An Efficient Vector System to Modify Cells Genetically

The transfer of foreign genes into mammalian cells has been essential for understanding the functions of genes and mechanisms of genetic diseases, for the production of coding proteins and for gene therapy applications. Currently, the identification and selection of cells that have received transferred genetic material can be accomplished by methods, including drug selection, reporter enzyme detection and GFP imaging. These methods may confer antibiotic resistance, or be disruptive, or require special equipment. In this study, we labeled genetically modified cells with a cell surface biotinylation tag by co-transfecting cells with BirA, a biotin ligase. The modified cells can be quickly isolated for downstream applications using a simple streptavidin bead method. This system can also be used to screen cells expressing two sets of genes from separate vectors.


Introduction
The transfer of foreign genes into eukaryotic cells, and in particular, into mammalian cells, plays an important role in our understanding of the function of coding genes and the regulatory components of non-coding sequences and also allows the production of coding proteins for therapeutic purposes and the development of strategies for gene therapy [1]. The rapid identification and selection of gene-modified cells are prerequisites for these applications.
Several systems have been developed for the selection of cells after gene transfer, including drug selection, GFP imaging and detection of other reporters [2,3,4,5]. Antibiotic selection is the most commonly used method and is based on the growth advantage of the transfectants in the presence of a cytotoxic agent along with the death of the non-transduced cells. The widely-used antibiotic resistance genes include aminoglycoside phosphotransferase, dihydrofolate reductase (DHFR), hygromycin B phosphotransferase, puromycin-N-acetyl-transferase, blasticidin S deaminase, and glutamine synthetase (GS), which confer resistance to G418, methotrexate, hygromycin, puromycin, blasticidin, and methionine sulfoximine respectively [6,7,8,9,10,11]. Isolation of gene-modified cells using this method requires several days to weeks and introduces undesirable drug resistance genes into the cells. Identification of transfected cells using reporter genes such as chloramphenicol acetyltransferase, alkaline phosphatase, b-galactosidase and firefly luciferase typically requires disruptive methods like cell permeabilization [7,9,12]. Green fluorescent protein (GFP) can be detected without cell permeabilization and is useful in fluorescence-activated cell sorting applications, but it can be toxic to cells.
Magnetic-activated cell sorting (MACS) is a simple solution for applications requiring the enrichment of cells of interest. MACS is dependent on the expression of a specific surface marker that can be recognized by a magnetic bead-tagged antibody. Gotoh et al. described eight streptavidin fusion genes as dominant selectable markers that can be combined with paramagnetic beads to select transfected cells [13]. Using MACS, it is possible to identify rare cell populations, separate large number of cells and large as many as 10 11 cells in approximately 1 hour [14]. The immunomagnetic selection procedure is simple and rapid. This method yields a highly pure population of transfected cells and can be used for a wide range of biological applications. Several approaches have been put forwarded to develop simpler and faster selection strategies. Kawahara et al. have proposed a novel selection system called the antigen-mediated genetically modified cell amplification (AMEGA) system, which employs an antibody/receptor chimera that triggers a growth signal in response to a cognate antigen without antibiotic selection [15,16,17,18,19].
The association of Streptavidin with biotin is the strongest known non-covalent bond, which is several orders of magnitude stronger than that of antigen-antibody interactions. The biotin ligase, BirA, can catalyze the biotinylation of the e-NH 2 of a 13amino acid peptide tag, a so-called minimal biotin acceptor sequence [20,21] and has been widely used for biotinylation of a protein of interest. In this study, we take advantage of a biotin ligase enzyme to catalyze the biotinylation of a cell-surface peptide tag co-transferred into the same cell to create an efficient vector system for modifying cells genetically.
The selection system consists of two vectors; one contains a target gene and the biotin ligase BirA as a reporter, and the other contains a second target gene and a BirA substrate peptide linked to a truncated form of human low-affinity nerve growth factor receptor (DLNGFR). The target gene cassette in each vector is used to express the genes of interest. Once the lentivectors enter cells through transient transfection or infection, one or more of the target proteins are expressed. BirA is retained in the endoplasmic reticulum (ER) after expression, and it efficiently biotinylates the BirA tags fused to DLNGFR when they pass through the ER. The biotinylated DLNGFR is transported to the cell surface where it labels the target cell. The biotinylated cells have a high affinity for streptavidin and can be quickly and efficiently captured in vitro with streptavidin-labeled beads [22]. Using this method, highly pure populations of transfected cells can be obtained. Lategeneration lentiviral vectors with advanced performance and safety profiles can efficiently transduce several types of nondividing and dividing cells. This method can be used to isolate mammalian cells modified by two sets of target genes from heterogeneous cell pools using magnetic streptavidin-labeled beads. This system has been used in multiple cell types, including the 293T, HCT-15, SupT1 and K42-91a, a mouse fibroblast cell line.

Cell surface expression of BirA enzyme substrate linked to a truncated LNGFR
To express a biotinylated tag on the cell surface, we constructed a tag protein composed of an ER leader sequence and a biotin tag peptide, the biotinylation of which can be catalyzed by coexpressed BirA, and a transmembrane protein fragment with a built-in Myc tag. Initially, both the HLA-A2 transmembrane protein and a truncated form of LNGFR were tested, and LNGFR was found to express the tag protein more efficiently [23]. Therefore, a truncated form of LNGFR containing the transmembrane domain and cytoplasmic tail was used as a fusion partner for the expression of the BirA substrate at the cell surface. As depicted in Fig. 1A, two lentivectors were constructed. Vector 1 contained a CMV promoter-driven transgene expression cassette and an EF-1alpha promoter-driven truncated LNGFR tagged with the BirA substrate peptide reporter. Vector 2 contained a second transgene cassette and the BirA enzyme to catalyze the biotinylation of the BirA peptide tag at the cell surface when coexpressed with the BirA substrate peptide in Vector 1. To test the vector system, a GFP reporter gene was inserted into the transgene expression cassette of Vector 1, termed 3.7lnGFP, and a lentivector containing the BirA expression cassette without a transgene was created and named 3.7CMVBirA. Both lentiviruses were produced in 293T cells using standard procedures.
To confirm that a biotinylated peptide tag was expressed on the cell surface, 293T cells were first transfected with the two lentiviruses, 3.7lnGFP and 3.7CMVBirA. The surface expression of biotinylated LNGFP was confirmed by binding to streptavidinlabeled beads 72 hours after transfection. As shown in Fig. 1B, cells expressing the biotinylated tag could be isolated with streptavidin-labeled beads, and almost all of the cells bound to the beads expressed GFP, this is further confirmed by FACS quantification (Fig. 2B). The surface expression of DLNGFR and Myc was also verified by FACS. As shown in Fig. 2A, cells transfected with 3.7lnGFP and cells transfected with both 3.7lnGFP and 3.7CMVBirA were stained with either anti-Myc or anti-LNGFR antibodies. The positive percentage stained with anti-Myc was 76.18% (3.7lnGFP) and 64.46% (both 3.7lnGFP and 3.7CMVBirA ) respectively, while the ones stained with anti-LNGFR was 69.38% (3.7lnGFP) and 66.15% (both 3.7lnGFP and 3.7CMVBirA) respectively. The cells transfected with 3.7CMVBirA alone were not stained by both antibodies above the background level of 7.33% and 6.99% respectively. Histogram analysis of flow cytometry data showed a broad range of staining intensity of non-isolated cells; however, the cells isolated with streptavidin-labeled beads formed a distinct population, the positive percentage being 98.05% (anti-Myc) and 97.07% (anti-LNGFR). The percentage of GFP positive cells is 68% (3.7lnGFP) or 70.16% (both 3.7lnGFP and 3.7CMVBirA) individually, and the percentage of streptavidin positive cells was 62.94%. Almost 87% GFP positive cells could be stained with streptavidin (Fig. 2B).
In the following experiments, western blots were performed to confirm whether biotinylation occured only when both BirA and the BirA substrate peptide tag were expressed by the same cell. As shown in Fig. 3A, cells that expressed the BirA substrate peptide tag alone or in combination with BirA could be stained with either anti-Myc or anti-LNGFR antibodies. In contrast, only cells expressing both proteins were stained with streptavidin, indicating that biotinylation took place only when both proteins were expressed in the same cell. FACS analysis (Fig. 3B) strengthened the results as 93.86% cells expressed biotinylated proteins on the surface of cells.

Isolation of the gene-modified cell lines with streptavidin-labeled magnetic beads
In the next set of experiments, the efficient isolation of clones containing transgenes was demonstrated. As shown in Fig. 4A, 293T clones that expressed GFP were quickly established by infection and sorting with magnetic beads. The 293T cells were first infected with 3.7lnGFP and 3.7CMVBirA lentiviruses. After infection, the GFP-expressing 293T cells were isolated using streptavidin-labeled beads. Only GFP-positive cells adhered to the streptavidin-labeled beads. Single 293T cells isolated with streptavidin-labeled beads were further cultured, and single cellderived clones were established within a period of a few days. Similarly, HCT-15 clones expressing GFP were established by infection and magnetic bead sorting. After infection with 3.7lnGFP and 3.7CMVBirA, GFP-positive HCT-15 cells were isolated with streptavidin-labeled beads, and a clone was established in one week (Fig. 4B). Long term and stable expression of transgene was investigated by FACS analysis of GFP expression, the percentage of GFP-positive cells at day 20 was 98%, and 88% GFP positive cells can be stained with streptavidin ( Fig. 4D).
One application for this system is the expression of biotinylated proteins in established 293T cells expressing the BirA enzyme. To establish a 293T cell line stably expressing the BirA enzyme, 293T cells were first infected with 3.7CMVBirA lentivirus. Next, the same cells were transiently transfected with 3.7lnGFP vector, and the modified cells were isolated with streptavidin-labeled magnetic beads (Fig. 4C).

HCT-15 cells modified with a single transgene (b2m-gp100) were specifically targeted by TCR-modified NK cells
To further demonstrate the efficiency of transgenic cell isolation, we transduced HCT-15 cells with b2m linked to a peptide derived from gp100 (b2m-gp100). HCT-15 cells are b2mdeficient human colorectal cancer cell line. Because they lack b2m, no stable HLA I molecules are expressed at the cell surface. As shown in Fig. 5B, non-transduced HCT-15 cells did not express HLA class I molecules at the surface. HCT-15 cells were infected with two lentiviral vectors: one expressing b2m-gp100 and the BirA substrate peptide tag and another expressing the BirA enzyme. Biotinylated cells were isolated with magnetic streptavi-din-labeled beads, and clones were established (Fig. 5A). The presence of b2m on the surface of HCT-15 cells was confirmed by FACS analysis using W6/32 or BBM.1 antibodies, wherein approximately 92.22% (BBM.1) and 95.18% (W6/32) of transduced cells were stained (Fig. 5B, top and middle panels). Biotinylated proteins at the cell surface were detected with streptavidin-PE-Cy5, and observed staining in approximately 87.13% of transduced cells (Fig. 5B, bottom panels). We further tested the stability of b2m/gp100 expression in HCT-15 cells using FACS analysis. 20 th passage of established b2m/gp100positive HCT-15 cells were stained with mAb W6/32 ( Fig. 5D) of which 92.64% were positive.
The established b2m/gp100-positive HCT-15 cells were further tested as target cells for NK92MI cells expressing GPA7A, a T cell receptor (TCR)-like antibody specific for a gp100 peptide-loaded HLA-A2 complex. As shown in Fig. 5C, the established b2m/ gp100-positive HCT-15 cells were effectively targeted by GPA7Atransfected NK92MI cells, which specifically recognized target cells expressing the gp100 peptide-loaded HLA-A2 complex. In contrast, parental NK92MI cells induced a much lower level of cytotoxicity.

The establishment of cell lines expressing two sets of desired transgenes
The potential of this system was further realized by the efficient establishment of cells modified with two sets of transgenes. The individual genes were cloned into separate lentiviral vectors and introduced into target cells, and stably transduced cell lines were then isolated. Two pairs of genes, a combination of human b2m and HLA-A2 and the combination of melanoma associated-Ag gp100-specific TCR-alpha and -beta chains, were used to establish the cell lines. The K42-91a cell line [24], a variant of the K42 mouse fibroblast cell line in which TAP2 has been knocked down with a targeted knockout of the CRT gene, was used to establish gene-modified cells expressing the HLA class I complex.
Lentiviruses expressing HLA-A2 and human b2m were used to infect K42-91a cells, the cells modified with both HLA-A2 and b2m were isolated using magnetic streptavidin-labeled beads, and stable clones were established. Staining with anti-HLA complex antibodies showed that human HLA I complexes were expressed on the surface of murine k42-91a cells (Fig. 6A, top two panels), the cells were also stained with anti-Myc antibody and streptavidin-PE-cy5 (Fig. 6A, bottom two panels). The surface expression of the human HLA I complex on k42-91a cells was further confirmed by flow cytometry, which showed a shift to the right in the human HLA I staining profile. In contrast, the staining profile of an irrelevant peptide did not change (Fig. 6B). To confirm stable expression of HLA-A2 and human b2m in K42-91a cells, 15 th passage of established HLA-A2 and b2m transduced K42-91a cells were analyzed using FACS ( Fig. 6C).
In another set of experiments, SupT1 cells were modified to express TCR genes. Wild-type SupT1 cells are TCR-deficient, and although all of the genes for the CD3 components are present, these cells do not express surface CD3 because of the lack of the TCR. The TCR is composed of alpha and beta chains and is MHC restricted. The CD3 complex is expressed on the surface of SupT1 cell when both of the TCR subunits are re-expressed. The TCR-alpha and -beta chains derived from melanoma-specific T cells were inserted into the transgene expression cassettes of the respective lentivectors, and viruses expressing the transgenes were produced [25]. SupT1 cells were infected with both a virus containing the TCR-alpha gene and a virus containing the TCRbeta gene. After isolation with magnetic streptavidin-labeled beads, the cells were further cultured, and the expression of the modified genes was confirmed by FACS analysis. As presented in Fig. 6D, CD3, LNGFR, TCR-alpha and -beta chains and biotinylated transmembrane proteins were clearly expressed on SupT1 cells due to the integration of both the TCR-alpha and - beta genes mediated by the lentiviral vectors. Homogenous SupT1 cells expressing TCR were obtained after infection (Fig. 6D, left panels) and magnetic bead separation (Fig. 6D, right panels).

Discussion
The development of cell-based gene therapy and the production of recombinant proteins for clinical applications both require genetically modified cells. These applications are largely hampered by the lack of protocols for the simple isolation of genetically modified cells. In this report, we have described a new method that permits the specific and efficient isolation of gene-modified cells based on the strong and specific interaction between biotin and avidin [13]. In our system, mammalian cells are modified with two separate vectors, one expressing the BirA enzyme, which has been engineered to be retained in the endoplasmic reticulum (ER), and the other expressing as a transmembrane protein with a BirA substrate peptide tag [26,27,28,29,30,31,32,33,34]. The BirA substrate peptide tag will be expressed on the surface of a cell when both genes are expressed along with two sets of target genes. The resulting biotin-labeled cells can be isolated using magnetic streptavidin-labeled beads. In this way, genetically modified cells can be obtained as early as 72 hours post-infection. The viability of the cell lines is not affected, and there is no toxic effect on cell proliferation [35,36,37].
Gene therapy depends on both efficient gene transfer into the appropriate target cells and the ability to enrich the transduced cells. Conventional selectable markers have limitations, including growth inhibition, mutations in antibiotic resistance genes, cell membrane permeability and morphological changes. DLNGFR is a truncated native receptor that is not immunogenic, does not mediate signal transduction and is already being used in clinical protocols [38]. DLNGFR is easily tracked in vivo, and cells expressing this transgene can be isolated using magnetic or fluorescence-activated cell sorting. The BirA enzyme is a foreign protein introduced into the cells and is designed as an ER-resident protein [22]. Although no cytotoxic effects have been reported upon introduction of BirA into cells, its potential effects on cell cycle and its long-term safety require further investigation [28].
The human immunodeficiency virus type-1(HIV-1) based lentiviral vectors provide high and stable transgene in both dividing and non-dividing cell types, the third generation selfinactivating (SIN) vectors minimize the risk of reconstituting a replication-competent lentivirus (RCL). In our lentiviral system, two ubiquitously internal promoters in lentivectors, human elongation factor-1a (EF-1a) promoter [39,40] and cytomegalovirus (CMV) promoter [35,41], were chosen to drive transgene expression, and stable cell lines were generated in different cell types. We also tested whether transgene expression could be stably maintained when transduced cells are expanded where we observed that transgenes were stable expressed after at least 15 passages in different cell types. However, silencing of vector transgene is an important factor to consider where long term stable expression of transgene is required in target cells. There are specific reports stating many transgenes driven by viral promoter including the CMV promoter have been found gradually silenced over period of months in certain cell types [42]. Replacement of CMV promoter by appropriate cell-type specific promoters when needed facilitate efficient transgene manipulation and isolation of gene-transducted cells.
Our results show that the use of this selective system allows for rapid, efficient, and nontoxic in vitro selection of gene-modified cell. This system has been applied in many different cell lines and will provide an option in isolating tranfected cells, and set to use the system to select the TCR-alpha and -beta genes derived from melanoma-specific T cells transduced human primary T cells which has future prospects to treat melanoma. More importantly, the selective system can be used to design vaccine, owing to genemodified DC vaccine expressing TAA, gene encoding cytotoxic T lymphocyte (CTL)-specific epitopes, and are promising vaccine candidates to be utilized for treatment of cancer [43,44,45], In summary, this methodology provides many potential of applica-tion in adoptive gene therapy, producing recombinant protein and studying the function of genes.

Cell lines and antibodies
The HEK293T cells and the HCT-15 colon adenocarcinoma cells used in this study were cultured in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (Gibco BRL, Caithersburg, MD, USA). The T lymphoid SupT1, NK92MI and K42-91a cell lines were cultured in RPMI-1640 medium supplement-ed with 10% fetal bovine serum (FBS) and 100 U/mL penicillin/ streptomycin. All cell lines were purchased from ATCC (Rockville, MD, USA) except HCT-15 cells (a gift from Chen Au Phe) and K42-91a cell line [24]. All transformants were supplemented with biotin at a final concentration of 100 mM. Cells were maintained in 5% CO 2   Production and concentration of lentiviral vectors, transduction and selection of transfected cells VSV-G pseudotyped lentiviral vectors stocks were prepared, concentrated, and titered as described previously [46]. Briefly, 293T cells were transiently transfected with 6 mg of packaging plasmids pLP1, pLP2 and pLP/VSVG (Invitrogen, Carlsbad, CA, USA) respectively and 6 mg of transfer vector. After 48 hours, virus-containing supernatants were harvested, passed through a 0.45 mm filter and frozen in aliquots at 280uC until used. The virus was used to transduce different cell lines in the presence of 4 mg/mL polybrene at different the multiplicity of infection (MOI). Adherent cells (293T, HCT-15, K42-91a) were transduced at an MOI of 1, suspension cells (supT1) were transduced at an MOI of 20. All cells were transduced by spin inoculation in a centrifuge at 1,8006 g for 90 minutes at 32uC in 6-well tissue culture plates (2 mL per well) and incubated at 37uC overnight; fresh medium was added after the overnight incubation. This process was repeated once after every 24 hours later. A mock infection (no vector-containing medium) was always performed in parallel under identical culture conditions. Magnetic bead-based cell separation was performed 72 hours post-infection. After the lentiviral infection and separation of transduced cell lines, the transduction efficiency was monitored by flow cytometry.

Magnetic bead-based cell separation
To avoid potential cleavage of surface proteins by trypsin, the adherent cells were released from the plates with 0.5 mM EDTA. The cell suspension was washed 3 times to remove unbound dbiotin. Next, Dynabead M-280 streptavidin-labeled beads (Invitrogen, Carlsbad, CA,USA) were mixed with the cell suspension at a ratio of 10 ml bead suspension per 1 mL (1610 7 ) target cell suspension. The resulting mixture was rotated for 20 minutes on ice. The tubes were placed into a magnetic device for 5 minutes, and the bound cells were collected. Tubes were removed from the magnet, and the beads were washed once with phosphate-buffered saline (PBS) / 2% FBS. The collected cells were harvested and resuspended in 1.5 mL PBS (pH 7.4). These procedures were repeated three or four times. The biotin-positive cells were separated from the suspension with a magnetic unit. The presence of the biotinylated surface marker on the harvested cells was confirmed either by additional staining of the cells with antibody or by observation of GFP in unstained cells under a fluorescence microscope.

SDS-PAGE and western blotting
One million cells were lysed and centrifuged, and the supernatants were boiled in SDS sample buffer for 5 minutes and separated by SDS-PAGE in a 12% polyacrylamide gel. The proteins in the gel were then transferred to Hybond C Extra membrane (Amersham) at 20 V for 2 hours. The membrane was blocked in 10 mL of PBS containing 2% Marvel milk (Premier International Foods, St Albans, UK) on a plate rocker (Stuart Gyro Rocker STR-9) for 1 hour and transferred to 10 mL of Phosphate Buffered Saline with Tween 20 (PBST) /2% Marvel milk containing a 1:1500 dilution of the primary antibody. The membrane was washed 3 times and transferred to PBST/2% Marvel milk containing a 1:5000 dilution of the HRP-conjugated secondary antibody. The membrane was washed twice in PBST and twice in PBS and dabbed dry with a paper towel. Two milliliters of enhanced chemiluminescence fluid were added evenly to the membrane. After 5 minutes, the fluid was pipetted off, and the membrane was dabbed dry which was then wrapped in Parafilm and attached to one side of an exposure box with tape. Photographic film (Kodak Biomax) was exposed to the membrane for 10 seconds to 30 minutes.
Flow cytometry analysis W6.32, BBM.1, anti-LNGFR and anti-Myc were used as primary antibodies. The cells were washed 3 times with PBS, and 1610 6 cells were incubated with mAbs on ice for 30 minutes and then incubated with or without secondary antibody (mouse IgG conjugated to FITC or PE-Cy5). The anti-human-TCR-a/b-PE antibody was used according to the manufacturer's instructions. After each incubation, the cells were washed twice with PBS. The labeled cells were analyzed by flow cytometry with a Guava Easycyte (Guava Technologies, Hayward, CA, USA). All FACS data were analyzed using FlowJo software (Tree Star, Ashland, OR, USA).

Cytotoxicity assay
Target cells were washed twice with Hank's buffered salt solution (HBSS) before staining. The 1-2610 7 cells were then incubated for 10 min with Hank's buffered salt solution (HBSS) supplemented with 2 mM carboxyfluorescein succinimidyl ester (CFSE, Invitrogen, Merelbeke, Belgium) at 37uC in the dark. Immediately following the addition of the CSFE, an equal volume of FBS was added to stop the reaction. After incubation with FBS for 1 min, the target cells were washed twice with HBSS and resuspended in complete medium at a concentration of 1610 5 cells /mL; 100 ml of each sample was added to each well of a 96well U-bottom plate. Triplicate samples containing mixtures of effector and target cells at E:T ratios of 1:1, 2:1, 4:1, 8:1, 16:1, and 32:1 were prepared. The cells were incubated at 37uC for 4 hours in the presence of 5% CO 2 . After incubation, the cells were transferred to a 1.5 mL tube, centrifuged at 2506 g for 5 minutes, washed once with ice-cold medium and resuspended in 500 ml icecold medium. The cells were incubated for 3 minutes on ice with 10 mg/mL propidium iodide ( PI; Sigma-Aldrich, Poole, England) and both green and red fluorescence were analyzed. Cells stained with both dyes represent dead target cells. Analysis was on gated CFSE positive cells only. The percentage of cytotoxic activity was calculated using the following formula: %specific death = (%total CFS + PI + (dead) cells2%spontaneous CFSE + PI + (dead) cells)/ (100%2%spontaneous CFSE + PI + (dead)cells)6100%. The percentage of spontaneous dead cells was measured in the tube containing target cells only [47].

Image Acquisition
Images were acquired on a ZEISS Axiovert 200 M microscope (Zeiss, Vienna, A) with a 406 NA 0.55 objective. The microscope and image acquisition were controlled by Axiovision software. For GFP observation we used the blue light excitation. Image processing, such as clipping of images, image montages etc. was performed in PowerPoint.