Disruption of MSSP, c-myc single-strand binding protein, leads to embryonic lethality in some homozygous mice


  • Mitsuaki Fujimoto,

    1. Graduate School of Pharmaceutical Sciences,
    2. CREST, Japan Science and Technology Corporation, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan
    Search for more papers by this author
  • Ken-ichi Matsumoto,

    1. Graduate School of Pharmaceutical Sciences,
    2. CREST, Japan Science and Technology Corporation, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan
    Search for more papers by this author
  • Sanae M. M. Iguchi-Ariga,

    1. College of Medical Technology, Hokkaido University, Kita 12, Nishi 6, Kita-ku, Sapporo 060, Japan
    2. CREST, Japan Science and Technology Corporation, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan
    Search for more papers by this author
  • Hiroyoshi Ariga

    Corresponding author
    1. Graduate School of Pharmaceutical Sciences,
    2. CREST, Japan Science and Technology Corporation, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan
    Search for more papers by this author

  • Communicated by: Hiroshi Handa

*: E-mail: hiro@pharm.hokudai.ac.jp


Background MSSP, c-myc single-strand binding protein, works as a factor for DNA replication, transcription, apoptosis induction, and myc/ras cooperative transformation. The cDNAs encoding four of the family proteins, MSSP-1, MSSP-2, Scr2 and Scr3, were cloned. These proteins possess two copies of putative RNA binding domains, RNP-A and RNP-B, and these RNA binding domains have been suggested to be indispensable to the functions of MSSP.

Results To elucidate its role in vivo, we generated Mssp knockout mice by homologous recombination in embryonic stem cells. Although intercrossing of Mssp+/– mice gave rise to mice homozygous to the mutant Mssp allele (Mssp−/–) and the Mssp−/– mice, once born, did not display an overt phenotype, the ratio of littermates born among Mssp+/+, Mssp+/– and Mssp−/– mice was 1 : 1.6 : 0.5, which is not a typical Mendelian ratio. When E2.5 embryos from the pregnant mice were cultured in vitro for 5 days, the inner cell mass and trophoblast giant cells in wild-type (Mssp+/+) E2.5 embryos developed normally. However, Mssp−/– E2.5 embryos displayed significant defects in growth and development. Since Mssp was expressed in uterine gland-transported glycogen, we evaluated the hormonal state of wild-type and Mssp−/– mice. The progesterone concentration of Mssp−/– mice was decrease to 6.5% of that of wild-type mice at E2.5.

Conclusions These results suggest that the deletion of the mssp gene results in both the growth defect in the embryo and the hormonal defect in adult female mouse. The embryonic defect and a decreased concentration of progesterone in female mice reflect a development defect of the pre-implantation embryo in Mssp−/– mice, thereby leading to embryonic lethality.


A sequence of 21 bp about 2 kb upstream of the human c-myc gene, has been identified as being of putative DNA replication origin and a transcriptional enhancer (Iguchi-Ariga et al. 1988; Negishi et al. 1992). Several proteins that bind directly to the single- or double-strand of 21 bp were identified and named MSSP. The human cDNAs encoding MSSP-1, MSSP-2 (Negishi et al. 1994; Takai et al. 1994), Scr2, Scr3 (Kanaoka & Nojima 1994), and RBMS3 (Penkov et al. 2000) have been independently cloned and shown to be MSSP family proteins (Haigermoser et al. 1996). MSSP-1, MSSP-2 and Scr2 are alternative splicing variants of the human MSSP gene2, which is comprises 15 exons distributed over more than 60 kilo base pairs (Haigermoser et al. 1996). Furthermore, the chromosome location of the human MSSP gene2 has been shown to be at 2q24 (Fujimoto et al. 2000). Each of these five proteins of the MSSP family contains two consensus motifs of an RNA-binding protein, RNP (Bandziulis et al. 1989; Dreyfuss et al. 1988). The MSSPs that bind to c-Myc via the RNP motif have been shown to stimulate the myc/ras cooperative transforming activity (Niki et al. 2000a), induce apoptosis (Iida et al. 1997), and stimulate DNA replication by binding to DNA polymerase alpha (Niki et al. 2000b). The expression of MSSP is induced by mitogens such as serum, and MSSP is expressed after the middle G1 phase of the cell cycle as is c-Myc. These findings suggest that MSSPs function in connection with c-Myc. Moreover, MSSPs suppressed transcription of the α-smooth muscle actin gene (Kimura et al. 1998) and inhibited the E-box-dependent transcription activity of c-Myc by converting the Myc/Max complex to an MSSP/Myc/Max ternary complex (Niki et al. 2000a). Thus, MSSPs have pleiotropic functions in cells.

In mice, two cDNAs encoding MSSP and Scr3 have been cloned. MSSP and Scr3 have homologous amino acid sequences around the N-terminus region containing the RNP, while there is little homology of the sequences near the C-terminus of the proteins (Fujimoto et al. 2000). The mouse Mssp gene was found to comprise at least 13 exons spanning 40 kilo bases (Fujimoto et al. 2000). Mouse MSSP was found to be expressed ubiquitously in all the tissues, but the shorter forms of MSSP mRNAs were identified in mouse testis. Shorter forms of the mRNAs expressed in the testis have also been reported in genes, including AMY-1 (Taira et al. 1998) and AKAP (Lin et al. 1995), and these were reported to be derived from the alternative usage of the poly A adenylation signal. AMY-1 and AKAP were suggested to be involved in spermatogenesis.

To elucidate the role of MSSP in vivo, we have generated MSSP-deficient mice by disrupting the Mssp gene in embryonic stem (ES) cells. The results suggest that the reduced ratio of littermates of homozygous Mssp (–/–) mice is due to the decreased concentration of progesterone in female mice, which affects the pre-implantation embryo.


Generation of Mssp−/– mice

The mouse mssp gene containing the entire coding region was isolated as a series of phage clones spanning ≈ 40 kb of DNA, and the exon/intron junction of the gene was determined (Fujimoto et al. 2000). To mutate the Mssp locus in ES cells, a replacement vector was constructed by the insertion of a neomycin resistance gene (neo) into the first coding exon of the Mssp gene. The diphtheria toxin A fragment (DT-A) gene was then inserted into the targeting vector as a negative selection marker to avoid non-homologous recombination (Fig. 1A). The targeting vector contained 5 kb regions homologous to the Mssp gene upstream and downstream of the neo gene. The ES cells were electroporated with the targeting vector and selected in a medium containing G418. A total of 238 G418-resistant colonies were selected and analysed by genomic Southern blot hybridization with 5′-Mssp and 3′-Mssp probes (Fig. 1B). The twelve clones obtained harboured the predicted targeted mutation at the Mssp locus. The targeting frequency was therefore 1/20 of G418r colonies, which is relatively high compared to that reported so far. After two ES cell lines, numbers 148 and 175 had been injected into eight-cell-stage embryos and then implanted into the uteri of pseudopregnant ICR female mice, 13 chimeras were generated. The chimeras from both line 148 and line 175 ES cells were confirmed to have been transmitted to germlines. After intercrossing the chimeras, viable and fertile heterozygous mice (Mssp+/–) were obtained, and these heterozygous mice were intercrossed to obtain homozygous mice deficient in the Mssp gene (Mssp−/– mice). Genotypes of the mice obtained were confirmed by Southern blot analysis using DNA isolated from tails of 3-day-old mice with a 3′-Mssp probe, and the results of typical examples are shown in Fig. 1C. The MSSP transcripts expressed in heterozygous and homozygous mice were also examined by RT-PCR using total RNAs with primers corresponding to the sequences upstream of exons 1 and 10. Compared to mRNA expressed in wild-type (Mssp+/+) mice, reduced and no RNA was expressed in heterozygous and homozygous mice, respectively, while mRNA of Scr3, another member of the MSSP family of proteins, and β-actin were expressed equally in all the mice (Fig. 1D). Furthermore, MSSP protein was detected in Mssp+/+ mice but not in Mssp−/– mice by Western blot analysis using a polyclonal rabbit anti-MSSP antiserum synthesized against the peptide numbers 343–370 of MSSP (Fig. 1E).

Figure 1.

Targeting disruption of the Mssp gene. (A) Schematic representation of a part of the wild-type Mssp gene (Wild-type locus), a linearized targeting vector, and a targeted Mssp allele. Exons are represented as black boxes, and the numbers above the boxes indicate the exon numbers for the Mssp gene transcripts. The exon containing ATG for the initiation methionine was named Exon 1. The arrows indicate the orientation of the neo and DT-A marker cassettes. (B) Southern blot analysis of an Mssp-targeted ES cell clone. HindIII/EcoRI (5′-Mssp), XbaI/SpeI (3′-Mssp) and NEO (derived from pKJ2) were used as probes on Southern blots to identify and characterize targeted clones. The predicted sizes of the expected SpeI or HindIII fragments of wild-type and targeted alleles are indicated. (C) Southern blot analysis of offspring from a mating of Mssp+/– and Mssp+/– mice. DNAs from tails of embryos were digested with SpeI and analysed as described in Experimental procedures using the 3′-Mssp probe. (D) RT-PCR analysis of MSSP mRNA in targeted mice. Total RNAs were extracted from the testes of wild-type (Mssp+/+), Mssp+/– and Mssp−/– mice and reverse-transcribed. The first strand of cDNA was synthesized by BcaBEST polymerase with the oligo dT primer and then amplified by PCR with sequence-specific primers for MSSP, Scr3 and β-Actin. (E) Western blot analysis of MSSP in targeted mice. Total proteins were extracted from the testes of wild-type and Mssp−/– mice and probed with an anti-MSSP polyclonal antibody against peptide numbers 343–370 of MSSP.

Placental defects in Mssp−/– embryos

The numbers of Mssp+/+, Mssp+/– and Mssp−/– littermates were not line with Mendelian frequencies, and their relative ratios were 1 : 2 : 0.5. We therefore investigated the time of the presumptive embryonic mortality in Mssp−/– mice by analysing the genotypes of embryos from Mssp+/– intercrosses at various stages of embryogenesis. Embryos dissected from the deciduous at embryonic days (E) 13.5 and 17.5 (with the morning of vaginal plug detection corresponding to E0.5) showed a decreased ratio of the Mssp−/– embryos, and finally the ratio of 1 : 1.6 : 0.5 among Mssp+/+, Mssp+/– and Mssp−/– was obtained from the mice at P5–P20 (Table 1). At E6.5, Mssp−/– mice were smaller than wild-type, and the uteri of Mssp−/– mice were retardedly developed (data not shown).

Table 1.  Genotypes of mice obtained from Mssp+/– breeding
Number of embryos (%)
age+/++/––/–empty deciduaetotal
  1. Embryos from Mssp+/– intercrosses were collected at the indicated embryonic day (E) or postnatal day (P), and the genotypes were determined by Southern blot analysis with a-3′-MSSP probe.

E13.513 (27)23 (47)7 (14)5 (10)48
E17.52 (20)6 (60)1 (10)1 (10)10
P5–P2064 (32)103 (53)27 (13) 194

To directly assess the growth capability of Mssp−/– embryos, eight-cell-stage embryos were prepared from Mssp−/– or Mssp+/+ intercrosses at E2.5 and cultured for 5 days in vitro, during which time they formed outgrowths. While Mssp+/+ E2.5 embryos hatched, attached to the culture dish and produced normal trophoblast giant cells during in vitro culture, about half of the embryos could not attach to the culture dish or died, and only 9.5% of the blastocysts of the Mssp−/– E2.5 embryos developed (Figs 2A, G, M). These findings are summarized in Table 2.

Figure 2.

Developmental defects in Mssp−/– embryos at the eight-cell stage in vitro. Wild-type (A) and Mssp–/– (G, M) eight-cell-stage embryos were removed at E 2.5 and cultured for 1 (B, H, N), 2 (C, I, O), 3 (D, J, P), 4 (E, K, Q) and 5 (F, L, R) days, during which time they developed outgrowths. The inner cell mass (ICM) grew normally and was surrounded by trophoblast giant cells (TG) in wild-type embryos. In contrast, Mssp−/– eight-cell-stage embryos did not attach to the culture dish.

Table 2.  Genotype and phenotype of eight-stage cells from Mssp−/– mice cultured in vitro.
 Genotype (%)
  1. The eight-stage cells were collected from Mssp+/+ and Mssp−/– mice at E2.5 and cultured individually for 5 days. Embryos were scored as defect when molula did not grow.


Expression of MSSP in testis and uterus

In order to clarify the reasons why a smaller percentage of Mssp−/– littermates was born after mating between Mssp−/– mice, the testis or uterus of the mice were analysed. The expression of MSSP during spermatogenesis was first examined by RT-PCR using mRNAs from testes as templates, with the primer corresponding to the sequence near the 5′-end of mMSSP. The results showed that MSSP mRNA was expressed in spermatogonia and spermatocytes of Mssp+/+ mice, but not of Mssp−/– mice (data not shown). The expression of MSSP in specific spermatogenic cells was then examined by Western blotting using an anti-MSSP rabbit polyclonal antiserum. In the testis, the particular cell-cycle stages can be recognized by their patterns which are characteristic of association within seminiferous tubes. Consistent with the results of RT-PCR, positive signals from MSSP were detected on spermatogonia and spermatocytes from Mssp+/+, but not from Mssp−/– mice (Figs 3A, B, D, E). No signals were detected on spermatid and sperm of Mssp+/+ and Mssp−/– mice. However, no morphological differences between the testes of Mssp+/+ and Mssp−/– mice were observed. In addition to the testis, an immunohistochemical analysis of stained sections prepared from the uteri of Mssp+/+ and Mssp−/– mice was carried out. Mssp was found to be expressed in uterine gland-transported glycogen of Mssp+/+ but not Mssp−/– mice, as was detected by the anti-MSSP antiserum (Figs 4A, C, E, G). Except for the smaller sizes of uteri of Mssp−/– mice than those of Mssp+/+ mice as described above, distinct morphological differences in Mssp−/– mice were not observed. Furthermore, in vitro fertilization using sperm and eggs from either Mssp−/– mice or Mssp+/+ mice was carried out to determine the reason for the reduced ratio of Mssp−/– mice littermates. Fertilization itself was, however, able to be performed both in Mssp−/– mice and Mssp+/+ mice (data not shown, see Discussion).

Figure 3.

Immunohistochemical analysis of testes in Mssp−/– mice. Testis sections from wild-type (A–C) and Mssp−/– (D–F) mice were prepared, reacted with an anti-MSSP antiserum (A, B, D and E) or a rabbit preimmune serum (C, F), and stained using an avidin-biotinylated peroxidase complex detection system as described in Experimental procedures. The sections were counterstained with haematoxylin and photographed using light-field microscopy. Bars in A and F indicate 50 µm.

Figure 4.

Immunohistochemical analysis of uteri in Mssp−/– mice. Uterus sections from wild-type (A–D) and Mssp−/– (E–H) mice were prepared, reacted with an anti-MSSP antiserum (A, C, E and G) or a rabbit preimmune serum (B, D, F and H), and stained using an avidin-biotinylated peroxidase complex detection system as described in Experimental procedures. The sections were counterstained with haematoxylin and photographed using light-field microscopy. Bars in A, C, E and G indicate 50 µm. Bars in B, D, F and H indicate 10 µm. UG = uterine gland.

Hormonal status of Mssp−/– mice

Since the sexual gland is thought to be the organ that connects with pre-implantation embryo and the maternal uterus and to end with embryo implantation, and since few morphological changes in the testis and uterus were observed in Mssp−/– mice, it is possible that changes in the hormonal state affecting the growth and development of the uterus in Mssp−/– mice are responsible for the reduced number of littermates born after the mating between Mssp−/– mice. To address this issue, levels of oestradiol and progesterone, both in the serum and uterus of adult females at 0.5, 1.5, 2.5 and 3.5 days after ovulation were measured, and the results of these levels in the serum were found to be similar to that in uterus. This period corresponds to one round of the ovulatory cycle of the mouse. The concentration of oestradiol in both Mssp+/+ and Mssp−/– mice was similarly changed during he ovulatory cycle; it increased by 3–4-fold at 1.5 days after ovulation, and decreased at 2.5 days and 3.5 days after ovulation (Fig. 5A). Progesterone levels in Mssp−/– mice were, on the other hand, dramatically decreased to 6.5% of those of Mssp+/+ mice after 2.5 days, the time that is known to be the very early stage of maintenance of progesterone secretion. These results suggest that different concentrations of progesterone causes the nutrient deficiency in Mssp−/– mice, leading to the reduced number of Mssp−/– littermates after mating between Mssp−/– mice.

Figure 5.

Concentrations of oestradiol and progesterone in serum from Mssp−/– mice. (A) Concentration of oestradiol. (B) Concentration of progesterone. The morning of the day when a vaginal plug was found was designated as E0.5. Shadow and black boxes represent concentrations of hormones in wild-type (Mssp+/+) and Mssp−/– mice, respectively.


The Mssp gene spans more than 40 kb and contains more than 13 exons, ranging from 50 to 530 base pairs. These structural features are similar to those of the human MSSP gene-2 (Haigermoser et al. 1996). In this study, we generated mice carrying a mutation in the Mssp gene. A replacement of the Mssp gene using two constructs of the targeting vector gave a very high frequency of heterozygously recombinating ES cells, about 1/20 G418r colonies targeted. It is possible that this high efficiency of recombination has resulted from a large insertion of the targeted Mssp gene into the Mssp locus. Two clones containing the targeted Mssp were used to generate chimeric mice. The Mssp−/– mice, which were generated by interbreeding the heterozygous mice, were normal in size and appearance, proved to be viable, and did not exhibit any overt phenotypic abnormalities. Testicular development in the Mssp−/– males was normal, and the Mssp−/– males were capable of mating with wild-type and Mssp−/– females. However, the newborn pups derived from intercrossing of Mssp−/+ mice showed that the Mssp−/–mice were apt to be in a state of perinatal lethality. The numbers of Mssp−/– mice or foetuses obtained were less than 14% of that of newborn pups, as calculated by the Mendelian rule for Mssp−/– mice. Histological examination of the testis, uterus and ovary, in which MSSP is expressed, revealed few obvious morphological abnormalities. These organs in Mssp−/– embryos at E6.5, however, were small and retarded. To determine the reason for this, we examined two points. First, E2.5 embryos from Mssp+/+ or Mssp−/– was cultured in vitro and we found that only 9.5% of the blastocysts of the Mssp−/– E2.5 embryos developed, while the Mssp+/+ E2.5 embryos developed normally. Second, we investigated whether eggs and sperm had abnormalities in in vitro fertilization experiments using fertilized eggs. It was found that the eggs and sperm from Mssp−/– mice were normal in fertilization, and the resultant fertilized eggs developed until the blastocyst stage. These results suggest that the embryos produced after mating between Mssp+/– mice gain the growth defect independently.

Furthermore, a reduced number of littermates after the mating of Mssp−/– mice were also evident, as well as in the case of mating between Mssp+/– mice. We therefore thought that the difference between the lethality of wild-type mice and that of Mssp−/– mice is due to a change in the balance of hormones in the uterus. The gland undergoes development in the uterus, both at puberty and during pregnancy. The essential hormonal factors regulating these two phases in mice have been identified; oestrogen, adrenocorticoid and growth hormone act during puberty, and oestrogen, progesterone and prolactine act during pregnancy (Nandi et al. 1995). Of these hormones, progesterone has been recognized as the mammalian ‘pregnancy hormone’ (Baulieu 1989; Clarke & Stutherland 1990), and the central role of progesterone in early pregnancy is to conduct a complex series of interactive steps that begin with the synchronized development of both the pre-implantation embryo and the maternal uterus and end with embryo implantation. Mssp was found to be expressed in uterine gland-transported glycogen and Mssp−/– females were found to have 6.5% of the progesterone levels of those in wild-type mice. These results suggest that the reduced progesterone concentration in Mssp−/– mice leads to undernutrition and that this state triggers suppression of the development of the embryos, which finally die, as was reported in the case of PRLR−/– mice (Ormandy et al. 1997).

MSSP-1, MSSP-2 and Scr2 are alternative splicing products of a single gene, the human MSSP gene2. Scr3, an MSSP family protein, is in a distinct gene, and there are highly homologous amino acid sequences in the N-terminal regions of MSSP and Scr3. It is therefore possible that an MSSP family protein such as Scr3 compensates for the loss of MSSP function in live Mssp−/– mice that have overcome the embryonic lethality that was observed in half of the embryos. The generation of mice carrying disruptions of both the MSSP and Scr3 genes should clarify these possibilities.

Experimental procedures

Generation of Mssp−/– mice

To construct the targeting vector to disrupt Mssp gene, a genomic DNA was isolated from a mouse genomic library of TT2 embryonic stem cells and a panel of restriction enzyme sites was mapped (Fujimoto et al. 2000). A gene-targeting vector was constructed as follows. A 5 kb EcoRI-XbaI genomic fragment containing intron 1 was subcloned. The 3′ flanking region spanning the internal EcoRI-XbaI sites and the neomycin-resistance gene (neo)-expression cassette in pKJ2 (Yagi et al. 1990) were first subcloned into pUC19. Then, both this 3′ flanking/neo fragment and the 5 kb BamHI-EcoRI fragment upstream of exon 1 were inserted into pMCDT-A (A+T/pau) containing DT-A. The targeting vector was linearized and electroplated into TT2 cells (Yagi et al. 1993) using a Bio-Rad Gene Pulsar at 250 V and 960 µF. ES cell clones positive for the desired homologous recombination were injected into embryos at the eight-cell stage from ICR mice, and then the resulting chimeric mice were crossed with ICR mice. Two ES clones with the germ-line transmission were obtained, and the genomic DNAs from ES cells or from mouse tails were analysed to examine whether or not the correct targeting event had occurred by Southern blot hybridization in which genomic DNAs digested with HindIII or SpeI were hybridized with either a 5′-Mssp (700 bp HindIII-EcoRI fragment) or a 3′-Mssp (1.4 kb XbaI-SpeI fragment) probe.

Cell culture

TT2 cells were cultured on feeder cells that had been pre-treated with mitomycin C in ES medium (Dulbecco's Eagle's medium) supplemented with 20% foetal calf serum, 104 units/mL leukaemia inhibitory factor, 10−4 m 2-mercaptoethanol, 0.1 mm nonessential amino acids, 4 mm nucleosides, and sodium pyruvate. The TT2 cells were seeded every day and passaged within 2 or 3 days.

RT-PCR analysis

Total testis RNAs were prepared by the acid guanidium thiocyanate-phenol-chloroform method (Chromcynski & Sacchi 1987), and cDNA was synthesized using the oligo dT primer and BcaBEST™ polymerase (Takara Co. Ltd). The first strand of cDNA products was amplified with specific primers for the first 5 min at 94 °C and then for 30 cycles of 1 min at 94 °C, 2 min at 55 °C and 3 min at 72 °C. The nucleotide sequences of the sense and anti-sense primers were 5′-AACAGTAGCAGCAGTAGCAAC-3′ and 5′-ATGGAAGCAGGTTGTAG-TGAC-3′, respectively. The amplified products were separated on 2% agarose gel and stained with ethidium bromide.


Uteri of wild-type and Mssp−/– mice were fixed in 4% paraformaldehyde and then embedded in paraffin. Tissue sections of 5 µm in thickness were thaw-mounted on to poly L-lysine-coated glass slides, deparraffinized by lemosol, dehydrated with ethanol, and then washed twice with distilled water. Endogenous peroxidase activity was abolished by incubating with 0.3% H2O2 for 30 min These sections were first stained for 30 min at room temperature with an anti-MSSP rabbit polyclonal antibody synthesized against the peptide corresponding to amino acid numbers 343–370 or non-immune antiserum at a dilution of 1/500 in PBS. The sections were then visualized using an avidin-biotinylated-peroxidase complex detection system (Vectastain Elite ABC kit, Vector Laboratories, Inc). Sections were also stained with by diaminobenzidine tetrahydrochloride (DAB) (Sigma) and with haematoxylin.

Hormone measurements

Mice were anaesthetized, and blood was collected by cardiac puncture. The concentrations of oestradiol and progesterone in serum obtained after centrifugation of blood were measured using EIA kits (Cayman Chemical). The levels of oestradiol and progesterone were measured at 0.5, 1.5 and 2.5 days after observation of a vaginal plug.

Culture of eight-stage cell embryos

Male and female wild-type or Mssp−/– mice were bred to obtain wild-type or Mssp−/– mouse embryos. The morning of the day when a vaginal plug was found was designated E0.5. Wild-type and Mssp−/– eight-stage cell embryos were collected by flushing oviducts with HEPES-buffered medium 2 and cultured for 5 days in ES medium containing 3% BSA without leukaemia inhibitory factor.


We thank Yoko Misawa and Kiyomi Takaya for their technical assistance. This work was supported by a grant from the Ministry of Education, Culture, Sports, Science and Technology of Japan.