Determination of the Optimal Concentration of Several Selective Drugs Useful for Generating Multi-Transgenic Porcine Embryonic Fibroblasts

Introduction

Genetically modified (GM) pigs are produced in many laboratories. A selectable marker positive for antibiotic resistance is essential for this procedure, although extensive manipulation of the genome, including disrup- tion of limited genes by gene targeting, has been performed (Ku¨holzer et al. 2000; Lai et al. 2002, Lai et al. 2006; Dai et al. 2002; Phelps et al. 2003; Ramso- ondar et al. 2003; Sharma et al. 2003; Harrison et al. 2004; Kolber-Simonds et al. 2004; Watanabe et al. 2005; Takahagi et al. 2005; Rogers et al. 2008).

This requires the derivation of stably transfected swine cells by using a selection agent. Most of these transfection experiments are involved in the process of gene targeting, in which one allele of the endogenous a-1,3-galactosyltransferase gene is disrupted, and the neomycin resistance gene (neo), encoding neomycin phosphotransferase II from trans- poson Tn5 (Schwartz et al. 1991), is introduced to confer resistance to the aminoglycoside antibiotic, G418 (geneticin).

These GM cells are next used for the generation of cloned GM piglets by somatic cell nuclear transfer (SCNT) technology (for a review by Lai and Prather 2003; d’Apice and Cowan 2008; Sachs and Galli 2009).

Unfortunately, the limited number of selectable markers available for pig cell transfection has hindered extensive manipulation of the porcine genome and the subsequent genetic analysis of this organism.

Currently, a few selectable markers are available for stable trans- fection of porcine cells (Takahagi et al. 2005; Watanabe et al. 2005; Brunetti et al. 2008; Caballero and Piedrah- ita 2009), but the conditions for drug selection (i.e. determination of optimal drug concentration and dura- tion of drug treatment) are undetermined.

The only exception is a report by Watanabe et al. (2005), who determined the optimal drug concentration of puromy- cin and succeeded in producing cloned GM piglets that express both enhanced green fluorescent protein (EGFP) cDNA and the puromycin-N-acetyltransferase gene (pac) isolated from Streptomyces alboninger that confers resistance to the aminonucleoside antibiotic puromycin.

There are 3 other pre-existing selectable markers, namely, bsr, hph and Sh ble, which have rarely been used for the genetic manipulation of the swine genome. The bsr gene encodes blasticidin S deaminase (Kamakura et al. 1987; Kobayashi et al. 1991) and confers resistance to blasticidin S, a peptidyl nucleoside antibiotic isolated from Streptomyces griseochromogenes that inhibits pro- tein synthesis in both prokaryotes and eukaryotes (Izumi et al. 1991).

The hph gene encodes hygromycin phospho- transferase (Lupton et al. 1991) and confers resistance to hygromycin B, an aminocyclitol antibiotic produced by Streptomyces hygroscopicus (Pettinger et al. 1953) that inhibits protein synthesis in both prokaryotes and eukaryotes. The Sh ble gene encodes a protein that binds to the antibiotic zeocin, a member of the bleomy- cin ⁄ phleomycin family, isolated from Streptomyces, and prevents it from binding to DNA.

In this study, we explored the possibility of using these other selectable markers (bsr, hph, and Sh ble) as useful markers for the isolation of GM swine transfectants. Furthermore, we constructed stable porcine embryonic fibroblasts (PEF) transfectants carrying a total of 5 selectable marker genes (neo, pac, hph, bsr and Sh ble) to test whether these markers confer resistance independently.

Materials and Methods

Cell lines and culture

The Clawn miniature pigs used in the present study were established at the Faculty of Agriculture of Kagoshima University (Nakanishi et al. 1991) and purchased from Japan Farm, Ltd (Kagoshima, Japan). Porcine embryonic fibroblasts were obtained from female foetuses on day 30 of pregnancy.

After removal of the head, intestine, liver, heart and limbs, the remaining tissues were finely minced into pieces using scissors and primarily cultured in PEF culture medium (containing Dulbecco’s modified Eagle’s medium (DMEM) ⁄ Ham’s F-12 (#124; Wako Pure Chemical Industries Ltd, Osaka, Japan), 10% foetal bovine serum, and antibiotics) at 38.4°C in a humidified atmosphere of 5% CO2 in air. They were passaged 3–4 times and then frozen. For each experiment, the frozen cells were thawed and passaged 4–8 generations prior to transfection.

Plasmids carrying selectable marker genes

The plasmid pAQC (Fig. 3A) is designed to express two transgenes: hph under control of the mouse phospho- glycerate kinase (PGK) promoter and tandem dimer Tomato (tdTomato) cDNA (kindly provided by Dr. Roger Tsien) under control of the chicken b-actin-based CAG promoter (Niwa et al. 1991). The 2 units are aligned in a head-to-head fashion.

The plasmid pKJ2 (Fig. 3A; Boer et al. 1990) is a vector for the expression of neo under the transcriptional control of the PGK promoter. The plasmid pCEIP-12 (Fig. 3A) was used as a plasmid simultaneously expressing both EGFP cDNA and pac under the transcriptional control of the CAG promoter.

This plasmid has a unique internal ribosomal entry site (IRES) sequence between the EGFP and pac genes on the cDNA to co-express both proteins.

The IRES itself functions to generate two or more unrelated reading frames from a single mRNA transcript (Fusse- negger et al. 1998a,b; Douin et al. 2004). The plasmid pCAG ⁄ bsr-7 (Fig. 3A) was constructed by inserting the bsr gene, isolated from pSV2bsr (Funakoshi, Tokyo, Japan), downstream of the CAG promoter.

The plasmid pSV2-zeo was purchased from Invitrogen (Carlsbad, CA, USA), and it includes the Sh ble gene under the control of the SV40 early promoter (Fig. 3A). Prior to transfection, the plasmid DNA was linearized by diges- tion with appropriate enzymes.

Selective drugs

The drugs used in this study were all purchased from Invitrogen Co., except for puromycin (InvivoGen Inc., San Diego, CA, USA). Working solutions for each drug were prepared by serial dilution prior to use.

Drug response assays

The susceptibility of PEFs to drugs was assayed by the calculation of viable cell numbers harvested at indicated time points, using the trypan blue dye exclusion method. One day before testing with drugs, cells (1 · 105) were plated into wells of a gelatin-coated 24-well plate (Iwaki, Tokyo, Japan), containing 1 ml of PEF medium.

Serially diluted drugs were added, and at indicated times, the viable cell number was estimated by harvesting cells in 0.25% trypsin ⁄ 0.1% EDTA and subsequently staining with trypan blue (Trypan Blue Stain 0.4%; Invitrogen Co.).

Cells grown in drug-free medium were simulta- neously harvested and used as control. Cells from triplicate wells were counted by a Burker–Turk haemo- cytometer (Erma, Tokyo, Japan), as described previously (Nakayama et al. 2007). Drug sensitivity was graphed as the percentage of growth, relative to the growth of PEFs in drug-free medium. We plotted the mean growth of each treatment group, as shown in Fig. 1.

Survival test after selection in the presence of drugs at optimal concentrations

One day before being tested with drugs, cells (1 · 105) were plated into wells of an Iwaki 24-well plate containing 1 ml of PEF medium. G418 (400 lg ⁄ ml), blasticidin S (8 lg ⁄ ml), puromycin (2 lg ⁄ ml), hygromy- cin B (40 lg ⁄ ml) and zeocin (800 lg ⁄ ml), all of which were known to kill intact PEFs within 10 days (see Fig. 1), were added to 1 ml of PEF medium.

Cells from triplicate wells were harvested at the indicated time points for counting cells, as described above. Cells grown in drug-free media were simultaneously harvested and used as control. We plotted the mean growth of each treatment group, as shown in Fig. 2.

Transfection for obtaining PEF transfectants carrying multiple transgenes

For acquiring stable transfectants carrying all the plasmids listed in Fig. 3A, the PEFs were serially transfected as schematically described in Fig. 3B. For transfection of cells, we used an electroporation-based Lonza Nucleofector system (Lonza Biologics, Cologne, Germany), because with this system, relatively high transfection efficiencies (>50%) were achieved in PEFs (Nakayama et al. 2007). At first, PEFs (~1 · 106) were electroporated with pAQC (6 lg).

After transfection, the cells were split into 3 60-mm gelatin-coated dishes (Iwaki) and further incubated in PEF medium contain- ing 40 lg ⁄ ml of hygromycin B for an additional 10–15 days. Colonies (containing 300–700 cells) were picked using a small paper disc (3MM Whatman paper, 5 mm in width, and 5 mm in length), which had been dipped in 0.125% trypsin ⁄ 0.01% EDTA.

Cells were directly transferred to wells of a gelatin-coated 48-well plate (Iwaki), with hygromycin B-containing PEF medium. These cells were cultured for 20–30 days until they reached confluency. These cells were further prop- agated in a stepwise manner and checked for the uniform expression of tdTomato-derived red fluores- cence under a fluorescence microscope, as described below.

One of the resulting clones (at least four colonies were isolated) was termed TH, as it was considered to possess tdTomato cDNA (T) and hph (H). To introduce pCEIP-12 and pKJ2 plasmid DNA into the TH clone simultaneously, cells were co-transfected with pCEIP-12 (2.5 lg) and pKJ2 (2.5 lg), as described above. Selec- tion was performed in the presence of puromycin (2 lg ⁄ ml) and G418 (400 lg ⁄ ml).

The surviving colonies were picked for further propagation and termed THEPN, because it contained EGFP cDNA (E) and pac (P) from pCEIP-12, and neo (N) from pKJ2, together with T and H. For simultaneously introducing pCAG ⁄ bsr-7 (2.5 lg) and pSV2-zeo (2.5 lg) into the genome of THEPN, one of the THEPN clones was subjected to electroporation, and cells were selected in the presence of blasticidin S (8 lg ⁄ ml) and zeocin (800 lg ⁄ ml). The resulting clone was termed THE- PNBS, because it possesses bsr (B) and Sh ble (S) genes together with T, H, E, P and N.

Detection of fluorescence

Fluorescence in the cells was examined with an Olympus BX60 fluorescence microscope (Olympus, Tokyo, Japan) with DM505 filters (BP460-490 and BA510IF; Olympus) and DM600 filters (BP545-580 and BA6101F; Olympus), which were used for EGFP-derived green fluorescence and tdTomato-derived red fluorescence, respectively.

Microphotographs were taken using a digital camera (FUJIX HC-300 ⁄ OL; Fuji Film, Tokyo, Japan) attached to the fluorescence microscope and printed using a Mitsubishi digital colour printer (CP700DSA; Mitsubishi, Tokyo, Japan).

Results

Drug sensitivity of PEFs

As the first step to determine the sensitivity of the PEFs to selectable drugs (including G418, blasticidin S, hygromycin B, puromycin, and zeocin), PEFs were incubated with medium containing various concentra- tions of each drug for up to 10 days.

Drug sensitivity was graphed as the percentage of growth, relative to the growth of PEFs in drug-free medium (Fig. 1). As a result, the optimal concentrations were determined for each drug, under which normal PEFs die within 10 days after treatment: G418, 400 lg ⁄ ml; blasticidin S, 8 lg ⁄ ml; hygromycin B, 40 lg ⁄ ml; puromycin, 2 lg ⁄ ml; and zeocin, 800 lg ⁄ ml.

Next, the duration of tolerance to each drug was examined at a defined concentration of PEFs. PEFs were cultured in the presence of each drug at the optimal concentration (indicated above) for up to 10 days.

As expected, puromycin treatment resulted in rapid cell death compared with treatment with other drugs (Fig. 2). Furthermore, zeocin treatment caused cell death more slowly than treatment with other drugs, although complete cell death was seen at 7 days after the treatment.

Isolation and characterization of multiple transgene-incorporating PEF transfectants

To show the usefulness of the above-described selection scheme for acquiring transfectants carrying multiple gene constructs, we employed step-by-step transfection protocols, as described in Materials and Methods. We first obtained TH clones by transfection with the pAQC plasmid that contains tdTomato and hph. These clones exhibited tdTomato-derived red fluorescence.

These TH cells were next transfected with pKJ2 and pCEIP-12 plasmids that contain neo and EGFP ⁄ pac genes, respectively. The resulting drug-resis- tant colonies (termed THEPN) exhibited EGFP-derived green fluorescence together with red fluorescence (d–f in Fig. 3C).

The THEPN cells were further transfected with pCAG ⁄ bsr-7 and pSV2-zeo that contain bsr and Sh ble genes, respectively. We finally obtained THEPNBS cells carrying bsr and Sh ble in their genome.

To examine whether the THEPNBS cells carry the expected transgenes, the presence of each gene was assessed by PCR analysis of genomic DNA isolated from THEPNBS cells. The results are shown in Fig. 3D. As expected, the THEPNBS cells had five unique drug- resistance genes in their genome.

Next, we tested whether THEPNBS cells can survive when cultured in medium containing all five selective drugs at optimal concentrations. TH and THEPN cells, together with intact PEFs, were also cultured as controls.

After 10 days of culture, cells were fixed and stained with Giemsa dye solution. The results are shown in Fig. 3E. As expected, THEPNBS cells survived after treatment with five selective drugs, while the control TH, THEPN and intact PEFs were killed by the treatment.

Discussion

To date, the most commonly used selectable marker in swine cells is neo, which confers resistance to G418 (Ku¨holzer et al. 2000; Lai et al. 2002; Lai et al. 2006; Dai et al. 2002; Phelps et al. 2003; Ramsoondar et al. 2003; Sharma et al. 2003; Harrison et al. 2004; Kolber- Simonds et al. 2004; Watanabe et al. 2005; Takahagi et al. 2005; Rogers et al. 2008).

However, other pre-existing positive selectable markers are not used, and if used, only few laboratories employ them (Takahagi et al. 2005; Watanabe et al. 2005; Brunetti et al. 2008; Caballero and Piedrahita 2009), and the precise condi- tions for the genetic engineering of swine cells, such as the optimal concentration of each drug for the selection of transfected cells, remain undetermined. This may be due to the difficulty of producing GM piglets.

In pig-to-human xenotransplantation, the necessity of introducing several xenotransplantation-related genes into the pig genome in order to avoid possible immu- nological attack from the host has long been warranted (for a review by Cozzi and White 1995).

Pronuclear injection of transgene-containing DNA into fertilized eggs is one of the chief methods to produce GM piglets. However, this method has a drawback, because the efficiency of gene transfer is very low, and the expression of the transgene is often suppressed, most likely by gene silencing (for a review by Chandler and Vaucheret 2001).

Lavitrano’s group recently succeeded in intro- ducing multiple transgenes (three genes, including EGFP and red fluorescence genes) into the pig genome by using sperm-mediated gene transfer (SMGT) technology, which was first developed by their group (Webster et al. 2005). They observed transgene expression in the resulting piglets.

However, this technology appears to not be widely used, probably because of the difficulty in reproducing the SMGT technique itself. In this context, it would be preferable to employ the SCNT-mediated production of GM animals, although its efficiency is still very low (for a review by Wilmut et al. 2002).

This is because the properties of the GM cells (e.g. expression levels of transgenes) are easy to analyse, and we can isolate high transgene-expressing cells with relative ease. Furthermore, it is possible to introduce multiple gene constructs into cells prior to SCNT. The application of other pre-existing selectable markers would allow us to realize such a concept.

In this study, we determined the optimal concentra- tions of selectable drugs suitable for the selection of GM PEFs. For example, we determined that 2 lg ⁄ ml of puromycin is preferable. This concentration appears to be in agreement with that reported by Watanabe et al. (2005) and Brunetti et al. (2008), who demonstrated that 1–2 lg ⁄ ml of puromycin is suitable for killing intact porcine cells.

The concentrations of G418 used for the selection of transfected swine cells differ among labora- tories, ranging from 100 to 600 lg ⁄ ml (Ku¨holzer et al. 2000; Dai et al. 2002; Lai et al. 2002; Ramsoondar et al. 2003; Sharma et al. 2003; Harrison et al. 2004; Kolber- Simonds et al. 2004; Takahagi et al. 2005; Watanabe et al. 2005; Lai et al. 2006; Brunetti et al. 2008; Rogers et al. 2008; Deng et al. 2011; Pan et al. 2010; Phelps et al. 2003).

The reason for the broad effective dose range of G418 may be that it exerts lower toxicity in eukaryotes compared with other drugs such as puromycin and hygromycin B (Glover and Hames 2011).

Takahagi et al. (2005) used hygromycin B (55–80 lg ⁄ ml) for the selection of gene-targeted swine cells. This value is similar to that (40 lg ⁄ ml) derived in this study. On the other hand, Brunetti et al. (2008) employed a relative high dose of hygromycin B (150 lg ⁄ ml). Concerning the use of blasticidin S, Caballero and Piedrahita (2009) employed 8 lg ⁄ ml of the drug for selection of GM PEFs, which is just the same used in this study.

In preliminary testing, the concentrations of each drug used for PEFs are also effective for other swine cell types such as epithelial cells isolated from the adult Clawn miniature kidney, and testicular cells probably derived from the Sertoli cells isolated from adult testes (data not shown).

In conclusion, intact PEFs are sensitive to each type of commercially available selective drug. Furthermore, we obtained stable PEF transfectants that confer resistance to each drug. Therefore, the information presented here concerning the optimal concentra- tions of each selective drug will be useful for the selection of GM porcine cells that harbour multiple transgenes.