The Pluripotency Regulator Zic3 Is a Direct Activator of the Nanog Promoter in ESCs§


  • Linda Shushan Lim,

    1. Stem Cell and Developmental Biology Group, Genome Institute of Singapore, Singapore
    Search for more papers by this author
  • Felicia Huimei Hong,

    1. Stem Cell and Developmental Biology Group, Genome Institute of Singapore, Singapore
    2. Department of Biological Sciences, National University of Singapore, Singapore
    3. NUS Graduate School for Integrative Sciences and Engineering, National University of Singapore, Singapore
    Search for more papers by this author
  • Galih Kunarso,

    1. Stem Cell and Developmental Biology Group, Genome Institute of Singapore, Singapore
    Search for more papers by this author
  • Lawrence W. Stanton

    Corresponding author
    1. Stem Cell and Developmental Biology Group, Genome Institute of Singapore, Singapore
    2. Department of Biological Sciences, National University of Singapore, Singapore
    • Stem Cell and Developmental Biology, Genome Institute of Singapore, 60 Biopolis Street, #02-01, Genome, Singapore 138672
    Search for more papers by this author
    • Telephone: 65-6478-8006; Fax: 65-6478-9051

  • Author contributions: L.S.L: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; F.H.H: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; G.K.: data analysis and interpretation; L.W.S: conception and design, financial support, data analysis and interpretation, final approval of manuscript. L.S.L. and F.H.H. contributed equally to this article.

  • First published online in STEM CELLS EXPRESS August September 24, 2010.

  • §

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


The transcription factor Zic3 is required for maintenance of ESC pluripotency. By genome-wide chromatin immunoprecipitation (ChIP-chip) in ESCs, we have identified 379 direct Zic3 targets, many of which are functionally associated with pluripotency, cell cycle, proliferation, oncogenesis, and early embryogenesis. Through a computational analysis of Zic3 target sequences, we have identified a novel Zic3 consensus binding motif (5′-CCC/TGCTGGG-3′). ChIP results and in vitro DNA binding assays revealed that Zic3 binds with high affinity and specificity on the Nanog promoter. Here, we demonstrate that Zic3 functions as a transcriptional activator of the Nanog promoter in three ways: (a) Nanog transcript levels are sustained with Zic3 overexpression in differentiating ESCs, (b) Zic3 depletion in ESCs downregulates Nanog promoter activity, and (c) Zic3 overexpression leads to increased Nanog promoter activity. Furthermore, the activity of a mutant Nanog promoter with ablated Oct4/Sox2 binding is rescued by Zic3 overexpression to nearly wild-type levels. This indicates that Nanog is a positive transcriptional target of Zic3 in a mechanism that is independent of Oct4/Sox2 binding. Hence, we demonstrate an important pathway for regulation of Nanog expression in pluripotent ESCs through direct activation by Zic3. STEM CELLS 2010;28:1961–1969


At the core of ESC transcriptional circuitry are the key pluripotent regulators Oct4, Nanog, and Sox2. These transcription factors maintain the stemness state by activating genes involved in self-renewal and pluripotency while repressing genes involved in lineage specification [1, 2]. We previously reported that Zic3, a zinc finger transcription factor, operates directly downstream of Oct4, Nanog, and Sox2, and maintains pluripotency by functioning as a gatekeeper preventing endoderm specification in ESCs [3]. In addition, gene expression profiling of fully and partially reprogrammed mouse somatic cells revealed an upregulation of Zic3 relative to their original differentiated fibroblast states [4]. Zic3 expression has also been found in regenerating Xenopus limbs although at levels lower than in pluripotent cells [5]. These data support the idea that Zic3 is involved in regulating stemness-related programs in the contexts of pluripotency, self-renewal, and dedifferentiation.

Our previous results demonstrated that downregulation of Zic3 resulted in a significant decrease in Nanog expression and loss of pluripotency in ESCs [3]. Nanog was identified as an important ESC transcription factor through gain-of-function studies demonstrating its ability to maintain mouse ESCs in the absence of leukemia inhibitory factor (LIF) and feeder cultures [6, 7]. Studies have shown that Oct4 and Sox2 co-occupy the promoter of Nanog and positively regulate its expression in ESCs [2, 8, 9]. Nanog is known to be essential for propagation of ESCs in an undifferentiated state, and loss of Nanog results in spontaneous differentiation into primitive endoderm. This is similar to the cell type formed on ectopic expression of Gata4 and Gata6 in ESCs [10] and it is thought that Nanog maintains pluripotency through repression of Gata4 and Gata6 pathways to prevent primitive endoderm differentiation in ESCs. During development, the primitive endoderm gives rise to the anterior visceral endoderm (AVE), which is critical in specification of the anterior-posterior embryonic axis [11, 12]. AVE abnormalities have been reported in Zic3-null embryos, with inappropriate distal localization of AVE markers where a proximal-distal (P-D) to anterior-posterior (A-P) rotation should have occurred [13, 14]. This failure to undergo a P-D to A-P rotation appears to be the main cause for the Zic3-null embryos to exhibit failure to gastrulate, establish the primitive streak, and to form mesoderm [14].

We previously observed that loss of Zic3 resulted in downregulation of Nanog and an increase in endoderm specification in ESCs [3]. We hypothesized that Nanog is a direct target of Zic3 in embryonic stem cells and further examined if the upregulation of endoderm genes was due to a decrease in Nanog expression or a direct result of silencing by Zic3 in ESCs. In this study, we sought to identify genes directly regulated by Zic3 in ESCs, and to investigate the mechanisms of Nanog transcriptional regulation by Zic3.


Cell Culture

E14 mouse ESCs were cultured on 0.1% gelatin-coated dishes with ESC medium comprising Dulbecco's modified Eagle's medium (DMEM), 15% stem cell-grade Fetal bovine serum (FBS), 0.1 mM MEM nonessential amino acids, 2 mM L-glutamine, 0.1 mM (β-mercaptoethanol, and mouse LIF (1,000 U/ml, Chemicon, Temecula, CA). All reagents were obtained from Gibco-Invitrogen (Carlsbad, CA unless otherwise stated. HEK 293T cells were cultured in DMEM supplemented with 10% fetal bovine serum (Gibco-Invitrogen) and 2 mM L-glutamine (Gibco-Invitrogen). The cells were maintained in an incubator at 37°C with 5% CO2.

Generation of Zic3-Inducible ESCs

The Ainv18 cell line was a gift from George Daley (Harvard Medical School, Boston, MA) [15]. The Zic3 transgene was polymerase chain reaction (PCR)-amplified from a mouse cDNA library and subcloned into the pLox-N-tag-hemagglutinin (HA) vector (George Daley, Harvard Medical School, Boston, MA) with the following primers, restriction sites indicated in uppercase: Fwd 5′-tat-tat-GGT-ACC-tac-gat-gct-cct-gga-cgg-ag-3′ and Rev 5′-tcg-gca-TCT-AGA-tca-gac-gta-cca-ttc-gtt-aaa-att-g-3′. An additional nucleotide was inserted in the forward primer to allow in-frame processing of the Zic3 insert with the N-terminal HA tags.

For targeted insertion at the Hypoxanthine-guanine phosphoribosyltransferase (HPRT) locus, 20 μg each of pLox-N-tag-HA-Zic3 and pSALK-Cre were transfected with Lipofectamine 2000 (Invitrogen, Carlsbad, CA) into 6 × 105 AINV18 cells in 10-cm2 dishes. Mouse ESC medium was supplemented with 350 μg/ml G418 solution (Invitrogen) 24 hours after transfection and maintained for 14 days. Colonies arising from G418 selection were individually picked and expanded on neomycin-resistant mouse embryonic fibroblast (MEF)s. Site-specific integration was confirmed by PCR analysis with the following primers: LoxinF 5′-cta-gat-ctc-gaa-gga-tct-gga-g-3′ and LoxinR 5′–ata-ctt-tct-cgg-cag-gag-ca-3′. To induce overexpression, AINV18-Zic3 cells were treated with 1.0 μg/ml doxycycline in mouse ESC medium.


The RNA-interference (RNAi) constructs were previously described: Zic3 and nontargeting shRNA [3]. Briefly, RNAi oligonucleotides were cloned into the BglII/HindIII sites of the pSuper vector (Oligoengine, Seattle, WA, Constructs were sequence-verified and transfected into E14 cells with puromycin selection as previously described [3].

Real-Time PCR

RNA was extracted and DNase-treated for removal of gDNA contamination using the RNeasy minikit (Qiagen, Hilden, Germany, cDNA was synthesized with 1.0 μg total RNA using the High Capacity cDNA Archive kit (Applied Biosystems, Foster City, CA, as per manufacturer's instructions, and synthesis reactions were performed in a thermocycler at 25°C for 10 minutes followed by 37°C for 2 hours. Reversed transcribed cDNA samples were diluted 10× in nuclease-free water and quantitative PCR (qPCR)was performed using the TaqMan Universal PCR Mastermix and ABI TaqMan probes (Applied Biosystems). PCR reactions were conducted in triplicate in 384-well reaction plates with a final volume of 10 μl on the ABI Prism 7900 machine. Threshold cycle (CT) values were processed using the Comparative CT Method as described in the ABI Prism User Bulletin #2 (2001) and the Student's t test was used to compute statistical significance.


Western blotting was performed with anti-HA (1:1,000; sc-805, Santa Cruz Biotechnology) and anti-βactin (Invitrogen, 1:3,000) followed by HRP-conjugated donkey-anti-mouse or donkey-anti-rabbit secondary antibodies (Santa Cruz Biotechnology) at a dilution of 1:5,000. Zic3 antibodies were custom produced as follows: A peptide (AIASANSKDTTKT) was designed against a C-terminal region not conserved among other Zic family members and, for further specificity, subjected to a protein BLAST search (NCBI). Peptide synthesis and antibody production were performed according to the custom polyclonal antibody protocol by Biogenes GmBH (Berlin, Germany). Briefly, rabbits were immunized and boosted every 28 days for 3 months. Subsequent monthly bleeds were affinity-purified for total IgG (tris-glycine buffer pH 7.5, 250 mM NaCl, 0.02% thimerosal) and their specificity was tested by Western blots.

Chromatin Immunoprecipitation and Data Analysis

E14 cells were cultured to a density of 1 × 108 cells for each immunoprecipitation (IP). Two biological replicates were performed per experiment. Cells were cross-linked for 10 minutes at room temperature with 1% (w/v) formaldehyde and the reaction subsequently quenched with 125 mM glycine. Nuclear fractions were isolated and the DNA sheared to average lengths of 200–400 bp. Antibodies were used for immunoprecipitation as previously described [1]. qPCR for chromatin immunoprecipitation (ChIP) enrichment was performed on the ABI PRISM 7900 machine with SYBR Green PCR mastermix (Applied Biosystems, Foster City, CA, Fold-enrichment was determined by normalizing threshold cycle values of ChIP samples against sonicated whole cell DNA extract, and subsequently to a nonenriched ChIP control region set at a value of 1. All primers gave a single product as confirmed by agarose gel electrophoresis and heat dissociation analysis.

For analysis on mouse promoter arrays, purified ChIP material was processed according to the Agilent ChIP-on-chip protocol, and labeled DNA was hybridized to Agilent mouse promoter ChIP-on-chip arrays for 40 hours at 65°C (G4490A; Agilent Technologies, Santa Clara, CA, Chips were washed and scanned as per manufacturer's protocol and the results were processed with Agilent's ChIP Analytics software v1.3. A p value cutoff < .001 was specified in our analysis. To further minimize false positives, we applied a “neighborhood voting” algorithm [2] to filter for high confidence Zic3-enriched sites, wherein binding was considered genuine only in the presence of a second, significantly enriched, neighboring probe (p < .005). A total of 33 genes were selected at random for PCR validation of enrichment. Of these, 32 of 33 genes showed enrichment greater than our 2.5-fold threshold for positive enrichment (False discovery rate (FDR) < 0.03; Supporting Information Table 1).

For the Zic3 de novo motif search, 332 high-quality-enriched probes (normalized log2 ratio > 2) were selected from the initial list of 665 enriched probes. The probe sequences were equally extended on both ends to a final length of 300 bp and uploaded to the Weeder motif discovery algorithm [16]. Ubiquitously occurring promoter motifs were filtered out [17] and the remaining motifs were assessed for frequency of occurrence within Zic3-bound sites as previously described [1]. We identified three highly similar binding motifs that correlated strongly with Zic3 enrichment. An independent search was conducted for these three motifs within all Zic3-enriched regions, which identified 212 positive-scoring sites. The native motif sequences were extracted from these 212 sites and aligned to generate the final position weight matrix (PWM).

ChIP PCR Primers


Luciferase Assays

The following constructs were transfected into 6.5 × 104 cells seeded in 24-well plates: 375 ng firefly luciferase reporter, 2.0 ng of the Renilla luciferase vector (pRL-TK; ProMega, Madison, WI,, and 500 ng of the respective knockdown or overexpression construct. The cells were cultured for 2 days post-transfection before harvesting for luciferase assays. Luciferase activity was measured using the Dual Luciferase System (Promega) in a Centro LB960 96-well luminometer (Berthold Technologies, Transfections were performed in triplicate and on three independent occasions.

Protein Expression and Purification

The DNA binding domain (DBD) of mZic3, encompassing five zinc fingers (amino acid residues 237–411 of the full-length protein, Swiss-Prot entry Q62521) was PCR-amplified, using the primers 5′-GGGGACAAGTTTGTACAAAAAAG CAGGCTTCGAAAACCTGTATTTTCAGGGCGCCTTCTTC CGTTACATGC-3′ and 5′- GGGGACCACTTTGTACAAGA AAGCTGGGTTTAAGATTCATGAACCTTCATGTG-3′, which contains attB sites and tobacco etch virus protease cleavage site preceding the Zic3 coding sequence at the N-terminus. Protein expression and purification was performed as previously described [18]. Briefly, the PCR product was introduced into pENTR/TEV/D-TOPO entry vector by Gateway BP reaction (Invitrogen) and the resultant construct was sequence-verified and recombined into the pDEST-histidine and maltose binding protein (HisMBP) expression plasmid (Addgene plasmid 11085, [19] by Gateway LR reaction (Invitrogen). The final expression plasmid pDEST-HisMBP-mZic3-DBD was transformed into Escherichia coli BL21 (DE3) cells (Invitrogen) and cultured in Terrific broth supplemented with 100 μg/ml ampicillin and 0.2% glucose at 37°C. Heterologous protein expression was induced at an OD600 nm of 0.5–0.7 with 0.5 mM isopropyl-β,D-thiogalactopyranoside at 18°C overnight. Cells were centrifuged and the pellet was resuspended in lysis buffer (10 mM Hepes pH 7.3, 100 mM NaCl, 30 mM imidazole) and sonicated. The crude bacterial lysate was centrifuged and the supernatant was incubated with Ni-Sepharose beads (GE Healthcare, Uppsala, Sweden, at 4°C for 2 hours. The mZic3-DBD-His6-MBP protein was eluted with elution buffer (10 mM Hepes pH 7.3, 100 mM NaCl, 300 mM imidazole) and treated with 0.1% polyethyleneimine in the presence of 1 M NaCl to remove contaminating DNA. The mZic3-DBD-His6-MBP protein was then further purified by size-exclusion chromatography using a HiLoad 16/60 Superdex 200 pg (GE Healthcare) in storage buffer (10 mM Tris pH 8, 100 mM NaCl). Fractions containing mZic3-DBD-His6-MBP proteins were pooled and concentrated to 20 μM. The protein identity of mZic3-DBD-His6-MBP was verified by Matrix-assisted laser desorption/ionization-Time of flight-Time of flight (MALDI-TOF-TOF).

Electrophoretic Mobility Shift Assay

Electrophoretic mobility shift assay (EMSA) was performed as previously described [18]. Briefly, protein was incubated with Cy5-labeled DNA probes in binding buffer (10 mM Tris pH8, 0.1 mg/ml bovine serum albumin, 50 μM ZnCl2, 100 mM KCl, 0.5 mM MgCl2, 10% glycerol, 0.1% NP-40, 2 mM β-mercaptoethanol) at 4°C for 1 hour. The reaction was then subjected to 10% native Tris-glycine polyacrylamide gel electrophoresis. Gel was analyzed with a Typhoon scanner (GE Healthcare). The affinity of protein to DNA was determined by titrating 0–500 nM (twofold serial dilution) of protein against 1 nM probe. The quantities of bound and unbound probes were obtained using the ImageQuant TL software (GE Healthcare). The bound fraction was then plotted against the protein concentration using IGOR Pro software (WaveMetrics, The curve was nonlinearly fitted and the dissociation constant (Kd) was determined as previously described [20]. Specificity of protein-DNA interaction was verified by competition assay, whereby 0–250 nM (twofold serial dilution) of competitor containing either the wild-type (WT) or mutated sequences of the consensus motif was titrated against 1 nM probe to compete for binding to protein at concentration equivalent to Kd (3 nM).


Identification of Zic3 Binding Site and Consensus Motif on the Nanog Promoter

Nanog expression is dependent on Zic3 in ESCs and depletion of Zic3 by RNAi results in a significant downregulation of Nanog [3]. Thus, we hypothesize that the Nanog promoter is occupied and directly regulated by Zic3. To determine if Zic3 physically occupies the Nanog promoter, we performed Zic3 ChIP in mouse ESCs. ChIP enrichment was assayed with 120-bp-tiled PCR Amplicons A–E (Fig. 1A) across the Nanog promoter and Zic3 enrichment of 21-fold was observed in Amplicon C, whereas the control ChIP with a nonspecific antibody raised in the same species demonstrated no enrichment (Fig. 1B).

Figure 1.

The Nanog promoter is physically occupied by Zic3. (A): Zic3 ChIP-polymerase chain reaction (PCR) was performed with 120-bp-tiled PCR regions (A–E) across the Nanog promoter. (B): Zic3 peak enrichment of 21-fold was observed in Amplicon C relative to a control ChIP with a nonspecific GST antibody. (C): A novel Zic3 consensus motif was defined through interrogation of Zic3 ChIP-enriched sequences using the WEEDER algorithm. This motif was present on the Nanog promoter between −154 to −146 of the transcription start site, corresponding to ChIP-PCR peak enrichment at Amplicon C. Abbreviation: ChIP, chromatin immunoprecipitation.

To determine if a Zic3 DNA binding motif was located within the ChIP-enriched region on the Nanog promoter, we derived a novel Zic3 binding sequence in ESCs through interrogation of Zic3 binding sites across ∼17,000 promoters of defined mouse transcripts, spanning −5.5 kb upstream to +2.5 kb downstream of transcription start sites (Agilent ChIP-on-chip arrays; G4490A; Agilent Technologies). This analysis identified 665 Zic3-enriched regions corresponding to 379 unique promoter regions that represent Zic3 regulatory targets. To understand the broad functions of Zic3 in ESCs, these target genes were annotated using the Panther [21] and Kyoto Encyclopedia of Genes and Genomes (KEGG) [22] pathway databases, and those that clustered under ESC-associated roles of pluripotency, cell cycle, proliferation, oncogenesis, and early embryogenesis were highlighted and summarized (presented in the Discussion as Fig. 4). Full details of enriched probes and associated genes are provided in Supporting Information Tables 2 and 3, and the original Agilent array dataset has been uploaded to the Gene Expression Omnibus database (accession number GSE22195).

DNA binding conservation for Zic3 was determined using high-quality-enriched sequences (normalized log2 > 2) in the Weeder motif discovery program [16]. This algorithm yielded three similar consensus sequences that were highly correlated with Zic3 ChIP enrichment. All 379 Zic3-occupied promoters were subjected to an independent scan with these three sequences, and 212 regions scored positive for at least one Zic3 motif. Native DNA sequences for these 212 regions were extracted, aligned, and scored for the final Zic3 consensus PWM in Figure 1C.

We scanned the Nanog promoter sequence for this novel Zic3 binding sequence and identified a close match to the consensus motif between −154 to −146 of the transcription start site (UCSC AK010332; Fig. 1A). This location corresponded to Zic3 peak binding within the 120-bp Amplicon C sequence, which was enriched 21-fold relative to a nonspecific antibody control (Fig. 1B). As the chromatin was sheared to lengths between 200 and 400 bp, Amplicon D which is 9 bp downstream of Amplicon C also demonstrated 14-fold enrichment. The subsequent tiled amplicon (Amplicon E) is 120 bp away from Amplicon C and was not enriched by PCR, indicating the specificity of our PCR assay for the Zic3 binding region. Only Amplicon C, which is 108 bp in length, contains the Zic3 motifs and therefore shows the highest enrichment by PCR. Taken together, these data indicate that Zic3 occupies the Nanog promoter at its binding motif where peak ChIP-PCR enrichment was detected.

Zic3 Binds to Its Consensus Motif with High Affinity and Specificity

To validate physical occupancy of Zic3 on the consensus motif, the Zic3 DNA binding domain was epitope-tagged at the N-terminus with histidine and maltose binding protein (mZic3-DBD-HisMBP), expressed in bacteria, and affinity purified. We examined the DNA binding affinity of mZic3-DBD-HisMBP by EMSA on a 16 bp Cy5-labeled Nanog promoter sequence comprising the putative consensus binding motif (Fig. 2A). The Nanog-Cy5 probe was incubated with 0.24–500 nM (twofold serial dilutions) of mZic3-DBD-HisMBP. We observed the ability of mZic3-DBD-HisMBP protein to bind and shift the probe in a dose-dependent manner (Fig. 2B; lanes 1–13), whereas the HisMBP tag alone did not shift the probe (Fig. 2B; lane 14).

Figure 2.

Zic3 binds with high affinity and specificity to its novel consensus motif. (A): Sequence of the Zic3 consensus motif on the Nanog promoter. Unlabeled sequences of the WT and mut sequences were used in the competitive gel shift assay. (B): EMSA demonstrating the binding affinity of Zic3 to the consensus binding motif on the Nanog promoter. Zic3-DNA binding domain (DBD) was incubated with the Cy5-labeled Nanog promoter sequence comprising the consensus motif, with HisMBP used as a negative control. (C): Competitive gel shift assay using WT competitors. Unlabeled WT competitor was allowed to compete for binding to Zic3-DBD against the Cy5-labeled Nanog sequence. (D): Competitive gel shift assay using mutant competitors. Unlabeled mutant competitor was allowed to compete against the Cy5-labeled Nanog sequence for binding to Zic3-DBD. (E): Plot of fraction of Cy5-Nanog bound to purified Zic3-DBD demonstrates that Zic3 binds to the consensus binding motif at a Kd of 2.4 nM. Error bars represent SEM of two biological replicates. Abbreviations: EMSA, electrophoretic mobility shift assay; HisMBP, histidine and maltose binding protein; Kd, dissociation constant; mut, mutated; WT, wild-type.

We further incubated mZic3-DBD-HisMBP with increasing concentrations of an unlabeled 16 bp competitor sequence (0–250 nM). The unlabeled competitor sequence effectively competed with the Nanog-Cy5 sequence for binding to mZic3-DBD-HisMBP (Fig. 2C). The Zic3 motif on the unlabeled competitor was then randomly scrambled to maintain the same GC content (Nanog mut sequence; Fig. 2A) and allowed to compete with the Nanog-Cy5 probe for binding to mZic3-DBD-HisMBP. The scrambled sequence did not compete with its corresponding probe for binding to mZic3-DBD-HisMBP (Fig. 2D) indicating that Zic3 occupancy on this novel consensus sequence is highly specific. Zic3 binds to the consensus motif with high affinity, yielding a Kd of 2.4 nM (Fig. 2E). These data collectively indicate that Zic3 recognizes and occupies its consensus motif on the Nanog promoter with high specificity and affinity.

Zic3 Is a Functional Activator of the Nanog Promoter

We previously demonstrated that loss of Zic3 expression in ESCs resulted in downregulation of Nanog [3]. Here, we were interested to test if the inverse is true, that is, whether Zic3 overexpression sustains Nanog levels in differentiating ESCs. Clonal mouse ESC lines were generated to express HA-tagged Zic3 from a doxycycline-inducible promoter [15]. Two HA-tagged Green fluorescent protein (GFP) overexpressing cell lines served as controls. Figure 3A demonstrates the specificity of doxycycline regulation in two Zic3-overexpressing clones. Nondoxycycline-induced clones (day 0) did not express exogenous HA-Zic3 protein, whereas strong expression was observed at day 2 and day 4 in doxycycline-supplemented ESC media (Fig. 3A). These data demonstrate the specificity and sensitivity of doxycycline induction in our Zic3-overexpressing cell lines.

Figure 3.

Zic3 functionally activates the Nanog promoter independently of the Oct4/Sox2 heterodimer. (A): Clonal ESC lines expressing HA-tagged Zic3 from a doxycycline-inducible promoter. Exogenous Zic3 protein was not expressed in the absence of doxycycline. (B): Zic3-overexpressing lines maintained Zic3 and Nanog transcript expression after 4 days of differentiation (p < .01), whereas downregulation was observed in Green fluorescent protein (GFP)-overexpressing controls. Oct4 and Sox2 transcript were downregulated in both Zic3-overexpressing and control cell lines. All quantitative polymerase chain reaction assays were performed in triplicate and significance was computed using the Student's t test. (C): Zic3-overexpressing lines showed a mixture of differentiating cells and ESC colonies after 4 days differentiation. (D): Nanog promoter-luciferase reporter (−212 to +117) comprising the Zic3 consensus motif and Oct4/Sox2-binding element (pNanog). (E): RNA-interference knockdown of Zic3 in ESCs resulted in downregulation of WT-pNanog activity (p < .01). This activity was further reduced with mutation of the Oct/Sox motif (mutOct/Sox-pNanog). (F): Zic3 overexpression resulted in upregulation of WT-pNanog activity in ESCs (p < .01) and is further able to rescue the activity of the Nanog promoter with a mutant Oct/Sox element. Luciferase experiments were performed in triplicate on three separate occasions, and significance was assessed by the Student's t test. Abbreviations: GFP, Green fluorescent protein; HA, hemagglutinin; mESCs, mouse ESCs; WT, wild-type.

LIF withdrawal is a well-characterized method of inducing ESC differentiation [23]. To determine if Zic3 overexpression sustains Nanog transcript levels in differentiating ESCs, Zic3- and GFP-inducible lines were exposed to doxycycline in media without LIF. Three biological replicates were performed per clonal line, and Nanog transcript expression was assayed to assess Zic3 regulation on the Nanog promoter. After 4 days of LIF withdrawal, we measured Zic3 transcript expression in these cells compared with WT undifferentiated mouse ESC levels. Zic3 levels were detected in GFP-overexpressing lines at 43% relative to undifferentiated ESCs (-LIF-GFP-O/E; Fig. 3B), in agreement with our previous report that Zic3 is downregulated on ESC differentiation [3]. In contrast, Zic3-overexpressing lines maintained Zic3 transcript at 94% compared with undifferentiated ESCs (-LIF-Zic3-O/E; Fig. 3B). Zic3-overexpressing lines expressed 2.1-fold higher Zic3 levels relative to GFP controls, at a level similar to WT undifferentiated ESCs (p < .01; Fig. 3B).

We next determined Nanog levels in response to sustained Zic3 expression in differentiating ESCs. Interestingly, Zic3-overexpressing lines maintained Nanog at 78% relative to undifferentiated ESCs, whereas in GFP-control lines Nanog transcript was reduced to 39% (Fig. 3B). Zic3 overexpression therefore sustained Nanog expression twofold higher than control cells undergoing differentiation (p < .01). In contrast, Oct4 and Sox2 transcript levels were downregulated to a similar extent in Zic3-overexpressing and GFP-control lines, indicating that Zic3 overexpression is insufficient to prevent downregulation of Oct4 and Sox2 expression in differentiating ESCs (Fig. 3B). In accordance with this data, Zic3-overexpressing lines showed a mixture of phase-dark differentiated cells and phase-bright ESC colonies that differentiated after repeated passage (Fig. 3C) and demonstrated a nonsignificant increase in Oct4 and Sox2 levels. These results indicate that Zic3 has the ability to significantly sustain Nanog expression in differentiating ESCs, albeit at levels insufficient to maintain pluripotency after extended periods in culture.

Zic3 Activates the Nanog Promoter Independently of Oct4 and Sox2

We were interested to determine the extent Zic3 functions as an activator of the Nanog promoter. A Nanog promoter-luciferase reporter construct was used comprising our validated Zic3 consensus motif and the Oct4/Sox2 consensus element (−212 to +117; Fig. 3D) [8, 9]. Luciferase assays were performed as previously described [3]. On RNAi knockdown of Zic3 in ESCs, Nanog promoter activity was downregulated 2.3-fold relative to the nontargeting control, further verifying transcriptional activation of Nanog by Zic3 (WT-pNanog; p < .01; Fig. 3E).

A well-characterized composite binding element for key ESC transcription factors Oct4 and Sox2 is located 12 bp upstream of our novel Zic3 element on the Nanog promoter (Fig. 3D) [8, 9]. We were interested to determine if functional activation of the Nanog promoter by Zic3 is dependent on the Oct4/Sox2 heterodimer. A Nanog promoter with mutated Oct/Sox binding sites was previously constructed, where Oct4/Sox2 binding is ablated and Nanog promoter activity is consequently downregulated [9]. We measured the activity of Zic3 on this mutant luciferase promoter construct, and consistent with results demonstrated by Rodda et al. [9], we observed an 8.5-fold decrease in mutant promoter intrinsic activity relative to the WT sequence (mutO/S-pNanog; Fig. 3E). RNAi-mediated knockdown of Zic3 greatly enhanced this effect, resulting in 19.5-fold downregulation relative to WT promoter activity (Fig. 3E). Thus, decrease in Zic3 expression levels further reduces Nanog promoter activity 2.3-fold in the absence of Oct4/Sox2 binding (mutO/S-pNanog; p < .01; Fig. 3E).

We next examined if the inverse effect with Zic3 overexpression was true. Corroborating with our earlier results, Zic3 overexpression resulted in a statistically significant threefold upregulation of WT-pNanog activity in ESCs (p < .01; Fig. 3F), and further confirms functional activation of the Nanog promoter by Zic3. The intrinsic activity of the sequence was reduced twofold relative to WT promoter (Fig. 3F). Interestingly, when mutOct/Sox-pNanog activity was assayed with Zic3 overexpression, a statistically significant fivefold increase was observed. This resulted in a rescue of mutOct/Sox-pNanog activity to almost WT promoter levels (p < .01; Fig. 3F). These data collectively indicate that Zic3 is a positive regulator of the Nanog promoter in ESCs and mediates this activation independently of Oct4 and Sox2.


The detailed molecular pathways by which Zic3 mediates its action on pluripotency remain as yet unknown. We hypothesized that Zic3 directly regulates Nanog in ESCs and were interested to determine its direct action on the Nanog promoter. To test this hypothesis, we defined a consensus Zic3 DNA-binding motif for specific recognition of its gene targets by establishing a global view of Zic3 occupancy in ESCs. Our ChIP analysis identified 379 unique Zic3-occupied promoter targets that classify under ESC-related functions of pluripotency, cell cycle, proliferation, oncogenesis and early embryogenesis. Figure 4 is a summary of Zic3-bound genes representing potential Zic3-regulated pathways in pluripotent ESCs and cells undergoing early differentiation. The complete gene list is provided in Supporting Information Table 3. (Note that the Agilent ChIP-chip array does not contain Nanog promoter probes and Zic3 ChIP enrichment on the Nanog promoter is therefore not reflected on this list.)

Figure 4.

Zic3 target genes identified by chromatin-immunoprecipitation. Zic3 binds to the promoters of genes involved in pluripotency, cell cycle control, oncogenesis, cell proliferation, chromatin remodeling, and early embryonic development.

Zic3 occupies the promoters of ESC genes Nanog, Oct4, Sox2, Rif1, and Phc1 and may signal ESC pluripotency through the upregulation of these genes. Here, we have demonstrated strong and specific activation of the Nanog promoter by Zic3. This accounts for the significant downregulation of Nanog induced by loss of Zic3 expression [3] and sustained expression of Nanog with Zic3 overexpression in differentiating ESCs. To further determine the specificity of Zic3 DNA target recognition in ESCs, we derived a novel Zic3 consensus binding sequence inferred from Zic3-bound DNA sequences. The Zic protein family was previously found to bind Gli DNA target sequences with low affinity in vitro, leading to the suggestion that Zic3 does not come in direct contact with DNA but is dependent on Gli as a transcriptional coactivator for target recognition [24, 25]. The novel Zic3 motif characterized in this study, 5′- CCC/TGCTGGG -3′, partially resembles the motif reported by Mizugishi et al. 5′-GGGTGGTC-3′ in that both have heavy preferences for guanine (G) and cytosine (C). However, we show that Zic3 binds to our novel consensus sequence with high affinity and specificity, suggesting that Zic3 binds directly to DNA and independently regulates transcription in ESCs.

Furthermore, we demonstrate that overexpression of Zic3 alone is capable of rescuing downregulated Nanog promoter activity with mutant Oct/Sox binding sites. This indicates that Zic3 has the capacity to function as a strong transcriptional activator in ESCs independent of the Oct4/Sox2 heterodimer, supporting a model that Zic3 and Oct4/Sox2 have distinct but complementary roles as transcriptional activators of Nanog (Fig. 5A).

Figure 5.

Transcriptional regulation of the Nanog promoter. (A): Nanog promoter transcriptional activation pathways in ESCs. Ovals represent genes and rectangles represent proteins. Oct4, Sox2, and Nanog are transcriptional activators of Zic3. Two pathways are illustrated for Nanog promoter activation. (1) The Oct4/Sox2 heterodimer binds to the Nanog promoter and positively regulates Nanog transcription [8, 9]. (2) In the absence of Oct4/Sox2 binding, we show that Zic3 is a strong activator of the Nanog promoter, suggesting that Zic3 positively signals Nanog transcription in a mechanism distinct from Oct4/Sox2 binding. (B): Multiple transcription factor binding on the Nanog promoter. Zic3 binds to the Nanog promoter within a densely occupied 100 bp MTL comprising Oct4, Sox2, Esrrb, and Klf4 [26] identified by chromatin immunoprecipitation-sequencing technology. Abbreviations: GCNF, Germ Cell Nuclear Factor; MTL, multiple transcription factor binding locus.

We previously demonstrated that Zic3 is directly regulated by Oct4, Nanog, and Sox2 and functions as a gatekeeper that blocks the specification of the endoderm lineage genes, thus preventing differentiation in ESCs [3]. Nanog is known to prevent differentiation of ESCs into endoderm and downregulation of Nanog in ESCs leads to an upregulation of Gata4 and Gata6 expression and endoderm-lineage specification [7, 27, 28]. We were interested to determine whether Zic3 mediates its ESC gatekeeper effects indirectly by sustaining Nanog expression or whether it also directly occupies and silences the promoters of endoderm-specific genes. The list of Zic3-bound target promoters did not include key endoderm genes that we previously reported to be highly upregulated with Zic3 knockdown, including Foxa2, Sox17, Gata4, and Gata6. It is therefore likely that Zic3 prevents differentiation of ESCs into the endoderm lineage through its maintenance of Nanog expression.

The regulation of Nanog expression has been attributed to several transcription factors (summarized in Fig. 5B). Nanog expression is positively regulated by Brachury [29], Klf2, Klf4, Klf5 [30], and Nanog-Sall4 heterodimer [31], which bind the distal enhancer, as well as FoxD3 [32] and Oct4-Sox2 heterodimer [8, 9], which bind the proximal promoter. Nanog expression has also been shown to be negatively regulated by Tcf3, Germ Cell Nuclear Factor (GCNF), and p53. Although the expression of Nanog is regulated by several transcription factors, it is the proximal promoter region (289 bp upstream of transcription start site (TSS)) that contains Oct4 and Sox2 binding sites that has been shown to contain sufficient cis-regulatory elements for Nanog expression [9]. However, it was also shown that Nanog expression could be maintained in the absence of Oct4 [6], suggesting that other transcription factors such as FoxD3 that bind the proximal promoter of Nanog contributes to the expression of Nanog.

More recently, a comprehensive ChIP-sequencing study on the binding profile of 13 transcription factors in mouse ESCs identified a large number of transcription factor binding peaks located within 100 bp of each other [26]. Loci co-occupied by four or more transcription factors were termed multiple transcription factor loci (MTL), and these were found distributed throughout the genome in intergenic, promoter, and gene coding regions. Interestingly, our newly identified Zic3 binding site on the Nanog promoter was located within an MTL located at chr6:122672970–122673163 comprising Oct4, Sox2, Esrrb, and Klf4 (Fig. 5B). None of the other ESC-related transcription factors, including Nanog and Stat3, were identified to bind within this densely occupied region and the closest binding transcription factor outside this region, nMyc, is located 86 bp downstream. A high occurrence of Esrrb and Klf4 binding sites were associated with Nanog-Oct4-Sox2 MTL, suggesting that these transcription factors operate in a coregulatory complex in ESCs [26]. Here, we have demonstrated that Zic3 is a strong positive regulator of the Nanog promoter and its regulatory activity is not dependent on the Oct4/Sox2 heterodimer. However, it is clear that their action on the Nanog promoter is additive, as both Zic3 and Oct4/Sox2 regulate the Nanog promoter in a similar manner and loss of either results in downregulation of Nanog promoter activity. It would be interesting to further examine the association of Zic3 with the Oct4-Sox2-Klf4-Esrrb regulatory unit to determine their interaction and potential for coregulation of downstream target genes.


Through a global promoter binding site survey of Zic3-bound sequences, we have identified a novel Zic3 consensus element that shows high specificity and affinity of binding on the Nanog promoter. Downregulation of Zic3 leads to a decrease of Nanog expression [3], whereas sustained expression of Zic3 prevents Nanog transcript levels from decreasing rapidly in differentiating ESCs. This observation is explained by our data indicating that Zic3 is a strong activator of the Nanog promoter independent of Oct4/Sox2 binding, and that Zic3 does not directly block endoderm specification but does so by maintaining Nanog expression in ESCs. These data shed light onto critical mechanisms in maintenance of pluripotency, and further, defines the role of Zic3 in relation to key transcription factors specifying the properties of pluripotency and self-renewal unique to ESCs.


We thank Dr. Ralf Jauch for helpful discussions and Yiting Lim and Siew Hua Choo for their invaluable assistance. We also thank Tapan Kumar Mistri for his advice on IGOR Pro software. This work was supported by the Biomedical Research Council and Agency for Science, Technology and Research (A*STAR). F.H.H. is supported by the NUS Graduate School for Integrative Sciences and Engineering Scholarship.


The authors indicate no potential conflicts of interest.