A Virus-Free Poly-Promoter Vector Induces Pluripotency in Quiescent Bovine Cells under Chemically Defined Conditions of Dual Kinase Inhibition

Authentic induced pluripotent stem cells (iPSCs), capable of giving rise to all cell types of an adult animal, are currently only available in mouse. Here, we report the first generation of bovine iPSC-like cells following transfection with a novel virus-free poly-promoter vector. This vector contains the bovine cDNAs for OCT4, SOX2, KLF4 and c-MYC, each controlled by its own independent promoter. Bovine fibroblasts were cultured without feeders in a chemically defined medium containing leukaemia inhibitory factor (LIF) and inhibitors of MEK1/2 and glycogen synthase kinase-3 signaling (‘2i’). Non-invasive real-time kinetic profiling revealed a different response of bovine vs human and mouse cells to culture in 2i/LIF. In bovine, 2i was necessary and sufficient to induce the appearance of tightly packed alkaline phosphatase-positive iPSC-like colonies. These colonies formed in the absence of DNA synthesis and did not expand after passaging. Following transfection, non-proliferative primary colonies expressed discriminatory markers of pluripotency, including endogenous iPSC factors, CDH1, DPPA3, NANOG, SOCS3, ZFP42, telomerase activity, Tra-1-60/81 and SSEA-3/4, but not SSEA-1. This indicates that they had initiated a self-sustaining pluripotency programme. Bovine iPSC-like cells maintained a normal karyotype and differentiated into derivatives of all three germ layers in vitro and in teratomas. Our study demonstrates that conversion into induced pluripotency can occur in quiescent cells, following a previously undescribed route of direct cell reprogramming. This identifies a major species-specific barrier for generating iPSCs and provides a chemically defined screening platform for factors that induce proliferation and maintain pluripotency of embryo-derived pluripotent stem cells in livestock.

Functionally, EpiSCs are poised to differentiate into primordial germ cells in vitro and neither contribute to all somatic cell lineages nor the germline in chimeras [13]. Under chemically defined conditions, murine ESCs will self-renew or differentiate, respectively, in response to leukemia inhibitory factor (LIF) and fibroblast growth factor (FGF)/extracellular signal-regulated kinase (ERK)-signaling [14]. In contrast, EpiSCs do not respond productively to LIF and can be stably propagated in the presence of FGF and activin [4]. Upon FGF/activin withdrawal and ectopic expression of either Klf2, Klf4 [10], Nanog [15] or Nr5a [16], EpiSCs can robustly revert to naïve 'ground state' pluripotency.
Over the past three decades, it has been difficult to derive PSCs from other mammalian species and all attempts to derive chimeracompetent bovine PSCs have failed [53,54]. Here we explore the possibility that 2i/LIF promotes pluripotency in cattle, a species previously considered non-permissive for PSC-derivation [50].

Results
Bovine iPSC-like colonies form in 2i/LIF medium In order to deliver all four factors into the same cell, we constructed poly-promoter plasmids containing the complete bovine cDNAs for OCT4, SOX2, KLF4, and c-MYC (Fig. 1A (Fig. 1C). Following passaging and re-plating after 48 h, SOX2 immunostaining showed that embryonic and fetal fibroblasts transfected most efficiently (16% and 26%, respectively, Fig. 1D). Lipofection was generally superior with no significant differences between the two plasmids (data not shown). We then passaged lipofected cells onto laminin-coated plates and replaced the somatic with 2i/LIF medium (Fig. 1E). Dome-shaped, tightly packed colonies with clear borders started to appear around 10 days post-transfection and continued to increase in size and number thereafter (Fig. 1F). With the exception of BFF-MBP, all cell lines gave rise to colonies (Fig. 1F). Transfection efficiency correlated well with colony formation (r 2 = 0.9591). We observed no significant differences between the two plasmids and only a small number of BEF40derived colonies after nucleofection (data not shown). Since BEF40 resulted in the highest yield of colonies, we concentrated on this line for subsequent experiments.
Following transduction, 2i/LIF medium has been shown to promote reprogramming into naïve pluripotency in both mouse and human iPSCs [11,50]. We therefore cultured mock-or pOSKM-transfected cells in N2B27 supplemented with or without LIF, PD or CHIR. PD applied alone or in combination with CHIR and LIF greatly decreased phospho-MEK1/2 levels, while total MEK1/2 protein was not affected (Fig. S1). CHIR alone did not modulate phospho-MEK1/2. BEF40 cultured in either N2B27 or N2B27/LIF retained the typical morphology of bovine fibroblasts serum-starved into quiescence [55] (Fig. 2A). Addition of PD or CHIR to non-transfected cells, either with or without LIF, was sufficient to induce formation of colonies, of which 7764% were AP+ ( Fig. 2A). These colonies were in size, morphology and AP staining intensity indistinguishable from transfected colonies cultured under the same conditions ( Fig. 2A). The average area and number of nuclei per AP+ colony on D16, determined from confocal sections of Hoechst-stained samples (n = 14), was 4742962839 mm 2 and 156621, respectively. iPSClike cell nuclei were significantly smaller than their ancestral BEF40 nuclei (2364 mm 2 vs 131615 mm 2 , P,0.001). The number of CHIR-induced AP+ single cells and colonies was higher than for non-supplemented N2B27 or PD alone but not significantly different from cells treated with PD/CHIR or PD/ CHIR + pOSKM (Fig. 2B). PD also induced AP+ colonies but this was not significant (P = 0.20) compared to non-supplemented medium. Culture in 2i/LIF was necessary to induce AP activity, as pOSKM-transfected BEF40 in N2B27 did not give rise to AP+ colonies. When we cultured non-transfected murine embryonic or human skin fibroblasts in LIF, PD, and CHIR, alone or in combination, we saw no morphological transformation into colonies or AP induction (n = 4 independent experiments). Overall, we transfected 2.35610 4 cells/cm 2 with a transfection efficiency of approximately 16% ( = 3.76610 3 transfectants/cm 2 ) and obtained about 15 AP+ colonies/cm 2 , resulting in a reprogramming efficiency of 0.4%.
To better characterize the response of bovine vs comparable murine and human cell types to 2i/LIF, we conducted noninvasive kinetic cell profiling using the xCELLigence TM . This integrated real-time system displays changes in cell proliferation, viability, morphology and adhesion as CI values. In bovine, addition of LIF, PD, or PD/LIF had no significant effect compared to non-treated controls. After normalization on LIFtreated controls, both CHIR and PD/CHIR significantly reduced the CI in different bovine cell types (Fig. 2C). In contrast, PD, CHIR and PD/CHIR all significantly increased the CI in both murine and human fibroblasts (Fig. 2C).
Bovine iPSC-like colonies do not expand in 2i/LIF Next we investigated the mechanism of bovine iPSC-like colony formation. As a proxy for cell proliferation, we quantified DNAsynthesis following different EdU-incorporation protocols. Only AP+ colonies were quantified (Fig. 3A). EdU was added to each transfected culture on D0, 2, 4, 6, 8, 10, 12 or 14 following the addition of 2i/LIF and cells fixed 48 h after labelling ('pulse-fix', Fig. 3B). When the first colonies become visible, only 561% of cells still synthesized DNA and this further declined to 260.6% after two weeks in culture. To determine what proportion of EdUincorporating nuclei became part of iPSC-like colonies, cells were labelled every two days ('pulse'), washed out of EdU ('chase') and further cultured until fixation on D16 ('pulse-chase', Fig. 3B). After a pulse during the first two days in 2i/LIF, 68612% of nuclei within each colony ended up being labelled with EdU on D16. When pulsed around the time of colony formation (D8), only 862% of cells within each colony still synthesized DNA. This proportion further declined to 262% after two weeks in culture, closely matching results from pulse-fix experiments. In order to detect slowly cycling cells, cultures were continuously kept in EdU until fixation at D16 ('cumulative label', Fig. 3B). This demonstrated that most cells in colonies (8160.02%) were cycling at least once during the culture period. However, during the time of primary colony formation only 10% of cells within each colony still synthesized DNA and, in agreement with the pulse-label experiments, colonies labelled towards the end of the culture period (.D12) contained less than 5% proliferating cells. Cumulatively labelled non-transfected control cells, cultured in either 2i/LIF or N2B27, also became nonproliferative over time (Fig. S2A). This indicates that culture conditions, not plasmid-induced reprogramming, induced quiescence. Using immunofluorescence, we further quantified the proportion of cells expressing cell proliferation markers Ki-67 and proliferating cell nuclear antigen (PCNA) in pKMOS-transfected BEF40 cells (Fig. S2B) and colonies (Fig. S2C). For both antigens, the proportion of positive cells progressively declined over time (Fig.  S2D). Using the Click-iT EdU assay, the proportion of apoptotic cells in primary colonies was small (,5% of total nuclei), indicating that most bovine iPSC-like cells in 2i/LIF were neither proliferating nor apoptotic but quiescent (Fig. S2E). Taken together, these results suggest that colonies primarily formed from non-proliferating cells and did not expand further through cell division.
Once colonies had formed, we determined whether their growth was influenced by the culture substrate. Colonies were manually picked and plated on feeder cells or laminin. Individual colonies were tracked for two weeks and their area was determined at regular intervals (Fig. 3C). All tracked colonies stained AP+ at the end of the tracking period. On both substrates there was no significant increase in colony area (n = 13 and n = 4 for feeders vs laminin, respectively). Regression splines plotted for each colony and colony averages showed no significant differences in shape (P = 0.2), indicating that colony growth was not differentially affected by substrate composition.
In order to establish bovine iPSC lines, we passaged colonies using either accutase or microblade dissociation. Initial plating efficiency was high (20/22 = 91% vs 22/22 = 100%, respectively). After first passaging, colonies growing on laminin were tracked for 65 days and their area determined (Fig. 3D). Tracked colonies stained AP+ at the end of the tracking period. In all cases (n = 5 and n = 12 for accutase vs blade, respectively), colony area significantly declined over time (P,0.001). Regression splines plotted for each colony and colony averages showed no significant differences in shape (P = 0.13), indicating that colony growth was not differentially affected by the passaging regime. In summary, colonies that formed from quiescent cells in 2i/LIF showed no signs of expansion for extended periods of time after passaging.

Bovine iPSC-like colonies differentiate in vitro and in teratomas
Bovine iPSC-like colonies formed solid and cystic EBs after 5 days to 3 weeks, respectively, in N2B27 suspension culture (Fig. 5A). These expressed ectoderm-(TUBB3, GFAP, NES), endoderm-(AFP) and mesoderm-(GATA4, MEF2C) markers ( Fig. 5B). Following injection of iPSC-like cells from two independent tranfections into SCID mice, large tissue masses, ranging from 8-20 mm in diameter and 1-5 g in weight, were harvested after seven weeks from two out of four hind legs. Histological examination of the two specimens showed differentiation into ectoderm (epidermis, neural tissue), endoderm (ciliated epithelium) and mesoderm (bone, cartilage) (Fig. 5C). These features were consistent with intramuscular grade 3 teratomas. Bovine-specific primers detected ACTB in both genomic DNA (Fig. S5A) and reverse transcribed cDNA (Fig. S5B) from bovine iPSC-, but not murine ESC-derived, teratomas. This confirms that the tumor originated from bovine cells. We conclude that iPSClike cells are capable of multi-lineage differentiation and production of complex teratomas. Lastly, we karyotyped bovine iPSC-like cells after 24 days in culture and all cells (n = 12/12) showed a normal number of 60 chromosomes (Fig. 5D).

Discussion
Here we report the first reprogramming of non-proliferating bovine cells into pluripotency under conditions of chemically defined signal inhibition. We refer to these cells as ''iPSC-like'' because their capacity for germline chimerism remains to be demonstrated. The first iPSC reports used integrating viruses to carry the pluripotency genes into cells, potentially disrupting endogenous genomic information and causing tumors [56]. Episomal viruses reduce the risk of insertional mutagenesis during iPSC generation [45,57,58]. However, viral DNA may still trigger the immune system [59]. This raises biosafety concerns for agricultural applications where the products (e.g. meat, milk) of iPSC-derived farm animals or their offspring ultimately enter the human food chain. Our constructs avoid lentiviral backbones or foot-and-mouth disease virus 2A oligopeptides [46,60,61] that in many countries would not be permissive for commercial livestock applications. Instead, we used single transfection of a novel expression vector where each bovine iPS factor was flanked by its own independent CMV promoter. Since our work in bovine started, iPSCs in other species have been produced using plasmidmediated approaches. This either involved serial co-transfection of two plasmids containing different sets of factors [44,46,49] or transfection of a polycistronic vector transcribing several factors from a single promoter [44,49,61]. In some cases, transfection approaches have resulted in iPSCs with no evidence of transgene integration [46,61]. As primary bovine iPSC-like colonies were likely of mixed origin, i.e. derived from more than a single reprogrammed fibroblast, and generation of clonal cell lines was unsuccessful, we did not attempt to identify non-integrative iPSC clones. Presence of the transgene in genomic DNA of D38 colonies indicated that at least some cells had stably integrated the vector. Compared to virus-mediated delivery methods, plasmids are technically simple and relatively cheap, eliminating the need for specialized biohazard containment facilities to produce viral stocks [21]. Overall, bovine reprogramming efficiency, corrected for transfection efficiency, was 0.4%. This was consistent with virusand plasmid-mediated approaches (ranging from 0-8% and averaging 1% [44,46,48,62]). It remains to be determined if the new poly-promoter design and use of isogenic bovine sequences results in higher reprogramming efficiency of bovine cells than using alternative vector designs carrying homologous mouse or human iPS genes.
Following transduction, 2i medium promotes reprogramming into naïve pluripotency by neutralizing inductive differentiation stimuli in both mouse and human iPSCs [11,50]. This effect appeared even more pronounced in bovine cells. CHIR and, to a lesser extent, PD induced colony formation and AP activity in bovine, but not murine or human, fibroblasts. How these two compounds elicit the transformation from single cells to compacted colonies, including changes in cell shape and size, migration and intercellular clustering, is not clear. It would be important to better characterize the dependency of bovine cells on MEK, GSK3B and components of their respective signaling pathways. Likewise, the signaling cascade leading to AP induction in bovine cells is not known. AP induction was unrelated to the morphological and molecular changes that caused colony formation, as it also occurred in single cells. High level of AP expression is a fairly non-specific marker for PSCs [63] that appears very early during the iPSC reprogramming [43]. More stringent molecular markers of pluripotency, such as telomerase activity, NANOG, and SALL4, were only induced when the cells were transfected with iPS vectors in the presence of 2i/LIF. This also applied to DPPA3, SOCS3 and ZFP42, all discriminatory markers of naïve pluripotency [10]. Silencing ectopic gene expression and activating endogenous iPS factors is considered another hallmark of full reprogramming [64]. We observed that SOX2, OCT4 and KLF4 transcripts predominantly originated from the endogenous loci. This indicates progressive epigenetic silencing of the CMV promoter, as previously described after transient transfection of ESCs [65]. Cell surface antigens SSEA-3/4 and TRA-1-60/81 also appear late during reprogramming and are considered among the most definitive markers of fully reprogrammed iPSCs [64]. Bovine iPSC-like cells strongly expressed both marker sets, similar to undifferentiated iPSCs in pig [35], monkey [39,40], and human [41] but different from mouse [17] and rat [32,33] which express SSEA-1 instead. Lastly, diagnostic markers of human embryo-derived stem cells and rodent EpiSCs (FGF5, T-BRACHYURY, and LEFTY2) were not detectable in 2i/ LIF-cultured bovine iPSC-like cells, providing additional molecular evidence of reprogramming into pluripotency.
A molecular link between pluripotency and the capacity for unlimited self-renewal is the presence of telomerase. This ribonucleoprotein is specifically active in immortal cells, such as cancer, germ cells and PSCs [66,67], stabilizing telomere length and extending cellular life span [68]. Bovine iPSC-like cells exhibited telomerase activity similar to mouse ESCs, suggesting that they were poised for long-term proliferation. The formation of large solid teratomas from a few thousand injected colonies indicates that bovine iPSC-like colonies did not irreversibly lose their proliferation potential in 2i/LIF and can resume cell division in the right environment.
Under serum-free 2i/LIF conditions, ERK-reliant cell types (e.g. MEFs) either die or become quiescent [50]. Murine and human iPSCs, on the other hand, arise from rapid proliferation [62,69,70]. In fact, it has been suggested that proliferation promotes pluripotency induction, whereas cell cycle arrest inhibits reprogramming and induces irreversible differentiation [70]. In mouse and human, individual reprogrammed cells first start to divide faster, getting smaller in the process and then giving rise to primary colonies through symmetric cell divisions [62,69,70]. As primary colonies expand, cells detach and form secondary colonies elsewhere [62]. Consequently, the number of primary colonies remains constant after some time, while the number of secondary colonies continues to increase [62]. We also observed a reduction in nucleus size as fibroblasts converted into iPSC-like cells, most likely during the first few days in 2i/LIF when most cells were still cycling. However, several lines of evidence suggest that proliferation was unlikely to play a major role during colony biogenesis: i) 95% of cells did not synthesize DNA when colonies formed (EdU pulse-fix); ii) .90% of cells present in colonies on D16 were outside S-phase one week earlier, around the time of colony formation (EdU pulse-chase); iii) .90% of cells within the population did not synthesize DNA during the week following colony formation (EdU cumulative labelling); iv) the proportion of Ki-67 and PCNA expressing cells decreased over time, correlating well with the decline in EdU-incorporation; v) primary colonies did not expand for several weeks before or after passaging; vi) the number of primary colonies stabilized after some time with no subsequent multiplication into secondary colonies. Irrespective of the EdU labelling protocol, ,2% of cells still synthesized DNA in D16 colonies. At that time, almost all cells within each colony were Figure 2. 2i/LIF affects colony formation and AP induction. (A) AP+ colonies after no, empty vector (pCMVe) or pOSKM-transfection. Scale bar = 100 mm; (B) Quantification of AP+ iPSC-like colonies, 15 days after passaging BEF40; ab treatments differ P,0.005 for colonies. (C) xCELLigence TM real-time kinetic profiling. Cell indices were determined for bovine and murine embryonic (BEF40, MEF, respectively), bovine fetal (BFF) and bovine, murine and human adult fibroblasts (AESF, MLF, and BJ, respectively). Curves were normalized after compound addition (blue vertical line) and slopes (red bars) determined during log phase (between blue and red vertical lines); * = slopes differ P,0.05 from LIF-treated control. doi:10.1371/journal.pone.0024501.g002 non-apoptotic and positive for SSEA-3/4 and TRA-1-60/81, providing direct evidence for the quiescence of molecularly reprogrammed iPSC cells. The kinetic response of bovine fibroblasts vs their murine and human counterparts was also characterized using an xCELLigence TM electronic cell sensor array. Based on their reduced CI values, bovine cells significantly diverged from mouse/human in their overall cell proliferation, morphology, and/or adhesion response to PD and CHIR, but not LIF, addition. This provides further evidence for previously unidentified differences in signaling pathways between these species. Varying PD and CHIR concentration may alleviate their anti-proliferative effect, perhaps through minimizing side effects on other kinases. Excluding cell proliferation as the main mechanism of colony formation, bovine iPSC colonies in 2i/LIF likely formed through migration and aggregation of individual cells that had become quiescent before or during re-acquisition of pluripotency. It is conceivable that quiescence may have even enhanced cell reprogrammability, similar to its beneficial effect in nuclear transfer-induced epigenetic reprogramming [71,72].
Several applications await bovine iPSC-like cells. First, they provide a chemically defined screening platform for candidate factors that maintain both proliferation and pluripotency of pluripotent stem cells in livestock. This is particularly relevant for establishing embryo-derived stem cells. Second, they may be converted into animals by using them as donors for somatic cloning with the prospect of significantly increasing cloning efficiency compared to conventional donors [73,74]. Bovine iPSC-like cells will also be tested for their ability to generate germline chimeras. Since germline transmission of superior genetics is the most important criterion for breeding, such animals would serve the same purpose as clones. Third, the pluripotent cell state facilitates transgenesis and homologous recombination [75,76,77]. Provided their proliferation block can be overcome, bovine iPSC-like cells would facilitate the precise genetic engineering of farm animals for improved production traits and biopharming. Beyond these agricultural applications, bovine iPSC-like cells would help to provide large animal models for human diseases [78,79,80,81,82,83,84], complementing research currently carried out with laboratory animals.

Plasmid construction
Complete protein coding cDNA sequences of bovine OCT4, SOX2, KLF4 and c-MYC were obtained from GeneBank (accession numbers NM_174580, NM_001105463.1, BC134523, and NM_001046074, respectively). After introducing silent mutations in KLF4 (TCG to TCC at position 1398) and OCT4 (GGT to GGA at position 967) and adding a 59-XhoI and 39-SalI site, respectively, the altered sequences were synthesized (GENEART, Germany). The CDS encoding a zinc finger nuclease was removed from pRK5.GZF1-N and the plasmid ligated using a linker (Fw 59-GGCTAGCTCGAGACGTG-39, Rv 59-GTCGACACGTCTC-GAGCTAGCCGC-39) that added an XhoI site downstream of the CMV promoter. The resulting plasmid was linearized with SacI, and ligated using a linker (Fw 59-CCTAGGGTACCACGT-GAGCT -39, Rv 59-CACGTGGTACCCTAGGAGCT-39) that added a KpnI site upstream of CMV promoter ('pCMVe-K). Following XhoI and SalI digest, the four factors were each inserted into pCMVe-K, resulting in single-factor plasmids (pO, pS, pK, pM). pO and pK were digested with KpnI, and the fragment containing both CMV promoter and vector sequence cloned into KpnI-linearized pS and pM, respectively, resulting in double-factor vectors (pKM, pSO). Both were partially digested with KpnI to isolate one fragment containing the vector, and one containing the two transcription factors with their CMV promoter. The fragments containing two factors were ligated into the linearized plasmids containing the other two, resulting in four-factor vectors (pOSKM, pKMOS). Plasmids were isolated using a PureLink TM HiPure Plasmid Filter Kit (Invitrogen).

Colony tracking and size determination
Colonies were either passaged by mouth pipette-assisted dissociation into large fragments in accutase TM (Millipore, New Zealand) or by cutting with a splitting blade (ESE 020, Bioniche Animal Health, USA) mounted to a micromanipulator (MO-188, Nikon Narishige, Japan). Fragments were cultured in 2i/LIF and photographed at regular intervals. For size determination, bright- . EdU+ nuclei were counted after fixation (solid blue vertical lines). Open arrow indicates first emerging colonies. (C) Prior to passaging, colonies from feeders (black triangles) or laminin (green circles) were tracked. Regression splines were plotted for each colony (left) and averages (right graph). (D) After accutase (black triangles) or blade (green circles) passaging, colonies on laminin were tracked. Regression splines were plotted for each colony (left) and averages (right graph). doi:10.1371/journal.pone.0024501.g003 field images were imported into Image J (http://rsb.info.nih.gov/ ij/) and areas measured using polygon selection.

Cell proliferation assays
Real-time changes in cell number, viability, and morphology were quantified using an RTCA-SP xCELLigence TM system (Roche, New Zealand). For each treatment, cells were seeded in triplicate in 100 ml somatic medium at 2.35610 4 cells/cm 2 onto laminin-coated 96-well E-Plates. In pilot experiments, the peak cell index (CI) for each cell line was determined. Each compound was diluted in pre-warmed medium and added at 1/4 to 1/3 of the peak CI, around 24 h after plating. CI readings were taken every 1 h. Curves were normalized on the respective CI values 30 min after compound addition when temperature of the fresh medium had equilibrated. Curve slope was determined during the interval between normalization time point and plateau phase (log phase), and normalized on the LIF-treated control.
DNA synthesis was assessed using a click-iTH EdU (5-ethynyl-29-deoxyuridine) proliferation assay (Invitrogen). Cells were: fixed with 4% paraformaldehyde (PFA) every two days ('pulse-fix'); washed out of EdU and cultured in medium supplemented with 10 mM thymidine quench ('pulse-chase'); or kept in EdU ('cumulative'). Cells stained without EdU labelling served as a negative control. In addition, we performed immunocytochemistry against Ki-67 and PCNA as described below.

Detection of apoptosis
Apoptosis was examined with the click-iTH TUNEL Alexa FluorH Imaging Assay (C10246; Invitrogen). Cells were fixed, permeabilized and stained according to the manufacturer's  instructions. Cells stained without EdU-incorporation and DNasetreated cells provide negative and positive controls, respectively.

DNA and RNA isolation
For genomic DNA isolation, cells or 50-100 mg of finely ground liquid nitrogen frozen tissue samples were lysed in 100 mM Tris pH 8, 200 mM NaCl, 5 mM EDTA, 0.1% SDS and 1 mg/ml proteinase K at 55uC. After 12-18 hours, samples were digested with RNAse A (10 mg/ml) at 37uC for 30 min, extracted twice with phenol/chloroform/isoamyl alcohol (25:24:1) and ethanol-precipitated. The air-dried pellet was resuspended in 50-100 ml H 2 O and used for PCR. For RNA isolation, cells or 50-100 mg of finely ground liquid nitrogen frozen tissue samples were lysed in TRIZOLH (Invitrogen)and cDNA synthesized as described [89]. Reverse transcriptase was omitted in one sample, each time a batch was processed for cDNA synthesis ('-RT').

PCR and RT-PCR
A Mastercycler Gradient (Eppendorf, Germany) or Thermal cycler (Bio-Rad, New Zealand) was used for PCR amplification using primers shown in Table S1. Ectopic forward primers bind at different positions of vector-specific 59-UTR, between pCMV and the first ATG of the insert (i.e. in pOSKM: pos. 618-711 for OCT4, pos. 2637-2738 for SOX2, pos. 4544-4631 for KLF4 and pos. 6902-6995 for c-MYC). Ectopic reverse primers bind in their respective target gene. Endogenous primers span specifically between in the 59-UTR (SOX2, OCT4) or 39-UTR (KLF4, c-MYC) and the adjacent exon of their respective target gene. The PCR was performed using the following conditions: one cycle denaturation at 95uC for 5 minutes, followed by 35 cycles of 30 seconds at 95uC, 30 sec at 52-63uC (see Table S1 for primerspecific annealing temperatures), 30 seconds at 72uC; 7 min extension at 72uC and cooling to 4uC.

Real-Time RT-PCR
A LightCyclerH (Roche, New Zealand) was used for qPCR amplification and data analysis. All reactions were performed with the LightCyclerH FastStart DNA MasterPLUS SYBR Green I Kit. Primers were designed using LightCyclerH Probe Design 2.0 or NCBI/Primer-BLAST. The ready-to-use ''Hot Start'' Light-CyclerH reaction mix consisted of 0.4 ul of each primer (10 mM), 2.0 ml LightCyclerH SYBR Green I master mix, 5.2 ml DEPC water, 1.0 ml DMSO if required and 1-2.0 ml cDNA template. The following four-segment program was used: 1) denaturation (10 min at 95 uC); 2) amplification and quantification (20 sec at 95 uC, 20 sec at 52-63 uC, followed by 20 sec at 72 uC with a single fluorescent measurement repeated 45 times); 3) melting curve (95 uC, then cooling to 65 uC for 20 sec, heating at 0.2 uC sec-1 to 95 uC while continuously measuring fluorescence); and 4) cooling to 4 uC. Product identity was confirmed by gel electrophoresis and melting curve analysis. For relative quantification, external standard curves were generated from serial 5-log dilutions for each gene in duplicate. One high efficiency curve (3.6$ slope $3.1, R 2 .0.99) was saved for each target gene and imported for relative quantification as described [89].

Telomerase activity
Telomerase activity was determined with the TRAPEZEH kit (Chemicon, USA). Heat-inactivated (85uC for 10 min) samples were used as internal negative controls. Reactions were separated on non-denaturing TBE-based 10% polyacrylamide (19:1) gels, stained with SYBR Gold (Invitrogen) and visualized on a Gel Doc TM 2000 documentation system (Bio-Rad).

Karyotyping
Bovine iPSC-like colonies (D24 post-transfection) were cultured for 2 days in 2i/LIF with 10% serum replacer (Invitrogen), treated with 1.67 mM nocodazole (Sigma) overnight, trypsinized and centrifuged at 1000 rpm for 5 min. The pellet was resuspended in 0.56% KCl solution, incubated at 38uC for 15 min and fixed in 220uC methanol: acetic acid (3:1) at 4uC for 30 min. Washing with fresh fixative was repeated twice before re-suspending the pellet in 500 ml of ice-cold fixative, spreading onto chilled microscope slides and staining with 10% KaryoMAXH Giemsa in Gurr buffer, pH 6.8 (BDH, New Zealand).

Embryoid body (EB) formation
Undifferentiated iPSC-like colonies were picked by mouth pipette and cultured on bacterial-grade Petri dishes (Falcon, USA) for 5-21 days in N2B27. The medium was changed every 2 days.

Teratoma formation
On D24 post-transfection, cells and colonies were harvested using a cell scraper, centrifuged and re-suspended in PBS +1% PVA (10-30k). Using a 23G needle, 100 ml of cell suspension (approximately 1-5610 6 cells per site) was injected intramuscularly into the quadriceps of adult immune deficient (SCID) male mice. After 5-7 weeks, tumors were graded [90], dissected, fixed overnight in Davidson's fixative, embedded in paraffin, sectioned, haematoxylin-eosin stained and analyzed by a pathologist service (Gribbles, New Zealand). Investigations complied with the New Zealand Animal Welfare Act 1999 and were approved by the Ruakura Animal Ethics Committee (AE Application 11849).

Statistical Analysis
All values are presented as mean 6 S.E.M, unless indicated otherwise. Statistical significance was accepted at P,0.05 and determined using the two-tailed t-test with equal variance (Fig. 1,  2, 4). Log ratios of the colony tracking data (Fig. 3) were analyzed using the residual maximum likelihood method in GenStatH (12 th Edition), with the treatments as fixed effects and individual colonies as random effects. Table S1 Primers used for end-point and/or quantitative (q) real-time RT-PCR; * = includes 4% or 10% DMSO for end-point or qPCR, respectively, ND = not determined.

Supporting Information
(DOC)