Proteomic Profiling of Rabbit Embryonic Stem Cells Derived from Parthenotes and Fertilized Embryos

Rabbit embryonic stem (rES) cells can be derived from various sources of embryos. However, understanding of the gene expression profile, which distincts embryonic stem (ES) cells from other cell types, is still extremely limited. In this study, we compared the protein profiles of three independent lines of rabbit cells, i.e., fibroblasts, fertilized embryo-derived stem (f-rES) cells, and parthenote-derived ES (p-rES) cells. Proteomic analyses were performed using two-dimensional gel electrophoresis (2-DE) and mass spectrometry. Collectively, the expression levels of 100 out of 284 protein spots differed significantly among these three cell types (p<0.05). Of those differentially expressed spots, 91% were identified in the protein database and represented 63 distinct proteins. Proteins with known identities are mainly localized in the cytoplasmic compartments (48%), nucleus (14%), and cytoskeletal machineries (13%). These proteins were majorly involved in biological functions of energy and metabolic pathways (25%), cell growth and maintenance (25%), signal transduction (14%), and protein metabolisms (10%). When protein expression levels among cell types were compared, six proteins associated with a variety of cellular activities, including structural constituents of the cytoskeleton (tubulins), structural molecule (KRT8), catalytic molecules (α-enolase), receptor complex scaffold (14-3-3 protein sigma), microfilament motor proteins (Myosin-9), and heat shock protein (HSP60), were found highly expressed in p-rES cells. Two proteins related to HSP activity and structural constituent of cytoskeleton in f-rES cells, and one structural molecule activity protein in fibroblasts showed significantly higher expression levels (p<0.05). Marker protein expressions in f-rES and p-rES cells were further confirmed by Western blotting and immunocytochemical staining. This study demonstrated unique proteomic profiles of the three rabbit cell types and revealed some novel proteins differentially expressed between f-rES and p-rES cells. These analyses provide insights into rES cell biology and would invite more in-depth studies toward rES cell applications.


Introduction
Rabbit embryonic stem (rES) cells are pluripotent cells derived from the blastocyst stage embryos [1][2]. Recently, more newly established rES cells are derived from fertilized embryos (f-rES) [3][4][5][6][7][8][9]. Parthenogenetically activated (PA) oocytes or embryos are subjected to artificial stimuli to initiate embryonic development without fertilization process or incorporation of sperm chromosomes. These parthenotes possess chromosomes entirely of the maternal origin and fail to develop to term due to a lack of paternal gene expressions or normal genomic imprinting [10][11]. Similar to f-rES cells, parthenote-derived rES (p-rES) cells can continuously proliferate in vitro, retain self-renewal capacity without differentiation [3][4]12], and also differentiate into cell lineages of the three germ layers both in vitro and in vivo. They expressed the same set of pluripotency marker genes (Oct4, Nanog, and Sox2), alkaline phosphatase (AP), and proteins such as cell surface markers including stage specific embryonic antigen-4 (SSEA-4), keratin sulfate antigens TRA-1-60, and TRA-1-81 as well as octamer-binding transcription factor (Oct4) [3][4]12], so did the f-rES cells. However, there are fundamental differences between f-rES and p-rES cells in that p-rES cells are generated from homozygous embryos consisting of only haploid female genome and lack of the expression of paternally imprinted genes such as Snrpn and Igf2 [3,12]. With these properties, p-rES cells have been proposed and proved to be useful in ameliorating or completely eliminating the risk of immunological rejection after cell transplantation [13].
Recently, proteomic analyses have been performed to monitor the global protein expression and post-translational modifications in mouse ES (mES) cells [14][15][16], and to determine the protein expression profiles in mouse [17][18], monkey [19], and human [20][21] ES cells undergoing chemically induced differentiation. So far, no authentic germline transmissible rES cell lines were reported and the molecular mechanisms superimposing distinct

Reverse Transcription-polymerase Chain Reaction (RT-PCR)
Rabbit ES cell colonies were lifted from feeders with dispase (1 mg/mL; Gibco17105041) at 37uC and fibroblasts were subject to trypsinization. The procedures for RNA extraction were performed as previously described [8,12]. Total RNAs of rabbit fibroblasts, f-rES, and p-rES cells were extracted using a total RNA extraction kit (Geneaid RT050, Taipei

Sample Preparation for Proteomic Analysis
To prevent contamination of feeder cells with rES cells, rES cell colonies were lifted from feeders by treatment with dispase (1 mg/ mL; Gibco17105041) at 37uC for 1-2 min for gentle cracking, and then rES cell culture medium was added to stop the enzymatic reaction. The colonies were collected into a 15-mL tube and stood for 3 min to settle and separate rES colonies from feeder cells. Old medium in the tube was replaced with fresh medium (10 mL) and then the tube was again left still for 3 min to set down rES cell colonies. The same procedures were repeated for three times to remove feeder cells from rES cells for analyses. For proteomic analysis, cultured rabbit fibroblasts and rES cells were washed with DPBS (Cat. No 21600-051, Gibco Products International) then trypsinized to single cells and centrifuged at 806g. The cell pellets were frozen in liquid nitrogen and stored at -80uC for further analysis. Cell samples (.10 6 cells per sample) were lysed in lysis buffer (9.5 M urea, 65 mM DTT, 2% Ampholyte pH 3-10, and 2% NP-40) and then frozen at -80uC for 20 min. After thawing and centrifugation at 19,0006g for 5 min, the supernatant was collected. Protein concentrations were determined by the Ettan 2-D Quant kit (GE Healthcare, Bio-Science AB, Uppsala, Sweden) using BSA as the standard. A total of 1,000 mg soluble proteins were subject to trichloroacetic acid (TCA) precipitation before analyses. Briefly, equal volume of 20% TCA was added to the sample and then incubated on ice for 1 h (vortexed every 15 min). The sample was then centrifuged and the supernatant was discarded. The pellets were washed twice with two volume of 90% ice-cold acetone and centrifuged at 19,0006g at 4uC for 10 min. The pellet was then lyophilized and dissolved in lysis buffer for protein analysis.

Protein Analysis by 2-DE
The 2-DE procedure was based on Görg et al. [22] with some modifications [23]. Briefly, 1,000 mg TCA-precipitated protein in 175 mL lysis buffer was mixed with an equal volume of rehydration buffer (8 M urea, 2% CHAPS, and 0.5% (v:v) Pharmalyte (pH 3 to pH 10), and then subject to IEF using 18-cm immobilized pH gradient (pH 3-10) strips on an Ettan IPGphor 3 (GE Healthcare, Bio-Science AB, Uppsala, Sweden) at 20uC. The strip was rehydrated at 30 V for 12 h and further focused for 64,000 voltage-hour. The focused strips were equilibrated first in equilibration solution (50 mM Tris-Cl (pH 6.8), 6 M urea, 30% (w/v) glycerol, 2% (w/v) SDS, and 0.002% (w/v) bromophenol blue) containing 100 mM DTT for 20 min, and then the strips were further equilibrated in equilibration solution containing 150 mM iodoacetamide for 20 min. After equilibration, proteins were separated by 12.5% SDS-PAGE using the Daltsix Vertical electrophoresis system (GE Healthcare, Bio-Science AB). The separation was run at 15uC with a condition of 2.5 Watts per gel for 25 min followed by 9 Watts per gel until the dye front reached the bottom of the gel (typically 7-7.5 h). The molecular weight (Mr) standards were purchased from Fermentas (#SM0661 Unstained Protein Ladder, Vilnius, Lithuania) containing synthesized peptides with the molecular weights ranging from 10 to 200 kDa. (Image Scanner III, GE Healthcare, Bio-Science AB) and saved as TIFF image files for further analysis.

Analyses of Protein Expression Levels in Different Cell Types
For each individual cell line, three 2-DE profiles were used for image analyses. Protein spots on the 2-DE gels were detected and analyzed using the Melanie 7 software package (GeneBio, Geneva, Switzerland). All the spots detected on the profiles were then grouped and quantified for spot volume by the software. Relative volumes of each spots were calculated to represent the expression levels. Ratios of the volume (RVol) of each spot to total volume of all quantified spots on each gel were generated by the software to correct variations in gel staining [25]. The RVol value thus represented the expression level of each protein spot. Expression ratios of spots among cell types were also calculated.

In-gel Digestion of Differentially Expressed Proteins
Differentially expressed protein spots were excised from the gels and placed in Eppendorf tubes. In-gel digestion was performed according to the procedure published by Havlis et al. [26] with minor modifications [23]. Gel plugs were washed twice with double distilled water followed by 50% acetonitrile in 50 mM ammonium bicarbonate and then with pure acetonitrile. Gel plugs were dried in a SpeedVac evaporator (Tokyo Rikakikai, Tokyo, Japan) and subjected to in-gel digestion or stored at -20uC. For digestion, gel plugs were re-swollen with 20 ng/mL trypsin (Promega, Madison, WI, USA) in 25 mM ammonium bicarbonate at 4uC for 30 min, and then at 56uC for 1 h. After digestion, peptide products were recovered by 97.5% acetonitrile and 2.5% trifluoroacetic acid (TFA).

Identification of Proteins by Matrix-assisted Laser Desorption/ionization-Time of Flight Mass Spectrometry (MALDI-TOF MS) and MALDI-TOF/TOF MS Analysis
Digested peptides were spotted directly onto a 600 mm/384 well AnchorChip TM sample target (Bruker Daltonics, Bremen, Germany), and then added with an equal volume of 1 mg/mL solution of alpha-cyano-hydroxycinnamic acid (CHCA) in 0.1% TFA/50% acetonitrile. The MALDI mass spectra were obtained using a Bruker autoflex TOF mass spectrometer equipped with a 384 sample Scout source (Bruker Daltonics  [26] for protein identification. The search criteria were taxonomy for Mammalia, fixed modification of carbamidomethyl modification, variable modifications of oxidation modification, mass accuracy of 50 to 300 ppm and maximally one missed cleavage site. Positive identification was achieved with minimum 5 peptides matched and with the set mass accuracy and modification when the score matched with significant probability to the protein or mixture of proteins in the database. The AnchorChip TM target were subjected to acquire MALDI-TOF/TOF spectra using a Bruker autoflex II TOF/TOF mass spectrometer (Bruker Daltonics) equipped with a delayed-extraction ion source. Metastable ions generated by laser-induced decomposition (LID) in the LIFT mode (Bruker Daltonics) were analyzed. The precursor ion and the corresponding fragment ions were selected in a time gate followed by further acceleration in the LIFT cell with 19.0 kV. The fragment ions were accelerated into the second field-free region and were separated in the two-stage gridless reflectron. The reflectron voltage was set at 27.4 kV. Mass spectra were processed using the FlexAnalysis 2.4 software (Bruker Daltonics). The proteins were identified by searching MS/MS spectra against NCBInr database (NCBInr 20110922 version) or SwissProt database (SwissProt 2012_04 version) using BioTools 3.0 software (Bruker Daltonics) in combination with the Mascot program [27]. The search criteria and positive identification were as described in the previous section for MALDI-TOF MS analysis.

Validation of Protein Expression by Western Blot Analysis
To further validate the differential expression of proteins among cell types, fibroblasts, f-rES, and p-rES cells were collected for Western blotting. Protein extracts were prepared by resuspending cells with 400 mL of sample buffer (0.15 mM NaCl, 5 mM EDTA, 1% Triton-100, 10 mM Tris-HCl, and 5 mM dithiothreitol) containing 0.1 mM protease inhibitor cocktail (Sigma P8340, St. Louis, MO, USA) on ice for 10 min, and then centrifuged at 18,6006g for 10 min at 4uC. The supernatant was stored at -80uC until use.
Prior to electrophoresis, the extracted protein was boiled for 5 min and then loaded on 10% (v/v) SDS-PAGE. The resolved proteins were transferred onto nitrocellulose membranes (Cat. No. HAHY00010, Millipore, Billerica, MA, USA) as described in previous study [8]. After blotting, the membrane was incubated in blocking buffer (5% chick serum in TBST) for 2 h at room temperature and then reacted with antibodies against a-tubulin

Bioinfomatic Analysis
The differentially expressed proteins among the three cell types were annotated for their subcellular distribution, biological processes, and molecular function by Gene Ontology (GO) database (http://www.geneontology.org/) and Biocompare (http://www.biocompare.com/). Briefly, unique proteins were uploaded to GO database and searched for their cellular component and molecular function. To simplify the classification, we chose the second or third level of tree browser in molecular function. The biological processes were classified by Biocompare because of their relatively simpler classification.

Statistical Analysis
The RVol of protein spots on all triplicate 2-DE gels of each cell types were analyzed using the least-squares means method in the general linear model procedure of Statistical Analysis System software [28]. The spots with significant probability (p,0.05) and with expression ratio higher than 1.5 were considered significantly different among cell types. The results of Western blot analysis were quantified by using the Image J software and the band density was normalized by using b-actin as an internal standard. The normalized data was analyzed using least-squares means method [28]. The data were presented as means 6 standard error and probability at p,0.05 was considered as significantly different among cell types.

Morphology of rES Cells and Comparison of Protein Profiles
To identify distinct protein expressions within rES cells of different origins, rabbit fibroblast, f-rES, and p-rES cells were collected and used for 2-DE analyses. Figure 2 shows the morphology of the fibroblast cells ( Fig Fig. 4. All gels showed a wide distribution of protein spots with pI ranging from 3.0 to 10.0 on 12.5% SDS-PAGE gels, and a mass ranging from 10 to 200 kDa. Of the 284 protein spots quantified among these three cell types, 100 showed distinguishable scale (p,0.05) greater than 1.5. To illustrate the differential expressions of proteins among cell types, spots with higher expression levels are circled and numbered in red and those with lower expression levels are in blue (Fig. 4). Fifty-four protein spots with similar expression levels between f-rES and p-rES cells are indicated by circles as shown in Fig. 4A. Among the 54 spots, the expression levels of 29 spots or proteins (in red) represent the proteins with expression levels that are higher in f-rES and p-rES cells than in fibroblasts ( Table 1). The other 25 spots (in blue), on the contrary, are those with lower expression levels lower in f-rES and p-rES cells. There were 14 proteins similar in expression levels between f-rES and fibroblast cells (Fig. 4B), and 23 protein spots between p-rES and fibroblast cells (Fig. 4C).

Identification of the Differentially Expressed Protein Spots
The identities of the differentially expressed protein spots among cell types were resolved by MALDI-TOF and MALDI-TOF/TOF MS. Among the 100 differentially expressed protein spots, 91 were successfully identified in the GenBank deposits which represented 63 different proteins. The detailed information on the identities of these proteins was shown in Table S1. for nuclear, 8% for endoplasmic reticulum, and 8% for mitochondrial proteins. The major biological processes in which these differentially expressed proteins participated are energy metabolisms (25%), cell growth and/or maintenance (25%), signal transduction (14%), and protein metabolisms (10%). In terms of Figure 4. Protein profiling of f-rES (A), fibroblast (B), and p-rES cells (C) grown to log phase. A total of 1,000 mg soluble proteins were TCA precipitated and separated by IEF using 180 mm IPG DryStrips at pH 3-10, followed by 12.5% SDS-PAGE gels. Proteins were visualized by colloidal Coomassie blue staining. Spots with higher expression levels are circled and numbered in red, and those with lower expression levels circled and numbered in blue. doi:10.1371/journal.pone.0067772.g004   their molecular functions, 18% of the proteins belong to functional or structural constituents of the cytoskeleton, 13% involve in structural molecule activity, 11% catalytic activity, 8% isomerase activity, 8% transporter activity, 6% heat shock proteins (HSP) or chaperones, and 17% with miscellaneous activities.

Grouping of the Differentially Expressed Proteins among Cell Types
The quantified protein spots are further divided into 7 groups based on their expression levels (Fig. 6). Group 0 (G0) contains 184 protein spots showing a similar expression level among the three cell types. Group 1 (G1) to Group 6 (G6) represent all differentially expressed protein spots ( Fig. 6 and Tables 1, 2, 3, 4). Group 1 consists of 54 protein spots with similar expression levels between f-rES and p-rES cells yet excluding those spots in G0 (Fig. 6A), of which 29 spots (Fig. 4A, red circles) had higher expression levels in both f-rES and p-rES cells than in fibroblast cells, and 25 spots (Fig. 4A, blue circles) showed lower expression levels in both f-rES and p-rES cells than in fibroblast cells (Fig. 4A and Table 1).
Group 2 represents 14 protein spots with similar expression levels between f-rES and fibroblast cells (Fig. 4B and Table 2); one spot has a higher expression level (in red) and 13 spots are with lower levels (in blue) in both f-rES and fibroblast cells than in those of p-rES cells. Group 3 contains 23 protein spots with similar expression levels in p-rES and fibroblast cells (Fig. 4C, Table 3); nine protein spots have higher expressions and 14 spots are lower expressed in p-rES and fibroblast cells than that in f-rES cells.

Validation and Confirmation of the Differentially Expressed Proteins by Western Blot Analysis and Immunocytochemistry
To corroborate the proteomic analysis, the relative expression of selected proteins differentially expressed was also compared among cell types by Western blot (Fig. 7) and immunocytochemistry (Fig. 8) analyses. Alpha-tubulin, peroxiredoxin 1, and TCP-1a protein were all expressed in the fibroblast, f-rES, and p-rES cells (Fig. 7A). Quantitative analysis showed that the a-tubulin had a higher expression in both f-rES and p-rES cells than in fibroblast cells (Fig. 7B). The a-tubulin was mostly localized in several cellular compartments of ES cells (Fig. 8). Similar to a-tubulin, TCP-1a showed a higher expression level (p,0.05) in f-rES cells than in fibroblasts by Western blotting (Fig. 7). Furthermore, TCP-1a was localized in the cytoplasm and cell cortex (Fig. 8).
These results indicate that TCP-1a is important for the development and function of cytoskeletal components of rabbit cells. Peroxiredoxin 1 was expressed in all three cell types (Fig. 7A,  B), as confirmed by the Western blot analysis and immunofluorescence staining (Fig. 8).
The differential expression of major HSPs was also observed by 2-DE. Western blot analysis confirmed that the expression levels of HSP60 and HSP90 in both f-rES and p-rES cells are higher than those in fibroblasts (p,0.05; Fig. 7). Immunofluorescent assay also confirmed that HSP60 and HSP90 are localized at the cytoplasm of rabbit cells (Fig. 8). Although immunocytochemical analysis indicated that HSP70 was localized in the cytoplasm of all three cell types (Fig. 8), result of Western blot analysis failed to confirm significant differences of this protein among f-rES, p-rES cells and fibroblasts (p.0.05).

Analysis of Imprinted Genes in rES Cells by RT-PCR
The expression of representative imprinted genes in rES cells were also determined (Fig. 9). The f-rES cells (R4) and fibroblasts expressed the paternally imprinted (Snrpn and Igf2) genes, but not in p-rES cell line (A2). This finding provided evidence for that our p-rES cell lines originated exclusively from the parthenogenetically activated oocytes without the involvement of paternal genome, and that the Snrpn and Igf2 genes are also imprinted in the rabbit.

Discussion
Rabbits are phylogenetically closer to humans than rodents and have been extensively used as model animals for the study of human physiology and diseases. In the present study, both f-rES and p-rES cells expressed the same pluripotency markers (Fig. 3), but they still differed in the expression profile of imprinted (Snrpn and Igf2) genes [4,12] (Fig. 9). While the gene and protein expressions underlying the pluripotency of f-rES and p-rES cells are largely unknown, this study investigated the protein profiles of these cell lines by a proteomics approach using rabbit fibroblast cells as the control. Among these cells, 100 out of 284 protein spots (,35%) differed in the expression levels, of which 91 protein spots representing 63 distinct proteins were identified. The proteins with known identities were mainly located in the cytoplasmic compartment and involved in energy and metabolic pathways. Some proteins were expressed exclusively in a specific cell type, indicating a specific nature or physiologic function of each cell type. For instance, at least six proteins including TUBB2A protein, KRT8 protein, a-enolase, 14-3-3 protein sigma, HSP60, and myosin-9 were expressed at significantly higher levels only in p-rES cells (Fig. 6). Two molecules (HSP60 isoform 4 and tubulin b-5 chain) and myosin light chain isoform LC17b protein showed significantly higher expression levels only in f-rES and fibroblasts, respectively.

Protein Profiles of p-rES Cells and Identification of the Differentially Expressed Proteins
Parthenote-derived ES cells are generated from embryos consisting of only the female genome. It has been reported that the absence of paternal alleles in parthenotes resulted in the lack of paternally imprinted gene expressions, which may have restricted the development of parthenotes [29][30]. We have already confirmed that there were no paternally imprinted genes such as Snrpn and Igf2 expressed in the p-rES cells [12]. In this study, 2-DE analysis showed that both upregulated and downregulated proteins were observed in p-rES cells. These proteins may be associated with the expression of maternally imprinted genes. Upregulated proteins of p-rES cells are presented in two groups (Group 2, Table 2; Group 5, Table 4). Group 2 consists of actin, Sadenosylhomocysteine hydrolase, carbonic anhydrase 2, VCP protein, ?-actin, cytoplasmic b-actin isoform 2, annexin A1, mitochondrial F1-ATPase, and ribosomal protein S2 ( Fig. 4B and Table 2). Group 5 consists of TUBB2A protein, KRT8 protein, aenolase, 14-3-3 protein Sigma, HSP60, and myosin-9 ( Fig 6A, 6D and Table 4). On the other hand, two downregulated proteins of p-rES cells were stress-induced-phosphoprotein 1 (STI1) at Group 2 (spot No. 219; Fig. 4B and Table 2) and heat shock protein 60 isoform 4 (Group 4, spot No. 116; Fig. 6D and Table 4). These proteins are likely, at least in part, associated with the expression of the maternally imprinted genes in p-rES cells, which require more studies to identify their physiological functions.

Bioinformatic Analysis of the Differentially Expressed Proteins
The 63 differentially expressed proteins among the three cell types were subjected to GO annotation. Most of these proteins involve in the biological processes of energy pathways, metabolisms and cell growth, and/or maintenance (Fig. 5). Proteins involved in energy pathways and metabolic processes partake in catalytic and antioxidant activities, such as a-enolase, glutathione gyclophilin 18, glyceraldehyde 3-phosphate dehydrogenase, and pyruvate kinase. Some differentially expressed proteins, including tubulin b-5 chain, TUBB2A, KRT8, myosin-9, and myosin light chain isoform LC17b, associated with the cytoskeletal infrastructure are involved in cell growth or maintenance, structural constituent of the cytoskeleton, and their functional molecule activity [31][32][33][34]. Fourteen percent of the differentially expressed proteins are related to signal transduction (Fig. 5) which may potentially play roles in maintaining the stemness and differentiation capacity of rES cells.

Putative Roles of Differentially Expressed Proteins in rES Cells
Representative proteins highly expressed in rES cells were selected for further validation by Western blot and immunocytochemical analyses. The expression patterns of TCP-1a, a-tubulin, HSP60, and HSP90 were the most prominent proteins in f-rES and p-rES cells revealed by both Western blot and immunocytochemical assays (Figs. 7 and 8). Of these, cytoskeletal proteins (e.g. actins and tubulins) are among the most abundantly expressed proteins in eukaryotic cells [35]. Tubulins are the major components of the filamentous structure of cellular microtubules with a-tubulin being the most common one [36]. The microtubule plays many crucial roles in intracellular transport, cell morphology, polarity, signaling, and division of the cell [37], which also Figure 7. Western blot analysis of total protein extracts from fibroblast, f-rES and p-rES cells. (A) The proteins of interest were analyzed using 50 mg of protein extracts from fibroblast, f-rES and p-rES cells for SDS-PAGE followed by Western blotting. Membranes were hybridized with antibodies against a-tubulin, peroxiredoxin 1, TCP-1a, HSP60, HSP70, and HSP90, using b-actin as a loading control. The protein bands in the blots were quantified by using an Image J software, and the quantitative data is shown in (B). Three replicates were performed in each cell type, and bars without the same alphabetic letters differ (p,0.05). doi:10.1371/journal.pone.0067772.g007 make it a target for the study of cancer therapy [35][36][37][38]. In this study, we found that a-tubulin or tubulin-b was upregulated in both f-rES and p-rES cells (Figs. 6,7,8), strongly suggesting that ES cells are one of the actively proliferating cell types compared to the terminally differentiated fibroblasts. Moreover, previous studies have also shown that a-tubulins in mES cells are downregulated along with vimentin, one of the intermediate filaments, during differentiation into neuronal cell lineages [16,39].
The TCP-1 complex is an oligomeric particle found in the eukaryotic cytosol consisting of four or five related polypeptides of a similar size (55-60 kDa) [40]. In vitro studies suggested that TCP-1 complex is a chaperonin in the eukaryotic cytosol participating in the correct folding of newly translated aand b-tubulins and refolding of urea-denatured tubulins and actins in rabbit reticulocyte lysates [41][42][43][44]. It is also functionally linked to cell growth and its expression decreases concomitantly with the growth arrest during differentiation [45]. Most interestingly, it has been reported that TCP-1 is related to the growth and survival during pig embryo development, and it is more drastically upregulated in pig parthenogenetic embryos than in fertilized embryos [46]. In this study, TCP-1a was found expressed in all the three cell types ( Fig. 8) with higher expression levels in rES cells detected by 2-DE (Table 1), particularly highly expressed in f-rES cells detected by Western blotting (Fig. 7). Although the exact cause for the slightly inconsistency between the two analyses is not clear, we infer that TCP-1a may play active roles in cell proliferation and/or cytoskeletal protein folding at least in f-rES cells. Further study is required to determine the precise role of TCP-1a in maintaining the stemness and undifferentiation of rES cells.
Peroxiredoxins are a family of small (22-27 kDa) nonseleno peroxidases in mammals with six isoforms (Peroxiredoxin 1-6) widely distributed in human cells including reproductive organs [47][48]. They function to serve as reactive oxygen species (ROS) detoxifiers in order to provide cytoprotection from internal and external environmental stresses by eliminating hydrogen peroxide from cells [49][50][51]. Peroxiredoxins 1 and 2 were highly expressed in ovary and testis [47,52]. In the female, peroredoxin 1 gene expresses in 3-day-old follicles and increases its expression in 21day-old during folliculogenesis in the rat [53]. In addition to being found in human endometrium and cervix-vagina fluid [54] and in the epithelium and the endometrial stroma of the uterus, peroredoxins 2 has been reported to be present in the mouse ooplasm of primary follicles, secondary follicles or premature follicles [55]. It is possible that peroxiredoxins also function similarly in rabbits. The results of this study showed that peroxiredoxins 1 and 2 were upregulated in rES cells (Table 1; spot No. 53, 209, and 250). The high peroxiredoxin levels in rabbit ES cells may be associated with their active proliferation capacity and at least partially linked to their future differentiation competency. However, Western blot analysis and immunocytochemistry failed to confirm a similar upregulation in rES cells. Reasons for this discrepancy are unclear, but it could be due to the stressful in vitro culture conditions, which might have leveled up its expressions in all three cell types, or the co-existence of other isoforms detected in this study which warrant further investigation.
Heat shock proteins are ubiquitously expressed, and are transcriptionally regulated [56] under physiological and stressful conditions such as elevated temperatures, oxygen tension and chemical insults [57][58][59]. The best characterized roles of HSPs are the involvement of chaperone-mediated protein folding [60]. In the mouse, downregulation of HSPs and the co-chaperones (HSP70 and HSP60) in mES cell lines upon differentiation was observed [61][62][63]. Proteomic analysis of this study showed that two protein spots (spot No. 240) of HSP60 were highly expressed in rES rather than in fibroblast cells (Table 4). Western blot and immunocytochemical analyses both confirmed the 2-DE results (Figs. 7B and 8). HSP60 has been known to bind directly with the Oct4 and Nanog genes which directly regulate Oct4 and other stemness genes involving the differentiation of adipose tissuederived stem cells (hATSC) in humans [64][65]. The finding in this study confirmed the observations in previous study that undifferentiated cells expressed more HSPs which might be attributable to their active protein synthesis or maintaining pluripotency-related cellular activities, compared to the terminally differentiated cells in the relatively quiescent state.
The expression of HSP90 was only upregulated in f-rES cell lines and showed similar lower levels in fibroblasts and p-rES cells based on our 2-DE analyses (Table 3). It has been reported that the chaperoning activity of HSP90 depends on its ability to hydrolyze ATP and its potential to form stable complexes with HSP70 and HSP70/HSP90/organizing protein (HOP) in fertilized mouse embryo-derived ES cells [64][65]. The HOP is a 60 kDa co-chaperone that binds and regulates the activity of chaperones [66,67]. Using RNAi to knockdown HOP in mES cell lines caused 68% depletion of STAT3 mRNAs, downregulated soluble phosphotyrosine-STAT3 levels, and leading to an extranuclear accumulation of STAT3, which ultimately reduced Nanog mRNA levels and lost the ability to form embryoid bodies [68]. These studies confirmed the previous work showing that HSP90 interacted with the JAK/STAT3 signaling molecules in somatic cells [69]. The HSP90 was reported to complex with STAT3 in human embryonic kidney carcinoma cells [70]. Western blot analysis did not completely confirm the differential expression of HSP90 observed in 2-DE analysis in different cell types (Fig. 7). The discrepancy was unclear. In general, HSP90 was upregulated in both f-rES cells and p-rES cells, but remained low in fibroblasts. It might be due to the passage number of rES cells used for analysis, or the commercially available HSP90 antibodies recognized different isoforms of HSP90. However, it has been reported that HSP90 was upregulated in both f-rES and p-rES cells, where the involvement of the proteostatic maintenance of onco-proteins [71] or stemness [64] were demonstrated. Recently, it was also reported that LIF promotes the interaction of HSP90 with STAT3 for maintenance of self-renewal in mES cells [64]. These works provide strong supportive evidence to our previous [8,12] and current proteomics findings that LIF addition was essential for maintaining self-renewal and the stemness of both f-rES and p-rES cells.
Based on the results from this study and the related reports, we propose an integrative model for the presumptive roles played by the representative molecules that distinguish these three rabbit cell types (Fig. 10). In this model, the a-tubulin and TCP-1a proteins may play an active role in cell proliferation and/or as regulative proteins for the cytoskeleton in rES cells [41][42][43][44]. Peroxiredoxin 1 functions to serve as a reactive oxygen species (ROS) scavenger or detoxifier to protect cytoplasm from internal and external environmental stresses [49][50][51]. It has been suggested that HSP60 is associated with the stemness of ES cells, in which it may bind some pluripotency proteins (e.g. Oct4 and Nanog) or genes (c-myc and Stat3) that are related to stress tolerance and/or maintenance of ES cell pluripotency [64][65]. In addition, HSP90 hydrolyzes ATP and forms a stable complex with HSP70 and HOP which binds and regulates the activity of chaperones related to the LIF/STAT signaling pathway [64]. In this context, some other key molecules, such as peroxiredoxin 2, cytoplasmic linker 2 and cofilin-1, which are all expressed in rabbit ES cells, may have actively taken part in maintaining normal cellular functions of rES cells.  Figure 10. An integrative diagram for presumptive cellular roles played by the representative molecules that distinguish the three rabbit cell types. (1) In the present study, proteins identified are mainly involved in biological functions of energy and metabolic pathways, cell growth and maintenance, signal transduction and protein metabolism. We hypothesize that the selected six proteins are associated with structural constituents of the cytoskeleton, catalytic activity, molecular chaperones, and heat shock proteins (HSP60, HSP70, and HSP90) which were all found expressed in many cell types. (2) Alpha-tubulin, and TCP-1a were expressed with higher expression levels in both rES than in fibroblasts. It is likely that these proteins may play an active role in cell proliferation and/or as regulative proteins for the cytoskeleton rES cells. (3) Peroxiredoxin 1 functions to serve as reactive oxygen species (ROS) detoxifiers to protect cytoplasm from internal and external environmental stresses. (4) HSP60 may bind some pluripotency proteins (e.g. Oct4 and Nanog) or genes (c-myc and Stat3) that are associated with stress tolerance and/or maintenance of stemness in ES cells. (5) On the other hand, HSP90 hydrolyzes ATP and forms a stable complex with HSP70 and HSP70/HSP90/organizing protein (HOP), which is a 60 kDa co-chaperone that binds and regulates the activity of chaperones related to the LIF/STAT signaling pathway. It has also been reported that HOP down-regulates soluble phosphotyrosine, leading to an extranuclear accumulation of STAT3 in LIF/STAT pathway, which ultimately reduced Nanog mRNA levels and lost the ability of ES cells to form embryoid bodies. (6) Some other molecules, such as peroxiredoxin 2, cytoplasmic linker 2 and cofilin-1, are actively expressed in rabbit ES cells. Peroxiredoxin 2 may play an antioxidant protective role in ES cells. Cytoplasmic linker 2 has been reported as a microtubule plus-end tracking protein that maintains microtubule dynamics by nucleation of noncentrosomal microtubules originated from the trans-Golgi network (TGN). Cofilin-1 is a membrane receptor for estrogen to regulate the mobility of ES cells. The corresponding groups (Table 1,

Conclusion
To the best of our knowledge, this study demonstrated, for the first time, that the differential proteomic profiles among rabbit ES cells and somatic fibroblasts. The results showed that the characterized proteins for cellular components, biological processes, and molecular functions are those mostly involved in the cytoplasmic functions, energy metabolisms, and structural constituents of the cytoskeleton. The results also provided substantial evidence for the fundamental differences between rabbit somatic cells and ES cells. Furthermore, some key proteins active in rabbit ES cells could be useful cues to re-design the derivation and culture systems for generation of the ground state or germline competent rabbit ES cell lines.

Supporting Information
Table S1 Protein identities of the differentially expressed proteins in fibroblast, f-rES and p-rES cells. & Spot numbers are the numbers labeled in Fig. 4. # Numbers in the column are the results from the MALDI-MS PMF analysis, i.e. the number of assigned peptides and percent sequence-coverage (in brackets); ns: no significant match. na: not analyzed. (DOC)