Human embryonic stem (hES) cells are capable of unlimited cell proliferation yet maintain the potential to differentiate into many cell types. Here we reported an Epstein-Barr virus (EBV)-based vector system used to improve transfection efficiency in hES cells. Plasmids containing oriP, the latent replication origin of EBV, can be propagated stably as episomal DNA in human cells that express the EBV nuclear antigen 1 (EBNA1), which binds to oriP and functions as the trans-acting replication initiator. It was reported that the EBV replicon could harbor a DNA fragment of up to 330 kilobase pairs. Plasmids containing an enhanced green fluorescent protein (EGFP)/puromycin resistance gene cassette along with or without oriP were used to transfect hES cells that stably express EBNA1. The presence of oriP moderately increased the transient transfection efficiency and more importantly it elevated the stable transfection efficiency by approximately 1,000-fold as compared with oriP-minus plasmids. The oriP plasmid as episomal DNA and green fluorescent protein expression in hES cells was maintained for months in the presence of drug selection and gradually lost (2%–4% per cell doubling) in the absence of selection. The presence of EBNA1 did not interfere with the hES cell properties or differentiation we tested and could maintain stable EGFP expression during differentiation. In addition to transgene expression, the EBV vector system could effectively enhance the RNA interference efficiency in hES cells. Thus, the EBV vector system that allows a large DNA insert and sustained expression of transgene or small hairpin RNA will enhance basic and translational research using hES cells.
Human embryonic stem (hES) cells are derived from the inner cell mass of in vitro fertilized blastocyst-stage early human embryos [1, 2] and are able to differentiate into all body tissues, making these cells important tools for developmental studies, drug discovery, and a promising source of cells for tissue replacement therapies. The development of hES cells provides unprecedented opportunities for biomedical research as well as regenerative medicine .
Efficient methods for transgene expression in hES cells are essential for studying functions of various genes in hES cells and for elucidating basic mechanisms that control hES cell self-renewal and differentiation. Several methods for the genetic modification of hES cells have now been reported, each with different advantages and disadvantages. It has been reported that a plasmid-mediated transfection method achieved transient and stable transfection frequencies of hES cells: no more than 10% and approximately 1/105 cells, respectively [4, 5]. These levels of efficiency are adequate for some applications but insufficient for specific experiments that require higher stable transfection efficiencies in hES cells that grow much poorer and slower at a clonal level than mouse ES cells. For example, when transgene expression causes hES cell differentiation, it is not possible to recover stable hES cell clones unless some conditional gene expression system is employed. However, if stable transfectants can be obtained at a high frequency, a useful number of transfectants can be obtained for analysis, even if the transgene induces differentiation . For these reasons, some researchers have previously explored the use of lentiviral vectors in hES cells [7–9]. Although lentiviral vectors allow relatively high transduction efficiencies of hES cells, they are fairly labor-intensive to use, limited by the packaging size, prone to disturbing transcriptional regulation of cellular genes by chromosomal integration of the proviral DNA, and variable in transgene expression levels due to uncontrollable variation in positions of the integration.
Episomal vectors may be more resistant to gene silencing or insertional mutagenesis because they should not be subjected to integration positional effects. Epstein-Barr virus (EBV)-based episomal expression vectors have been used successfully in vitro and in vivo [10–14]. The ability to introduce and maintain very large human genomic DNA fragments (>100 kb) as episomes in human cells is one of the significant advantages of these vectors [15–19]. As a multicopy, circular episome, EBV is maintained extrachromosomally in latently infected B lymphocytes naturally, replicates just once each cell cycle, and is partitioned faithfully to daughter cells, akin to cellular chromosomes [20–22]. Synthesis and maintenance of the EBV vectors require both a cis-acting component, the latent origin oriP, and a trans-acting component, the EBV nuclear antigen 1 (EBNA1) [23–25]. Through the binding to oriP, EBNA1 facilitates the retention and replication of the oriP-containing plasmid, ensuring its long-term persistence in human cells . In addition, EBNA1 also increases transcription activities from promoters near oriP and enhances nuclear uptake of the oriP plasmid . Both the oriP and the EBNA1 gene can be on the same vector, but the stability of the vector can be improved if the EBNA1 gene integrates into the cellular genome, followed by introduction of an oriP-containing plasmid .
Here we first demonstrate uses of an EBV-based expression system to achieve stable gene expression in EBNA1-expressing hES cell lines that harbor an oriP-containing plasmid expressing enhanced green fluorescent protein (EGFP). In EBNA1-expressing hES cells, the presence of oriP moderately increased the transient transfection efficiency. More importantly, it elevated the stable transfection efficiency by approximately 1,000-fold as compared with plasmids without oriP. The presence of EBNA1 expression did not interfere with the differentiation of hES cells and maintained stable EGFP expression during differentiation. Moreover, we found that the EBV vector system also effectively enhanced the RNA interference (RNAi) efficiency in hES cells by using an oriP-containing plasmid expressing small hairpin RNA (shRNA).
Materials and Methods
H1 and H9 hES cells (WiCell Research Institute, Madison, WI, http://www.wicell.org) were maintained as undifferentiated cells on inactivated mouse embryonic fibroblast (MEF) feeder cells in Dulbecco's modified Eagle's medium (DMEM)-F12 supplemented with 20% knockout serum replacement (Knockout SR; Gibco, Grand Island, NY, http://www.invitrogen.com), 1 mM l-glutamine (Gibco), 0.1 mM nonessential amino acids (Gibco), 4 ng/ml basic fibroblast growth factor (bFGF) (Gibco), 0.1 mM β-mercaptoethanol (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) as described previously  or on growth factor reduced Matrigel basement membrane matrix (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com) in MEF-conditioned medium (CM) supplemented with 4 ng/ml bFGF . Cell morphology was recorded using an inverted microscope (TE2000U; Nikon, Tokyo, http://www.nikon.com).
Plasmid Construction and Transfections
pCMV-EGFP-oriP was provided by Dr. Bill Sugden (University of Wisconsin, Madison, WI). It contains an oriP element and a cytomegalovirus (CMV) promoter-driven EGFP followed by an encephalomyocarditis virus-derived internal ribosome entry site (IRES)-puromycin resistance gene (puroR) cassette. The sequence of pCMV-EGFP-oriP is accessible at http://mcardle.oncology.wisc.edu/sugden/. To replace the CMV promoter with other promoters that are more active in hES cells, we made the pEF1-EGFP-oriP plasmid using the human elongation factor 1α (EF1α) promoter. It was excised from pBudCE4.1 (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) to replace the CMV promoter in pCMV-EGFP-oriP. A control (oriP-minus) plasmid, pEF1-EGFP, was created by religating the pEF1-EGFP-oriP after the oriP element was deleted by removing the NdeI fragments. The pEB plasmid that contains a phosphoglucokinase promoter, EBNA1 cDNA, IRES, and neomycin resistance gene (neoR) was a gift of Dr. D. W. Burt (Roslin Institute, Edinburgh, U.K.). For cloning pshEGFP, a vector expressing shRNA targeting against EGFP, an oligonucleotide encoding a stem-loop structure targeting EGFP with the targeting sequence CCACTACCAGCAGAACACC was designed using Oligo-Engine workstation software (OligoEngine Inc., Seattle, http://www.oligoengine.com) and then was subcloned into the pSUPER.retro vector (OligoEngine) under the control of the human H1 promoter between the HindIII and BglII sites. Subsequently, pshEGFP was reconstructed into pshEGFP.oriP1 or pshEGF-P.oriP2 by inserting the oriP element (with two opposite directions) into the XbaI and XhoI sites of pshEGFP. pMD18-T-oriP, a control vector containing oriP, was also constructed by inserting an oriP element amplified by polymerase chain reaction (PCR) into pMD18-T vector (TaKaRa Biotechnology Co., Dalian, China, http://www.takara.com). Key regions in all constructs were verified by DNA sequencing.
Transfections were performed using FuGENE 6 transfection reagent (Roche Diagnostics, Basel, Switzerland, http://www.roche-diagnostics.com), as described by the manufacturer's protocol. Stable EBNA1-expressing hES cell lines were established by transfecting hES cells with the pEB vector and selecting for the growth of neoR cells with G418 (100 μg/ml). Then EBNA1 hES cell clones or untransfected hES cells at the same density were transiently or stably transfected with 2 μg of pEF1-EGFP-oriP or the control pEF1-EGFP plasmid. Transfected cells were observed by microscopy followed by flow cytometric analysis at 48 hours post-transfection or further selected for stable transfectants with puromycin (1 μg/ml). After 7 days of puromycin selection, the surviving colonies were recorded by the Fujifilm imaging system (FLA-5100; Fuji Photo Film Co., Ltd., Tokyo, http://home.fujifilm.com) with or without crystal violet staining for live cells. The EBNA1-expressing hES cells that were also successfully and stably transfected by the pEF1-EGFP-oriP plasmid were named as EBEG cell lines, denoting the presence of both EBNA1 (EB) and the EF1-EGFP (EG) cassette. For RNAi, ∼1.0 × 105 EBNA1 hES cells were co-transfected transiently with 8 μg of pshEGFP.oriP, pshEGFP, or pMD18-T-oriP, together with 2 μg of the pEF1-EGFP-oriP vector. Forty-eight hours after transfection, the hES cells were counter-stained for DNA with Hoechst 33342 and observed under an inverted fluorescence microscope to examine the RNAi-mediated EGFP knockdown in EBNA1 hES cells.
EBNA1 and EGFP Gene Expression Analysis
Total RNA of G418-resistant hES cells was isolated with Trizol Reagent (Gibco) followed by treatment with DNase I (Promega, Madison, WI, http://www.promega.com) at 37°C for 1 hour to remove the residual genomic DNA, and then 4 μl of digested RNA was used for the reverse transcription (RT) reactions using the Reverse Transcription System (Promega). One-tenth of the RT mixture was placed in a total volume of 50 μl containing 1.5 mM MgCl2, 1× PCR buffer, 0.2 mM deoxyribonucleotide triphosphates (dNTPs), 10 pM each of the two EBNA1 primers or glyceraldehyde-3-phosphate dehydrogenase primers , and 1 unit of Taq DNA polymerase. The PCR parameters were as follows: 94°C for 5 minutes; 35 cycles through 94°C for 45 seconds, 60°C for 45 seconds, and 72°C for 45 seconds; and then extension at 72°C for 10 minutes. One-fifth of the PCR mixture was electrophoresed on 1.5% agarose gel.
For Western blots, total isolated protein from EBNA1 cells, EBEG cells or teratoma tissues was fractionated on an 8% (for EBNA1) or 10% (for green fluorescent protein) polyacrylamide gel and transferred to a nitrocellulose membrane, blocked with 0.5% nonfat dry milk in TBST (1× Tris-buffered saline, 0.05% Tween-20) overnight. The primary antibody used in this assay was a 1:800 dilution of monoclonal anti-EBNA1 (ABI, Columbia, MD, http://www.abionline.com) or 1:4,000 dilution of rabbit anti-GFP polyclonal antibody (Sigma-Aldrich). The secondary antibody was a 1:1,000 dilution of peroxidase-labeled anti-mouse antibody or anti-rabbit antibody (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com). The blot images were developed in SuperSignal West Pico Stable Peroxide Solution and SuperSignal West Pico Luminol/Enhancer Solution (Pierce, Rockford, IL, http://www.piercenet.com) and captured with Fujifilm (FLA-5100; Fuji Photo Film Co., Ltd.).
For further detection of EBNA1 expression in EBNA1 cells and EBEG cells, immunofluorescence analysis was carried out. EBNA1 cells and EBEG cells were stained with the primary antibody against EBNA1 (Abcam, Cambridge, U.K., http://www.abcam.com) at 1:100 dilution. Cy3-conjugated anti-mouse IgG1 antibody (Sigma-Aldrich) was used as the secondary antibody at 1:100 dilution. For observing the red fluorescence of EBNA1 and the green fluorescence of EGFP in EBNA1 cells and EBEG cells, fluorescence microscopy was used.
Flow Cytometric Analysis
For flow cytometric analysis, the stable EBEG cells under G418 selection were harvested at different time points after puromycin withdrawal. Nontransfected EBNA1 hES cells and EBEG cells were washed with phosphate-buffered saline (PBS) and carefully trypsinized to yield single cell suspensions, followed by several washes in PBS. The single cell suspension was used for detecting EGFP by flow cytometric analysis and cells were analyzed live (without fixation) with propidium iodide to exclude dead cells on a FACScan (Becton, Dickinson and Company) with CELLQUEST software.
Identification of Stem Cell Markers on EBNA1 hES Cells
EBNA1 hES cells were cultured for 5 days prior to analyzing alkaline phosphatase activity. On the 5th day, EBNA1 hES cells were fixed with a fixative (90% methanol-10% formaldehyde) for 1–2 minutes and rinsed with 1× rinse buffer (20 mM Tris-HCl, pH 7.4, 0.15 M NaCl, 0.05% Tween-20). Naphthol-fast red violet (FRV) solution was prepared by mixing FRV with Naphthol AS-BI phosphate solution and water in a 2:1:1 ratio for alkaline phosphatase staining; enough staining solution was added to cover each well; cells were incubated in dark at room temperature for 15 minutes. After washing, the stained cells were observed under an inverted microscope. The cell surface antigen expression of EBNA1 hES cells could be analyzed by using immunofluorescence techniques. Briefly, cells were fixed in 4% paraformaldehyde-PBS for 15–20 minutes at room temperature, rinsed twice (5–10 minutes each) with 1× rinse buffer, permeabilized in 0.1% Triton X-100-PBS for 10 minutes at room temperature, and incubated in a blocking solution (4% normal goat serum-PBS) for 30 minutes at room temperature. Then, primary antibodies against SSEA-3, SSEA-4, TRA-1-60, and TRA-1-80 (Chemicon, Temecula, CA, http://www.chemicon.com) were added for 1 hour at room temperature, and secondary antibodies (fluorescein isothiocyanate [FITC]-labeled goat anti-mouse IgM or IgG) were added for 30–60 minutes at room temperature (Chemicon). Finally, the stained cells were photographed under an inverted fluorescence microscope. Moreover, we detected OCT4 and Nanog expression in EBNA1 hES cells by reverse transcription-polymerase chain reaction (RT-PCR). The PCR parameters were as follows: 94°C for 5 minutes; 35 cycles through 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 30 seconds; and then extension at 72°C for 10 minutes.
OCT4 primers: forward, 5′-TGGGGGTTCTATTTGGGAAGG-3′; reverse, 5′-GTTCGCTTTCTCTTTCGGGC-3′.
Nanog primers: forward, 5′-TGCCTCACACGAGACTGTC-3′; reverse, 5′-TGCTATTCTTCGGCCAGTTG-3′.
Telomerase activity of EBNA1 hES cells was assayed by TRAPEZE Telomerase Detection Kit (S7700; Chemicon).
Embryoid Body (EB) Formation
EBNA1 hES cells were dissociated into small clumps by 1 mg/ml collagenase IV (Gibco) and cultured in suspension as EBs in differentiation medium on gelatin-coated plates for 7 days. Differentiation medium contained 80% DMEM-F12, 20% defined FBS, 1 mM l-glutamine, 1% nonessential amino acids, and 0.1 mM β-mercaptoethanol.
Nontransfected hES cells, EBNA1-expressing hES cells, or EBEG cells (∼5 × 106) from approximately 60% confluent six-well plates were injected into the rear leg muscles of 4-week-old male severe combined immunodeficient-beige mice (Shanghai Slac Laboratory Animal Co., Ltd., Shanghai, China). Eight to 10 weeks after injection, the resulting teratomas were examined histologically by sectioning followed by H&E staining. The animal care and use protocol was approved by the institutional animal care and use committee of Central South University.
Isolation of Episomal DNA
The EBEG cells under puromycin selection for at least 12 weeks were pelleted, washed, and lysed in 0.6% SDS-1 M NaCl and then stored at 4°C for 24–48 hours. High-molecular-weight DNA was removed by centrifugation at 17,000g for 45 minutes at 4°C . The supernatant containing the low-molecular-weight DNA was digested with RNase A (100 μg/ml) at 37°C for 1 hour followed by digestion with proteinase K (200 μg/ml) for 1 hour at 50°C. Then, the free episomal DNA was recovered with QIAprep spin columns (Qiagen, Hilden, Germany, http://www1.qiagen.com) and used to transform the HB101 Escherichia coli cells with Gene Pulser (Bio-Rad, Hercules, CA, http://www.bio-rad.com) using standard electroporation protocols. Twelve independent ampicillin-resistant E. coli colonies were randomly picked and expanded in 3 ml of LB-Amp, and minipreps were prepared. Plasmid DNA from each of colonies was cut by KpnI and run on 1% agarose gel to test for vector stability.
Quantitative PCR Analysis
To quantify the copy number of pEF1-EGFP-oriP in the EBEG cells, the LightCycler PCR and Detection System (Roche) with SYBR Premix Ex Taq Kit (TaKaRa Biotechnology Co.) was used for amplification and quantification of oriP DNA. The episomal plasmids (together with the genomic DNA) from EBEG cells were extracted with the Wizard Genomic DNA Purification Kit (Promega) for PCR detection. PCRs were run in triplicate with 20-μl reaction volumes containing 2×SYBR Premix Ex Taq, 0.2 μM each of primers (forward, 5′-GGGTTCAGTGGTGGCATT-3′; reverse, 5′-ATCCAGTCTTTACGGCTTGT-3′), 200 ng of genomic DNA. PCR parameters were as follows: 95°C for 10 seconds; 40 cycles through 95°C for 5 seconds, 55°C for 10 seconds, and 72°C for 10 seconds. Melting curve analysis (from 65°C to 95°C, followed by cooling to 40°C) was also performed to exclude nonspecific PCR products. pEF1-EGFP-oriP was used as a template in standard curve analysis at concentrations of 5 × 10−1 ng/μl to 5 × 10−6 ng/μl, corresponding to 5.26 × 107 to 5.26 × 102 copies of pEF1-EGFP-oriP. Double-stranded PCR product (144 base pairs) was quantified by monitoring fluorescence of the DNA-binding SYBR Green I dye.
Induction of hES Cell Differentiation by Bone Morphogenetic Protein 4
EBNA1 hES cells or EBEG cells were cultured in CM supplemented with 4 ng/ml bFGF on Matrigel-coated plates. Bone morphogenetic protein 4 (BMP4) (100 ng/ml) (R&D Systems Inc., Minneapolis, http://www.rndsystems.com) was added to the culture according to Xu et al. . Cell morphology was photographed by inverted microscope. For syncytium formation and analysis, EBNA1 hES cells or EBEG cells were individualized with trypsin-EDTA (Gibco) for 15 minutes at 37°C and plated at low density subject to treatment daily with BMP4. Some of the individualized BMP4-treated cells formed syncytial cells within 2 weeks of the treatment. These cells were treated with the Golgi blocker brefeldin A (Sigma-Aldrich) at 1.25 μg/ml for 4 hours at 37°C, fixed with 2% paraformaldehyde for 10 minutes, and immunostained with a monoclonal mouse anti-human chorionic gonadotropin (CG)-β antibody or EBNA1 antibody (Abcam) at 1:100 and Cy3-labeled or FITC-labeled anti-mouse IgG1 antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com) at 1:100. The cells were finally stained for the nuclei with Hoechst 33342 (Sigma-Aldrich) and photographed by phase contrast and fluorescence microscopy.
Stable Expression of EBNA1 in hES Cell Clones Increases the Transient and Stable Transfection Efficiencies of oriP-Containing Plasmids
With H9 hES cells, we compared several commercial reagents, including FuGENE 6, Lipofectamine 2000, and Exogen 500, as well as the CaPO4 precipitation method to transfect hES cells using the EGFP reporter gene. We found that FuGENE 6 had some advantages in transfection of hES cells over the other transfection methods, resulting in lower toxicity to hES cells, higher efficiency for introducing foreign genes into hES cells , convenience of use, and short handling time (data not shown). Subsequently, we compared the activity of different promoters in hES cells using the EGFP reporter gene and found that the regulatory function of the EF1α promoter was obviously stronger than that of the CMV promoter in undifferentiated hES cells (data not shown), so we used the EF1α promoter in the following experiments to drive EGFP reporter expression.
H1 and H9 hES cells grown on Matrigel in CM were transfected with the pEB plasmid and were selected for stable EBNA1 and neoR expression using G418. Individual colonies were picked and expanded in medium containing G418 for 8 weeks. Total RNA of the clones was then isolated and analyzed for EBNA1 mRNA expression. The expected EBNA1 PCR product (244 base pairs) was found in all the selected clones (Fig. 1A). To confirm the EBNA1 protein expression, Western blots and immunostaining were performed. The expected 87-KDa wild-type EBNA1 protein was detected by Western blots (Fig. 1B). In addition, in situ staining revealed that the EBNA1 protein could be found in EBNA1 hES cells (Fig. 1C, f).
We next examined the biological characteristics of EBNA1-expressing H1 and H9 hES cells. We found that stable transfectants of EBNA1-expressing hES cells were similar to the parental hES cells in hES cell phenotype (SSEA-3+, SSEA-4+, TRA-1-61+, TRA-1-80+, AKP+, OCT4+, Nanog+, telomerase+), the ability to differentiate into EB-like spheres in vitro and form teratoma tissues in vivo (Fig. 1C–1G), and cell cycle and growth properties (data not shown). In the resulting teratomas from EBNA1 hES cells, tissues from three embryonic germ layers that were typically represented at this time point in untransfected hES cells-formed teratomas, such as the neural epithelium, retina-like epithelium, cartilage, muscle, bone, and pseudostratified ciliated columnar epithelium, could be found histologically (Fig. 1G). Furthermore, the expression of EBNA1 did not interfere with the differentiation of EBNA1 hES cells induced by BMP4 (Fig. 1H).
Multiple individual EBNA1 or parental hES cell clones were transfected with pEF1-EGFP-oriP or pEF1-EGFP (lacking oriP) vectors. EGFP expression in hES cells was detected by fluorescence microscopy and flow cytometric analysis. If not using EBV-based vector system, FuGENE 6-mediated transient transfection efficiency was just 3%–8%. However, the transient transfection efficiency could be enhanced up to 30%–40% with the help of EBV-based vector system (data not shown), similar to the report of Tomiyasu et al. . To obtain stable transfection by the EGFP expression vectors, we selected stable transfectants based on the co-expression with the puroR gene. On the 7th day after puromycin selection, there appeared numerous puroR colonies in the wells where EBNA1 hES cells were transfected with the pEF1-EGFP-oriP plasmid, but just a few puroR colonies in the wells where EBNA1-expressing hES cells were transfected with the pEF1-EGFP plasmid or where un-transfected hES cells were transfected with the pEF1-EGFP-oriP or pEF1-EGFP plasmid (some representative data are shown in Fig. 2A, 2B). By flow cytometric analysis, pEF1-EGFP-oriP plasmid resulted in stable transfection efficiency approximately 1,000-fold higher in EBNA1 hES cells than that by transfection with pEF1-EGFP plasmid in EBNA1 hES cells or that by transfection with pEF1-EGFP-oriP plasmid in un-transfected hES cells (data not shown). Fluorescence microscopy and Fuji imaging system confirmed the stable EGFP expression from the pEF1-EGFP-oriP vector. The transfection efficiency detected by crystal violet staining on H1 hES cells was consistent with H9 hES cells (Fig. 2B). Our data demonstrated that only the EBNA1 hES cells supported replication of oriP plasmids and the presence of EBNA1 could significantly increase the stable transfection efficiency of oriP plasmids in hES cells, as previously reported in other cell types [28, 33].
The results of Western blots confirmed the expression of EBNA1 and EGFP in teratomas derived from EBEG cells (Fig. 2C), demonstrating that the expression of EBNA1 and EGFP did not interfere with the basic biological characteristics or differentiation of hES cells and the transgene expression remained after differentiation of hES cells (see more below).
The Presence of EBNA1 Did Not Interfere with the Differentiation of hES Cells Induced by BMP4 and Could Maintain Stable EGFP Expression During Differentiation
To examine stable transgene expression in EBNA1-expressing hES cells, as well as potential adverse effects of EBNA1 expression in hES cell differentiation, we first demonstrated BMP4-induced differentiation from adherent EBNA1 cells or EBEG cells. As previously reported , morphological changes became obvious on day 2 (Fig. 3D). A synchronous wave of differentiation occurred, characterized by flattened, enlarged cells with reduced proliferation. Moreover, we plated the EBNA1 cells and EBEG cells as single cells at low density and treated them with BMP4. Syncytial cells appeared after 2 weeks of treatment (Fig. 3G). The differentiated cells still could emit green fluorescence, which was seemingly stronger in the nuclei (Fig. 3E, 3H) and express EBNA1 (Fig. 3F, 3I). The syncytial cells contained different numbers of nuclei (from two to 100) and were positive for CG-β on immunostaining (Fig. 3J).
The oriP Plasmid Was Maintained as Episomes During Drug Selection, Could Be Recovered After Prolonged Periods of Culture, and Was Lost at Approximately 2%–4% per Cell Generation in the Absence of Selection
After culture and double selection in G418 and puromycin for 12 weeks, the episomal DNA of EBEG cell clones was isolated and used to transform E. coli cells. Twelve colonies were randomly picked and expanded in 3 ml of LB-Amp, and minipreps prepared. Each of the 12 plasmids isolated from EBEG cell clones showed the same banding pattern as the original pEF1-EGFP-oriP input plasmid DNA used in transfection (Fig. 4A, 4B). This provides strong evidence that the oriP plasmids were maintained as episomes during drug selection and could be recovered after prolonged periods of culture with no gross rearrangement occurred.
We also attempted to quantify the copy number of pEF1-EGFP-oriP in the EBEG cells by quantitative real-time PCR. According to the standard curve using the same pEF1-EGFP-oriP containing the oriP fragment, there were approximately two to eight copies of pEF1-EGFP-oriP per cell in EBEG hES cell population.
We also examined how stable the pEF1-EGFP-oriP episomal DNA in EBEG cell clones could be without selection. Twelve clones were picked and expanded with G418 and puromycin selection for 12 weeks. Then, puromycin selection was withdrawn, and EGFP expression was analyzed. Flow cytometric analysis of four clones is shown. We observed an average loss rate per generation of 2%–4% among all the different EBEG cell clones except clone 4 (Fig. 4C). This loss rate of oriP plasmids in hES cells in the absence of selection pressure is similar to that reported in other human cell types [34, 35]. In one exception, clone 4 demonstrated a much slower decline in the percentage of EGFP-expressing cells, presumably because it integrated into a chromosomal site favorable for expression, although we did not obtain direct evidence since this clone was lost before the planned Southern blot analysis.
Application of RNA Interference in hES Cells on the Basis of EBV Vector System
We also explored whether we could combine the RNAi technique with the EBV vector system in hES cells. A vector expressing shRNA against EGFP, pshEGFP.oriP, was constructed by inserting the oriP fragment into a vector (pshEGFP) that we made previously. The pEF1-EGFP-oriP vector, which provides an EGFP target transcript, pshEGFP.oriP, or pshEGFP were cotransfected into EBNA1 hES cells (Fig. 5). The results showed that the RNAi efficiency was enhanced notably using the EBV vector system and that the direction of oriP did not play a crucial role in RNAi-mediated EGFP knockdown in hES cells. In addition, pMD18-T-oriP, a cloning vector containing oriP, was constructed and used in the RNAi experiment to rule out the possibility that the RNAi effect mediated by pshEGF-P.oriP was due to the dilution or other nonspecific effects caused by the presence of oriP (Fig. 5).
Eukaryotic viral vectors have played an important role in the expression and analysis of genes. These vectors can be categorized as nonchromosomal and self-replicating (episomal) vectors and nonreplicating (integrating) vectors . Nonreplicating vectors have been used widely in gene therapy, but their integration into host chromosomes may cause insertional mutagenesis or subject them to position of integration effects, a phenomenon in which the transcription of an integrated gene will be affected by its integration site. Although lentiviral vectors can transduce hES cells efficiently and maintain long-term expression of transgenes, positional effects due to random and uncontrollable integration remain a concern for these vectors, and small packaging size limits the utility of these vectors for some specific applications. As episomal vectors, EBV replicon-based plasmids can avoid position of an integration effects or insertional mutagenesis. In addition, the large potential DNA insert size (up to 330 kb has been reported using a human artificial episomal chromosome system in human cells ), the ease of recovering these shuttle vectors in E. coli, and the lack of labor-intensive packaging procedures make EBV vectors particularly attractive. Furthermore, EBNA1 can increase transfection (up to 1,000-fold) of oriP vectors [28, 33] and enhance gene expression by 100-fold in some cell lines . For long-term gene transfer and expression, the EBV replicon has been successfully used to deliver the cystic fibrosis transmembrane conductance regulator gene to transformed human airway epithelial cells defective in cAMP-dependent chloride transport , to induce significant growth suppression and apoptotic tumor cell death in prostate cancer in vivo , and to target therapeutic agents at EBV-related tumor cells . Most previous studies of EBNA1 effects on cells in culture are consistent with EBNA1 having no effect on cell growth or survival. Also, a recent report shows that EBNA1 alone does not induce lymphoma in transgenic FVB mice . These data have not revealed an adverse effect of EBNA1 continuous expression in tumor formation, unlike the SV40 T antigen that serves as both a replication initiator and transforming protein. Thus, we set out to test the EBV vector system in hES cells, expecting a high transfection efficiency in hES cells and investigating its suitability in stable gene expression in undifferentiated and differentiated hES cells.
We generated stable EBNA1 hES cells in which EBNA1 expression and functions were confirmed by RT-PCR, Western blots, and luciferase reporter assays as shown in a related study . The EBNA1 hES cells maintained several key stem cell characteristics and had the ability to differentiate, judged by the formation of EB-like spheres and BMP4-induced differentiation in vitro and the formation of teratoma tissues in vivo, which were identical to the parental hES cells. Although we did not observe any overt changes in developmental potential of EBNA1 hES cells by the evaluation of EB-like sphere formation, teratoma development, and BMP4 differentiation to trophoblasts, we could not rule out the possibility that EBNA1 might interfere with cell differentiation or functions of a certain lineage or stage derived from hES cells in some unanticipated way.
The high stable transfection efficiencies we obtained in both H1 and H9 hES cells using the EBV vector system represented a significant improvement over those previously reported in hES cells with plasmid vectors [4, 5]. The 2%–4% loss rate of the episomes in hES cells in the absence of continuous selection is also observed in other human cell lines [34, 35]. Of our 12 stable EBEG cell clones, only one (clone 4) could stably express EGFP without selection. We postulate that the episomal vector in this clone may also integrate into the chromosomes of EBEG cells. The requirement for the continuous use of drug in dividing human cells to obtain stable gene expression could be a disadvantage of the EBV vectors. However, in some other types of cells, long-term stable extrachromosomal persistence of oriP vectors can be observed after 1 year of continuous in vitro cultivation or even without any selection for a period of 3 months [16, 41]. Until now, the selection of individual stable oriP episomes has not been completely understood, but it may represent a random epigenetic change, as episomes recovered from stable cells (and expanded in bacteria) do not transfect other cells with a higher efficiency, and cells stably maintaining an EBV episome are not transfected at a higher efficiency with a second EBV episome . A 6-month study indicated that the isoforms of EBNA1 differed with respect to their efficiency of plasmid maintenance; for example, the oriP maintenance mediated by Raji-form EBNA1 was more stable than that mediated by wild-type EBNA1, whereas the truncated EBNA1(IR3del)-mediated oriP maintenance was unstable . In addition, sequences adjacent to oriP can improve the persistence of EBV-based episomes in B cells . There have some data showing that transfection using chemical transfection reagents resulted in inefficient homologous recombination in hES cells , which may be overcome by EBV-based episomal vectors, for more nuclear import of EBV vectors can occur  and much more stable transfectants can be obtained using EBV vector system ([28, 33]; our results), making it attractive to apply EBV vectors for homologous recombination in hES cells. Therefore, a better understanding of these epigenetic changes or the mechanisms underlying how to stably maintain the oriP vectors might further broaden the applications of EBV vectors.
In the EBV system we used, the EMCV-IRES was used to ensure the EBNA1 and neoR genes, the EGFP and puroR genes to be translated from a common dicistronic mRNA under the control of the PGK promoter in pEB or EF1α promoter in pEF1-EGFP-oriP/pEF1-EGFP, respectively. Therefore, G418 resistance or puromycin resistance should be coupled with the expression of EBNA1 or EGFP and so can facilitate identification of positive clones, providing a very quick and useful way for isolation of transfected hES cell clones. According to our results, all the G418-resistant cell clones were positive for EBNA1, and all the puromycin-resistant cell clones expressed EGFP.
RNAi may be faster and more convenient to knock down the expression of a specific gene than by traditional gene targeting methods in hES cells. RNAi technique can also be used to modulate gene expression and induce ES cells to differentiate directionally, complementary to cDNA expression. Recently, RNAi methodology has been described for high-efficiency loss of gene function in hES cells merely based on retroviral and lentiviral vectors [45, 46]. However, the low transfection efficiency of RNAi vectors mediated by most commonly used methods still restrains the widespread use of RNAi in hES cells. Through the combination of EBV vector system and RNAi technique, we were delighted to find that the EBV vector system can markedly enhance the RNAi efficiency in both H1 and H9 hES cells. The present work should also help to use RNAi more efficiently in hES cells, as well as stable transgene expression.
In conclusion, the EBV-based episomal vector system should enhance our ability and applications of both gain- and loss-of-function analyses in hES cells. It will help us to reveal the developmental roles of specific human genes and modulate hES cell differentiation.
The authors indicate no potential conflicts of interest.
We sincerely thank Dr. J. A. Thomson for guidance, especially at the initial stage of this project. We gratefully acknowledge WiCell Research Institute for providing H1 and H9 hES cells, Dr. D. W. Burt for providing the pEB vector, and Dr. Bill Sugden for the pCMV-EGFP-oriP plasmid used in vector construction. We thank three anonymous reviewers for excellent suggestions that led to substantial improvement. This work was supported by a grant from the National Natural Science Foundation of China (30200140). C.R. and M.Z. contributed equally to this work.