Simple and efficient method for generation of induced pluripotent stem cells using piggyBac transposition of doxycycline-inducible factors and an EOS reporter system

Authors

  • Tomoyuki Tsukiyama,

    1. Laboratory of Reproductive Biology, Graduate School of Agriculture, Kyoto University, Kitashirakawa Oiwake-Cho, Sakyo-Ku, Kyoto 606-8502, Japan
    Search for more papers by this author
    • Present address: Laboratory for Pluripotent Stem Cell Studies, RIKEN Center for Developmental Biology (CDB), Kobe, Hyogo 650-0047, Japan.

  • Ryota Asano,

    1. Laboratory of Reproductive Biology, Graduate School of Agriculture, Kyoto University, Kitashirakawa Oiwake-Cho, Sakyo-Ku, Kyoto 606-8502, Japan
    Search for more papers by this author
  • Takamasa Kawaguchi,

    1. Laboratory of Reproductive Biology, Graduate School of Agriculture, Kyoto University, Kitashirakawa Oiwake-Cho, Sakyo-Ku, Kyoto 606-8502, Japan
    Search for more papers by this author
  • Narae Kim,

    1. Laboratory of Reproductive Biology, Graduate School of Agriculture, Kyoto University, Kitashirakawa Oiwake-Cho, Sakyo-Ku, Kyoto 606-8502, Japan
    Search for more papers by this author
  • Masayasu Yamada,

    1. Laboratory of Reproductive Biology, Graduate School of Agriculture, Kyoto University, Kitashirakawa Oiwake-Cho, Sakyo-Ku, Kyoto 606-8502, Japan
    Search for more papers by this author
  • Naojiro Minami,

    1. Laboratory of Reproductive Biology, Graduate School of Agriculture, Kyoto University, Kitashirakawa Oiwake-Cho, Sakyo-Ku, Kyoto 606-8502, Japan
    Search for more papers by this author
  • Yasuhide Ohinata,

    1. Precursory Research for Embryonic Science and Technology, Japan Science and Technology Agency, 5 Sanban-Cho, Chiyoda-Ku, Tokyo 102-0075, Japan
    Search for more papers by this author
  • Hiroshi Imai

    Corresponding author
    1. Laboratory of Reproductive Biology, Graduate School of Agriculture, Kyoto University, Kitashirakawa Oiwake-Cho, Sakyo-Ku, Kyoto 606-8502, Japan
    Search for more papers by this author

  • Communicated by: Hideyuki Okano

imai@kais.kyoto-u.ac.jp

Abstract

PiggyBac (PB) transposition of reprogramming factors (Oct3/4 (O), Sox2 (S), Klf4 (K) and c-Myc) is a safe, nonviral method for generating induced pluripotent stem cells (iPSCs). However, compared with retroviral methods, the reprogramming efficiency of the PB-mediated methods is relatively low. In this study, we describe a simple and efficient system for generating high-quality iPSCs by a single transfection of multiple plasmids that does not require the use of a virus, special instruments or skilled techniques. To improve reprogramming efficiency, we modified the components of the polycistronic 2A vectors used in this study and also investigated the combination of another reprogramming-related factor (L-Myc). By simultaneous transposition of multiple PB vectors containing an EOS (early transposon promoter and Oct3/4 and Sox2 enhancers) reporter and modified polycistronic doxycycline (Dox)-inducible factors, we reprogrammed mouse somatic cells with an efficiency higher than is usually obtained with retroviral methods and we established some iPSC lines that contributed highly to chimeras. By using the Dox-inducible system, we also showed that the appropriate elimination of exogenous-factor expression at appropriate time accelerated the induction of Oct3/4 when a combination of OKS and c-Myc vectors were used.

Introduction

Takahashi & Yamanaka (2006) previously showed that four defined transcription factors (reprogramming factors; Oct3/4, Sox2, Klf4 and c-Myc) can reprogram somatic cells into induced pluripotent stem cells (iPSCs). iPSCs have great potential in human regenerative medicine. However, preclinical studies using large animals are necessary before iPSCs can be applied to humans. Although disease models in large animals were expected to be good models for preclinical studies, iPSC lines that can maintain pluripotency without the support of exogenous factors and that can contribute to germline chimera are not yet available for these species. The goal of this study was to establish a system for screening culture conditions that can be used to generate iPSCs in various mammalian species. For this purpose, we chose a doxycycline (Dox)-inducible system (Brambrink et al. 2008; Stadtfeld et al. 2008) and EOS reporter system (Hotta et al. 2009) for generating iPSCs.

The Dox-inducible system can control the expression of reprogramming factors with or without Dox. Therefore, this system could be useful for examining whether the reprogrammed cells can maintain pluripotency without the support of exogenous factors. For example, ‘pre-’iPSCs that were maintained their pluripotency by the exogenous reprogramming factors could be generated by the Dox addition and could be passaged in different culture conditions containing various combinations of growth factors, serum replacements, small-molecule compounds, peptides, feeder layers and extracellular matrices. After the Dox withdrawal, the dependency of these pre-iPSCs to exogenous factors and an optimal culture condition for iPSCs could be analyzed.

In mice, iPSC colonies can be selected by the expression of reporter genes that are inserted into pluripotent genes such as Oct3/4 or Nanog by gene targeting although the procedure is rather complicated (Maherali et al. 2007; Okita et al. 2007; Wernig et al. 2007). For most mammalian species other than mice, however, no ESC lines or iPSC lines leading to the production of transgenic animals have been established. Therefore, the use of large genomic fragments such as bacterial artificial chromosome (BAC) is difficult and the production of transgenic animals having a pluripotent reporter is time-consuming. The EOS contains conserved regions that act as enhancers of reporter genes and is functional in mouse and human ES/iPSCs as a pluripotent reporter (Hotta et al. 2009). Therefore, the EOS reporter system may be useful for selecting pluripotent-stem-cell colonies in various mammalian species as well as for generating mouse iPSCs from genetically unmodified cells.

To simultaneously introduce the Dox-inducible reprogramming factors and the EOS reporter into cells, we used piggyBac (PB) transposition. The PB transposon system (Kaji et al. 2009; Woltjen et al. 2009; Yusa et al. 2009) is a vector system for generating iPSCs, in which integrated constructs including reprogramming factors can be removed by reexpression of transposase (PBase). However, in our preliminary experiments, PB transposition was not as effective at generating iPSCs in mice as the usual method that uses retroviral vectors. To improve the efficiency of iPSC generation by using PB transposition, we focused on the order of introducing genes in the polycistronic 2A vectors used in this study. Okita et al. (2008) reported that the order of genes in polycistronic retroviral vectors, Oct3/4, Klf4 and then Sox2 (OKS), can affect the efficiency of reprogramming and generation of iPSCs. In addition, replacing one (c-Myc) of the four factors with L-Myc was found to result in the generation of a large number of high-quality iPSCs (Nakagawa et al. 2010).

In this experiment, we constructed PB vectors with a Dox-inducible cassette consisting of OKS, c-Myc (cM), L-Myc (LM) or OKS and L-Myc (OKSLM) and compared the reprogramming abilities of these PB vectors on mouse embryonic fibroblasts (MEFs) with the reprogramming ability of a previously reported polycistronic vector containing c-Myc, Klf4, Oct3/4 and Sox2 (MKOS) (Woltjen et al. 2009). We showed that the Dox-inducible PB vector could efficiently reprogram mouse somatic cells compared with retroviral methods and that appropriate shutoff of exogenous genes by the withdrawal of Dox could be accelerated the endogenous Oct3/4 expression.

Results

Improvement of reprogramming efficiency

Fourteen days after the Dox addition, GOF18 MEFs that were transfected with Dox-inducible MKOS, OKS, OKS+cM, OKS+LM or OKSLM vectors formed many mouse ESC-like colonies. Cells transfected with OKS+cM produced far more colonies than cells transfected with other combinations (Fig. 2). Most of these colonies, however, were negative for Oct3/4-GFP and the positive rate was only 0.73 ± 0.23%. By contrast, most of the colonies from cells that were transfected with MKOS (90.61 ± 6.31%), OKS+LM (57.07 ± 5.53%) or OKSLM (66.39 ± 15.16%) were positive for Oct3/4-GFP (Fig. 2).

Figure 2.

 Reprogramming efficiency of MEFs by various PB vectors. (A) The numbers of ESC-like and Oct3/4-GFP-positive colonies at days 14 and 21 in the presence or absence of Dox. (B) The average numbers of ESC-like and Oct3/4-GFP-positive colonies at days 14 and 21 in the presence or absence of Dox. Open bars show the number of total ESC-like colonies, and closed bars show the number of Oct3/4-GFP-positive colonies. +, Dox was withdrawn at day 14; −, Dox was supplemented for the whole term.

Cells were further cultured for 7 days with or without Dox, and the number of colonies was counted at day 21. To avoid overestimating the reprogramming efficiency that was caused by splitting of the culture, Dox treatment was continued for one of the culture duplicates and withdrawn from the other (Fig. 1B). Surprisingly, most of the GFP-negative colonies from cells that were transfected with OKS+cM became GFP-positive when Dox was withdrawn and the positive rate reached to 89.83 ± 4.98% at day 21 (Fig. 2). However, most of the colonies remained GFP-negative when the Dox treatment was continued, while the number of colonies increased. The positive rate was only 4.53 ± 0.45% at day 21 (Fig. 2).

Figure 1.

 Reprogramming of mouse somatic cells and establishment of iPSCs using PB vectors. (A) Construction of PB vectors used in this study. (B) Timeline for the analysis of reprogramming efficiency of cells. (C) Timeline for the establishment of iPSC lines.

Interestingly, when the original MKOS was used as a vector, the number of colonies tended to reach a plateau at around 14 days of culture. In contrast, the colony number continued to increase at least to 21 days when OKS+cM was used (Fig. 2). In the case of OKS+cM combined with the Dox withdrawal, the numbers of Oct3/4-GFP-positive colonies in three trials were 321, 169 and 324 at 21 days of culture and the reprogramming efficiencies were 0.32%, 0.17% and 0.32%, respectively. This efficiency (0.27 ± 0.05%) was higher than that obtained by retroviral vectors [generally 0.01%–0.1% (Hochedlinger & Plath 2009)].

Establishment of iPSC lines

Fourteen days after the Dox addition, MEFs that were transfected with OKS+LM formed many EOS-GFP-positive colonies with mouse ESC-like morphology, whereas MEFs transfected with OKS+cM formed a few EOS-GFP-positive colonies, and many EOS-GFP-negative and ESC-like colonies. In the case of OKS+LM (n = 3), the average numbers of EOS-GFP-positive colonies and ESC-like colonies were 28.33 ± 2.19 and 48.33 ± 6.64, respectively, and the rate of EOS-GFP-positive colonies was 59.68 ± 4.06%, whereas in the case of OKS+cM, the average numbers of EOS-GFP-positive colonies and ESC-like colonies were 25.00 ± 3.61 and 305.33 ± 17.57, respectively, and the rate of EOS-GFP-positive colonies was 8.10 ± 0.75%. These results were consistent with those obtained from the previous section: Improvement of reprogramming efficiency.

To confirm the similarity of Oct3/4-GFP reporter and EOS reporter, we reprogrammed GOF18 MEFs which were simultaneously transfected with EOS-RFP reporter. In the case of OKS+LM, we confirmed that all of EOS-RFP-positive colonies were also Oct3/4-GFP-positive (Fig. 3A–C and S1J in Supporting Information) and most of Oct3/4-GFP-positive colonies were also EOS-RFP-positive (Fig. S1J in Supporting Information). In the case of OKS+cM, although some EOS-RFP-positive and Oct3/4-GFP-negative colonies were appeared at day 14 (Fig. S1A–C in Supporting Information), the appearance of many EOS-RFP-positive colonies after the withdrawal of Dox was the same with that of Oct3/4-GFP (Fig. S1G–I in Supporting Information). All EOS-RFP-positive colonies were also Oct3/4-GFP-positive 7 days after Dox withdrawal (Fig. S1J in Supporting Information).

Figure 3.

 Characterization of established iPSC lines. The morphology of an EOS-RFP- and Oct3/4-GFP-positive colony (A, RFP-positive image; B, GFP-positive image; C, phase contrast image) derived from cells that were transfected with OKS+LM. The morphology of an EOS-GFP-positive colony (D, GFP-positive image; E, phase contrast image) derived from cells that were transfected with OKS+LM. The established iPSC line (F, GFP-positive image; G, phase contrast image) transfected with OKS+LM. Karyotyping of OKS+LM-induced iPSCs (H). Alkaline phosphatase activity in OKS+LM-induced iPSCs (I, activity in a culture dish; J, activity in colonies). OCT3/4 (K, OCT3/4 staining; L, EOS-GFP-positive image; M, Hoechst staining), NANOG (N, NANOG staining; O, EOS-GFP-positive image; P, Hoechst staining) and SSEA-1 (Q, SSEA-1 staining; R, EOS-GFP-positive image; S, Hoechst staining) expression in OKS+LM-induced iPSCs.

Two EOS-GFP-positive colonies from OKS+cM and four EOS-GFP-positive colonies from OKS+LM were picked up (Figs 3D,E and S2A,B in Supporting Information), harvested by trypsinization and transferred to a new SNL feeder layer. Half of these colonies were derived from MEFs that were transfected with TagRFP simultaneously. After one passage, Dox was withdrawn and cells were maintained with N2B27-mESM supplemented with LIF/2i on fibronectin-coated dishes.

All the resulting cell lines showed several iPSC-like cell characteristics: mouse ESC-like morphology, including round shape, large nuclei and scant cytoplasm, and strong expression of EOS-GFP (Fig. 3F,G). These cells showed normal karyotype (Figs 3H, S2G and S3E in Supporting Information, 2n = 40). EOS-GFP expression was low in cells from colonies with partial EOS-GFP expression (Fig. S2A–D in Supporting Information). However, the EOS-GFP-expressing cells were readily concentrated by puromycin selection, indicating that selection by the EOS reporter was also functional (OKS+cM pu, Fig. S3A,B in Supporting Information). The EOS-GFP-expressing cells grew well beyond several months without serum or a feeder layer and showed strong alkaline phosphatase activity (Figs 3I,J, S2E,F and S3C,D in Supporting Information). Immunocytological analysis also showed that the cells were positive for mouse ESC markers, such as OCT3/4, NANOG and SSEA-1 (Figs 3K–S, S2H–P and S3F–N in Supporting Information). Semiquantitative RT-PCR analysis showed that the iPSCs expressed many undifferentiated ESC-specific genes, including Oct3/4, Sox2, Klf4, Nanog, Rex1 (also known as Zfp42), Fgf4, Lefty1 and ERas (Fig. 4A), and did not express fibroblast-specific genes such as Fibrillin2 (Fbn2) and collagen, type V, alpha 2 (Col5a2) (Fig. 4A). Furthermore, quantitative real-time RT-PCR analysis showed that exogenous transcription factors were down-regulated in the iPSCs when Dox was withdrawn from the culture medium (Fig. 4G). Interestingly, endogenous Oct3/4, Sox2 and Klf4 were up-regulated in the absence of Dox (Fig. 4B,D), whereas endogenous c-Myc and L-Myc were down-regulated in the same condition (Fig. 4E,F).

Figure 4.

 Endogenous and exogenous gene expression specific for undifferentiated ESCs in iPSCs. To distinguish the amplification from genomic contamination, primers for the expression of Nanog, Rex1, Fgf4, ERas, Fbn2 and Col5a2 were designed to span intron–exon boundaries. A OKS+cM-induced iPSC line after puromycin selection was denoted by pu. iPSC lines simultaneously transfected with a TagRFP were denoted by RFP. Quantitative real-time RT-PCR analysis for endogenous reprogramming factors and transgenes (B-G). Relative gene expressions of endogenous reprogramming factors (B, Oct3/4; C, Sox2; D, Klf4; E, c-Myc; F, L-Myc) to ESCs. Relative transgene expressions in iPSCs (G) to the cells that were cultured continually in the presence of Dox. Error bars indicate SE between three technical replicates. *< 0.05, Student’s t-test.

To analyze pluripotency of the iPSC lines, we assayed cell differentiation in vitro and in vivo. First, these cells formed floating EBs, which differentiated into cells expressing glial fibrillary acidic protein (GFAP) (an ectoderm marker), actin smooth muscle (ASM) (a mesoderm marker) and α-fetoprotein (AFP) (an endoderm marker), indicating that the iPSCs differentiated into three germ layers (Figs 5A,C and S5A–C in Supporting Information). Second, we injected the iPSCs subcutaneously into nude mice. Four weeks after injection, these cells formed teratomas containing cells from all three germ layers, including skin epithelium structures (ectoderm), cartilage structures (mesoderm) and ciliated epithelium structures (endoderm) (Figs 5D–F and S5D–F in Supporting Information). We confirmed that all iPSC lines could form teratomas (Fig. S4 in Supporting Information). Finally, the iPSCs were injected into mouse blastocysts and the resulting embryos were transferred into five pseudopregnant mice. A RFP OKS+cM clone and a RFP OKS+LM clone that showed high EOS-GFP and RFP fluorescence were used for injections to host blastocysts. Twenty-eight live OKS+LM embryos and 24 live OKS+cM embryos were obtained from two pseudopregnant mice at E13.5. Twenty-six of the OKS+LM embryos and five of the OKS+cM embryos were chimeras judging by their RFP fluorescence and eye color. The OKS+LM chimeras showed markedly high chimerism (Figs 5G,H and S5G,H in Supporting Information) and the genital ridges showed high RFP fluorescence, indicating that the iPSCs differentiated to genital tissues (Figs 5I–K and S5I–K in Supporting Information). Live pups provided further evidence of chimerism as shown by their coat colors from three pseudopregnant mice (Figs 5L and S5L in Supporting Information). Finally, we confirmed the germline transmission of the OKS+LM iPSCs judged by coat color in the progenies of chimeras (Fig. 5M, 7, 2 and 6 germline chimeras per 11, 9 and 11 newborn mice, respectively). Although germline progeny was not obtained from OKS+cM chimeras to date, RFP-positive elongated spermatids were observed in the lumina of the seminiferous tubules of OKS+cM chimeras (Fig. S5M–R in Supporting Information). These data clearly showed that our established iPSC lines were authentic iPSC lines. However, we could not detect clear EOS-GFP fluorescence in the genital ridges (Figs 5I–K and S5I–K in Supporting Information), indicating that an EOS reporter might not be functional in E13.5 primordial germ cells (PGCs).

Figure 5.

 Pluripotency of established iPSC lines. Immunohistochemical staining by three-germ-layer markers in differentiated cells derived from OKS+LM-induced iPSCs. Glial fibrillary acidic protein (A, ectoderm), actin smooth muscle (B, mesoderm) and α-fetoprotein (C, endoderm) were used for the markers. The derivative of three germ layers (D, skin epithelium structures; E, cartilage structures; F, ciliated epithelium structures) in teratomas derived from OKS+LM-induced iPSCs. Contribution of OKS+LM-induced iPSCs to mouse embryonic development (G, Bright-field image; H, RFP image). Embryos were analyzed by fluorescence microscope at E13.5. Contribution of OKS+LM-induced iPS cells to genital ridges (I, Bright-field image; J, GFP image; K, RFP image) at E13.5. Live chimeric mice (L) derived from OKS+LM-induced iPSCs. Germline-transmitted F1 progenies (M) derived from OKS+LM-induced iPSCs.

Discussion

The preceding results show that it is possible to use PB transposition of Dox-inducible reprogramming factors and an EOS reporter to establish iPSC lines with high efficiency and without the use of viral vectors.

Interestingly, in the case of OKS+cM, many Oct3/4-GFP-negative and ESC-like colonies appeared when exogenous factors were expressed. However, these colonies turned out to be Oct3/4-GFP-positive when Dox was removed from the culture and the expression of exogenous factors was turned off. Possible explanations of this phenomenon include the following: (i) intensive expressions of exogenous factors could suppress endogenous Oct3/4 expression directly by negative feedback regulation (Pan et al. 2006), (ii) residual expressions of exogenous factors in reprogrammed cells could impede the correct transcriptional network in these cells and result in suppression of endogenous Oct3/4 indirectly (Sridharan et al. 2009), and (iii) in the case of OKS+cM, partially reprogrammed cells could be generated more easily when exogenous factors were expressed and acquisition of full pluripotency was accelerated by the shutoff of exogenous factors.

Reprogramming by OKS+cM requires introduction of at least four vectors (OKS, cM, rtTA and PBase) into each cell, whereas MKOS can be transfected by three vectors (MKOS, rtTA and PBase). Therefore, we expected that MKOS would be more efficient at transfecting MEF cells than OKS+cM. However, the efficiency of reprogramming by OKS+cM was much higher than that by MKOS (Fig. 2). This indicates that OKS+cM is more effective at reprogramming than MKOS and that the expression levels of each reprogramming factor might be better with OKS+cM. It should be mentioned that OKS+LM was much more efficient at reprogramming than OKSLM (Fig. 2), which emphasizes the importance of the levels of each reprogramming factor. However, increasing the number of insertions of PB vectors into a genome makes it difficult to remove these vectors by reexpression of PBase (Woltjen et al. 2009). Two strategies have been proposed to eliminate inserted PB vectors efficiently. One is negative selection of vector-free cells by herpes virus thymidine kinase (HSVtk)-fialuridine (FIAU) selection system (Yusa et al. 2009) and the other is reducing the number of vectors (Woltjen et al. 2009). To reduce the number of vectors, further studies are needed to determine the optimal levels of each reprogramming factor and the optimal order of the four reprogramming factors in a single polycistronic vector.

The iPSCs used for chimera production in this experiment have a hygromycin resistance gene behind the TagRFP gene (Fig. 1A). Therefore, differentiated cells that are derived from the iPSCs can be recovered from the chimera by hygromycin selection. The selected cells already have the Dox-inducible factors and can be used to induce reprogramming by only the addition of Dox. The ‘secondary’ reprogramming system is useful for analyzing reprogramming under genetically identical conditions (Brambrink et al. 2008; Stadtfeld et al. 2008).

Here, we showed that only one transfection of multiple plasmids containing the minimal and essential reporter and reprogramming factors is sufficient for production of iPSCs. We constructed a PB vector containing an EOS reporter that was originally used in the lentiviral vector system (Hotta et al. 2009). We confirmed in the present experiment that the PB-EOS vector was functional in generation of mouse iPSCs and that additional reporters and selection markers were not necessary. EOS reporter contains short gene regulatory regions that are conserved in various mammalian species (Hotta et al. 2009). EOS may also be used as a universal pluripotent reporter in various mammalian species. Therefore, the system could be used to generate iPSCs in various mammalian species. Indeed, we have observed EOS reporter expression during iPSC generation in pig, dog, rabbit and cattle (data not shown). In lentiviral system, however, it is essential to use special instruments for biological containment and it requires complicated procedures, whereas PB system does not require high containment level and it requires only one transfection. Because the facility for generation of transgenic large animals that has high containment level is extensive and expensive, PB system may be useful for generation of transgenic large animals.

In large domestic animals, some iPSC lines have been established but all of them maintain their pluripotency under the support of exogenous factors (Esteban et al. 2009; Ezashi et al. 2009; Shimada et al. 2009; Wu et al. 2009). It is necessary to determine an optimal culture condition for the establishment and maintenance of transgene-independent iPSCs. In previous study, Yusa et al. (2009) also successfully reprogram mouse somatic cells with high efficiency by PB system. However, because they used a constitutive CAG promoter to drive the reprogramming factors, it is difficult to regulate transgene expressions. In contrast, our system can easily regulate these expressions and enable us to analyze whether reprogrammed cells can maintain pluripotency without the expression of exogenous transcription factors. Therefore, our system can be applied to screen an optimal culture condition for generating authentic iPSCs in various mammalian species.

Experimental procedures

Vector construction

To prepare the OKS PB vector (PB-TET-OKS, Fig. 1A), pCX-OKS-2A was digested by EcoRI, blunted and introduced into blunt-ended PB-TET-MKOS at the BsrGI site. To prepare the OKSLM PB vector (PB-TET-OKSLM, Fig. 1A), PB-TET-OKS was amplified by PCR using a reverse primer eliminating the Sox2 stop codon, and L-Myc containing a 2A sequence in the upstream was amplified by PCR using a forward primer containing a 2A sequence (Table S1 in Supporting Information). Then, the ligation of these DNA fragments was performed. To prepare the L-Myc PB vector (pPB-TET-L-Myc, Fig. 1A), the protein coding regions of these genes were amplified by PCR using oligonucleotide primers as described in Table S1 in Supporting Information. The amplified DNA fragments were digested by XhoI and NotI and introduced into pPBhCVM*1-cHApA at the XhoI and NotI sites. To prepare the EOS PB vector (pPB-EOS-C(3+)-EiP, Fig. 1A), PL-SIN-EOS-C(3+)-EiP were digested by BamHI and ClaI, blunted and introduced into the blunt-ended empty PB vector. To prepare the TagRFP PB vector (pPBCAG-TagRFP-IH, Fig. 1A), pTagRFP-N were digested by XhoI and NotI and introduced into pPBCAG-cHA-IH at the XhoI and NotI sites. To prepare the EOS-RFP PB vector, pPBCAG-cHA-IP were digested by BglII and BamHI and introduced into pPB-EOS-C(3+)-EiP vector at the BglII and BamHI site. pTagRFP-N were digested by BglII and NotI and introduced into pPBCAG-cHA-IP at the BglII and NotI sites. Then, two resulting vectors were digested by BglII and ClaI and ligated.

The original vectors (PB-TET-MKOS, PB-CAG-rtTA Adv, pCX-OKS-2A and PL-SIN-EOS-C(3+)-EiP) were obtained from Addgene (plasmid 20959, plasmid 20910, plasmid 19771 and plasmid 21313) (Okita et al. 2008; Hotta et al. 2009; Woltjen et al. 2009). The empty PB vector, the c-Myc PB vector (pPB-TET-c-Myc), pPBhCVM*1-cHApA, pPBCAG-cHA-IH and pPBCAG-cHA-IP were kindly gifted from Dr Hitoshi Niwa at RIKEN Center for Developmental Biology. The PBase vector (pCAG-PBase) was kindly gifted from Dr Austin Smith at University of Cambridge (Guo et al. 2009).

Preparation of MEFs

For preparation of MEFs, embryos were collected at embryonic day (E) 13.5–15.5. After removal of the heads and visceral tissues, the remaining bodies were washed in fresh PBS and minced and the isolated cells were maintained in Dulbecco’s modified Eagle medium (DMEM, Invitrogen) containing 10% fetal bovine serum (FBS, JRH Biosciences), penicillin (Sigma) and streptomycin (Sigma).

Analysis of reprogramming efficiency

To examine the effect of vector constructions on the efficiency of iPSC generation, MEFs that originally having Oct3/4-GFP reporter (GOF18 MEFs) were used. GOF18 MEFs were plated at 1 × 105 cells per 35-mm dish and incubated overnight. Cells were then transfected using 16 μL of Lipofectamine LTX transfection reagent (Invitrogen) and 2 μL of Plus reagent (Invitrogen) mixed with 1 μg of PBase vectors (pCAG-PBase) and 1 μg of reprogramming PB vectors (PB-TET-MKOS, PB-TET-OKS, the combination of PB-TET-OKS and pPB-TET-c-Myc (OKS+cM), the combination of PB-TET-OKS and pPB-TET-L-Myc (OKS+LM) or PB-TET-OKSLM) and rtTA PB vectors (PB-CAG-rtTA Adv).

The transfected cells were cultured in 10% FBS DMEM supplemented with Primocin (InvivoGen). One day after transfection, the medium was replaced by Glasgow minimum essential medium (GMEM, Sigma) containing 10% FBS, 2 mm l-glutamine (MP Biomedicals), 1 mm sodium pyruvate (Sigma), 1× MEM nonessential amino acids (NEAA, Invitrogen), 0.1 mm 2-mercaptoethanol (2-ME, Wako), penicillin and streptomycin (10% FBS GMEM ESM) supplemented with 1.5 μg/mL doxycycline (Dox, Sigma). Seven days after the Dox addition, the medium was replaced by GMEM containing 15% Knockout Serum Replacement (KSR, Invitrogen), 0.3% FBS, 2 mm l-glutamine, 1 mm sodium pyruvate, 1× NEAA, 0.1 mm 2-ME, penicillin and streptomycin (mouse ESC medium, mESM) supplemented with leukemia inhibitory factor (LIF). As a source of LIF, we used the commercially available LIF (ESGRO (1000 U/mL), Millipore) or LIF-conditioned medium (1 : 1000 dilution) from COS-7 or 293FT cell cultures that had been transduced with a mouse LIF-encoding vector (Niwa et al. 1991; Yoshida-Koide et al. 2004). The medium was changed every day.

The number of ESC-like colonies and the expression of Oct3/4-GFP were compared 14 days after the Dox addition (d14) (Fig. 1B). In this experiment, to measure reprogramming efficiency precisely, these cells were not passaged during reprogramming procedure because reprogramming efficiency was overestimated by splitting of culture (Smith et al. 2010).

Establishment of iPSC lines using the EOS reporter

Reprogramming of GOF18 MEFs with EOS-RFP reporter was performed according to Analysis of reprogramming efficiency.

To efficiently establish mouse iPSC lines from MEFs that did not have reporter genes, 1 × 105 cells of 129B6F1 MEFs were transfected with OKS+cM or OKS+LM and simultaneously with PB-CAG-rtTA Adv, pCAG-PBase and the EOS PB vector (pPB-EOS-C(3+)-EiP). For the analysis of teratoma formation and chimera production, a part of cells were also simultaneously transfected with a constitutive TagRFP-expressing PB vector (pPBCAG-TagRFP-IH).

One day after transfection, Dox was supplemented into the culture. Four days after the Dox addition, 1 × 105 cells (approximately 1/8 of dish) were reseeded on the SNL feeder layer and the medium was replaced by mESM supplemented with LIF at the next day (Fig. 1C). Fourteen days after the Dox addition, primary colonies were picked up and chemically dissociated and transferred on the SNL feeder layer in 48-well plates. Following the passage, Dox was withdrawn and the established iPSCs were maintained with 50% Neurobasal medium (Invitrogen) and 50% DMEM/F12 (Invitrogen) containing 0.5× N2 (Invitrogen), 0.5× B27 (Invitrogen), 1×l-glutamine, penicillin and streptomycin (Invitrogen), 0.05% BSA (Sigma) and 0.15 mm 1-thioglycerol (Sigma) (N2B27-mESM) supplemented with LIF, 1 μm PD0325901 and 3 μm CHIR99021 (2i, Stemgent) on human plasma fibronectin (Millipore)-coated dishes. The medium was changed every day.

Alkaline phosphatase staining and immunohistochemical staining

Alkaline phosphatase staining was performed using the Vector Alkaline Phosphatase Substrate kit (Vector) according to the manufacturer’s instructions.

For immunofluorescence analysis, cells were fixed with PBS containing 3.7% paraformaldehyde for 10 min at room temperature. After washing with PBS, cells were blocked with 10% goat serum and 0.1% Triton X-100 for 45 min at room temperature and were treated with primary antibodies overnight at 4 °C. Primary antibodies included Oct3/4 (1 : 25, SC-5279, Santa Cruz), Nanog (1 : 250, AB5731, Millipore), SSEA-1 (1 : 25, MC-480 (SSEA-1)-s, Developmental Studies Hybridoma Bank), glial fibrillary acidic protein (GFAP, 1 : 100, Z0334, DAKO), actin smooth muscle (ASM, 1 : 1000, MS-113-P0, Thermo) and α-fetoprotein (AFP, 1 : 100, MAB1368, R&D Systems). Alexa Fluor 594 goat anti-mouse IgG or IgM (1 : 500, Molecular Probes) and Alexa Fluor 594 goat anti-rabbit IgG (1 : 500, Molecular Probes) were used as secondary antibodies. Nuclei were stained with 1 μg/mL Hoechst 33342 (Sigma).

Semiquantitative RT-PCR and quantitative real-time RT-PCR

Total RNAs of cells were prepared using an RNeasy Mini Kit (QIAGEN). DNase (TURBO DNA-free Kit, Ambion) was added to preparations to avoid genomic contamination. For reverse transcription, ReverTra Ace (Toyobo) and Random Primer (Invitrogen) were used, according to the manufacturer’s instructions. PCR was carried out with ExTaq (Takara). For real-time PCR, Power SYBR Green PCR Master Mix (Applied Biosystems) was used. Transcription levels were determined in triplicate reactions and normalized to Gapdh. Primer sequences are shown in Table S1 in Supporting Information.

Differentiation of iPSCs in vitro

To form embryoid bodies (EBs), established iPSCs were harvested by trypsinization and transferred to MPC-treated round-bottom dishes (Nunc) in Iscove’s modified Dulbecco’s medium (IMDM, Invitrogen) containing 15% FBS, 2 mm l-glutamine, 1 mm sodium pyruvate, 1× NEAA, 0.1 mm 2-ME, penicillin and streptomycin. The medium was changed every other day. After 3 days of culture, the floating EBs were transferred to gelatin-coated plates and cultured in the same medium for another 3 days. Immunohistochemistry was analyzed as described in Alkaline phosphatase staining and immunohistochemical staining.

Teratoma formation

iPSCs were harvested by trypsinization, transferred into tubes and centrifuged, and the resulting cell pellets were suspended in 10% FBS DMEM. One × 106 cells were injected subcutaneously into the dorsal flank of nude mice (BALB/c nude (nu/nu), Oriental BioService). Four weeks after injection, tumors were dissected, weighted and fixed with PBS containing 3.7% paraformaldehyde. Paraffin-embedded tissues were sliced and stained with hematoxylin and eosin.

Chimera production

To generate chimeric mice, iPSCs were injected into embryos (CD1, Oriental BioService) at the blastocyst stage. The blastocysts were transferred into the uterine horns of five pseudopregnant mice (CD1 background). Chimerism of live embryos and pups were judged by RFP fluorescence, eye color and coat color. After mating with C57BL/6 mice, germline transmission of iPSCs was judged by coat color of progeny. The contribution of RFP-positive cells into the lumina of seminiferous tubules was analyzed by confocal fluorescence microscopy.

Acknowledgements

We thank the following for kindly providing materials for this study: Dr Allan Bradley (Sanger Institute) for the SNL feeder cells, Drs. Takashi Yokota (Kanazawa University) and Jun-ichi Miyazaki (Osaka University) for the mouse LIF-encoding vector, Dr Hitoshi Niwa (RIKEN Center for Developmental Biology) for the empty PB vector, pPB-TET-c-Myc, pPBhCVM*1-cHApA, pPBCAG-cHA-IH and pPBCAG-cHA-IP, and Dr Austin Smith (University of Cambridge) for the PBase vector (pCAG-PBase). This study was supported by a grant from the Japan Society for the Promotion of Science (Research Fellowship Program) to T.T.

Ancillary