hESC by flow cytometry. Consistent with our previous results, overexpression of POLR3G did not affect hESC expression of pluripotency markers, as we detected similar levels of SSEA4+ or TRA181+ cells in hESC with or without ectopic POLR3G expression (Fig. 4F). Similarly, using TUNEL assays we found that hESC with or without ectopic POLR3G expression displayed a similar level of apoptosis (Fig. 4G). Finally, we showed that ectopic expression of POLR3G in undifferentiated hESC does not modulate expression of Pol III transcribed genes, including tRNA Leu, 7SL, and 5S (Supporting Information Fig. 1B).

Figure 4. hESC with elevated levels of POLR3G are more resistant to differentiation. (A): Schematic diagram of the lentiviral vector used to overexpress FLAG-tagged POLR3G. (B): Representative phase contrast images showing the morphology of hESC transduced with lentiviruses with (bottom panel) or without POLR3G overexpression (top panel). (C): Quantitative PCR (qPCR) analysis of total and endogenous POLR3G expression in hESC transduced with lentiviruses with or without POLR3G overexpression. Average relative expression levels and standard deviations are shown. (D): Western blot analysis of ectopic FLAG-tagged POLR3G expression in hESC with or without POLR3G overexpression. β-Actin was used as protein loading control. (E): qPCR analysis of OCT4, SOX2, and NANOG in hESC transduced with lentiviruses with or without POLR3G overexpression. Average relative expression levels and standard deviations are shown. (F): Flow cytometry analysis of SSEA4+ or TRA181+ cells in hESC transduced with lentiviruses with or without POLR3G overexpression. Average percentage and standard deviations are shown. (G): Flow cytometric analysis of TUNEL+ cells in hESC transduced with lentiviruses with or without POLR3G overexpression. Average percentage and standard deviations are shown. (H): Morphology of day 7 EB with or without POLR3G overexpression. Representative phase contrast images are shown. (I): qPCR analysis of POLR3G expression in day 7 EB with or without POLR3G overexpression. Average relative expression levels and standard deviations are shown. (J): qPCR analysis of OCT4, SOX2, and NANOG expression in day 7 EB with or without POLR3G overexpression. Average relative expression levels and standard deviations are shown. (K): qPCR analysis of markers from the endodermal (AFP, SOX17, PDX1, and GATA4), mesodermal (BRACHYURY, MIXL1, and EOMES), ectodermal (PAX6, MAP2, and SOX1), and trophectodermal (CGB7) lineages in day 7 EB with or without POLR3G overexpression. Average relative expression levels and standard deviations are shown. All experiments were repeated at least three times. Abbreviations: cPPT, central polypurine tract; EB, embryoid bodies; EF1α, elongation factor 1 alpha; hESC, human embryonic stem cell; IRES, internal ribosome entry site; LTR, long terminal repeat; Neo, neomycin resistance gene; RRE, Rev response element; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling; WPRE, Woodchuck hepatitis virus posttranscriptional regulatory element.

POLR3G Overexpression Renders hESC More Resistant to Differentiation During EB Formation

We reasoned that the lack of an apparent phenotype in hESC overexpressing POLR3G could be due to the fact that undifferentiated hESC are enriched with POLR3G transcripts. Therefore, we subjected the POLR3G overexpressing hESC to EB differentiation for 7 days. We confirmed that POLR3G overexpressing EB retain a high level of POLR3G expression (6.6-fold increase, Fig. 4I). Although the morphology of EB generated from POLR3G overexpressing hESC is comparable to those derived using the empty vector control (Fig. 4H), they showed a mild increase in OCT4, SOX2, and NANOG transcripts (168%, 189%, and 204%, respectively, Fig. 4J). Also, POLR3G-overexpressing EB possess much lower levels of various differentiation markers compared to the control, including SOX17 (42%), GATA4 (35%), BRACHYURY (35%), MIXL1 (35%), EOMES (26%), PAX6 (58%), SOX1 (54%), and MAP2 (63%, Fig. 4K). Taken together, these results suggest that ectopic expression of POLR3G renders hESC resistant to spontaneous differentiation by EB formation.

POLR3G Is a Downstream Target of NANOG and OCT4 and Erk1/2 Signaling in hESC

As our results indicate that POLR3G plays a role in modulating pluripotency in hESC, we asked whether POLR3G could be regulated by OCT4, SOX2, and NANOG. Using the Genomatix software, we identified two upstream NANOG binding sites and one OCT4 binding site downstream of the transcription start site (Fig. 5A). Using ChIP, we detected NANOG association with the POLR3G gene in binding site two but not binding site one in hESC (Fig. 5A). Moreover, our results demonstrated OCT4 association with a downstream binding site at the POLR3G gene (Fig. 5A). Furthermore, we carried out siRNA-mediated knockdown of NANOG or OCT4 in hESC using nucleofection. Knockdown of NANOG or OCT4 in hESC resulted in downregulation of POLR3G transcripts to 41% and 17%, respectively (Fig. 5B, 5C). Such POLR3G downregulation is not observed in our negative control hESC transfected without siRNA or with siRNA targeting β2M. Taken together, our experimental results support the notion that POLR3G is a downstream target of NANOG and OCT4 in hESC.

Figure 5. POLR3G is a downstream target of OCT4, NANOG, and the Erk1/2 signaling pathway. (A): Top panel: Schematic diagram of the POLR3G gene structure showing the presence of two NANOG binding sites and one OCT4 binding site near the transcription start site. Bottom panel: Chromatin IP results of an H9 sample immunoprecipitated with antibody to NANOG or OCT4. Negative controls were performed with samples immunoprecipitated in the absence of antibody or with an isotype-matched antibody. The input is a positive control from samples prior to immunoprecipitation. (B): Quantitative PCR (qPCR) analysis of POLR3G and NANOG expression in hESC with siRNA-mediated knockdown of NANOG levels. hESC nucleofected with siRNA against β2M or without siRNA were used as negative control. Average relative expression levels and standard deviations are shown. (C): qPCR analysis of POLR3G and OCT4 expression in hESC with siRNA-mediated knockdown of OCT4 levels. hESC nucleofected with siRNA against β2M or without siRNA were used as negative controls. Average relative expression levels and standard deviations are shown. (D–F): qPCR analysis of POLR3G expression in hESC treated with (D) rapamycin, (E) LY294002, and (F) U0126 for 6 or 24 hours. hESC treated with DMSO or U0124 were used as negative controls. Average relative expression levels and standard deviations are shown. All experiments in this figure were repeated at least three times. Abbreviations: DMSO, dimethyl sulfoxide; IP, immunoprecipitation; siRNA, small interfering RNA.

We also explored if POLR3G expression could be regulated by other intracellular signaling mechanisms, namely the extracellular signal-regulated kinase 1/2 (Erk1/2), the mammalian target of rapamycin (mTOR), and the phosphatidylinositol 3-kinases (PI3K)/Akt, all of which have been previously described to regulate Pol III transcription (reviewed in ref. [26]). Using qPCR, we determined that treatment with the mTOR inhibitor rapamycin (Fig. 5D) or the PI3K inhibitor LY294002 (Fig. 5E) had no significant effect on POLR3G expression. In contrast, inhibition of the Erk1/2 pathway using U0126 resulted in a significant decrease in POLR3G expression to 28% compared to the control (Fig. 5F). Western blot analysis also confirmed downregulation of POLR3G protein following U0126 treatment (Supporting Information Fig. 2). Therefore, our results suggest that expression of POLR3G is regulated by the Erk1/2 pathway but not by mTOR or PI3K/Akt.

DISCUSSION

This article represents the first study on the functional role of POLR3G in hESC. We provide evidence that POLR3G expression is enriched in undifferentiated hESC compared to their differentiated derivatives (Fig. 1A–1C), a result consistent with previous reports [7, 12]. Our results indicate that POLR3G is predominately localized to the nucleocytoplasm and is excluded from the nucleoli (Fig. 1D). Moreover, a similar enrichment of POLR3G is also observed in undifferentiated hiPSC (Fig. 1E and 1F). Therefore, POLR3G can be used as a molecular marker to readily distinguish undifferentiated hESC or hiPSC from their differentiated counterparts.

Little is known about the developmental role of the RNA Pol III complex. Only one previous study addressed this question and showed that mutation of POLR3B, the second largest subunit of Pol III, disrupts digestive organ development in zebrafish [27]. Here we characterize the expression of POLR3G in early mouse embryos. Interestingly, we show that POLR3G is absent in unfertilized oocytes and present predominately in the cytoplasm in fertilized zygotes and two cell embryos (Fig. 2A–2C). As gene transcription is largely silenced at the zygotic stage (reviewed in ref. [28]), it is possible that POLR3G expression is subject to translational regulation in fertilized zygotes. However, the localization pattern of POLR3G in fertilized zygotes and two-cell embryos suggested that it may not be participating in Pol III-mediated transcription in the nucleus during these developmental stages. Nuclear localization of POLR3G is first observed during the 8–16 cell stages and the blastocyst stage (Fig. 2D and 2E). The presence of POLR3G in the inner cell mass indicates that it may also play a role in regulating pluripotency in the inner cell mass.

Next, we used an inducible shRNA lentiviral system to knockdown POLR3G expression in hESC. Interestingly, downregulation of POLR3G resulted in loss of pluripotency in hESC, as indicated by morphological changes and loss of OCT4, SOX2, and NANOG expression (Fig. 3B–3E, 3G). Moreover, shRNA-mediated suppression of POLR3G in hESC resulted in an increased expression of differentiation markers from all three germ layers, suggesting that POLR3G plays a role in the maintenance of hESC pluripotency (Fig. 3F). Downregulation of POLR3G also modulates Pol III transcription, resulting in decreased 7SL expression together with increased tRNA Leu and 5S expression compared to control (Supporting Information Fig. 1A). However, decreased level of POLR3G had no significant effect in hESC apoptosis (Fig. 3H). On the other hand, overexpression of POLR3G in undifferentiated hESC did not impact their pluripotency or apoptosis (Fig. 4E–4G). Interestingly, when the cells were promoted to differentiation during EB formation, we observed a significant decrease in the expression of differentiation markers, indicating that elevated levels of POLR3G render hESC more resistant to differentiation (Fig. 4K). Hence, our functional studies of POLR3G demonstrate an important role of POLR3G in maintaining the pluripotent state of hESC. In this regard, it is interesting to note that Haurie et al. [12] demonstrated that POLR3G overexpression contributes to cell tumorgenicity and oncogenic transformation. Recent findings have pointed to a contribution of the stem cell regulatory transcriptional network in malignant transformation and establishment of cancer stem cells [29, 30]. Thus future studies that focus on the role of POLR3G in cancer stem cells will be of immense interest.

Given the important role of POLR3G in maintaining hESC pluripotency, we also studied the regulation of POLR3G expression by other known stem cell factors. Our results demonstrated that the level of POLR3G does not regulate its own expression at the transcriptional level, as ectopic expression of POLR3G did not impact the levels of the endogenous transcripts (Fig. 4C). Our ChIP results suggest that NANOG and OCT4 bind to the POLR3G gene (Fig. 5A). Moreover, knockdown of OCT4 or NANOG also resulted in POLR3G downregulation in hESC (Fig. 5B and 5C), supporting the notion that POLR3G is a downstream target of OCT4 and NANOG. Additionally, previous studies have also that indicated signaling via PI3K/Akt, mTOR, and Erk1/2 can regulate Pol III transcription (reviewed in ref. [26]). Using pharmacological inhibitors, we found that inhibition of the Erk1/2 pathway caused downregulation of POLR3G expression, whereas inhibition of the PI3K/Akt pathway or the mTOR pathway had no effect on POLR3G expression (Fig. 5D–5F). Therefore, our results identified a role of Erk1/2 pathway in regulating POLR3G expression. Interestingly, signaling via Erk1/2 has been reported to play an important role in maintenance of hESC [15, 31, 32], whereas Erk1/2 signaling negatively regulates self-renewal in mouse embryonic stem cells (mESCs) [33]. We speculate that the regulation of POLR3G expression by Erk1/2 signaling may be different in mESC compared to hESC. Further studies on the links between Erk1/2, Pol III, and pluripotency in the context of hESC and mESC may be warranted.

Finally, a recent study has identified a paralog of POLR3G based on amino acid similarity, termed POLR3G-like (POLR3GL, aka RPC32β) [12]. In this study, the authors described the existence of two isoforms of Pol III with differential transcription activities; one that contains POLR3G and another configuration that contains POLR3GL [12]. In this regard, we observed that the POLR3GL transcripts are markedly upregulated on hESC differentiation, an expression pattern that is opposite to that of POLR3G (unpublished data). This leads to an exciting hypothesis that two different Pol III isoforms, POLR3G-containing Pol III or POLR3GL-containing Pol III, may be operating in undifferentiated or differentiated hESC, respectively, serving as a molecular switch to regulate cell fates of hESC. In this regard, it is interesting that knockdown of POLR3G did not lead to decrease in global Pol III transcription, but upregulation of tRNA Leu and 5S expression and downregulation of 7SL expression in hESC (Supporting Information Fig. 1A). We hypothesize that upon POLR3G knockdown in hESC, a different isoform of Pol III complex that contains POLR3GL operates in hESC to allow for differential Pol III transcriptional outputs. Indeed, previous in vitro transcription assays had indicated that POLR3GL-containing Pol III transcribes tRNA and 5S more efficiently compared to the POLR3G-containing Pol III [12]. However, the transcriptional efficiency of 7SL by the two isoforms of Pol III complex remained unknown. Further studies will be necessary to confirm the existence of the POLR3GL-containing Pol III in hESC and investigate how this Pol III isoform impacts on the cell fate of hESC. In addition, the transcriptional initiator factor (TFIIIB) may serve as another means to regulate Pol III transcriptional output in hESC, as previous studies suggest that TFIIIB activity is downregulated during mouse embryonal carcinoma differentiation [34, 35]. Future studies that focus on understanding the interplay between TFIIIB and the two Pol III isoforms in regulating Pol III transcription would be interesting.

CONCLUSIONS

In this article, we show a novel role of POLR3G in the maintenance of hESC. We demonstrate that POLR3G expression is highly enriched in undifferentiated hESC and hiPSC. Moreover, knockdown of POLR3G resulted in hESC differentiation, whereas increased levels of POLR3G caused hESC to become more resistant to differentiation. We further show that POLR3G is a downstream target of OCT4 and NANOG and its expression can be regulated by the Erk1/2 signaling pathway based on our pharmacological study. Taken together, our results demonstrate an important role for POLR3G in regulating hESC pluripotency.

Acknowledgements

We are grateful for financial support from the California Institute of Regenerative Medicine (Grant: TG1-00008 to R.C.B.W. and H.F., RCI-00110 to P.J.D., and TG2-01152 to E.L.S.) and the Australian Stem Cell Center (to A.L.L.). We would also like to thank Prof. Robert Blelloch (University of California San Francisco) for helpful discussions and Prof. James Thomson (University of Wisconsin) for the provision of hiPSC cell lines to A.L.L. R.C.B.W. is currently affiliated with Laboratory of Genetics, National Institute on Aging, National Institutes for Health, Baltimore, MD, USA.

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

The authors indicate no potential conflicts of interest.