Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Oct 1.
Published in final edited form as: Cell Stem Cell. 2015 Jul 30;17(4):462–470. doi: 10.1016/j.stem.2015.07.005

Ground state conditions induce rapid reorganization of core pluripotency factor binding before global epigenetic reprogramming

Christina Galonska 1,2,3,4, Michael J Ziller 1,2,3,4, Rahul Karnik 1,2,3, Alexander Meissner 1,2,3
PMCID: PMC4592414  NIHMSID: NIHMS708991  PMID: 26235340

SUMMARY

Mouse embryonic stem cells (mESCs) cultured under serum/LIF conditions exhibit heterogeneous expression of pluripotency-associated factors that can be overcome by two inhibitors (2i) of the MEK and GSK3 pathway, respectively. Several studies have shown that the “ground state” induced by 2i is characterized by global hypomethylation and specific transcriptional profiles, but little is known about the contributing effectors. Here we show that 2i conditions rapidly alter the global binding landscape of OCT4, SOX2 and NANOG. The dynamic binding influences enhancer activity and shows enrichment for regulators linked to Wnt and Erk signaling. Epigenomic characterization provided limited insights to the immediate transcriptional dynamics suggesting these are likely more secondary effects. Likewise, knock-out of the PRC2 component EED to prevent H3K27me3 deposition had minimal effect on the transcriptome, rendering it largely dispensable for continued repression of bivalent genes and de novo silencing in 2i.

Graphical Abstract

graphic file with name nihms708991u1.jpg

INTRODUCTION

Under common culture conditions in serum plus Leukemia Inhibitory Factor (LIF) mouse embryonic stem cells (mESCs) receive partially opposing signaling cues, as serum/LIF induced STAT3 signaling favors the pluripotent state, but does not block the activation of Fgf/Erk signaling involved in cellular differentiation (Kunath et al., 2007). To suppress these differentiation cues, mESCs can be maintained in the absence of serum using defined conditions that include two kinase inhibitors (2i) (Ying et al., 2008). PD0335901 (PD) inhibits the function of mitogen-activated protein kinase kinase (MAPKK), thereby inhibiting the Erk signaling cascade, and CHIR99021 (CHIR) selectively targets glycogen synthase kinase-3 (Gsk3), resulting in an enforcement of canonical Wnt signaling. This “2i state” of so-called naïve pluripotency has been shown to be very distinct in its transcriptional and epigenetic profiles from ESCs cultured in serum/LIF (Habibi et al., 2013; Marks et al., 2012). The state is marked by the specific up-regulation of genes involved in metabolism and stem cell maintenance, while developmental and differentiation associated genes become more repressed (Marks et al., 2012). The observed differences in the transcriptional state further coincide with drastic changes in the epigenetic landscape of the cells, i.e. a reduction of H3K27me3 at bivalent gene promoters and a global loss of DNA methylation (Ficz et al., 2013; Habibi et al., 2013; Marks et al., 2012).

Another hallmark of the 2i transition is the more homogenous distribution of NANOG protein, switching from fluctuating (lower and higher) levels under serum conditions to an increased expression and more uniform protein distribution under 2i conditions (Silva et al., 2009), pointing to a possible stabilization and reinforcement of the core pluripotency circuitry in 2i. Although it is known that the pluripotent state is highly dependent on the intrinsic gene regulatory network centered around the core pluripotency factors OCT4, SOX2, and NANOG (OSN) (Chen et al., 2008; Kim et al., 2008) and that cell state transitions involve the remodeling of the core network (Buecker et al., 2014; Chan et al., 2013), little is known about the specific impact that the two inhibitors have on the OSN target spectrum.

Given the broad interest in the “2i state” and the dramatic cellular state changes that are induced through the inhibitors, we decided to investigate the effects on OSN binding and study their behavior early under the distinct culture conditions. To capture putative regulatory events, upstream of major transcriptional and epigenetic dynamics, we focused on mESCs 24 hours post switch (serum/LIF to 2i/LIF, hereafter referred as “serum” and “2i”) and measured gene expression, DNA methylation, selected histone modifications and OSN binding. Integrative analysis of this multi-layered data set together with select functional validation experiments including enhancer activity assays and PRC2 mutant ES cells provide new mechanistic insights that continue to advance our understanding of this unique cell state.

RESULTS

Reorganization of core pluripotency factor binding in the early 2i transition

Several recent studies have highlighted the transcriptional and epigenetic dynamics between cells in serum versus 2i (Ficz et al., 2013; Marks et al., 2012). Notably, the induction and repression of selected factors (e.g. Prdm14, Dnmt3a/3b) has been reported as early as 24h after addition of the two inhibitors, pointing to a rapid transition of the ESC transcriptome (Ficz et al., 2013). Using RNA sequencing we were able to identify 1013 genes that significantly changed their expression in the initial 2i transition (24h 2i) and partially overlapped with the 1413 altered genes in the 2i population after 3 passages (3P) (Figure S1A, Table S1). We grouped the differentially expressed genes across all three conditions into six clusters (Figure 1A, Figure S1B). This captures recently described changes such as the down regulation of Otx2 and up regulation of Klf2 (Yang et al., 2014; Yeo et al., 2014). Genes with moderate but not yet significant expression changes in the initial transition (cluster 2 and 5) encompass factors such as Prdm14 and Eras, which become highly divergent later in the transition (Figure S1B). Overall, nearly half of the genes that become repressed in 2i and a quarter of genes that become up regulated in 2i show already altered expression within the first 24h (Figure 1B). Oct4 and Sox2 show little expression change upon switching conditions, while Nanog expression increases by around 1.5-fold (Figure S1C). It is worth noting that besides Nanog, we find several other pluripotency related genes that are differentially expressed in the initial 2i transition, including Dppa3 and Utf1 at the extreme ends (Figure 1C). Utilizing our previously established Utf1-mCitrine reporter line (Galonska et al., 2014), we confirmed that Utf1 is rapidly and uniformly repressed upon switching to 2i (Figure 1D). Additionally, STELLA and NANOG appear more homogenously distributed across the population within the first 24h in 2i (Figure 1E).

Figure 1. Globally reconfigured binding of the core pluripotency factors in the early 2i transition.

Figure 1

Open in a new tab

A. Expression data (z score log2 count) for cells cultured in serum, 24h 2i or 3 passages 2i (3P). Only genes with FPKM>1 in at least one of the conditions are shown.

B. Venn diagrams showing the overlap between all significantly downregulated and upregulated genes in 24h 2i vs. serum as compared to 3P 2i vs. serum. Diagrams are drawn to scale.

C. Expression (log2 fold change) of pluripotency associated transcription factors in serum vs. 24h 2i. The error bars represent one standard error over two replicates.

D. FACS sorting showing the fluorescent intensity of Utf1-mCitrine reporter cells in serum vs. 24h 2i. On the right: Images of the reporter cell line in serum and after switching cells 24h to 2i. Scale bar: 150μm.

E. Immunofluorescent pictures of STELLA and NANOG in serum vs. 24h 2i. Protein distributions of STELLA and NANOG are more homogenous after switching 24h to 2i. Scale bar: 80μm.

F. Heatmap showing all differential binding sites (z score log2 count) for SOX2, OCT4, NANOG, H3K4me3 and H3K27me3 in serum and 24h 2i. Numbers of higher (up) and lower (down) enriched sites are shown below.

G. Pie charts illustrating the proportion of differentially enriched SOX2, OCT4, NANOG, H3K4me3 and H3K27me3 sites that overlap with promoters (±1kb of TSS) or enhancers based on previously published p300 binding sites in mESCs in serum (Creyghton et al., 2010). All regions not overlapping with either one of these two were classified as other regions.

H. Scatterplot showing all significantly up- and downregulated genes after 24h 2i that are (brown dots) or are not (black dots) differentially bound by OCT4, SOX2 or NANOG. 29% of the upregulated genes in 2i become differentially bound by the core factors (left pie chart) and 46% of all downregulated genes in 2i become differentially bound by the core factors (right pie chart).

To gain more mechanistic insights into the early regulatory events we analyzed the genome-wide binding patterns of OCT4, SOX2 and NANOG (OSN) under both conditions. To be able to quantify both new and altered binding patterns we utilized established state of the art analytical methods based on a negative-binomial model and variance pooling as implemented in the publically available software packages diffbind and DESeq2. Using this approach we find that all three factors become subject to widespread binding rearrangements, with 14,796 sites exhibiting differential NANOG binding, 1,352 sites with differential SOX2 binding and 503 sites with differential OCT4 binding with a minimal two-fold change at an FDR of 0.1 (NANOG, SOX2) or a p-value ≤0.05 (OCT4) (Figure 1F, Table S2). Clustering these regions reveals mostly coordinated binding changes for all three factors (Figure 1F). It should be noted that many of the observed binding changes were subtler for OCT4 and SOX2, thereby frequently falling below our threshold criteria for differential binding sites (Figure S1D). Importantly, the majority of differentially enriched binding sites are shared between 24h and P3 (Figure S1E), suggesting that early binding changes are largely maintained. Gene ontology analysis shows differentially bound sites are enriched for biological processes associated with development and differentiation (Table S3), in line with OSNs known targeting of developmental genes in serum (Kim et al., 2008).

Analysis of H3K4me3 and H3K27me3 patterns showed that the majority of genomic sites with altered OSN binding are distinct from the sites that show dynamics for these two histone modifications (Figure 1F). This is largely expected based on our observations that the core pluripotency factors are more frequently found at distal putative regulatory sites (Gifford et al., 2013; Ziller et al., 2013) and the fact that H3K4me3 and H3K27me3 can be found primarily at promoters or intragenic regions (Figure 1G). While initially less dynamic, H3K4me3 and H3K27me3 are further reorganized at later stages of the transition (Figure S1F). To evaluate if the observed OSN binding changes in 2i are indeed state-specific rather than the result from transitioning to a more homogenous cell population, we utilized a NANOG GFP reporter line whose serum population consists of a heterogeneous mixture of GFP negative (NANOG not present) and GFP positive cells (NANOG present). Next, we isolated the GFP positive population in serum (Figure S1G), and compared its NANOG binding pattern to the 2i population. Although a fraction of binding sites changed between the whole serum population and the GFP positive subpopulation, the global rewiring that occurs in the 2i transition was not seen. Thus, the GFP positive fraction exhibited equally many binding changes to the 2i state as the whole serum population (Figure S1H). This is in agreement with previous results showing that the transcriptome of REX1 positive sorted serum cells is still highly divergent from the homogenous REX1 positive 2i population (Marks et al., 2012).

Lastly, to assess the effect of the reconfigured binding on transcription in more detail, we correlated differential binding with expression of the nearest gene (≤ 50K of TSS). A large fraction of all differentially expressed genes in 2i are bound by OSN in serum (Figure S1I). Specifically, 46% of all up regulated genes and 29% of all down regulated genes exhibit a differential OSN binding pattern after 24h in 2i (Figure 1H). However, considering the number of differentially bound sites and differentially expressed genes, this association is not statistical significant in all cases: While genes repressed upon transition to 2i were enriched for OSN sites that show decreased binding in 2i, up regulated genes were equally enriched for sites with lower and higher OSN binding (Figure S1J).

Differentially bound sites are enriched at distal gene regulatory elements

Our analysis points towards an enrichment of dynamic regions at distal putative regulatory sites, which is in line with two recent studies showing the transition to alternative pluripotent states (e.g. 2i to EpiSC cells) is accompanied by remodeling of the cellular enhancer landscape (Buecker et al., 2014; Factor et al., 2014). A well-known example for differential enhancer use in primed vs. naïve pluripotency is the switch from the Oct4 proximal (serum) to the distal (2i) element (Tesar et al., 2007). Although we do not observe a complete switch in binding between these elements, we find the proximal enhancer shows decreased NANOG binding upon switching to 2i, while the distal enhancer showed a small increase in SOX2 (Figure 2A).

Figure 2. Differential OSN binding can be linked to differential enhancer activity in 2i.

Figure 2

Open in a new tab

A. ChIP-seq tracks showing the binding of NANOG, SOX2, OCT4 and p300 in serum and 24h 2i (p300 only in serum) along the Oct4 proximal (PE) and distal (DE) enhancer elements. Serum and 24h 2i ChIP-seq tracks are scaled to the same height for each factor and locus. Differential binding for NANOG and SOX2 is highlighted in grey. Regions tested in the luciferase assay (panel B) are displayed as red boxes. Identified OSN binding motifs are shown as circles green circles: SOX2, blue circles: OCT4. Number of circles indicates the frequency of the motif.

B. Top: Relative Luciferase activity of the Oct4 proxmial (PE) and distal (DE) enhancer elements in serum and 24h 2i (n=3). The empty pGL4.24 plasmid (Promega) was used as control (ctrl). Bottom: Relative Luciferase activity of the Oct4 proximal (PE) and distal (DE) enhancer elements in serum as compared to the scrambled elements in 24h 2i (n=3). The empty pGL4.24 plasmid (Promega) was used as control (ctrl). The number of scrambled nucleotides targeting the OSN motifs and total length of the total tested elements are shown on the right.

C. ChIP-seq tracks showing the binding of NANOG, SOX2, OCT4 and p300 along the Utf1, Jarid2 (enhancer 1 and 2), Fgf5 and Prdm14 enhancer elements in serum and 24h 2i (p300 only in serum). Serum and 24h 2i ChIP-seq tracks are scaled to the same height for each factor and locus. Regions tested in the luciferase assay (panels D and E) are displayed as red boxes. Differential binding for NANOG, SOX2 and OCT4 is highlighted in grey. Identified OSN binding motifs are shown as green circles: SOX2, blue circles: OCT4 and yellow circles: NANOG. Number of circles indicates the frequency of the motif.

D. Relative Luciferase activity of the Utf1, Jarid2 (enhancer 1 and 2), Fgf5 and Prdm14 enhancer elements in serum and 24h 2i (n=3). The empty pGL4.24 plasmid (Promega) was used as control (ctrl).

E. Relative Luciferase activity of the Utf1, Jarid2 (enhancer 1 and 2), Fgf5 and Prdm14 enhancer elements in serum as compared to the disrupted elements in 24h 2i (n=3). The empty pGL4.24 plasmid (Promega) was used as control (ctrl). The number of scrambled nucleotides and total length of the tested elements are shown on the right.

To determine the possible effects of differential binding, we first assessed the activity of several enhancer elements in serum and 2i using a luciferase based reporter assay. As expected, the proximal enhancer shows diminished activity in 2i, while the distal enhancer element clearly gained activity (Figure 2B). We then replaced identified OSN motifs with a scrambled sequence (see Methods) and found the activity of the elements was greatly impaired (Figure 2B, Figure S2A). To expand these findings, we next tested additional targets including Utf1 and Jarid2 (1) as examples with no differential binding, Fgf5 as an example of lower NANOG binding in 2i, and the Prdm14 as well as a second Jarid2 (2) region representative of increased OSN binding (Figure 2C). In general we found that differential OSN binding was associated with altered activity of the tested elements in 2i (Figure 2D) and all elements exhibited impaired activity upon disruption of the identified binding motifs (Figure 2E). Notably, control regions (no p300 enrichment in serum) did not alter their activity in 2i and in the case of the Otx2 promoter the disruption of the OSN binding sites had no apparent effect on the activity either (Figure S2B and S2C).

OSN rewiring precedes the global loss of DNA methylation in 2i

We next wanted to explore possible mechanisms that could lead to the global reorganization in transcription factor binding. As shown earlier H3K4me3 and H3K27me3 dynamics showed little overlap with changes in the OSN binding pattern (Figure 1F). As expected OSN sites that are bound in serum already exhibited low levels of DNA methylation (Figure 3A), however sites that specifically gain OSN binding in 2i exhibit higher methylation levels in serum. To test whether the change in methylation is a cause or consequence of the differential binding, we measured global DNA methylation in serum vs. 2i (24h). Importantly, in agreement with previous studies (Habibi et al., 2013; Ficz et al., 2013), we could not yet detect the global decrease in DNA methylation (Figure 3B). As a more site-specific loss of DNA methylation might be obscured in the global DNA methylation analysis, we next assessed the methylation status of regions with differential OSN binding after 24h in 2i but also did not detect any change in DNA methylation in this analysis (Figure 3C). In contrast, when we assess the intermediate transition state (7 days in 2i) and the established 2i population (3P), especially sites with a gain in OSN binding become subject to a gradual loss of methylation (Figure S3A).

Figure 3. OSN rewiring precedes global DNA demethylation and overlaps with motifs of factors linked to Wnt and Erk signaling.

Figure 3

Open in a new tab

A. DNA methylation levels of differentially bound OCT4, SOX2 and NANOG sites under serum conditions.

B. Histograms showing the global DNA methylation distribution in mESCs (serum and 24h 2i).

C. DNA methylation levels of differentially bound OCT4, SOX2 and NANOG sites in serum versus 24h 2i.

D. Transcription factor motifs that are enriched in regions with increased binding for NANOG or SOX2 in 24h 2i. Transcription factors that are implicated in Wnt signaling are highlighted in green. Motifs are shown exemplary for OTX2, LEF1 and DLX4.

E. Transcription factor motifs that are enriched in regions with increased binding for NANOG or SOX2 in serum. Transcription factors that are implicated in Erk signaling are highlighted in red. Motifs are shown exemplary for ZFP423 and EGR1/2.

F. Expression (log2 fold change) of transcription factors in serum vs. 24h 2i that were identified as co-factors for sites with higher NANOG and SOX2 binding in 2i. The error bars represent one standard error over two replicates.

G. Expression (log2 fold change) of transcription factors in serum vs. 24h 2i that were identified as co-factors for sites with higher NANOG and SOX2 binding in serum. The error bars represent one standard error over two replicates.

Differentially bound OSN sites co-occur with motifs linked to Wnt and Erk signaling

Given the limited role of epigenetic modifications we next looked into the occurrence of motifs that might point to relevant co-factors. We found that sites with increased binding in 2i are enriched for motifs corresponding to the transcription factors LEF1 and TCF7, which are direct effectors of the Wnt pathway (Reya and Clevers, 2005), as well as several HOX proteins whose expression is modulated through this pathway (Deschamps and van Nes, 2005) (Figure 3D). On the other hand, sites with decreased binding in 2i were specifically enriched for motifs of Erk pathway modulators, such as EGR1, SP1 and KLF7 (Black et al., 2001; Gregg and Fraizer, 2011; Merchant et al., 1999) (Figure 3E). Notably, several of the identified co-factors change their expression upon addition of the two kinase inhibitors (Figure 3F and 3G). This change in available co-factors might therefore impact the propensity of OSN to preferentially engage certain binding sites over others.

Transcriptional silencing of bivalent genes is independent of PRC2 function in 2i

In line with previous reports ((Marks et al., 2012) we find several bivalent loci such as the HoxA cluster or Otx2 that exhibit a gradual loss of H3K27me3 (Figure 4A and 4B). However, both also show decreased H3K4me3 (Figure S4A and S4B) as reported by others (Gafni et al., 2013). When we plotted the global enrichment level of both histone modifications in serum versus 2i, we observed a general shift to lower enrichment for H3K4me3 at bivalent gene loci in the initial 2i transition, and this shift is largely maintained (Figure 4C and 4D). On the other hand, the repressive H3K27me3 mark shows a more subtle shift to lower enrichment within the first 24h in 2i which becomes more pronounced after 3 passages. ChIP-qPCR further confirmed these subtle differences for H3K27me3 within the first 24h in 2i (Figure S4C). The faster H3K4me3 dynamics compared to H3K27me3 in this context are similar to observations in reprogramming (Koche et al., 2011). Overall, only 55% of all bivalent genes in serum retain their bivalent status after 24h in 2i and even fewer (40%) after 3 passages (Figure 4E, Table S4). However, similar to the HoxA cluster and Otx2, the differential enrichment of these histone marks seems to have little impact on the expression of associated genes (Figure S4D). This is consistent with previous reports showing that the basal expression level of many bivalent genes remains the same or is even further down regulated in 2i (Marks et al., 2012).

Figure 4. Transcriptional silencing of bivalent genes is independent of PRC2 function in 2i.

Figure 4

Open in a new tab

A. ChIP-seq tracks for H3K27me3 along the HoxA cluster in serum, 24h 2i and 3P 2i as well as RNA-seq tracks for serum and 3P 2i. Read counts are normalized to 1 million reads.

B. ChIP-seq tracks for H3K27me3 along the Otx2 gene locus in serum, 24h 2i and 3P 2i as well as RNA-seq tracks for serum and 3P 2i. Read counts are normalized to 1 million reads.

C. Differentially enriched 1kb tiles for H3K27me3 and H3K4me3 in serum versus 24h 2i. Bivalent genes are highlighted in black.

D. Differentially enriched 1kb tiles for H3K27me3 and H3K4me3 in serum versus 3P 2i. Bivalent genes are highlighted in black.

E. Number of bivalent genes in serum, 24h 2i and 3P 2i. Only 55% of genes that are bivalent in serum retain their bivalent status after 24h in 2i which further decreases to 40% after 3P in 2i.

F. Expression (log2(FPKM)) for a selection of lineage markers in wildtype (WT) and Eed mutant cells in serum vs. 3 passages 2i (3P). The transcriptional changes coincide with a decrease of H3K27me3 deposition along the promoter regions of the genes, exemplary shown for the Otx2, Gata4 and Bmp6 locus. The error bars represent one standard error over two replicates.

G. Morphologic stabilization of Eed mutant cells in 2i. The cells exhibit a tight morphology together with the abrogation of spontaneous differentiation in 2i. Scale bar: 100μm.

H. ChIP-seq tracks for H3K27me3 and H3K4me3 along the Utf1 and Tcf15 locus in serum and after 24h in 2i. Read counts are normalized to 1 million reads.

I. Expression (log2(FPKM) of Utf1 and Tcf15 in serum vs. 3P 2i in the WT and Eed mutant cells. The error bars represent one standard error over two replicates.

J. Violin plot showing the average expression (log2(FPKM) for bivalent genes that lose their bivalent state in 3P 2i (Table S4) in the WT in serum and 2i (3P), as well as in the Eed mutant line in serum and 2i (3P). Note that the average expression of bivalent genes decreases for the WT and the Eed mutant cells after three passages in 2i.

To better understand the relevance of H3K27me3 repression, we used an Eed mutant ES cell line that is devoid of H3K27me3 (Montgomery et al., 2005) (Figure S4E). Under serum conditions, Eed deficient cells are extremely unstable and display aberrant expression of differentiation-associated genes (Figure 4F), highlighting the importance of PRC2 function as a balancing factor in the metastable serum state (Boyer et al., 2006; Montgomery et al., 2005). However, when switching to 2i, the cells exhibited a complete morphologic and molecular stabilization, reflected by the abrogation of spontaneous differentiation and associated expression of differentiation genes (Figure 4F and 4G). Additionally, we find that Utf1 and Tcf15 de novo silencing – two genes that gain H3K27me3 early in the 2i transition – is unaffected despite the absence of H3K27me3 (Figure 4H and 4I). The same independence of PRC2 mediated silencing was observed in Eed knockdown cells (Figure S4F and S4G). Therefore the establishment of a repressed chromatin structure at the Utf1 and Tcf15 loci, albeit rapid in wild type cells, can be interpreted as a secondary event in their transcriptional down-regulation. To assess the generality of this finding, we generated RNA-seq data for Eed mutant cells in serum vs. 2i (3P) and compared the expression profiles to the wild type cells in serum vs. 2i (3P). Interestingly, we find that PRC2-independent gene regulation is not limited to the loci mentioned above as the majority of differentially expressed genes becomes similarly induced or repressed in the absence of H3K27me3 (Figure S4H and S4I). Focusing on the differential expression of bivalent genes confirms this trend even further. Especially the maintained or enforced repression of bivalent genes that lose their bivalent state in 2i (Table S4) highlights that many of these genes do not rely on a functioning PRC2 complex in 2i (Figure 4J).

Lastly, it has been shown that the protein PRDM14, a PR-domain-containing transcriptional regulator, is upregulated under 2i conditions and subsequently negatively regulates Dnmt3a and Dnmt3b expression (Leitch et al., 2013; Yamaji et al., 2013). Moreover, a previously published expression dataset suggests a negative correlation between the presence of PRDM14 and Utf1 expression (Yamaji et al., 2013), which potentially accounts for the latter’s repression in 2i. Again, as for WT cells, we detected a clear up-regulation of Prdm14 in the Eed mutant cells switched to 2i (Figure S4J). The previously postulated model suggested that PRDM14 exerts its repressive function through a direct recruitment of PRC2 to its target sites. However, we find that H3K27me3 deficient cells demonstrate uncompromised repression of Dnmt3a/3b as well as of Utf1 suggesting that PRDM14 can either exert its function directly as a signaling regulator or operate through a different repressive mechanism (Figure 4I and S4J). Taken together, these findings render PRC2 dependent dynamics of H3K27me3 a negligible variable in the explanation of transcriptional reconfiguration upon the transition to 2i.

DISCUSSION

The stable transition from serum/LIF to the 2i conditions is associated with global transcriptional and epigenetic changes in mouse ESCs. Key events in this transition constitute the up-regulation of metabolism associated genes, the down-regulation of developmental associated genes and global hypomethylation of the cells as well as a decline of the repressive histone mark H3K27me3 along bivalent gene loci (Marks et al., 2012). Although a more homogenous distribution of NANOG points to a possible stabilization and reinforcement of the core pluripotency circuitry in 2i, details on the binding of the core pluripotency factors, OCT4, SOX2 and NANOG in the transition to 2i have not been studied. To fill this mechanistic gap, we focused on ESCs 24h post switch and mapped the binding pattern of the core pluripotency factors OSN as well as the differential enrichment of the activating and repressive histone marks H3K4me3 and H3K27me3. While the core pluripotency factors show only moderate (Nanog) to no (Oct4 and Sox2) differential expression in 2i, we detected a globally reconfigured binding pattern for all three factors after only 24h in 2i, highlighting their high sensitivity to altering ESC culturing conditions. Importantly, the early rewiring of OSN is largely maintained throughout the 2i transition and precedes the majority of changes in the enrichment of the bivalent chromatin marks H3K4me3 and H3K27me3 as well as the global loss of DNA methylation.

Although we find over one-third of all differentially expressed genes to be differentially bound in 2i, the overall correlation is rather weak as the majority of differential binding events are not associated with expression changes. This result may not be that surprising as only a small subset of OSN engaged genes are known to be bona fide targets while the majority is not affected by the depletion of any of the factors (Loh et al., 2006). Notably, this is a common phenomenon observed for transcription factors in higher eukaryotes where usually only 10–25% of TF occupancy can be linked to expression of neighboring genes (Spitz and Furlong, 2012). Attributes that enhance the predictive power (functional binding events) are the presence of strong binding motifs and TF binding events at active enhancer elements (Cusanovich et al., 2014). Indeed, we find that the differential binding of the core pluripotency factors along key enhancer elements – but not at non-enhancer elements – correlates with their activity in 2i and that the disruption of identified OSN motifs results in a complete loss of enhancer activity. Although TF binding alone is not necessarily indicative of transcriptional activity, it might label regions that will be used subsequently, e.g. by stage specific TFs (Spitz and Furlong, 2012). These priming events may therefore facilitate developmental transitions by reducing the rate-limiting step of chromatin remodeling. Interestingly, we find that differential binding sites are enriched for TF motifs associated with canonical Wnt signaling and the Erk pathway, the main signaling cascades perturbed through the inhibitors CHIR and PD respectively. Specifically sites with higher NANOG and SOX2 binding in 2i are enriched for motifs of TFs implicated in the 2i-enhanced Wnt signaling pathway, while sites with lower binding share motifs of cofactors involved in the differentiation-associated Erk signaling pathway which is repressed in 2i. Thus, binding/priming of genes connected to differentiation is maintained in a metastable serum state, yet seems to be less crucial under enforced pluripotency conditions where the engagement shifts to maintaining an enforced pluripotency state.

Besides the rewiring of the core pluripotency network in 2i, we find that the divergent enrichment of the histone modifications H3K4me3 and H3K27me3 does not account for the majority of observed transcriptional variance in 2i. Moreover, H3K27me3 deficient Eed mutant cells induce normal repression of genes such as Utf1 and Tcf15 in 2i and maintained the silencing of bivalent loci. This dispensable function of PRC2 function in 2i is in contrast to the metastable serum/LIF state, which is highly dependent on proper PRC2 for maintenance of the pluripotent state (Geula et al., 2015). This apparent PRC2 independence in 2i is even more intriguing considering that this transition is also accompanied by global DNA hypomethylation (except for imprinted genes) (Ficz et al., 2013; Habibi et al., 2013). Therefore, 2i pluripotency – a state shielding ESCs from differentiation-inducing signals – appears to function in the absence of H3K27me3 and DNA methylation, both of which are generally regarded as key repressive epigenetic regulators in the metastable serum state. While more work remains to be done, our results further contribute to the understanding of the 2i state and regulation of pluripotency in general.

EXPERIMENTAL PROCEDURES

Cell Culture

Mouse embryonic stem cells (mESCs) were cultured in regular media containing 15% FBS, 1% PEN/STREP, 1% Glutamine, 1% NEAA and LIF. For ES cell maintenance dishes were coated with 0.2% gelantin and irradiated CF1 mouse embryonic fibroblasts (MEFs) were plated as confluent layer of feeder cells. ESCs were seeded in a density of 50000 cells/ 6well and were splitted every 2–3 days. The 2i media was prepared according to the following protocol (Ying et al., 2008) and was supplemented with LIF.

Antibodies

ChIP was performed using the following antibodies at a final concentration of 1μg/1x106 cells: H3K4me3 (Millipore, 07-473), H3K27me3 (Millipore, 07-449), OCT4 (Santa Cruz, sc-8628 X), SOX2 (Santa Cruz, sc-17320 X), NANOG (Cosmo Bio Ltd, REC-RCAB0002P-F). Immunostaining was done with the following primary antibodies, all diluted 1:1000: NANOG (BD, 560259), STELLA (Santa Cruz, sc-67249). Western Blots were done using the following antibodies: UTF1 (1:1000; Abcam, ab24273), Histone H3 (1:5000; Abcam, ab61251), EED (Abcam, ab4469).

Chromatin Immunoprecipitation

Cells were crosslinked in 1% formaldehyde for 15 minutes (5min for UTF1 Antibody and UTF1-BioChIP) at room temperature, with constant agitation, followed by quenching with 125mM Glycine for 5 minutes at room temperature. Nuclei were isolated and chromatin was sheared using Branson sonifier until the majority of DNA was in the range of 200–700 base pairs. Chromatin was incubated with antibody overnight at 4°C, with constant agitation. Immunoprecipitation of antibody-protein complexes was completed using Protein A or Protein G Dynabeads (Invitrogen) for 2–3 hour at 4°C, with constant agitation.

Supplementary Material

1
2
3
4

HIGHLIGHTS.

  • Early and widespread reconfiguration of OCT4, SOX2 and NANOG binding in 2i

  • Differential OSN binding correlates with enhancer activity in 2i

  • Dynamic targets show co-occurring motifs of factors linked to Wnt and Erk signaling

  • PRC2 appears dispensable for continued silencing and de novo repression in 2i

Acknowledgments

We would like to thank Robbyn Issner, Holly Whitton, and Hongcang Gu for their advice and support regarding the ChIP-seq library preparation. AM is a New York Stem Cell Foundation, Robertson Investigator. This work was funded by the Human Frontiers Science Program, NIH grants 1P50HG006193, P01GM099117 and the New York Stem Cell Foundation.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Black AR, Black JD, Azizkhan-Clifford J. Sp1 and kruppel-like factor family of transcription factors in cell growth regulation and cancer. Journal of cellular physiology. 2001;188:143–160. doi: 10.1002/jcp.1111. [DOI] [PubMed] [Google Scholar]
  2. Boyer LA, Plath K, Zeitlinger J, Brambrink T, Medeiros LA, Lee TI, Levine SS, Wernig M, Tajonar A, Ray MK, et al. Polycomb complexes repress developmental regulators in murine embryonic stem cells. Nature. 2006;441:349–353. doi: 10.1038/nature04733. [DOI] [PubMed] [Google Scholar]
  3. Buecker C, Srinivasan R, Wu Z, Calo E, Acampora D, Faial T, Simeone A, Tan M, Swigut T, Wysocka J. Reorganization of enhancer patterns in transition from naive to primed pluripotency. Cell stem cell. 2014;14:838–853. doi: 10.1016/j.stem.2014.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Chan YS, Goke J, Ng JH, Lu X, Gonzales KA, Tan CP, Tng WQ, Hong ZZ, Lim YS, Ng HH. Induction of a human pluripotent state with distinct regulatory circuitry that resembles preimplantation epiblast. Cell stem cell. 2013;13:663–675. doi: 10.1016/j.stem.2013.11.015. [DOI] [PubMed] [Google Scholar]
  5. Chen X, Xu H, Yuan P, Fang F, Huss M, Vega VB, Wong E, Orlov YL, Zhang W, Jiang J, et al. Integration of external signaling pathways with the core transcriptional network in embryonic stem cells. Cell. 2008;133:1106–1117. doi: 10.1016/j.cell.2008.04.043. [DOI] [PubMed] [Google Scholar]
  6. Creyghton MP, Cheng AW, Welstead GG, Kooistra T, Carey BW, Steine EJ, Hanna J, Lodato MA, Frampton GM, Sharp PA, et al. Histone H3K27ac separates active from poised enhancers and predicts developmental state. Proceedings of the National Academy of Sciences of the United States of America. 2010;107:21931–21936. doi: 10.1073/pnas.1016071107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cusanovich DA, Pavlovic B, Pritchard JK, Gilad Y. The functional consequences of variation in transcription factor binding. PLoS genetics. 2014;10:e1004226. doi: 10.1371/journal.pgen.1004226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Deschamps J, van Nes J. Developmental regulation of the Hox genes during axial morphogenesis in the mouse. Development. 2005;132:2931–2942. doi: 10.1242/dev.01897. [DOI] [PubMed] [Google Scholar]
  9. Factor DC, Corradin O, Zentner GE, Saiakhova A, Song L, Chenoweth JG, McKay RD, Crawford GE, Scacheri PC, Tesar PJ. Epigenomic comparison reveals activation of “seed” enhancers during transition from naive to primed pluripotency. Cell stem cell. 2014;14:854–863. doi: 10.1016/j.stem.2014.05.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Ficz G, Hore TA, Santos F, Lee HJ, Dean W, Arand J, Krueger F, Oxley D, Paul YL, Walter J, et al. FGF signaling inhibition in ESCs drives rapid genome-wide demethylation to the epigenetic ground state of pluripotency. Cell stem cell. 2013;13:351–359. doi: 10.1016/j.stem.2013.06.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gafni O, Weinberger L, Mansour AA, Manor YS, Chomsky E, Ben-Yosef D, Kalma Y, Viukov S, Maza I, Zviran A, et al. Derivation of novel human ground state naive pluripotent stem cells. Nature. 2013;504:282–286. doi: 10.1038/nature12745. [DOI] [PubMed] [Google Scholar]
  12. Galonska C, Smith ZD, Meissner A. In Vivo and in vitro dynamics of undifferentiated embryonic cell transcription factor 1. Stem cell reports. 2014;2:245–252. doi: 10.1016/j.stemcr.2014.01.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Geula S, Moshitch-Moshkovitz S, Dominissini D, Mansour AA, Kol N, Salmon-Divon M, Hershkovitz V, Peer E, Mor N, Manor YS, et al. Stem cells. m6A mRNA methylation facilitates resolution of naive pluripotency toward differentiation. Science. 2015;347:1002–1006. doi: 10.1126/science.1261417. [DOI] [PubMed] [Google Scholar]
  14. Gifford CA, Ziller MJ, Gu H, Trapnell C, Donaghey J, Tsankov A, Shalek AK, Kelley DR, Shishkin AA, Issner R, et al. Transcriptional and epigenetic dynamics during specification of human embryonic stem cells. Cell. 2013;153:1149–1163. doi: 10.1016/j.cell.2013.04.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gregg J, Fraizer G. Transcriptional Regulation of EGR1 by EGF and the ERK Signaling Pathway in Prostate Cancer Cells. Genes & cancer. 2011;2:900–909. doi: 10.1177/1947601911431885. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Habibi E, Brinkman AB, Arand J, Kroeze LI, Kerstens HH, Matarese F, Lepikhov K, Gut M, Brun-Heath I, Hubner NC, et al. Whole-genome bisulfite sequencing of two distinct interconvertible DNA methylomes of mouse embryonic stem cells. Cell stem cell. 2013;13:360–369. doi: 10.1016/j.stem.2013.06.002. [DOI] [PubMed] [Google Scholar]
  17. Kim J, Chu J, Shen X, Wang J, Orkin SH. An extended transcriptional network for pluripotency of embryonic stem cells. Cell. 2008;132:1049–1061. doi: 10.1016/j.cell.2008.02.039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Koche RP, Smith ZD, Adli M, Gu H, Ku M, Gnirke A, Bernstein BE, Meissner A. Reprogramming factor expression initiates widespread targeted chromatin remodeling. Cell stem cell. 2011;8:96–105. doi: 10.1016/j.stem.2010.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kunath T, Saba-El-Leil MK, Almousailleakh M, Wray J, Meloche S, Smith A. FGF stimulation of the Erk1/2 signalling cascade triggers transition of pluripotent embryonic stem cells from self-renewal to lineage commitment. Development. 2007;134:2895–2902. doi: 10.1242/dev.02880. [DOI] [PubMed] [Google Scholar]
  20. Leitch HG, McEwen KR, Turp A, Encheva V, Carroll T, Grabole N, Mansfield W, Nashun B, Knezovich JG, Smith A, et al. Naive pluripotency is associated with global DNA hypomethylation. Nature structural & molecular biology. 2013;20:311–316. doi: 10.1038/nsmb.2510. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Loh YH, Wu Q, Chew JL, Vega VB, Zhang W, Chen X, Bourque G, George J, Leong B, Liu J, et al. The Oct4 and Nanog transcription network regulates pluripotency in mouse embryonic stem cells. Nature genetics. 2006;38:431–440. doi: 10.1038/ng1760. [DOI] [PubMed] [Google Scholar]
  22. Marks H, Kalkan T, Menafra R, Denissov S, Jones K, Hofemeister H, Nichols J, Kranz A, Stewart AF, Smith A, et al. The transcriptional and epigenomic foundations of ground state pluripotency. Cell. 2012;149:590–604. doi: 10.1016/j.cell.2012.03.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Merchant JL, Du M, Todisco A. Sp1 phosphorylation by Erk 2 stimulates DNA binding. Biochemical and biophysical research communications. 1999;254:454–461. doi: 10.1006/bbrc.1998.9964. [DOI] [PubMed] [Google Scholar]
  24. Montgomery ND, Yee D, Chen A, Kalantry S, Chamberlain SJ, Otte AP, Magnuson T. The murine polycomb group protein Eed is required for global histone H3 lysine-27 methylation. Current biology : CB. 2005;15:942–947. doi: 10.1016/j.cub.2005.04.051. [DOI] [PubMed] [Google Scholar]
  25. Reya T, Clevers H. Wnt signalling in stem cells and cancer. Nature. 2005;434:843–850. doi: 10.1038/nature03319. [DOI] [PubMed] [Google Scholar]
  26. Silva J, Nichols J, Theunissen TW, Guo G, van Oosten AL, Barrandon O, Wray J, Yamanaka S, Chambers I, Smith A. Nanog is the gateway to the pluripotent ground state. Cell. 2009;138:722–737. doi: 10.1016/j.cell.2009.07.039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Spitz F, Furlong EE. Transcription factors: from enhancer binding to developmental control. Nature reviews Genetics. 2012;13:613–626. doi: 10.1038/nrg3207. [DOI] [PubMed] [Google Scholar]
  28. Tesar PJ, Chenoweth JG, Brook FA, Davies TJ, Evans EP, Mack DL, Gardner RL, McKay RD. New cell lines from mouse epiblast share defining features with human embryonic stem cells. Nature. 2007;448:196–199. doi: 10.1038/nature05972. [DOI] [PubMed] [Google Scholar]
  29. Yamaji M, Ueda J, Hayashi K, Ohta H, Yabuta Y, Kurimoto K, Nakato R, Yamada Y, Shirahige K, Saitou M. PRDM14 ensures naive pluripotency through dual regulation of signaling and epigenetic pathways in mouse embryonic stem cells. Cell stem cell. 2013;12:368–382. doi: 10.1016/j.stem.2012.12.012. [DOI] [PubMed] [Google Scholar]
  30. Yang SH, Kalkan T, Morissroe C, Marks H, Stunnenberg H, Smith A, Sharrocks AD. Otx2 and Oct4 drive early enhancer activation during embryonic stem cell transition from naive pluripotency. Cell reports. 2014;7:1968–1981. doi: 10.1016/j.celrep.2014.05.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Yeo JC, Jiang J, Tan ZY, Yim GR, Ng JH, Goke J, Kraus P, Liang H, Gonzales KA, Chong HC, et al. Klf2 is an essential factor that sustains ground state pluripotency. Cell stem cell. 2014;14:864–872. doi: 10.1016/j.stem.2014.04.015. [DOI] [PubMed] [Google Scholar]
  32. Ying QL, Wray J, Nichols J, Batlle-Morera L, Doble B, Woodgett J, Cohen P, Smith A. The ground state of embryonic stem cell self-renewal. Nature. 2008;453:519–523. doi: 10.1038/nature06968. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Ziller MJ, Gu H, Muller F, Donaghey J, Tsai LT, Kohlbacher O, De Jager PL, Rosen ED, Bennett DA, Bernstein BE, et al. Charting a dynamic DNA methylation landscape of the human genome. Nature. 2013;500:477–481. doi: 10.1038/nature12433. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS708991-supplement-1.xlsx (681.9KB, xlsx)
2
NIHMS708991-supplement-2.xlsx (523.1KB, xlsx)
3
NIHMS708991-supplement-3.xlsx (31KB, xlsx)
4
NIHMS708991-supplement-4.xlsx (172KB, xlsx)

RESOURCES