Generation and Characterization of a Novel Mouse Embryonic Stem Cell Line with a Dynamic Reporter of Nanog Expression

Background The pluripotent state in embryonic stem (ES) cells is controlled by a core network of transcription factors that includes Nanog, Oct4 and Sox2. Nanog is required to reach pluripotency during somatic reprogramming and is the only core factor whose overexpression is able to oppose differentiation-promoting signals. Additionally, Nanog expression is known to fluctuate in ES cells, and different levels of Nanog seem to correlate with ES cells’ ability to respond to differentiation promoting signals. Elucidating how dynamic Nanog levels are regulated in pluripotent cells and modulate their potential is therefore critical to develop a better understanding of the pluripotent state. Methodology/Principal Findings We describe the generation and validation of a mouse ES cell line with a novel Nanog reporter (Nd, from Nanog dynamics), containing a BAC transgene where the short-lived fluorescent protein VNP is placed under Nanog regulation. We show that Nanog and VNP have similar half-lives, and that Nd cells provide an accurate and measurable read-out for the dynamic levels of Nanog. Using this reporter, we could show that ES cells with low Nanog levels indeed have higher degree of priming to differentiation, when compared with high-Nanog cells. However, low-Nanog ES cells maintain high levels of Oct4 and Sox2 and can revert to a state of high-Nanog expression, indicating that they are still within the window of pluripotency. We further show that the observed changes in Nanog levels correlate with ES cell morphology and that Nanog dynamic expression is modulated by the cellular environment. Conclusions/Significance The novel reporter ES cell line here described allows an accurate monitoring of Nanog’s dynamic expression in the pluripotent state. This reporter will thus be a valuable tool to obtain quantitative measurements of global gene expression in pluripotent ES cells in different states, allowing a detailed molecular mapping of the pluripotency landscape.


Introduction
Embryonic Stem (ES) cells are characterized by their selfrenewal capacity and pluripotenciality [1,2]. These cells can be derived from the inner cell mass (ICM) of the mammalian blastocyst and can be maintained in vitro under very specific culture conditions ( [3,4], reviewed in [5]). Due to their properties, ES cells constitute a promising resource for the next-generation of cellular therapies; however, scientific, technological and ethical questions are still preventing the development of ES cell-based techniques. One of the major bottlenecks has been the lack of a conceptual understanding of the pluripotent state, which has not emerged yet from the systematic molecular characterization of various pluripotent stem cells.
Recent work has led to a novel view of pluripotency in ES cells as a self-maintaining and intrinsically-controlled ''ground state'' [6,7], regulated by a gene regulatory network (GRN) in which the transcription factors (TFs) Nanog, Oct4 and Sox2 (NOS network) play a central role [7][8][9]. Extensive characterization of the transcriptional program elicited by these three TFs revealed that they function in concert to sustain the ES cell state by activating other pluripotency genes while, simultaneously, repressing differentiation-promoting genes [7,9,10]. This repression is thought to play a central role in maintaining the pluripotent state, reducing its vulnerability to the myriad of extrinsic signals that promote differentiation along the various embryonic lineages. However, recent work has shown that both Oct4 and Sox2 can also function as lineage specifiers, assisting the emergence of mesendodermal and neuroectodermal fates, respectively [11,12]. These findings support a different view of the pluripotent state, as a highly unstable and transient cellular state, driven by the competing lineage-promoting activities of the different ''pluripotency'' factors [13], instead of a ground state implemented and maintained by the NOS circuitry. This scenario emphasizes the precarious and volatile nature of this state and challenges the idea of an intrinsic ability of ES cells to sustain their state based on a dedicated genetic network. The question therefore remains as to which functions do the pluripotency factors play in establishing and maintaining the pluripotent state of ES cells.
One feature that distinguishes Nanog from its partners Sox2 and Oct4 is the reported heterogeneous expression of this TF in ES cell cultures (and also in the blastocyst's ICM), with some cells showing high levels of Nanog expression while others exhibit reduced levels [14,15]. Furthermore, cells with low or no Nanog expression can evolve into a high-expression state, implying that Nanog levels fluctuate in individual ES cells (contrarily to Oct4 and Sox2) [16,17]. Nanog was initially discovered by virtue of its capacity to oppose differentiation-promoting signals, being essential to maintain ES cells in the absence of LIF/STAT3 signalling [18,19]. This led to the hypothesis that fluctuating levels of Nanog confer different degrees of responsiveness to differentiation signals in individual ES cells, resulting in distinct cellular outputs upon differentiation stimuli [16]. The resulting population heterogeneity might be central to the pluripotent state: on one side, it ensures that there is always a fraction of cells primed to differentiate within a stem cell population, a condition of pluripotency; on the other side, it ensures that other cells are resistant to differentiation cues and keep their pluripotency condition, thereby supporting the selfrenewal property.
It is therefore crucial to understand how fluctuations in Nanog levels are generated, and how these fluctuations endow ES cells with various degrees of responsiveness to differentiation cues. To address these questions, a method to monitor in real time the dynamic Nanog expression is required, and we report here the generation of a mouse ES cell line comprising a novel Nanog:VNP reporter (Nd, from Nanog dynamics) that faithfully reflects Nanog expression in ES cells. This reporter provides a powerful tool to dissect the molecular mechanisms regulating Nanog functions in pluripotent cells.

Generation of a Nanog:VNP Reporter ES Cell Line
To monitor the dynamic Nanog expression in ES cells, we have generated a novel ES cell line containing a transgenic Bacterial Artificial Chromosome (BAC) with a fluorescent reporter inserted under the control of the Nanog regulatory regions. This strategy leaves intact the two endogenous Nanog alleles, precluding in this way possible haploinsufficiency phenotypes in ES cells. In addition, since BACs contain large regions of the genome including most, if not all, of the gene-controlling regions, transgene expression is expected to occur in a pattern identical to that of the endogenous gene [20]. A BAC centred around the Nanog gene was chosen (clone RP24-464B14) and the Nanog coding sequence was replaced with a cDNA encoding the VNP fluorescent protein [21,22], using recombineering technology ( Figure 1A) [20]. VNP is a short-lived reporter consisting of a very bright and fast maturating yellow fluorescent protein (Venus), fused to a nuclear localization signal (NLS) and to a degradation signal (PEST). The recombined BAC was electroporated into E14tg2a ES cells, and neomycin selection was applied to select clones with BAC transgene integration. Eight independent VNP-expressing clones were isolated, expanded and characterized. All clones showed similar morphology, viability and growth rate ( Figure S1A). However, due to likely differences in the BAC integration site, the percentage of Nanog:VNP-expressing cells was different in each clone, ranging from 4% to 60% as measured by flow cytometry (FC) ( Figure S1B).
To select an ES cell clone whose levels of Nanog:VNP transgene expression closely mimic endogenous Nanog expression, we first determined the percentage of Nanog-positive cells in the parental E14tg2a ES cells, as well as in the TNG-A ES cell line (which contains a stable GFP inserted into one of the Nanog alleles) [16], grown in standard self-renewal conditions (serum/LIF). This was done by immunofluorescence (IF) in fixed ES cells with an anti-Nanog antibody (Figure 2A), and also by flow cytometry analysis following intracellular staining (FC-IS) with the same antibody ( Figure 2B). Our results indicate that, at each given time, approximately 55% of ES cells express Nanog, detected exclusively in the nucleus by IF ( Figure 2B, Table 1). We have also quantified the number of Oct4 and Sox2-positive cells in E14tg2a and TNG-A ES cells, and found that more than 90% of cells contain these two proteins in the nucleus ( Figure 2C, Table 1).
From the various BAC ES clones (Figures S1B), only the ''Nd'' clone showed a percentage of Nanog:VNP positive cells (57±13%, Figure S1B) similar to that of Nanog-positive cells in E14tg2a and TNG-A cell lines. This clone was thus selected for further characterization. DNA analysis of this clone shows that only one copy of the BAC was integrated ( Figure S2), and mRNA analysis by Northern blot reveals that the transcript encoding the VNP reporter has the expected size, indicating that BAC integration did not disrupt the reporter cassette. In addition, the levels of VNP and Nanog mRNAs are regulated in parallel, decreasing after LIF removal and increasing after culture in ground state (2i) conditions ( Figure 1B), showing that reporter expression is properly controlled by the upstream Nanog regulatory regions.
We next analysed the percentage of cells positive for Sox2, Oct4 and endogenous Nanog in the selected Nd clone. The values were very similar to those obtained for both E14tg2a and TNG-A cells ( Figure 2B,C, Table 1), showing that the integration of the Nanog:VNP BAC does not interfere with the normal expression of these pluripotency genes. More importantly, both double-IF analysis ( Figure 2A) and FC-IS analysis ( Figure 2C) reveal a strong correlation between Nanog and VNP expression in Nd ES cells ( Figure 2D), indicating that the Nanog:VNP reporter provides an accurate fluorescent read-out for Nanog expression in ES cells. In contrast, TNG-A ES cells show more GFP-expressing cells than Nanog-positive cells (81.7±1.1% compared with 54.8±13.2%, Figure 2B,E, Table 1), a result most likely due to perdurance of the stable GFP protein used as a reporter in this ES cell line.
To further assess the adequacy of the VNP reporter, we measured the protein half-life times for VNP, Nanog, Oct4 and GFP in the various ES cell lines ( Figure 3A): while GFP is stable for at least 6 hours in TNG-A cells, without any decay after translational inhibition, both Nanog and VNP have similar halflives (2.3 hours and 1.8 hours, respectively), justifying the use of VNP as an adequate reporter for Nanog expression. As expected, Oct4 is stable during the studied period of time, for all the ES cell lines.
Lastly, we have measured the VNP, Nanog, Oct4 and Sox2 mRNAs' half-lives, in E14tg2a and Nd cell lines ( Figure 3B). While Oct4 mRNA is stable in both cell lines, all other transcripts show fast decay rates, with Sox2 mRNA being the more short-lived, while VNP and Nanog mRNAs have very similar half-lives (,4hours).

Growth and Differentiation of Nd ES Cells are Similar to the Parental Cell Line (E14tg2a)
To further validate the Nd ES cell line, we compared its growth characteristics with the parental E14tg2a cell line, in serum/LIF conditions. Both ES cell lines grew as typical ES colonies throughout several passages (n$8), with similar proliferation rates (shown as fold increase in Figure 4A) and viability (above 95%, Figure 4A). In addition, both ES lines maintained a normal diploid karyotype ( Figure S3).
We also assessed the differentiation potential of the Nd cell line, using two different methods: formation of embryoid bodies (EBs) and monolayer neural differentiation [23]. In both conditions, Nd cells show normal differentiation patterns, generating EBs composed of cells from the three germ layers, and forming typical neuroepithelial rosette-like structures when grown in monolayer conditions in the absence of LIF ( Figure 4B). Comparing RT-PCR results obtained for ES cells and day 4 EBs, we observed the expected downregulation of known pluripotency genes (Nanog, Oct4, Sox2 and Rex1) along differentiation, concomitantly with upregulation of ectoderm (Fgf5), mesoderm (T-brachyury) and endoderm (Gata6) lineage-markers ( Figure 4C). Overall, these data confirm the pluripotent characteristics of the Nd cell line.

Monitoring Nanog Heterogeneity and Expression Dynamics with Nd ES Cells
Previous work indicates that Nanog expression in ES cells is highly dynamic, with fluctuating levels in individual cells [16,24], resulting in the observed heterogeneous distribution of Nanoghigh and Nanog-low cells in populations of ES cells grown in serum/LIF conditions. To test whether the Nd reporter cell line is adequate to monitor these reversible alterations in the levels of Nanog expression, we followed the growth of FACS-purified subsets of Nd cells with either low VNP (VNP L ) or high VNP (VNP H ) levels ( Figure 5A), in identical self-renewal conditions (serum/LIF). Samples were taken every 24 hours, for 4 days, and analysed for the distribution of Nanog:VNP expressing cells. The results ( Figure 5B, Figure S4) show that both subsets of Nd ES cells can re-establish the normal heterogeneous distribution of VNP levels after 2-4 days in culture. These results are comparable to those reported for TNG-A ES cells containing a stable GFP knocked-in the Nanog locus [16], although the restoration of the reporter mRNAs, in self-renewal (serum/LIF and 2i) and differentiation (serum) conditions. Transcript sizes are indicated on the right. Higher levels of Nanog RNA are detected for both E14 and Nd ES cells in 2i conditions, compared to serum/LIF, while Oct4 is highly expressed in both cases. VNP mRNA is only detected in Nd ES cells. In differentiation conditions (serum alone), Nanog, Oct4 and VNP mRNA expression is strongly reduced. GFP* refers to the fusion transcript present in TNG-A ES cells, arising from the recombination of a GFP cDNA into the Nanog locus [16]. This fusion transcript has a similar size to the endogenous Nanog mRNA expressed from the other allele. doi:10.1371/journal.pone.0059928.g001 original distribution was considerably faster and more complete in Nanog:VNP Nd ES cells. We have also analysed by RT-PCR the expression of various pluripotency and lineage-specific markers in the initial sorted populations, as well as in the populations formed after 4 days in culture. The results ( Figure 5C) indicate that the initial VNP H and VNP L subsets have similar levels of Oct4 and Sox2 mRNA expression, with an expected lower expression of Nanog in the VNP L subset. However, the expression of lineage markers (Fgf5, Gata6 and T-brachyury) is strikingly up-regulated in the sorted VNP L subset, implying that these cells display a higher degree of priming to differentiation. After 4 days in culture of this VNP L subset, the expression of these lineage markers seems to reach even higher levels, but this is most likely due to the relatively high number of VNP L cells still present in 4-day cultures (note the lower VNP expression in lane 6), as a result of the slower rate of restoration of Nanog heterogeneity in these cells, when compared with VNP H -derived cells ( Figure 5B).
Overall, these results indicate that the Nanog:VNP reporter provides a good read-out for the dynamic levels of Nanog expression in ES cells, allowing us to correlate these levels with the variations in differentiation potential that ES cells are constantly experiencing. In addition, our experiments reveal that although VNP L cells show higher levels of priming to differentiation, they remain within the ground state of pluripotency [7]. This is shown by the persistent high levels of Oct4 and Sox2 expression, and by the capacity of VNP L cells to revert to a state of high Nanog expression, re-establishing the typical expression profile of ES cells in self-renewal conditions.

Nanog Levels in ES Cells Grown in ''Ground State'' Conditions
Maintenance of the pluripotent state in ES cells is known to be dependent on inhibition of differentiation stimuli encoded by FGF/ERK signalling [25,26]. This finding was explored to define improved culture conditions for ES cells, which can be efficiently maintained, without serum, in the presence of highly specific FGF/ERK inhibitors, together with GSK3ß inhibitors (2i media) [6]. To test the behaviour of the Nanog:VNP reporter in these conditions, we compared the growth of Nd cells in serum/LIF versus 2i, to which we added LIF to get optimal culture conditions. FC analysis reveals that 56.268.0% of Nd cells show VNP expression in presence of serum/LIF ( Figure 2E), and that this percentage increases to 91.163.1% ( Figure 6A,B, Table1) when Nd cells are grown for 48 hours in 2i. This increase is maintained Figure 2. NOS network and reporters expression for cell lines grown in serum/LIF conditions. (A) Immunofluorescence detection for Nanog (red) and VNP/GFP (green) in E14tg2a, Nd and TNG-A cell lines; nuclei counterstained with DAPI (blue). Nanog expression is observed in approximately 50% of cells for all cell lines, while VNP is expressed in the same percentage of Nd cells, with a high degree of co-localization. Higher number of GFP positive cells than Nanog-positive is observed for TNG-A cells. Scale-bar = 50 um. (B) Representative dot blots for FC-IS analysis of Nanog and VNP/GFP for E14tg2a, Nd and TNG-A cells. Negative controls (samples with no primary antibody) for both Nanog and VNP/GFP staining's are shown, based on which the positive gate regions were designed. Nanog expression is similar between different cell lines grown in the same media. In Nd cells, VNP levels mimic Nanog expression in both culture conditions, while in TNG-A cells, GFP levels are much higher than Nanog levels, most likely due to reporter stability. (C) Representative dot blots for FC-IS analysis of Oct4 and Sox2 for E14tg2a and Nd. Negative controls (samples with no primary antibody) for both Oct4 and Sox2 staining's are shown, based on which the positive gate regions were designed. Oct4 and Sox2 levels are always expressed in more than 90% of the cells in both media conditions. (D) Correlation plot for Nanog and VNP proteins expression in Nd cells. Graph depicts data from 24906 cells from three biological replicates obtained by FC-IS. Statistical analysis indicates a high degree of correlation between the expression of both proteins (Pearson correlation = 0.647). (E) Representative dot blots for FC analysis of live VNP/GFP for E14tg2a, Nd and TNG-A cells grown. E14tg2a cells were used as negative controls, based on which positive gate regions were designed. Obtained data is similar to that obtained by FC-IS (C). Quantifications of (A-C,E) may be observed in Table 1 through three serial passages in 2i media ( Figure 6C) and seems to be accompanied by slightly lower, but not statistically different, proliferation rates (data not shown). We next assessed the expression of the Oct4, Sox2 and Nanog proteins in Nd and E14tg2a cells grown for 48 hours in serum/LIF or 2i. Analysis by FC-IS shows an increase in Nanog-expressing cells in both ES lines (as well as in TNG-A cells), confirming the results obtained with the Nanog:VNP reporter in Nd cells ( Figure 2B,C versus Figure 6D,E, Table 1). As expected, the percentage of Oct4 and Sox2 positive cells is similar in both culture conditions. These results were replicated at the mRNA level, as quantitative PCR analysis reveals a statistically significant increase in the expression levels of VNP and Nanog mRNAs when ES cells are grown in 2i, while Oct4 and Sox2 expression does not vary significantly ( Figure 6F). Concomitantly, a clear decrease was observed in the expression of Dusp6, a known FGF/ERK target [27,28], and of Fgf5, a marker for epiblast progression [29]. Together, these results reveal an effective increase in the number of ES cells expressing high levels of Nanog (and of the reporter Nanog:VNP) when FGF/ERK signalling is blocked. Nonetheless, a small population of ES cells with no or low levels of Nanog can be consistently detected in 2i cultures ( Figure 6B, asterisk), suggesting that although population heterogeneity is reduced, fluctuations in Nanog expression may still occur in the absence of FGF/ERK signalling.

Nanog Levels Correlate with ES Cell Colony Morphology
ES cell cultures in serum/LIF and 2i media reveal considerably different morphologies ( Figure 6G), with ES cell colonies showing a more tightly packed morphology in 2i conditions, and a concomitant reduction in flattened differentiated-like cells at the colonies' periphery. When we reduced the concentrations of FGF/ ERK and GSK3ß inhibitors present in 2i media, we could observe that ES cells colonies become less tightly packed and more adherent to the gelatin substrate ( Figure 7A). These morphological alterations correlate strikingly with the reduction of Nanogexpressing cells in the whole population (monitored by the Nanog:VNP reporter, Figure 7B,C), suggesting a possible role of Nanog in coordinating the differential expression of cell adhesion molecules in ES cells, as part of its function in modulating the pluripotency state. Actually, it has been shown that spontaneous differentiation of ES cells is accompanied by a decrease in the presence of E-cadherin at the cell surface, probably as part of the epithelial-mesenchymal transition (EMT) necessary for entry into differentiation [30]. Moreover, it has been suggested that EMT inhibition in human ES cells favors the induction of the ground state pluripotency [31], and that Nanog represses expression of Eomes, preventing it from cooperating with SMAD2/3 to activate the transcriptional network directing EMT [32]. Thus, the different cell adhesion behaviour of high-Nanog ES cells might be important to inhibit EMT and to resist entry into differentiation. In addition, this possible role of Nanog in controlling cell adhesion properties might be very important in vivo, for instance during the segregation of epiblast and endoderm lineages in the ICM, where differential profiles of cell adhesion molecules in high-Nanog epiblast progenitors and low-Nanog endoderm progenitors might underlie the cell sorting process that correctly positions the two lineages in the blastocyst [33,34].

Conclusions
In this paper, we describe a novel mouse ES cell line, engineered to have a short-lived fluorescent protein (VNP) as a reporter of Nanog expression. This reporter provides an accurate read-out of Nanog expression, and was used to monitor the behaviour of purified sub-populations of ES cells in different Nanog expression states. Our data confirms that low-Nanog cells have a higher degree of priming to differentiation, while ES cells with high Nanog are in a pristine pluripotent state. Using this novel Nanog reporter, we could observe heterogeneity of Nanog expression even when ES cells are cultured in ground state conditions (2i), in contrast to what has been previously suggested [35].
Finally, we observed that the levels of Nanog expression correlate with different ES cell morphologies, suggesting that Nanog might have a role in coordinating the differential expression of cell adhesion molecules in ES cells.
This novel Nanog reporter ES cell line will therefore be a powerful tool to perform a quantitative and systematic analysis of the pluripotent state, and to dissect the molecular mechanisms that regulate Nanog function.

Maintenance and Differentiation of Mouse ES Cells
The ES cell lines used for this study were E14tg2a and TNG-A (both a kind gift from Austin Smith's lab, University of Cambridge, UK) [16] and Nd, a novel transgenic ES cell line. TNG-A ES cells were derived from E14tg2a ES cells and contain a stable GFP reporter fused to the puromycin resistance gene inserted into the Nanog locus [16]. The Nd cell line is a BAC-transgenic ES cell line for VNP-tagged Nanog gene, derived from E14tg2a ES cells (see below for BAC-transgenic ES cell line generation).
For the EB formation assay, ES cells were seeded on 60-mm bacterial-grade petri dishes at 3610 4 cells/cm 2 , in GMEM medium in the absence of LIF. EBs formed within 24 hours and medium was changed every 2 days. Aggregates growth was monitored under an inverted microscope for 8 days, RNA was collected at day 0 and day 4.
For the neural differentiation assay, the monolayer protocol described in Abranches et al, 2009 was used [23]. Briefly, ES cells were plated at specific cell densities and grown for 4 days in RHB-A medium (Stem Cells Inc.) on gelatin-coated plates. On day 4 cells were dissociated and re-plated onto laminin-coated tissue culture plastic in RHB-A medium supplemented with 5 ng/ml murine bFGF (Peprotech). Medium was changed every 2 days and cells were monitored under an inverted microscope for 8 days for the appearance of neural rosettes.

BAC-transgenic ES Cell Lines Generation
BAC clone RP24-464B14 bearing 213-kb insert containing the Nanog gene was ordered from CHORI database (http://bacpac. chori.org). The 59HA for Nanog was PCR amplified from BAC DNA using forward primer 59-CGGGATCCGCTGAAAG-GAAAGCCGTGTA- 39 and reverse primer 59-CCTCGCCCTTGCTCACCATAGAAAGAAGAGT-TAAATGTC-39. VNP sequence followed by rabbit beta-globin polyA was amplified from VNP-pCAGGS vector [22] using forward primer 59-ATGGTGAGCAAGGGCGAGG-39 and reverse primer 59-CAGGTCGAC GGATCTCCATAAG-39. Both amplification products were joined together by PCR ligation using Phusion DNA Polymerase (Finnzymes), and the resulting 1.9-kb PCR fragment was cloned into pBlueScript KS (Stratagene) and sequenced. The 39HA for Nanog was PCR-amplified from BAC DNA using forward primer 59-GAGCGGCCGCGACTTACG-CAACATCTGGGC-39 and reverse primer 59-GCTCTAGAG-CATGTTCTAAGTCCTAGGTTTG-39, and cloned downstream of floxed PGK-E7-neoR cassette from PL451 vector (Pentao Liu, Addgene) between NotI and XbaI sites. 59-and 39parts of the final NanogVNP-neoR cassette were joined together by 3-way ligation of BamHI-SalI and SalI-XbaI DNA fragments into pBlueScritp KS digested BamHI-XbaI and end-sequenced. Nanog BAC clone RP24-464B14 was transfected with defective  Figure S4. Statistically significant differences (p-value ,0.05) observed between ''All'' and VNP sub-populations are denoted with (*), while statistically significant differences between VNP L and VNP H are denoted with (**). (C) mRNA expression analysis by RT-PCR of cells collected immediately after sorting (day 0) and four days after replating (day 4). The purified VNP L subpopulation (immediately after sorting) shows lower Nanog mRNA levels and higher expression of lineage-affiliated genes (Fgf5, Gata6 and T-brachyury). After culture for 4 days, both VNP L and VNP H subsets show similar Nanog mRNA expression. Expression of lineage markers in the VNP L subpopulation is still more elevated after 4 days of culture, most likely reflecting the slower reversion to heterogeneity ( Figure S4). Similar levels of Oct4 and Sox2 expression are observed for all analysed samples. doi:10.1371/journal.pone.0059928.g005 Figure 6. NOS network and reporters expression for cell lines grown 2i conditions. (A) Representative dot blots for FC-IS analysis of Nanog and VNP/GFP for E14tg2a, Nd and TNG-A cells grown in 2i. Negative controls (samples with no primary antibody) for both Nanog and VNP/GFP staining's are shown, based on which the positive gate regions were designed. Nanog expression is similar between different cell lines grown in the lambda prophage bearing RET-ET recombineering system under thermolabile repressor [36]. Resulting Tet-resistant bacteria were subjected to first recombineering with PISCE plasmid to remove the LoxP site in the backbone of pTARBAC1 vector (reviewed in [20]), and to second recombineering with Nanog-VNP cassette. Tetracycline, ampicillin and neomycin-resistant recombinant BAC clones were selected by PCR screening and pulse-field gel electrophoresis of analytical digests with XhoI, PI-SceI, AscI and SnaBI enzymes. Final recombinant BAC clone was digested with AscI, ethanol precipitated and microdialysed on 0.01 mm filter (Millipore) prior to electroporation in ES cells. The linearized BAC was electroporated into E14tg2a cells (electroporation conditions: 250 V and 500 mF). Electroporated cells were plated in serum/LIF conditions and incubated at 37uC, 5% CO2. After 24 h, neomycin selection was started (200 ug/mL G418) and individual clones were picked after 12 days. Clones were expanded and analysed.
To determine the number of BAC copies integrated in the Nd ES cell line, standard curves for the relative quantities of BAC (pBAC) and of Nanog (pNanog) were determined using a dilution series of DNA plasmids containing each sequence. Threshold cycle (Ct) values were plotted against log-transformed concentrations of plasmid DNA, trend lines were calculated and quantities of pBAC relative to pNanog were determined for the wild-type E14tg2a ES cell line and for the Nd ES cell clone.

Immunocytochemistry
Fixed cells (15 min, 4%PFA) were blocked with 10% (v/v) FBS and 0.05% (v/v) Tween in phosphate buffered saline (PBS) for 1 hour, followed by overnight incubation with primary antibodies (Table S1). Cells were washed 3 times in PBS followed by incubation for 1-2 hours with AlexaFluor-conjugated secondary antibodies (Molecular Probes) and DAPI nuclear stain (1:10000, Sigma). Images of fixed cells were obtained with a DM5000B microscope and a DC350F camera (Leica Wetzlar, Germany). Living cells were photographed under an inverted microscope Leica DMIL with a DC200 camera. Images were processed using Photoshop CS (Adobe, San Jose, CA).
The number of Nanog, GFP/VNP, Oct4 and Sox2 expressing cells was quantified as a proportion of the total number of cells in culture, counted with the ImageJ Cell Counter software. The number of positively labelled cells was quantified by counting 5 to 10 randomly selected fields per coverslip, corresponding to a minimum 1000 cells, counted as DAPI nuclei. Two coverslips were counted per each condition and the analysis was repeated for at least three independent experiments for E14tg2a and Nd ES cell lines.
For FC-IS analysis, 2 ug of each antibody was used per 10 6 cells. Briefly, fixed cells were blocked for 1 hour with 0.25% (w/v) saponin and 5% (v/v) sheep serum in PBS, followed by incubation for 1 hour with primary antibodies. Cells were washed 2 times in 0.25% (w/v) saponin in PBS followed by incubation for 1 hour with AlexaFluor-conjugated secondary antibodies (Molecular Probes).

Flow Cytometry (FC) Analysis
For live cells FC analysis and sorting experiments, cells were dissociated and resuspended in 4% (v/v) FBS in PBS. Nanog:GFP and Nanog:VNP analysis was performed on a FACS Calibur cytometer (Becton Dickinson) and cell sorting experiments were done on a FACS Aria cell sorter (Becton Dickinson). Live cells were gated based on forward scatter and side scatter and/or by propidium iodide dye exclusion. For sorting, VNP positive and VNP negative Nd ES cell populations were collected and cell viability at the end of the FACS sorting procedure was determined using trypan blue dye exclusion method. FACS sorted cells were processed for RNA extraction and/or plated on gelatin-coated dishes in serum/LIF conditions. For FC-IS, cells were re-suspended in PBS after the immunocytochemistry procedure (see Immunocytochemistry section) and samples with no primary antibody were used as negative controls.

RNA Extraction and RT-PCR
Total RNA was extracted from 10 6 cells using High Pure RNA Isolation kit (Roche Diagnostics), with the inclusion of DNAseI treatment according to manufacturer's instructions. The first strand cDNA was synthesized from 0.5 ug of total RNA using SuperscriptII Revese Transcriptase (Invitrogen) and random hexamers. After synthesis, each cDNA was diluted 5-fold and 5 ul of diluted cDNA used in PCR reaction with gene-specific primers (Table S2). The absence of contaminating genomic DNA was confirmed for each RNA extraction by PCR amplification of Gapdh specific product from RT negative samples. The relative amount of each transcript was normalized to the level of Gapdh.
Quantitative PCR reactions were carried out in ABI Prism 7000 detection system (Applied Biosystems) in a 20 ul volume containing Itaq SYBR green Supermic mix (Bio-Rad) and 0.2 mM of each primer. Forty cycles were performed, each consisting of 10s denaturation at 95uC and 1 min primer annealing and elongation at the respective primers annealing temperature. The specificity of the reactions was confirmed using melting curve and gel same media, but it is highly increased (,90%) when compared to serum/LIF conditions (,55%, Figure 2B). In 2i media, both Nd and TNG-A cells reporter levels mimic Nanog expression. (B) Representative histogram showing Nanog:VNP expression for the Nd cell line grown in serum/LIF (blue) and in 2i (green). The negative control cells (E14tg2a) grown in the same conditions are represented in gray and black. In serum/LIF conditions, approximately 55% of the cells express Nanog:VNP, while this value increases to around 90% in 2i conditions. Despite this change in expression levels, a population of ES cells with no or low levels of Nanog is always observed (*). (C) Graph depicting Nanog:VNP expression data from three different biological replicates during three serial passages (P1, P2 and P3), for the Nd cell line grown in serum/LIF (blue) and in 2i (green). Statistically significant differences were consistently observed between serum/LIF and 2i conditions for all tested passages (*p-value ,0.003), while no differences were detected for cells grown in the same culture media (p-value .0.20). (D) Representative dot blots for FC-IS analysis of Oct4 and Sox2 for E14tg2a and Nd. Negative controls (samples with no primary antibody) for both Oct4 and Sox2 staining's are shown, based on which the positive gate regions were designed. Oct4 and Sox2 levels are always expressed in more than 90% of the cells in both media conditions. (E) Representative dot blots for FC analysis of live VNP/GFP for E14tg2a, Nd and TNG-A. E14tg2a cells were used as negative controls, based on which positive gate regions were designed. Obtained data is similar to that obtained by FC-IS (A). (F) Quantitative RT-PCR data for VNP, Nanog, Oct4, Sox2, Fgf5 and Dusp6 expression in E14tg2a and Nd cells. Expression levels in 2i conditions were normalized to Gapdh gene and are relative to expression levels in serum/LIF. While, for both cell lines, Oct4 and Sox2 expression does not change significantly (p-value .0.10), both VNP and Nanog expression levels show a 2-3 fold increase (*p-value ,0.03). This change is accompanied by a decrease in the expression levels of Fgf5 and Dusp6 when FGF/ERK signaling is abolished (**p-value ,5.8610 26 ). No statistically significant differences were observed between E14tg2a and Nd cell lines (p-value .0.08). (G) Representative bright field images of Nd cell line grown in serum/LIF and in 2i. In 2i conditions, ES cells grow in more tightly packed colonies with reduced flattened differentiated-like cells. Quantifications of (A,D,E) may be observed in Table 1. doi:10.1371/journal.pone.0059928.g006 electrophoresis analysis to confirm the presence of a single band. Control assays containing no templates were also performed. 2 2DDCt method was used to calculate expression levels [37] and results were subsequently analysed for statistical significance using a t-test.

Northern Blot Analysis
Total RNA (5mg/lane) was separated on formaldehyde gel electrophoresis and transferred onto Hybond N+ membrane (GE Healthcare) in 106SSC, UV-crosslinked and hybridized at 42uC with [a-32P]-dCTP (3000 Ci/mmol; Perkin-Elmer/NEN)-labeled DNA probe in UltrahybH ultrasensitive hybridization buffer (Ambion). Radioactive probes were cleaned from unincorporated radionuclide on Ilustra Microspin G-50 columns (GE Healthcare) prior to hybridization. In all experiments, probes with greater than 0.5610 9 dpm/mg activity were used. Approximate lengths of different mRNAs were estimated by comparison with the migration distance of rRNAs. (The same blots were hybridized sequentially with probes for different transcripts, after stripping in 0.5% SDS at 65uC).

Protein and RNA Half-lives
To determine protein half-life, protein synthesis was inhibited by treating cells with 100 ug/ml of cycloheximide [38] and aliquots were taken at different time points (0 h, 0.5 h, 1 h, 3 h and 6 h). 6 h DMSO treated cells were used as a control and the decay of the targeted proteins was determined by western blot analysis.
To determine RNA half-life, transcription was blocked by treating ES cells with 5 ug/ml of actinomycin D [39] and samples were collected at different time points (0 h, 0.25 h, 0.5 h, 0.75 h, 1 h, 2 h, 3 h, 4 h, 5 h and 6 h). 6 h DMSO treated cells were used as a control. Total RNA was extracted, cDNA was synthesised and quantitative PCR was performed as described in the ''RNA extraction and RT-PCR'' Materials and Methods section.

Karyotypic Analysis
Karyotypic analyses were performed for E14tg2a, Nd and TNG-A ES cell lines. Mitotic chromosome preparations were carried out using standard procedure available at the Jackson laboratory website [40]. Metaphase chromosomes were stained with DAPI, photographed and counted using ImageJ software. At least 60 metaphase nuclei were examined for each cell line ( Figure  S3). (n = 2). Trend lines were inserted and used to obtain values for slope (y = slope.x+y-intercept) and correlation coefficients (R 2 ). (B) Quantities of pBAC relative to pNanog were determined for the wild-type E14tg2a ES cell line and for the Nd ES cell clone (n = 2). Plot shows that, as expected, no pBAC integration exists in wildtype ES cells, while the Nd clone has one copy of pBAC per two of pNanog. (TIF)  Figure S4 Representative dot blots for Nanog:VNP expression after re-plating of Nd ES cells subpopulations in serum/LIF. FACS sorted populations of Nd ES cells (VNP L and VNP H ) were replated in serum/LIF and reporter's expression was measured for four consecutive days (see Figure 5B). After 2-4 days, heterogeneity is re-established and expression of Nanog:VNP is similar between the three populations, either derived from the sorted VNP L and VNP H subsets, or from the whole population (''All''). The observed rate of reversion is faster and more robust for the VNP H than for the VNP L subpopulations (respectively, 2 and 4 days). E14tg2a cells were used as a negative control, to obtain the positive gate region. (TIF)