Subcellular Localization of ENS-1/ERNI in Chick Embryonic Stem Cells

The protein of retroviral origin ENS-1/ERNI plays a major role during neural plate development in chick embryos by controlling the activity of the epigenetic regulator HP1γ, but its function in the earlier developmental stages is still unknown. ENS-1/ERNI promoter activity is down-regulated upon differentiation but the resulting protein expression has never been examined. In this study, we present the results obtained with custom-made antibodies to gain further insights into ENS-1 protein expression in Chicken embryonic stem cells (CES) and during their differentiation. First, we show that ENS-1 controls the activity of HP1γ in CES and we examined the context of its interaction with HP1γ. By combining immunofluorescence and western blot analysis we show that ENS-1 is localized in the cytoplasm and in the nucleus, in agreement with its role on gene's promoter activity. During differentiation, ENS-1 decreases in the cytoplasm but not in the nucleus. More precisely, three distinct forms of the ENS-1 protein co-exist in the nucleus and are differently regulated during differentiation, revealing a new level of control of the protein ENS-1. In silico analysis of the Ens-1 gene copies and the sequence of their corresponding proteins indicate that this pattern is compatible with at least three potential regulation mechanisms, each accounting only partially. The results obtained with the anti-ENS-1 antibodies presented here reveal that the regulation of ENS-1 expression in CES is more complex than expected, providing new tracks to explore the integration of ENS-1 in CES cells regulatory networks.


Introduction
Endogenous retroviruses (ERVs) are known to play an important role in the expression of their host genome, notably during the first developmental stages when totipotent [1] or pluripotent cells [2] adopt new cell fates. ERV's promoters act as enhancers in different cellular models and lineages [3]. More rarely, ERVs introduce coding sequences that are adopted by the host genome to play a major role in host survival. This is the case for Syncitin, involved in formation of the mammalian placenta [4], and for ERNI, also called ENS-1 [5], that controls the timing of the neural plate emergence during chick embryonic development [6]. More precisely, ENS-1/ERNI acts as a boundary element between the epigenetic regulator HP1c and the protein complex that is recruited to the promoter of the neural plate inducer Sox2 before its expression. This linker property involves two distinct motifs in the protein, the HP1box engaged with HP1c and the coiled-coil domain interacting with other proteins recruited to the promoter. The repression, mediated by HP1c, on the promoter of Sox2 is released in the prospective neural plate due to competition between ERNI and the newly synthesized protein BERT, another coiled-coil domain protein that does not bind HP1 [7].
In addition, Erni is expressed earlier during the chick developmental process notably in the hypoblast, in the pluripotent epiblast [8] and in its derived embryonic stem cells, cultivated in vitro [9], where it was called Ens-1. Silencing of the gene occurs later, as final differentiation is achieved [8,9]. In the epiblast, this expression pattern is managed by the pluripotency transcription factor Nanog and by a combination of Gata and Ets transcription factors that are expressed in the epiblast, in the hypoblast and in the prospective neural plate [8]. It is probable that the ERV Ens-1/Erni controls transcription of the host's genome in pluripotent cells either directly with its promoter sequences spread over the genome [5], or indirectly by controlling HP1c on its target genes. Among them Sox2 is also known as a key-player in the maintenance of pluripotent cells in mammals [10,11], but other genes, mainly involved in cell proliferation control, have also been described in mouse ES cells as HP1c target genes [12].
Despite its important role during chick development, demonstrated using transient transfections, the expression of the Ens-1 gene has only been followed at the transcriptional level and the endogenous protein ENS-1 has never been observed.
In this study, we raised ENS-1 specific antibodies in order to follow ENS-1 expression in CES cells and during their differentiation. Our results reveal that the distribution of ENS-1 is a new level of its regulation.

Materials and Methods
Cell culture and DNA transfection CES are pluripotent stem cells that were isolated from chick embryonic epiblast [13] and expressing the pluripotency supporting genes [14]. CES were cultivated as indicated previously [15]. Plasmid DNA transfections were performed using Lipofectamine 2000 (Invitrogen) in cell cultures at 80% confluency according to the manufacturer's instruction. Stable overexpression was obtained using linearized DNA plasmids, and stably transfected cells were selected by culture in medium supplemented with puromycin (200 mg/ml). COS7 cells were cultivated with DMEM supplemented with 6% fetal calf serum, 1% penicilin/streptomycin, 1% glutamin.

Antibodies production
The anti-Ens-1 antibody used for immunofluorescence experiments was obtained by immunization of mice with the recombinant ENS-1 protein produced in Escherichia coli as follow. The full-length coding sequence of the ENS-1 gene was fused in Cterminal with a 6x-histidine Epitope tag in a pET-22 expression system and transformed in Rosetta (DE3)pLys Competent Cells (Novagen). Protein production was induced with 1 mM IPTG for 3 hours at 37uC and immediately checked by direct loading of cells on a SDS-polyacrylamide gel followed by Coomassie blue staining. Cells were then lysed by sonication and the different fractions (soluble and insoluble cell extracts) were kept for further analysis. As ENS-1 is predominantly detected in inclusion bodies, the insoluble fraction was isolated by centrifugation and solubilized in 8 M urea, 100 mM NaH 2 PO 4 , 10 mM Tris HCl (pH 8.0). ENS-1 His tag protein was purified on Ni-NTA columns (Qiagen) and refolded by overnight dialysis in 50 mM NaH 2 PO 4 , 300 mM NaCl, 10 mM Tris HCl (pH 8.0). Purification efficiency was checked by Coomassie blue staining of a SDS-polyacrylamide gel loaded with the purified ENS-1 protein before its use as immunogen for antibody production. Production of the monoclonal antibody was carried out by Covalab (France) and immunizations were performed using complete Freund adjuvant according to the relevant legislation and following protocols approved by the ethic committee of the University of Bourgogne (France). Antibody producing hybridomas were first selected by Covalab for their reactivity with ENS-1 coated wells in ELISA tests. Positive hybridomas were next screened for use in western-blotting and in immunostaining. Their reactivity was tested in CES cells and in cells devoid of the gene Ens-1 as negative control. The selected IgG (from clone 16h4g6e4) was purified on protein A-coupled to Sepharose and called 16h4 in the results part. This unique clone worked in immunostaining and its specificity is demonstrated in the results section.
As none of the monoclonal antibodies produced worked in western-blot, we developed polyclonal antibodies in rabbits using peptides of 16-20 amino acids isolated from the ENS-1 sequence as immunogens. Design and purification of the peptides, rabbit immunizations and antibody purification from the serum were carried out by Eurogentec using its proprietary adjuvant combination and following protocols approved by the ethic committee of the CER groupe (Belgium) according to the European legislation. One of the three tested peptides (named 3807: C-DRIRVLQNEARTRAGK-CONH 2 ) gave a specific polyclonal antibody working in western blot as described in the results section. This antibody did not work in immuno-staining. The 3807 peptide (DRIRVLQNEARTRAGK) is localized upstream the coiled-coil domain in the N-terminal part of the protein and spans amino-acids in position 47 to 62 of the protein (GenBank: AAK06824.1).

DNA constructs
For transient overexpression experiments, chick HP1c and Ens-1 (Genbank: NM_001080873) were cloned in a pCi expression vector (Promega) modified to introduce a FLAG tag at the Nterminal part of the protein. HP1c (Genbank: NM_204643) was amplified from chicken ES cDNA. For in vitro translation, Ens-1 was cloned in a pGBKT7 expression vector (Clontech). Genes expressed in fusion with a GFP protein at their N-terminal part were cloned in pEGFP-C1 vector (Clontech).
For stable overexpression, Ens-1 was cloned in a bicistronic construct (pCX Ens-1-HA-ires-puro) to express ENS-1 in fusion with two HA-tags at the C-terminal part of the protein. In this construct Ens-1 was separated from the puromycin resistance gene by an IRES, allowing translation of both proteins from a common mRNA transcribed under the control of a CAG promoter (a chicken b-actin promoter combined with a CMV enhancer) to ensure a strong expression [16].
To test the transcriptional activity of proteins using the CAT reporter assay, the genes were cloned in the pM vector (Clontech) to express HP1c or ENS-1 in fusion with a Gal4 DNA binding domain at their N-terminal part. The reporter plasmid pG4-TK-CAT (Chloramphenicol Acetyl Transferase) contains six copies of the Gal4 binding sites upstream of a TK promoter to control CAT expression. pG4-TK-CAT was a gift from Dr E. Manet and has been described elsewhere [17]. The renilla luciferase reporter vector pRL-CMV, used as control, was from Promega.
Targeted mutations of ENS-1 were performed using the Quick Change Mutagenesis Kit (Stratagene) following the manufacturer's instructions. Mutation of the HP1 box (PxVxL) was performed using the following primers: Sense:TTGATGAATGGATTAGCCACAGCCAGAGCCGA-GAAATTAGTTAAC; antisense:GTTAACTAATTTCTCGGC-TCTGGCTGTGGCTAATCCATTCATCAA. The control mutation was performed on a PxVxL motif that was not involved in the interaction of ENS-1 with HP1c as assessed using a two-hybrid assay in yeast described previously [7]. The primers used were: sense:TTGATGAATGGATTAGCCACAGCCAGAGCCGA-GAAATTAGTTAAC, antisense: GTTAACTAATTTCTCGGCTCTGGCTGTGG-CTAATCCATTCATCAA.
In both PxVxL mutants the proline, valine and leucine amino acids were replaced by an alanine.
Deletion of the coiled-coil domain from position 238 to position 411 of the Ens-1 coding sequence was obtained by PCR amplification in two fragments of the sequence flanking the coiled-coil domain. Primers were designed to introduce an EcoRI motif that was used to ligate both amplified fragments. The fragment upstream of the coiled-coil domain was amplified using the following primers: sense:ACTCGAGCCACCATGAGCAACAGTATGGCC; antisense:ACCGGTGTCCTTACACGTGAACTCCACAGC-TGCTTT. To amplify the fragment downstream of the coiled-coil domain, the primers were: sense: ACGTGTAAGGACACCGGT; antisense: TACGCGTTCAGCTCCCCTTGAGCTT. The deletion of the coiled-coil domain lowers the molecular weight of the ENS-1 protein by 5.5 kDa.
To measure the activity of both start codons in the ENS1 59 UTR, fragments were cloned downstream to the T7 promoter in the p0renilla vector previously described [18] as illustrated in the results. The human b-globin 59-UTR with the authentic initiation codon was obtained by hybridizing two synthetic oligodeoxyribonucleotides (Eurogentec) downstream to the T7 promoter to generate the pGlobin-renilla vector used as control (CTR) [18].

Pull-down experiments
Chicken HP1c cloned in the pCi-flag plasmid was produced in COS7 cells upon transfection with Exgene 500 (Fermentas) following the manufacturer's instructions. The next day, lysis of COS7 cells was performed (lysis buffer: 50 mM Tris-HCl pH 7.4; 150 mM NaCl; 1 mM EDTA; 1%Triton X-100; 1X Complete protease inhibitor cocktail (Roche); 1X Halt phosphatase inhibitor cocktail (Thermo Scientific)), the lysate was centrifuged and the supernatant was incubated with [ 35 S]-Met (Amersham) to label ENS-1 synthesized in vitro using a TnT Coupled Reticulocyte Lysate Systems kit (Promega) and pGBKT7-Ens1 plasmid. The mixture was incubated for 5 hours at 4uC on a rotating shaker. The flag-HP1c protein was precipitated using Anti-flag M2 affinity gel (Sigma) incubated with the protein mixture overnight at 4uC on a rotating shaker. After centrifugation the agarose beads were washed four times in washing buffer (50 mM Tris-HCl pH 7.4; 150 mM NaCl). Bound proteins were eluted with loading buffer at 100uC and analyzed by SDS-PAGE and fluorography using a STORM Phosphorimager.

Immunofluorescence labeling
Cells were seeded on glass cover slides previously coated with gelatin 1% in 24 wells plates and cultivated to reach 80% confluency. CES cells were seeded with irradiated feeder cells and cultivated on the cover slides 48 h before analysis except for differentiated CES. In this case differentiation was induced 24 h after seeding without feeder cells, and using 10 26 M retinoic acid as previously described [8] for the indicated period of time. Culture medium was removed and fixation was performed with 2.5% paraformaldehyde in PBS for 20 min at room temperature. After washing tree times with PBS, the cells were saturated and permeabilized with saturation buffer (PBS, 0.1% Triton X-100, 10% fetal calf serum). Primary antibodies diluted in saturation buffer were incubated 1 h at room temperature. After washing three times with PBS, the secondary antibody diluted in saturation buffer was added for 1 h at room temperature. Following washing three times with PBS, cover slides were mounted on a slide using Gelmount (Biomeda) supplemented with Draq5 (Biostatus) diluted 1/1000. For co-localization studies, cells were incubated with a mixture of the anti-ENS1 monoclonal antibody 16h4 (5 mg/ml) produced in mice and the rabbit anti-human HP1c antibody from Abcam (Ab10480, 1/200). After washing, the cells were incubated with both Alexa secondary antibodies (1/1000) simultaneously. F(ab')2 goat anti-mouse IgG conjugated with Alexa 488 or Alexa 555 or F(ab')2 goat anti-rabbit IgG conjugated with Alexa 555 were from Molecular Probes. In Fig. S1 immunostaining of CES was performed using the 42s2 mouse monoclonal antibody (Upstate) directed against HP1c in addition to the Ab10480 (Abcam).
Acquisition of the images was performed with a Zeiss LSM510 Confocal microscope using a 636 (NA 1.4) Plan NeoFluor objective (PLATIM, UMS 3444 Biosciences Gerland-Lyon Sud). Each channel was imaged sequentially using the multi-track recording module before merging. Experiments were performed at least three times and gave similar results.

Electron microscopy
CES cells dissociated with trypsin-EDTA (Gibco) were pelleted by centrifugation and fixed with PFA 4% in PBS for 30 min in ice. These fixative conditions decreased the detection of the ENS-1 protein observed by immunofluorescence but were required to preserve most of the cell structures upon the following treatment.
Cells were then washed in PBS, mixed with 1% low melting point agarose and aspirated into a syringe. After agarose solidification, cells embedded into agarose were extruded from the syringe and the resulting extruded rods were cut into 3-5 mm 3 cylinders. Samples were washed in PBS at room temperature for 30 min, dehydrated in 70% ethanol and embedded in LR White resin (Electron Microscopy Sciences). Ultrathin sections were mounted on nickel grids and immunolabelling performed by floating the grids onto drops of the different solutions. We used PBS-1% BSA as saturation buffer. The anti-ENS-1 antibody (clone 16h4) and the anti-human androgen receptor antibody (441, Santa Cruz) used as IgG1 negative control were diluted in saturation buffer and incubated 1 h with the cells. After washing in PBS the cells were incubated for 1 h at room temperature with donkey anti-mouse IgG antibody conjugated with 6 nm diameter gold particles (EM grade, Jackson Immunoresearch) diluted in saturation buffer. After washing in PBS, the labeling was fixed with PBS-2% glutaraldehyde for 10 min, washed three times with PBS and once with water. Ultrathin sections were then contrasted with uranyl acetate and lead citrate, and examined with a Philips CM 120 electron microscope equipped with a Gatan Orius 200 2Kx2K digital camera (Centre Technique des Microstructures, Université Lyon 1, Villeurbanne, France).

Protein extractions and western blot
Whole cell lysates were prepared using RIPA lysis buffer (1% Triton X-100; 0.5% Sodium deoxycholate; 0.1% SDS; 150 mM NaCl; 0.1 mM DTT; 1 mM EDTA; 1 mM N-ethylmaleimide, Complete proteases inhibitor (Roche) in Tris-HCl 50 mM, pH 7.4). Following dissociation using trypsin-EDTA and washing in PBS, cells were suspended in cold RIPA lysis buffer (40610 6 cells/ml) and incubated 30 min in ice. The lysate was centrifuged at 10000 g for 30 min at 4uC. The supernatant was harvested and protein concentration was measured using Bradford reagent (Sigma) and BSA in the standard curve. The indicated protein quantities were mixed with loading buffer (10% glycerol, 2% SDS, 0.7 M b-mercaptoethanol, 0.01% bromophenol blue in 62.5 mM Tris-HCl pH 6.8) and denaturation was performed 3 min at 100uC.
Fractionation experiments were performed using two protocols. In the first one, cells were disrupted using detergent before the cytoplasm and the nucleus were separated. Dissociated cells were suspended (60610 6 /ml) in cold buffer A (340 mM sucrose, 60 mM KCl, 15 mM NaCl, 2 mM EDTA, 0.5 mM EGTA, 0.1 mM DTT, 0.65 mM spermidine, 0.1% Triton X-100, Complete EDTA (Roche), 10 mM N-ethylmaleimide in 15 mM Tris-HCl pH7.4) and incubated 1 min in ice. The cytoplasm and the nuclei were separated by centrifugation 1500 g for 5 min at 4uC. The cytoplasm corresponding to the supernatant was clarified from residual nuclei by centrifugation 13000 g at 4uC for 5 min and the supernatant was harvested (cytoplasmic fraction, C). The nuclei containing pellets were suspended in buffer A and washed by deposition on a buffer A cushion in a 14 ml falcon tube followed by centrifugation at 9400 g for 10 min at 4uC. The nuclei were disrupted using the hypotonic buffer B (2 mM EDTA, 0.5 mM EGTA, 0.1 mM DTT, Complete EDTA (Roche), 10 mM N-ethylmaleimide in water). After 30 min incubation in ice, nuclei disruption was checked using a microscope. Next the soluble and insoluble nuclei components were separated by centrifugation 18000 g for 5 min at 4uC. Protein concentration was measured using Bradford reagent (Sigma) in the cytoplasmic fraction and in the soluble fraction of the nucleus containing fractions and the indicated quantities were mixed with protein loading buffer. The insoluble nucleus fraction was suspended in buffer A and mixed with the loading buffer (N) without measurement of the concentration but the volume corresponding to 6610 6 cells was loaded to the gel.
In the second fractionation protocol the cells were disrupted without detergent according to the protocol provided by Abcam with some modifications. Briefly, cells were dissociated in PBS using a scraper, rapidly pelleted by centrifugation, re-suspended in cold fractionation buffer (250 mM sucrose, 20 mM Hepes, 10 mM KCl, 1.5 mM MgCl2, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 1 mM N-ethyl maleimide, Complete protease inhibitor (Roche)) and passed through a 26G needle before 30 min incubation on ice. Cell disruption was checked using a microscope. The nuclei were separated from the cytoplasm by centrifugation 1500 g for 5 min at 4uC. The supernatant contained the cytosol, the cytoplasmic membrane and mitochondria, the pellet contained nuclei. The supernatant was clarified by centrifugation 18000 g, 5 min at 4uC to separate the mitochondria in the pellet from the cytosol and the cytoplasmic membrane in the supernatant. The cytosol and the cytoplasmic membrane were separated by ultracentrifugation 100000 g for 1 h at 4uC. The supernatant corresponding to the cytosol (Cs) was harvested and the pellet was washed once in the fractionation buffer before suspension in RIPA lysis buffer (Cp). The nuclei were washed as indicated in the first protocol on a cushion of fractionation buffer, suspended in RIPA lysis buffer and sonicated (N).
For protein dosage and loading, all the fractions were treated as indicated for whole cell lysates.
Proteins separated on 10% SDS-polyacrylamide gels were transferred on Hybond ECL membranes (GE Healthcare). Transfer efficiency of the proteins was checked by transient staining of the membrane using Ponceau's red (Sigma). Blots were submitted to Western analysis using the following antibodies: anti-HP1c (clone 2MOD-1G6AS, Euromedex), anti-HA (HA.7, Sigma), anti-b tubulin (Sigma), anti-HSP90 (AC88, Enzo), anti-LaminB1 (Abcam), 2MeH3K4 (Upstate), 3807 anti-ENS-1 polyclonal antibody. Anti-mouse and anti-rabbit secondary antibodies coupled with peroxidase were from Sigma. Saturation of the membrane was performed in saturation buffer (0.1% Tween, 5% low fat milk in PBS) 1 h at room temperature. Antibodies were diluted in saturation buffer and incubated with the membrane at room temperature for 1 h. After washing 3 times for 10 min in PBS-0.1% Tween, the membrane was incubated 1 h at room temperature with the HRP-conjugated secondary antibodies diluted in saturation buffer. Anti-mouse and anti-rabbit HRP conjugates were from Promega. After washing, peroxydase activity was revealed using the Amersham ECL western blotting detection reagent (GE Healthcare) as substrate and X-ray films (Fuji).
CAT assay CES cells seeded in 12-well plates were co-transfected at 70% confluency using Lipofectamine 2000 (Invitrogen) as described by the manufacturer with a mixture of the following plasmids: pM plasmid (20 ng) encoding for proteins in fusion with Gal4 DNA binding domain (GBD-proteins), pG4-TK-CAT reporter plasmid (300 ng) allowing the tethering of the GBD-proteins upstream the TK promoter that controls the expression of the CAT gene, pCi plasmid (200 ng) allowing the expression of additional proteins and pRL-CMV plasmid (20 ng) as control to normalize transfection efficiency. After 24 h culture, the medium was removed and the cells were washed twice with PBS. Cell lysates were prepared and CAT expression was measured using the CAT ELISA kit from Roche according to the manufacturer's instructions. Renilla luciferase luminescence was measured on templates from the same lysates using the Renilla Luciferase Assay System (Promega) according to the manufacturer's protocol and using a Mithras LB940 microplate reader (Berthold technologies).
In silico analysis of the Ens-1 gene copies in the chicken genome We used the sequence of Soprano [19] as a query to search the chicken genome (version WASHUC 2, retrieved from Ensembl http://www.ensembl.org/index.html) using BLASTN [20]. We recovered 78 copies among which 25 were very conserved compared to the reference sequence. We used ORF Finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html) on the 25 conserved sequences to determine the presence of ORFs and to retrieve the subsequent protein sequences. The coiled coil domains were predicted on the protein sequences using the MultiCoil program [21].

In vitro transcription
Capped RNAs were obtained by using 2 mg of linear DNA template mixed with 20 U of T7 RNA polymerase (Promega Co., Madison, WI, USA), 40 U of RNAsin (Promega Co, Madison, WI, USA), 10 mM rATP, rUTP, rCTP, 0.48 mM rGTP, 30 mM DTT in transcription buffer [40 mM Tris-HCl (pH 7.9), 6 mM MgCl2, 2 mM spermidine and 10 mM NaCl], the m7GpppG cap analogue (Invitrogen, Co) was added to a final concentration of 1.92 mM as previously described [22]. The transcription reaction was carried out at 37uC for 1 h and the RNAs were precipitated with LiCl at 2.5 M final concentration. The integrity of the RNAs was checked by electrophoresis on non-denaturating agarose gels and their concentration was quantified by spectrophotometry at 260 nm using Nanodrop (NanoDrop Technologies, Wilmington, Delaware, USA).

In vitro translation
In vitro transcribed RNAs were translated in 10 ml of the supplemented untreated RRL 50% (v/v) (Promega Co., Madison, WI, USA) in the presence of KCl (75 mM), MgCl 2 (0.5 mM), 20 mM of amino acids mix minus methionine and 0.6 mCi of [ 35 S]-methionine (GE Healthcare Life Sciences) for 30 min at 30uC. Reactions were stopped with 26 SDS-loading buffer and the products were resolved by 15% SDS-PAGE. Gels were dried and subjected to autoradiography using Biomax films (Eastman Kodak Co.).

ENS-1 can control HP1c activity in ES cells
The control exerted by ENS-1 on the transcriptional regulation mediated by HP1c has been demonstrated in the neural plate [7].
Here we investigated whether this activity might also occur in chicken ES cells (CES). To this end we used a CAT reporter gene placed under the control of a minimal VP16 promoter downstream of the DNA motif recognized by the Gal4 protein.
Cells were co-transfected with a pM vector encoding for HP1c in fusion with the DNA binding domain of Gal4 (GBD) to target its interaction with the CAT reporter plasmid. In line with results The Retroviral Protein ENS-1 in Chick ES Cells PLOS ONE | www.plosone.org from another study [23], Figure 1A shows that transcription was repressed when tethering HP1c to the promoter. Transfection of CES cells with the previous cocktail supplemented with an ENS-1 expression vector restored the transcriptional activity to levels obtained in the absence of HP1c while ENS-1 mutated in the HP1 box had no effect (Fig. 1A). To test whether ENS-1 acts directly on the promoter as a transcriptional activator, ENS-1 in fusion with the GBD was co-transfected with the CAT reporter plasmid. Transcription was not affected by the recruitment of ENS-1 to the promoter (Fig. 1B) confirming that the effect of ENS-1 on transcription was dependent on its interaction with HP1c.
Therefore ENS-1 can directly interact with HP1c and modulate its function as a transcriptional regulator in CES.

ENS-1 is both a cytoplasmic and nuclear protein
Since HP1c is a nuclear protein, we looked for the distribution of the endogenous protein ENS-1 in these cells. To this end a monoclonal antibody was produced against the recombinant protein ENS-1 (as described in Material and Methods) and tested for use in confocal microscopy. Among a dozen of clones reacting with the immunogen, only one was suitable for immunostaining. Specificity for ENS-1 was first demonstrated by transfection of CES with the ENS-1-GFP fusion protein. As shown in Figure 2A, the antibody co-localized with ENS-1-GFP but not with GFP alone used as control. Secondly, as Ens-1 is restricted to the galliform species [5], we used STO murine embryonic fibroblasts as negative control. No signal was observed with the antibody in these cells (Fig. 2B). Therefore the antibody is specific for ENS-1 and represents a relevant tool to study expression and distribution of this protein in CES. As shown in Figure 2C almost all the CES cells expressed ENS-1 but inside the cells the labeling was not uniform, notably in the cytoplasm when compared with the nucleus (see Fig. 2D). In the cytoplasm the distribution of ENS-1 appeared as organized, forming a scaffold of spots and stretches that sometimes crossed-over the nuclear membrane. In the nucleus that was labelled with Draq5, a stoechiometric DNA-intercaling agent, the distribution of ENS-1 was more uniform. Nucleolus that were not labeled with Draq5 were poorly stained with the anti-ENS-1 antibody (Fig. 2D).
The distribution of ENS-1 was further examined by electron microscopy with the same antibody (Fig. 3). To preserve the intracellular structures this approach required fixative conditions that were stronger than those used for confocal analysis and thus altered the ENS-1 epitope recognized by the 16h4 antibody. A compromise was found using 4% PFA only. Most of the cell structure was preserved as well as the ENS-1 epitope but some cytoplasmic structures were lost, leaving empty areas in the cytoplasm that likely correspond to the endoplasmic reticulum (Fig. 3). ENS-1 was abundantly detected at the periphery of these areas (Fig. 3A) and in the nucleus (Fig. 3B). Examination of the cytoplasmic compartment shows that ENS-1 accumulates at the cytoplasmic membrane (Fig. 3A). In the nucleus, the protein was found in regions that were more or less dense to electrons, and corresponding to chromatin and to non-chromatin spaces respectively, suggesting that ENS-1 location in the nucleus is not restricted to protein complexes formed on DNA (Fig. 3B). No particles were observed in controls where anti-ENS-1 antibody was replaced by an irrelevant antibody (Fig. 3D) or in the absence of primary antibody (Fig. 3C).
Therefore, we conclude that the protein ENS-1 is expressed in all CES cells and is largely present in the cytoplasm and in the nucleus.

ENS-1 sparsely co-localizes with HP1c in the nucleus and in the cytoplasm
To explore the context in which the interaction between the endogenous proteins ENS-1 and HP1c may occur, co-localization experiments were performed in CES. Consistent with a previous study in mouse ES cells [24], immunostaining of HP1c heterogeneously formed a speckled and diffuse pattern in the nucleus of CES (Fig. 4A). Yellow spots in the orthogonal sections of the cells (Fig. 4B, panel a) indicated that both proteins colocalized in interphasic nuclei while other ENS-1 and HP1c molecules were visualized as neighboring green and red spots respectively. Merging with the DNA staining, in blue, produced white spots reflecting co-localizations of ENS-1 with HP1c in chromatin (Fig. 4B, panel b) and a majority of purple spots corresponding to HP1c on chromatin. These results are in agreement with the ability of ENS-1/HP1c heterodimers in the The Retroviral Protein ENS-1 in Chick ES Cells PLOS ONE | www.plosone.org embryo to specifically regulate gene transcription [7]. Panels c, d and e, showing each of the labeling that were merged in panels a and b (Fig. 4B), revealed that ENS-1 was present in all the thickness of the nucleus (panel c) while HP1c and the chromatin had a more restricted distribution. Notably HP1c proteins were concentrated in the center of the cells (panel d) while chromatin was mostly observed in the upper half (panel e). Measurement of the fluorescence intensity along the longitudinal axis of two cells confirmed that most of the HP1c dots did not overlap with those of ENS-1 (Fig. 4C, upper panel) and that co-localization between both proteins was a restricted event that also occurred in regions with low DNA density (Fig. 4C, lower panel). These results suggest that ENS-1/HP1c heterodimers can exist independently on the interaction with chromatin. In agreement with this hypothesis, pull-down experiments performed between in vitro translated 35 S-ENS-1 and cell lysates from COS7 cells overexpressing chicken HP1c demonstrate that the interaction between both proteins is direct, involving the HP1 box of the ENS-1 protein but not its coiled-coil domain as illustrated by the use of mutated ENS-1 proteins (Fig.4D). During mitosis, HP1c is released from the chromatin and is distributed diffusely throughout the cell [25]. Accordingly, in prometaphasic cells when the nuclear membrane was dissolved, HPc and ENS-1 were progressively released from the condensing chromatin and accumulated in the cytoplasm (Fig. 4E). Co-localizations were observed (Fig. 4E panel b) but most of the proteins were maintained separated as illustrated by the few number of yellow spots in comparison with the red and green ones (see panels a and c, Fig. 4E). In interphasic cells some HP1c molecules were also detected in the cytoplasm as confirmed by the use of two distinct antibodies (Fig. S1) and a partial colocalization with ENS-1 was found in this compartment (Fig. S2 a).
Therefore the interaction between ENS-1 and HP1c does not require any CES specific co-factor nor chromatin binding and might occur in the cytoplasm and in the nucleus. However despite a close contact between both proteins outside the nucleus during mitosis, the majority of them co-localize in the nucleus of interphasic cells.

ENS-1 protein is differently regulated in the cytoplasm and in the nucleus
To better characterize the protein ENS-1 in the cytoplasm and in the nucleus, we performed western blot analysis of whole cell lysates. To this end we developed another ENS-1 specific antibody called 3807 (see Material and Methods). As shown in Figure 5A, this antibody gave a unique band of about 55 kDa in CES, but not in mouse ES cells, indicating its specificity for chicken cells. This molecular weight is in accordance with the 54 kDa prediction made from the protein sequence (490aa, Genebank: AAK06824). To confirm that the protein recognized by the 3807 antibody is ENS-1, CES cells were transfected with an expression vector (pCx ENS-1-HA-ires-puro) encoding for an HA-tagged ENS-1 protein.
To be sure that antibiotic resistant cells will also express Ens-1 transcripts, both genes were placed under the control of the same promoter and were separated by an IRES sequence to generate a unique transcript for both genes that are next translated independently. Indeed, culture with puromycin of CES cells transfected with pCx ENS-1 but not with the empty vector, induced the expression of a unique HA-tagged protein as observed with the anti-HA antibody in Figure 5B. This protein is also recognized by the 3807 antibody in addition to the endogenous ENS-1 protein detected below and in control cells. These results confirm the specificity of the 3807 antibody for ENS-1.
In a previous report we have shown that the promoter of Ens-1 is repressed while the expression of Sox2 is induced when CES differentiate [8]. These results are in agreement with the idea that differentiation of CES in vitro may mimic the release of the repression mediated by the dimer ENS-1/HP1c on the promoter of Sox2 before the emergence of the neural plate. To explore the relevance of this hypothesis, variations of ENS-1 expression during differentiation were examined at the protein level. Western-blot results in Figure 5C show that the protein content decreased to become quite undetectable 24 h after retinoic acid addition and was maintained at this residual level after 48 h. In the same samples the expression level of HP1c was not affected by differentiation (Fig. 5C) in agreement with previous observations in mouse ES cells [12]. These results are in accordance with the known regulation of the promoter of Ens-1 [8]. To confirm with another approach, ENS-1 expression was analyzed during CES cells differentiation by immunofluorescence experiments using the 16h4 antibody. This approach confirmed the decrease of the cytoplasmic ENS-1 during differentiation but with changes in the distribution of ENS-1 (Fig. 6A). In contrast to undifferentiated cells where the protein was mostly cytoplasmic, the localization was mostly nuclear in differentiated cells (Fig. 6A, Fig. S2). Notably, the organized distribution of ENS-1 in the cytoplasm was lost, giving a diffuse signal after 24 h differentiation that became undistinguishable from the back-ground after 48 h while the signal in the nucleus was still over the background (Fig. 6A). These results suggest that variations in ENS-1 contents are not similar in the cytoplasm and in the nucleus.
To further explore these changes, enrichments in the nuclear and in the cytoplasmic proteins were obtained by fractionation of cell lysates before analysis by western blot using the 3807 antibody.
Results in Figure 6B show that the 55 kDa ENS-1 protein is present in the cytoplasm and in the nucleus of undifferentiated cells. However in the nucleus only, this band is flanked by two other ones of 58 and 47 kDa that were not detected when using whole cell lysates. During differentiation, the 55 kDa protein was decreased in the cytoplasm but maintained in the nucleus. This was also generally true of the two other proteins. More precisely, the relative proportions of the three ENS-1 related bands in the nucleus were modified during differentiation. In undifferentiated cells, the 58 kDa nuclear protein was more abundant than the two others. After 24 h differentiation, the proportion of the 55 and the 47 kDa bands increased relatively to the 58 kDa band and became more abundant. Later, in 48h-differentiated cells, all three proteins were equally and strongly detected in the nucleus (Fig. 6B).
Altogether these results confirm the observations made by immunostaining with the 16h4 antibody and indicate that the 55 kDa protein ENS-1 is maintained in the nucleus but not in the cytoplasm upon differentiation. They also reveal that ENS-1-like It is of note that all these proteins are associated with the insoluble components of the nucleus (Fig. 6B) and were not found in the soluble fraction (Fig. S3). The decrease observed in Figure 5C with whole cell lysates reflects the cytoplasmic ENS-1 protein only since insoluble components were discarded by centrifugation before analysis.
To confirm that the cytoplasm does not contain the additional 47 and 58 kDa proteins detected in the nucleus, a fractionation protocol disrupting the cells without detergent was managed to preserve the soluble and the insoluble parts of the cytoplasm that were next separated by centrifugation. Results in Figure 6C show that ENS-1 was restricted to the pellet confirming that ENS-1 is not a soluble protein. Contamination with nuclear components was excluded since a unique band of 55 kDa was recognized with the anti-ENS-1 antibody and only traces of histone (2MeH3K4) and laminB1 coming from the nucleus were found. In parallel, analysis of the nuclear content showed the additional bands already observed with the previous protocol. The homogenous protein loading rendered possible the quantitative comparisons between the distinct sub-cellular fractions. The 55 kDa protein ENS-1 was equivalently distributed between the cytoplasm and the nucleus in CES cells. In the nucleus only, the 47 kDa was present but poorly detectable and the abundance of the 58 kDa was confirmed. Interestingly, HP1c was found in all fractions, in agreement with immunofluorescence data. Therefore, it seems that differently from HP1c, ENS-1 is not a free protein in the cytoplasm nor in the nucleus but is rather associated with insoluble structures.
Altogether the results obtained by a combination of immunofluorescence and western blot approaches using two distinct ENS-1 specific antibodies support the conclusion that the endogenous protein ENS-1 is differently regulated in the cytoplasm and in the nucleus and that ENS-1-like proteins also contribute to this difference between both compartments.
In silico analysis of Ens-1 gene copies Due to its retroviral origin, distinct copies of the gene Ens-1 exist [5] that may generate distinct ENS-1 like proteins depending on the conservation of their ORF sequence. Those with the most conserved ORF in an updated sequence of the chicken genome are listed in Table 1. Seven copies encode for a 54 kDa ENS-1 protein and one copy for an ENS-1 like protein of 47 kDa but none for a 58 kDa protein. Based on a full sequence homology with the 3807 peptide and on conservation of the promoter active domains [8], the 47 kDa and five of the 54 kDa proteins encoding copies may account for the protein pattern detected with the 3807 antibody. Copies encoding for lower or for higher molecular weight proteins ranking from 34 to 69 kDa fulfill the same criteria but were not detected by this antibody (see Fig.5A, 6C and S3 to compare distinct cell lysis protocols). Of note, all the ENS-1 like proteins with an intact 3807 sequence have a conserved coiled-coil domain but only the 54 kDa proteins have an intact HP1box. Thus, it can be assumed that ENS-1-like proteins would be inactive towards HP1c if generated from distinct gene copies.
Alternatively, the 58 kDa ENS-1 like protein may result from translation initiation starting from a start codon distinct from the ATG that accounts for the 54 kDa protein. This hypothesis is supported by the existence of a GTG sequence in the 59UTR of the gene Ens-1 that is located 96 nucleotides upstream the ATG (Fig. 7A) and may extend the Ens-1 coding sequence, as already reported for human genes [26]. Both codons are in frame and the N-terminal extension generated from the GUG would overweight the ENS-1 protein of 3 kDa (Fig. 7B), corresponding to the difference between the 55 and the 58 kDa proteins detected by western blot. We checked the translation efficiency from the GUG codon alone or in the presence of the ATG. Using the constructs containing the 59UTR of ENS-1 upstream of a luciferase reporter  gene under the control of a T7 promoter (Fig. 7C), we generated RNAs that were used for in vitro translation. Results in Figure 7D show that the GUG start codon efficiently initiated the translation of the luciferase gene (as indicated by the arrow (LUC) in lanes II and I). When the ATG start codon was also present, translation was initiated from both codons to yield two proteins that differ in size by 3 kDa. Luciferase was produced from the ATG while the upstream GTG was responsible for the synthesis of the extended protein by translation of the 99 nucleotides segment from GTG to ATG (in lanes IV and III). It is noteworthy that protein expression from the GTG was less efficient when the latter was placed upstream of the ATG. Interestingly, when the Ens-1 59UTR was driving protein synthesis (constructs I and III), expression was severely reduced from both start codons and translation from GUG became residual. These data indicate that an N-terminal extension of ENS-1 during RNA translation may account for the 58 kDa ENS-1 like protein but the low activity of the GUG start in the context of the Ens-1 59UTR suggests that an additional regulation of translation would be involved in CES for a real contribution to the protein pattern revealed with the 3807 antibody.
Finally, it cannot be ruled out that ENS-1 like proteins result from post-translational modifications of the ENS-1 protein itself. In yeast two-hybrid assay ENS-1 interacts with the conjugating enzyme Ubc9 (unpublished data) that transfers a SUMO protein to its protein substrate [27]. SUMO are about 10 kDa proteins in size, that is compatible with a modification of the 47 kDa ENS1like protein. We found in the sequence of ENS-1 two sumoylation motifs that are conserved in all the copies with an intact 3807 peptide sequence (Table 2), thus supporting the potential involvement of post-translational modifications in the production of the 5 kDa protein. It is of note that one of the SUMO binding sites is located inside the coiled-coil domain involved in dimerization of the protein and in its recruitment to Sox2 promoter [7]. In addition, as SUMOylation has been reported to regulate intracellular distribution of proteins [28], this may explain the nuclear localization of ENS-1 in spite of lack of nuclear localization signal.

Discussion
This paper describes the endogenous form of the protein ENS-1 that controls the timing of the neural plate emergence in chicken by interacting with the epigenetic regulator HP1c. Using different approaches, based on new antibodies raised against ENS-1, we show that this protein is strongly expressed in pluripotent CES cells isolated from the epiblast, in agreement with its transcrip-tional pattern of expression [8]. These antibodies are specific for ENS-1 since they react with a tagged-ENS-1 transgene but not with cells devoid of the Ens-1 gene. Each antibody was screened for immunofluorescence and for western-blot applications but none worked in both approaches. Therefore both antibodies were used in parallel in western blot or in immunofluorescence to characterize the protein ENS-1. They gave converging results about localization of the protein both in the nucleus where HP1c is concentrated and in the cytoplasm, but with distinct features as revealed by western blot. In the nucleus, but not in the cytoplasm, three anti-ENS-1 reactive bands were found. Our results support the idea that the relative proportion of these ENS-1 like proteins is part of the regulation of ENS-1.
Indeed, differentiation of ES cells using retinoic acid promotes their commitment preferentially toward the neurectoderm [12]. Under these conditions, the expression of ENS-1 at the transcriptional level is only partially repressed [8]. With both antibodies we show in parallel, that the ENS-1 protein content is decreased in the cytoplasm but not in the nucleus where it is maintained but with deep changes in the relative proportions of all three proteins.
In the prospective neural plate the recruitment of ENS-1 to the promoter of Sox2 represses the gene, and activation [29] [30] is induced by competition with the neo-synthesized protein BERT that releases ENS-1 [7]. In mice, Sox2 [31] is known to be, along with Oct4 [32] and Nanog [33], one of the key players of pluripotency that are repressed during differentiation. Differently, in CES cells we reported that Sox2 is induced upon differentiation [8], suggesting that ENS-1 is fully active to repress Sox2 in CES. In agreement, we show here that ENS-1 can control the function of HP1c in a reporter system and that both proteins locally colocalize on chromatin in the nucleus of CES cells.
Identification of genes targeted by the ENS-1/HP1c heterodimer in CES represents a future challenge to understand integration of these proteins in the pluripotency network of CES cells, and the data presented here strongly support the development of approaches directly addressing this question at the protein level. The development of additional antibodies suitable for immune-precipitation will be required, this cannot be achieved with the antibodies presented here. Targeting other epitopes of the protein for antibody preparation will also be essential to better characterize the nature of the ENS-1 like proteins. The analysis of the different Ens-1 gene sequences suggests that distinct copies of the gene Ens-1 or another translation initiation start could at least partially explain the ENS-1 protein pattern observed, but the contribution of post-translational modifications cannot be excluded. In conclusion our results demonstrate that understanding the regulation of the protein ENS-1 is more complex than could be anticipated from analysis of its transcripts and this study provides new tools to track this important protein during chick development. Figure S1 Immunostaining of the chicken HP1c with the commercial antibody. (A) CES cells transiently transfected with HP1c in fusion with GFP (green) were labeled with the Ab1080 (Abcam) antibody used in Fig. 4. (B) CES cells were labeled with a mixture of the rabbit Ab1080 (red) and the mouse 42s2 (green, Upstate) anti-HP1c antibody. The secondary antibodies used were an anti-rabbit conjugated with Alexa 555 (red) and an anti-mouse antibody conjugated with Alexa 488 (green). Overlapping signals gave a yellow color. Results from two independent experiments are presented. In the nucleus both antibodies gave similar staining and both detected traces of HP1c in the cytoplasm but with more or less intensity depending on the experiment. (TIF) Figure S2 Distribution and colocalization of ENS-1 with HP1c during differentiation of CES cells. Images are those presented in Figure 5C complemented with fluorescence intensity for the three labeling along the line represented in the merged image by the red arrow. CES cells (a) or cells differentiated for 4 h (b), 24 h (b) or 48 h (d) with retinoic acid were stained as in Figure 4 with anti-ENS-1 (green) and anti-HP1c antibody (red). Signal intensities are presented as arbitrary units. (TIF) Figure S3 Western blot of soluble and insoluble protein fractions in the nucleus of CES cells. The whole blot corresponding to Fig. 5D is represented. N1 to N3 and S1 to S3 are loading replicates of respectively the insoluble and the soluble fractions of the nucleus. The S fractions had proteins concentrations lower (5 mg) than the N fractions (15 mg) but even in the S3 fraction that was slightly colored by Ponceau's red, no ENS-1 protein was detected. (TIF)