Zfp296 negatively regulates H3K9 methylation in embryonic development as a component of heterochromatin

The Cys2/His2-type zinc finger protein Zfp296 has been implicated in stem cell pluripotency and tumor pathogenesis. However, its mechanisms remain elusive. Here, we demonstrated that a Zfp296 deficiency in mice impairs germ-cell development and embryonic growth. Zfp296 was intracellularly localized to heterochromatin in embryos. A GST-Zfp296 pull-down experiment using ES cell nuclear extract followed by LC-MS/MS showed that Zfp296 interacts with component proteins of heterochromatin (such as HP1, Dnmt1, Dnmt3b, and ATRX) and the NuRD complex. We focused on H3K9 methylation as a hallmark of heterochromatin, and found that Zfp296 overexpression in cultured cells reduces the Suv39h1-mediated H3K9 methylation. Consistent with this finding, in Zfp296 −/− mouse embryos, we observed a global increase in H3K9 methylation in a developmental stage-dependent manner, and showed, by ChIP-qPCR, that the H3K9me3 levels at major satellite repeats were elevated in Zfp296 −/− embryos. Our results demonstrate that Zfp296 is a component of heterochromatin that affects embryonic development by negatively regulating H3K9 methylation.

appears to be involved in tumorigenesis or tumor progression. Zfp296 is also highly expressed in human and mouse embryonic stem (ES) cells and in induced pluripotent stem (iPS) cells, but its expression progressively decreases during differentiation [21][22][23] . In addition, the expression of Zfp296 in combination with Yamanaka factors Oct3/4, Sox2, Klf4, and c-Myc, enhances the efficiency of iPS cell generation 24 . These observations suggest that Zfp296 regulates pluripotency in ES cells and iPS cells. Zfp296 has been proposed to function as a transcription factor in pluripotent cells 24 , but its roles in tumorigenesis and other biological processes remain undefined.
In the present study, we generated Zfp296 knockout mice and found that Zfp296 is required for proper germ-cell development and embryonic growth. We also found that Zfp296 is localized to the DAPI (4′,6-diamidino-2-phenylindole)-dense heterochromatin foci in embryonic somatic cells. In addition, we showed that Zfp296 binds to components of heterochromatin and the nucleosome remodeling and deacetylase (NuRD) complex in vitro and that Zfp296 overexpression decreases Suv39h1-mediated H3K9 methylation in HEK293T cells. We observed that the H3K9 methylation levels in Zfp296 −/− embryonic cells increase in a developmental stage-dependent manner and that Zfp296 deficiency affects the H3K9me3 levels at major satellite repeats. Based on these results, we discuss the roles of Zfp296 as a temporal repressor of H3K9 methylation during mammalian embryogenesis and as a chromatin effector at the H3K9me-dependent heterochromatin in embryonic somatic cells.

Results
Zfp296 −/− mice are healthy but partially infertile, with small testes and ovaries. We generated Zfp296 −/− mice from Zfp296 +/− intercrosses, and confirmed the loss of Zfp296 protein in the testis by Western blotting (Fig. S1D; see Figs 1A and S1A-S1C). Zfp296 −/− mice were viable and healthy with a normal body weight ( Fig. 1B and C), and showed no signs of abnormal hematopoiesis or neurogenesis (data not shown). We also detected no changes in the frequency of spontaneous tumorigenesis (data not shown) or in lifespan (Fig. S1E). Bcl11a/Ctip1 and Bcl11b/Ctip2, putative paralogs of Zfp296 (see Fig. S1F for homology), are required for lymphoid development 25,26 . However, flow cytometry analysis of the thymocytes and peripheral blood cells from Zfp296 −/− mice revealed no abnormalities in the expression of lymphocyte surface markers ( Fig. S1F and S1G). This result indicated that Zfp296 may not be involved in lymphocyte development, and that its molecular role is probably distinct from that of its putative paralogs Bcl11a/Ctip1 and Bcl11b/Ctip2.
Previous studies suggested that Zfp296 is mainly expressed in germ-cell lineages, which include post-meiotic spermatids 15 and germ-line stem cells 24 . We therefore investigated the effects of Zfp296 ablation on fertility. When Zfp296 −/− mice were intercrossed (9 mating pairs), no progeny were obtained. In addition, mating experiments between Zfp296 −/− and wild-type mice revealed that 70-80% of the male and female Zfp296 −/− mice were infertile, and that fertile Zfp296 −/− mice tended to lose their fertility at a younger age than control mice (Fig. 1D). The testis and ovary of Zfp296 −/− mice were significantly smaller than those of Zfp296 +/+ and Zfp296 +/− mice ( Fig. 1E-H). In contrast, fertility defects were not detected in Zfp296 +/− mice; we therefore used not only wild-type mice but also Zfp296 +/− mice as controls. Histological and immunofluorescence analyses showed that the testis (2 months postnatal) of Zfp296 −/− mice was mostly occupied by degenerated tubules lacking germ-line cells ( Fig. 1I; see Fig. S1H for postnatal development of the testis). Such degeneration in the tubules is considered to represent an impairment in spermatogenesis, and in fact, the production of mature sperm was greatly reduced (Fig. 1J). Similarly, we observed that oocytes at all stages of development were decreased in the ovary (3 months postnatal) of Zfp296 −/− mice ( Fig. 1K and L; see Fig. S1I for postnatal development of the ovary). These results indicated that a severe reduction in the number of sperm and oocytes led to impaired fertility in Zfp296 −/− mice. In contrast to adult mice, the embryonic gonads at E17.5 appeared to be the same size in Zfp296 −/− and control mice (Fig. 1M). However, immunofluorescence analysis for the germ-cell markers PGC7 and TRA98 showed decreased numbers of germ cells in the E17.5 embryonic gonads of Zfp296 −/− mice compared with control mice (Fig. 1N). Collectively, these findings demonstrated that a deficiency in Zfp296 caused a reduction in germ-cell numbers in the embryonic gonad, resulting in a degenerated testis and ovary during postnatal development.
Loss of Zfp296 causes impaired primordial germ-cell development. We next assessed the effects of Zfp296 ablation on earlier germ-cell development. In the mouse, primordial germ cells (PGCs) are specified by E7.25, and then migrate into the developing hindgut at E8.5 27 . From E7.25, PGCs stain positively for alkaline phosphatase (AP) 27 . AP staining of E8.5 hindguts revealed no difference in the number of PGCs between Zfp296 −/− and control mice ( Fig. 2A). In addition, AP staining of Zfp296 −/− versus control embryos at around E9.5 revealed no difference in PGC migration ( Fig. S2A and S2B). These findings indicated that Zfp296 does not affect the specification and migration of PGCs. We next examined whether Zfp296 influences the post-migratory PGC biology. Immunostaining of E11.5 genital ridge sections for Oct3/4 showed that the number of PGCs was markedly decreased in Zfp296 −/− embryos compared with controls ( Fig. 2B and C). Analysis of the mitotic marker phospho-histone H3 (H3S10ph) showed that the proliferation of PGCs was significantly decreased in the E11.5 Zfp296 −/− genital ridge ( Fig. 2D and E). However, we observed no difference in the apoptosis of PGCs between Zfp296 −/− and control embryos at E11.5 as assessed by cleaved caspase-3 staining ( Fig. 2D and F).
We next assessed the differentiation status of the PGCs in E12.5 Zfp296 −/− embryos. Nanog 28 , MVH 29 , and Kit 30 begin to be expressed in PGCs soon after their specification and are highly expressed in almost all of the PGCs at E12.5. Notably, immunofluorescence analyses of E12.5 Zfp296 −/− genital ridges revealed Oct3/4-positive cells that were negative for Nanog, MVH, or Kit ( Fig. 2G-I), and the percentages of Oct3/4-positive cells that were positive for these proteins were significantly lower in the Zfp296 −/− genital ridges than that in the controls (Fig. 2L). Considering the previous report that Zfp296 acts as a transcriptional activator of Oct3/4 through its germ cell specific conserved region 4 and that overexpression of Zfp296 in ES cells leads to Nanog upregulation 24 , Oct3/4 and Nanog might be downregulated in the PGCs of Zfp296 −/− mice. However, we detected no considerable changes in Oct3/4 expression in comparison with the other PGC markers such as E-cadherin and PGC7 by immunofluorescence (Fig. 2J-L). On the other hand, Nanog was clearly downregulated in the PGCs of E12.5 Zfp296 −/− embryos (Fig. 2L). However, immunostaining analyses of E13.5 embryonic gonads showed a partial recovery of PGC marker expression in Zfp296 −/− mice (Fig. S2C), indicating that the repression of Nanog, MVH, and Kit expression observed in the PGCs of E12.5 Zfp296 −/− embryos was developmental stage-dependent. Collectively, these results suggested that Zfp296 more strongly affects the differentiation status of PGCs in the post-specification phase than in the specification and migration phase.
Zfp296 −/− mice exhibit partial embryonic lethality and growth retardation. While analyzing the role of Zfp296 in germ-cell development, we noticed that the Zfp296 deficiency sometimes caused embryonic death at around E9.5-E14.5, and that from E12.5 on, the proportion of Zfp296 −/− mice was significantly lower than that expected from Mendel's rule (Fig. 3A). We also found a growth retardation of Zfp296 −/− embryos during E9.75-E12.5, which was not apparent before E9.5 or after E14.5 ( Fig. S3A and S3B). Although at E12.5 some of the Zfp296 −/− embryos were similar in size to controls, the average body weight of the Zfp296 −/− embryos was significantly lower than that of controls ( Fig. 3B). At E11.5, the Zfp296 −/− embryos exhibited various degrees of growth retardation (Fig. 3C). At E9.75, almost all of the Zfp296 −/− and a few Zfp296 +/− embryos exhibited slowed growth (Fig. 3D). We confirmed that placental defects were not seen with any of the growth-retarded Zfp296 −/− embryos from intercrosses of Zfp296 +/− mice (Fig. 3E). Western blot analysis revealed that, in wild-type at E9.75, the Zfp296 expression was much higher in the embryos than in the placenta (Fig. 3F), consistent with the finding that Zfp296 ablation did not affect placental development. We next analyzed the change in Zfp296 protein levels in wild-type embryos from E8.5-E10.5 by Western blotting, and found that the Zfp296 protein level began to increase from E9.75 (Fig. 3G). Consistent with this observation, whole-mount in situ hybridization analysis revealed that the Zfp296 mRNA expression was ubiquitously increased in E9.75 embryos (Fig. 3H). These findings suggested that upregulation of Zfp296 from E9.75 might be related to the growth defects seen in the Zfp296 −/− embryos. On the other hand, qRT-PCR analysis of various tissues at E12.5 demonstrated that Zfp296 was preferentially expressed in the liver, testis, and ovary (Fig. 3I). Although Zfp296 −/− mice showed no abnormalities in the liver during pre-and postnatal development, high-level expression of Zfp296 in the testis and ovary at E12.5 might be correlated with the aberrant germ-cell development phenotypes observed in the Zfp296 −/− mice.

Zfp296 interacts with components of heterochromatin in vitro and affects H3K9 methylation in cultured cells.
To address the potential functions of Zfp296 in further depth, we analyzed the intracellular localization of Zfp296 by immunofluorescence analysis of E9.75 embryos using confocal laser scanning microscopy, and found that Zfp296 was localized to the DAPI-dense heterochromatin foci (Fig. 4A). Such Zfp296 staining was lost in the Zfp296 −/− embryos (Fig. S4A). We next determined the domains responsible for its heterochromatin localization. We performed immunofluorescence studies in primary MEF cells (which were derived from E14.5 mouse embryos and showed very weak or no endogenous expression of Zfp296 protein) transfected with plasmid vectors expressing Flag-Zfp296 or its deletion mutants shown in Fig. 4B. Zfp296 contains six ZF domains composed of a single CCHC-ZF and five C2H2-ZF domains. The results revealed that the heterochromatin localization of Zfp296 was dependent on its second and third ZF domains (Fig. 4C). We also found that the fourth to sixth ZF domains of Zfp296 were required for its nucleus-specific localization (Fig. 4C).
Bcl11b/Ctip2, a putative paralog of Zfp296, physically recruits chromatin-modifying complexes and establishes the heterochromatic environment 31,32 . Thus, we hypothesized that Zfp296 also binds to heterochromatin components. To examine this possibility, we carried out GST-pull down experiments using ES cell nuclear extract and recombinant GST-Zfp296 or GST (as a control). The captured proteins were subsequently analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS). We identified 170 proteins that showed significant binding to GST-Zfp296 compared with GST alone (2-fold change cut-off; p < 0.05, Fisher's exact test) ( Fig. 4D and Table S1). Among these Zfp296-interacting proteins, we found components of heterochromatin and the NuRD complex (Fig. 4E). These findings were consistent with a previous study indicating that ZNF296 (the human homolog of Zfp296) interacts with MBD3L1 33 . Interestingly, DNA repair proteins such as H2AX, MSH6, DDB1, LIG3, and MDC1 were also shown to significantly bind to GST-Zfp296 (Table S1). In addition, GST-Zfp296 appeared to pull down endogenous Zfp296 (Table S1). Another coimmunoprecipitation analysis in human embryonic kidney (HEK) 293 T cells revealed that Zfp296 might form oligomers (Fig. S4B and S4C). Similarly, Bcl11a/Ctip1 is reported to form oligomers in vivo 34 .
H3K9 methylation is involved in heterochromatin formation and heterochromatic gene silencing. Suv39h is known to induce H3K9 trimethylation at pericentromeric heterochromatin 35 . Therefore, we next assessed the effect of Zfp296 expression on Suv39h-dependent H3K9's methylation. The transient transfection of HEK293T cells with a plasmid vector expressing Myc-Suv39h1 induced an increase in H3K9me2 and H3K9me3, which was inhibited by the coexpression of Zfp296 ( Fig. 4F and G). We also observed by live-cell imaging that Zfp296-GFP colocalized with Suv39h1-DsRed at heterochromatin foci in HEK293T cells (Fig. S4D). Taken together, these findings indicated that Zfp296 targets heterochromatin and represses Suv39h-dependent H3K9 di-and trimethylation.
Zfp296 −/− embryos exhibit whole-body H3K9 hypermethylation. Having found that heterochromatin-localized Zfp296 could repress H3K9 methylation in cultured cells, we next examined whether a Zfp296 deficiency could affect H3K9me3 levels during embryogenesis. We first compared the H3K9 methylation levels in E9.75 Zfp296 −/− versus control embryos by quantitative Western blotting, and notably, we detected a global increase in H3K9me2 and H3K9me3 (Fig. 5A and B). To confirm this result, we immunostained hindgut sections of E9.75 Zfp296 −/− and control embryos for H3K9me1, H3K9me2, and H3K9me3. Consistent with the Western blotting results, we found that H3K9me3 was elevated in both the somatic cells and the PGCs, that H3K9me2 was elevated, particularly in the PGCs, and that no considerable changes were observed for H3K9me1, in the Zfp296 −/− embryos compared to wild-type (Figs 5C and S5A). PGCs are known to exhibit decreased H3K9me2 and increased H3K27me3 during their migration phase 36 . By measuring the immunofluorescence intensity, we found that the Zfp296 deficiency caused an upregulation of H3K9me2 but did not affect the H3K27me3 level in PGCs (Figs 5D and S5B). This finding suggested that the Zfp296 loss could disturb the epigenetic regulation in PGC development. On the other hand, after E11.5, we found no difference in the H3K9 methylation levels between Zfp296 −/− and the control embryos (data not shown). These observation appeared to be correlated with the appearance of the Zfp296-deficient phenotypes.
Zfp296 has been regarded as a transcription factor in previous studies [21][22][23] . To explore the cause of the increased H3K9 methylation observed in Zfp296 −/− embryos, we analyzed the expression of several key genes involved in H3K9 methylation in E9.75 Zfp296 −/− and control embryos, by qRT-PCR. We detected significant transcriptional differences in Eset and Jmjd2c between the Zfp296 −/− and control embryos (Fig. S5C). However, these changes were relatively small (within 2-fold). These findings suggested that Zfp296 suppresses H3K9 methylation in the developing embryos by a mechanism other than the direct transcriptional regulation of H3K9-modifying enzymes.
The pericentromeric region is enriched with large clusters of repetitive satellite sequences and transposable elements, which are characterized by highly methylated H3K9. We next examined whether Zfp296 ablation affected the H3K9me3 levels at repetitive DNA elements. For this analysis, we performed a chromatin immunoprecipitation (ChIP) assay of E9.75 embryos with anti-H3K9me3, followed by qPCR. The H3K9me3 levels at major satellite repeats were significantly elevated in the Zfp296 −/− embryos compared with controls (Fig. 5E). Taken together, these results indicate that Zfp296 temporally modulates H3K9 methylation levels at heterochromatin during embryogenesis.

Discussion
Zfp296 has been described as a DNA-binding transcription factor 24 , but here we showed that it can directly bind to heterochromatin-related proteins. Similarly, Bcl11b/Ctip2, a putative paralog of Zfp296, interacts with HP1 and Suv39h1. A Bcl11b/Ctip2-containing complex was reported to promote H3K9 methylation 31,32 . In contrast, here we observed that Zfp296 decreased the H3K9 methylation levels in HEK293T cells. This finding was supported by the observation that the H3K9me3 levels in Zfp296 −/− embryos were globally elevated. Together, these results suggest that Zfp296 and Bcl11b/Ctip2 may play opposing roles in the H3K9 methylation of heterochromatin. (G) Quantification of the data shown in (F). The density of each band relative to that obtained with anti-histone H3 was determined. The level of each histone modification was expressed relative to the level in HEK293T cells transfected with empty vector. Data represent mean + SD of three independent experiments. *p < 0.05, **p < 0.01 by two-tailed Student's t-test.
H3K9 modification appears to have critical roles in embryogenesis; this importance is highlighted by the finding that mutations in H3K9-modifying enzymes cause severe embryonic growth defects 37 . Consistent with this point, we found that Zfp296 ablation in mice resulted in partial embryonic lethality and growth retardation at E9.5-E14.5. Recent genome-wide profiles of chromatin modifications in various embryonic cells and tissues have revealed that these modifications are temporally regulated during development 38 . In Zfp296 −/− embryonic cells, elevated levels of H3K9 were seen at heterochromatin, but not after E13.5 (data not shown), and the defective phenotypes of Zfp296 −/− embryos coincided with the temporal expression pattern and levels of Zfp296. This finding leads to the hypothesis that Zfp296 serves as a temporal repressor of H3K9 methylation during embryogenesis. Epigenetic reprogramming in PGCs involves a widespread loss of both H3K9 methylation 36,39,40 . We showed that the H3K9me2, H3K9me3 were elevated in the Zfp296 −/− PGCs. In addition, we observed that the expressions of Nanog, MVH, and Kit were reduced in Zfp296 −/− PGCs compared with those in control PGCs. The expression of these genes in PGCs is known to be upregulated through epigenetic reprogramming 39,41 . Therefore, it is likely that the inappropriate differentiation of PGCs seen in Zfp296 −/− mice resulted from H3K9 hypermethylation and incomplete epigenetic reprogramming. . The level of each histone modification normalized to that of H3 is shown for individual embryos. *p < 0.01 by two-tailed Student's t-test. (C) Immunofluorescence staining of genital ridge sections from E9.75 wild-type and Zfp296 −/− embryos for H3K9me3 (left) or H3K9me2 (right). PGC7 served as a marker for PGCs. DNA was counterstained with DAPI. Scale bar = 10 μm. (D) Signal intensity of H3K9me2 and H3K27me3 in PGCs relative to somatic cells. Signal intensity in PGCs was measured with software and normalized to that of three neighboring somatic cells. Data represent mean + SD of 20-25 PGCs from two embryos for each genotype. *p < 0.01 by two-tailed Student's t-test. (E) H3K9me3 ChIP-qPCR at the indicated loci in E9.75 wild-type and Zfp296 −/− whole embryos. Chromatin was prepared from a pool of four embryos for each genotype and subjected to ChIP assay using anti-H3K9me3 antibody. The resulting DNA samples were assayed by qPCR in triplicate. Data represent mean + SD. *p < 0.05 by two-tailed Student's t-test. The Tpsg1 gene locus was chosen as a non-target control locus, as previously described 55 .
Scientific RepoRts | 7: 12462 | DOI:10.1038/s41598-017-12772-y Chromosome instability is linked to tumor initiation and progression. Suv39h deficiency impairs H3K9 methylation at pericentromeric heterochromatin and leads to chromosome instability. For instance, Suv39h-deficient mice exhibit spontaneous B cell lymphoma and meiosis defects 35 . We found that the overexpression of Zfp296 protein reduced Suv39h1-mediated H3K9 di-and trimethylation in vitro. Therefore, Zfp296 appears to act as a negative regulator of chromosome stability, which may explain the tumorigenic potential of Zfp296 in acute leukemia. This possibility is supported by our finding that Zfp296 deficiency did not affect the frequency of spontaneous tumorigenesis and lifespan; however, the role of Zfp296 loss in the pathogenesis of oligodendroglioma and prostate carcinoma in which Zfp296 is epigenetically silenced is still not known 19,20 . On the other hand, DNA double strand breaks (DSBs) can cause genome rearrangements and impair genomic stability. γH2AX, the phosphorylated form of H2AX, is regarded as a central component of the damaged chromatin 42 , and DNA repair protein MDC1 is known to directly bind γH2AX to regulate responses to DBSs 43 . HP1 also accumulates at DNA damage sites [44][45][46] . In this study, we showed significant binding of Zfp296 to H2AX, MDC1, and HP1 (Table S1). Its binding to other DNA repair proteins such as MSH6, DDB1, and LIG3 was also shown. These interactions may suggest a possible role of Zfp296 in DNA repair and shed light on the mechanisms of Zfp296-associated tumorigenesis. The suppression of H3K9 methyltransferases, such as Suv39h1, Eset, and G9a, promotes transcription factor accessibility and increases reprogramming efficiency [47][48][49] . Thus, Zfp296 might accelerate iPS reprogramming through the repression of H3K9 methylation.
In conclusion, we demonstrated that Zfp296 has a pivotal role in germ-cell development and embryonic growth. We also showed that Zfp296 functions as a chromatin effector. The epigenetic regulation mediated by Zfp296 is also likely to play important roles in cancer pathogenesis and the induction of pluripotency as well as in embryogenesis.

Methods
Targeted disruption of the Zfp296 locus. The genomic region containing the Zfp296 gene was amplified by long PCR and cloned into a plasmid. The targeting vector was designed to insert an IRES-βgeo-pA cassette into exon 3 of the Zfp296 gene, which encodes all of the zinc-finger domains. The Zfp296 targeting vector was linearized and electroporated into EB3 embryonic stem cells (129/Sv) 50 , and selection was performed with G418. Genomic DNA from G418-resistant colonies was analyzed for targeted disruption of the Zfp296 gene by long PCR. The targeted ES cells were injected into C57BL/6 blastocysts to generate chimeras. The resulting chimeric mice were bred with C57BL/6 mice, and germline transmission of the knockout allele was confirmed by PCR of the genomic DNA from tail tips. Heterozygous mice were backcrossed to C57BL/6 mice. Our studies used mice obtained after backcrossing for at least six generations. To stage the embryonic age, noon on the day of vaginal plug detection was defined as embryonic day 0.5 (E0.5). Mice were housed and maintained in a controlled environment according to the institutional guidelines. All animal experiments were performed in accordance with the institutional guidelines (protocols #21-089 and #26-066), which were reviewed and approved by the Animal Care and Use Committee of the Osaka University Graduate School of Medicine. Mice were euthanized with an intraperitoneal injection of pentobarbital sodium at 180 mg/kg body weight.
Genotyping PCR. Genotypes were determined by PCR analysis of the genomic DNA isolated from tail-biopsies or embryonic fragments (yolk sac, head, or tail). The gene-specific primers are listed in Table S2. The PCR conditions were as follows: initial denaturation 5 min at 95 °C, followed by 30-35 cycles of denaturation 30 sec at 94 °C, annealing 30 sec at 64 °C, and extension 1 min at 72 °C. PCR products were separated by electrophoresis in a 2% agarose gel and visualized by ethidium bromide staining. Antibody generation. Rabbits were immunized with a synthetic peptide corresponding to the C-terminal region of Zfp296 (amino acid residues 432 to 445: TLDKHLRQKHPEMA). Antisera were collected and then affinity-purified with an agarose column coupled with the immunizing peptide.
Western blot analysis. Samples (cultured cells or whole embryos) were lysed in lysis buffer containing 50 mM Tris-HCl (pH 6.8), 2% SDS, and 10% glycerol supplemented with protease inhibitor cocktail (Calbiochem), followed by sonication. After centrifugation and protein quantification, the lysates were boiled for 5 min, loaded onto a 10-15% polyacrylamide Tris-glycine gel, separated by electrophoresis, and transferred onto a polyvinylidene difluoride (PVDF) membrane (Millipore). After washing, the membranes were blocked with 3% (w/v) skim milk in Tris-buffered saline (TBS) with 0.1% Tween-20 (TBST) or Blocking One (Nacalai) for 1 h at room temperature, and then incubated with primary antibodies in TBST for 1 h at room temperature. Primary antibodies were detected with horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. The blot was developed using ECL or ECL Prime detection reagents (GE Healthcare), and visualized using a luminescent image analyzer LAS-4000 (Fujifilm) or X-ray film. For quantification, signal intensities on Western blots were measured with an LAS4000 imager and the instrument's software. Antibodies and their dilutions are listed in Table S3.
Immunofluorescence and histological analysis. Embryonic tissues (whole embryo and genital ridge) were fixed with 2% paraformaldehyde (PFA) in PBS for 2 h at 4 °C. Postnatal tissues (testis and ovary) were fixed with 4% PFA in PBS overnight at 4 °C. After washing, the tissues were incubated in 10% (w/w) sucrose in PBS for 1 h at 4 °C, and then transferred to 20% sucrose in PBS and incubated overnight at 4 °C. The tissues were frozen in OCT compound (Sakura), and then 10-μm-thick sections were cut, adhered onto MAS-coated slides (Matsunami), and rehydrated before staining. For immunofluorescence staining with anti-Zfp296 antibody, antigen retrieval was performed using Target Retrieval Solution pH 6.1 (DAKO). The sections were blocked in 10% goat or donkey serum plus 3% bovine serum albumin (BSA) in PBS for 1 h at room temperature, and then incubated with primary antibodies overnight at 4 °C. The sections were then incubated with Alexa Fluor-conjugated secondary antibodies for 1 h at room temperature. The slides were counterstained with 1 μg/ml DAPI (Molecular Probes) for 1 h at 4 °C prior to mounting. For histological analysis, tissues were fixed in 4% PFA overnight at 4 °C and then embedded in paraffin. The tissues were cut into 5-μm-thick sections and stained with hematoxylin and eosin. The antibodies and their dilutions are listed in Table S3. All imaging was performed using an FV1000 confocal microscope (Olympus) or a BZ-9000 multifunctional microscope (Keyence).
Alkaline phosphatase staining and whole-mount in situ hybridization. Embryos were fixed with 4% PFA in PBS and, after washing three times with PBS, were stained using an alkaline phosphatase detection kit (Sigma) according to the manufacturer's instructions. Whole-mount in situ hybridization was performed as described 51 . Digoxigenin-UTP-labeled riboprobes for the 3′UTR sequence of Zfp296 were synthesized according to the manufacturer's instructions (Roche).
qRT-PCR. RNA was extracted from whole embryos using an RNeasy Mini kit (Qiagen) according to the manufacturer's instructions. The RNA was then used for cDNA synthesis using ReverTraAce-α (Toyobo) and dT 20 primer. Quantitative PCR analyses were performed using a DNA Master SYBR Green kit (Roche) and transcript-specific primer pairs (Table S2). Reactions were run on a Step One Plus real-time PCR system (Applied Biosystems). Each sample was analyzed in triplicate. Expression levels were determined by a standard curve for each primer pair and normalized to the level of the housekeeping gene Tbp.

GST pull-down experiments and mass spectrometry. Recombinant GST-Zfp296 or GST proteins
were purified in hypertonic buffer using glutathione magnetic beads (Pierce). EB3 mouse embryonic stem cells were harvested with a cell scraper (Sumilon) into hypotonic buffer containing 10 mM Tris-HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl 2 , and 0.1% NP-40 supplemented with protease inhibitor cocktail (Calbiochem). After vortexing briefly, the lysates were left on ice for 10 min, and then centrifuged at 14,000 g for 10 min at 4 °C. The supernatant was removed, and the pellet was resuspended in high-salt buffer containing 20 mM Tris-HCl (pH 8.0), 500 mM NaCl, 3 mM MgCl 2 , and 0.5% NP-40 supplemented with 100 U/ml Cryonase (cold-active nuclease; Takara) and protease inhibitor cocktail. After vortexing briefly, the lysates were incubated for 20 min at 4 °C for nuclease digestion, and then sonicated on ice using a Sonifier 250 (Branson) with a microtip. The sonication conditions were as follows: power level 2 for two 10-sec pulses, with 2-min rests on ice between pulses. After centrifugation to remove debris, the supernatant was collected and then diluted 3-fold in buffer containing 20 mM Tris-HCl (pH 8.0) supplemented with protease inhibitor cocktail. The lysates were incubated with purified recombinant GST-Zfp296 or GST proteins bound to glutathione magnetic beads overnight while rotating at 4 °C. The beads were washed three times with PBS plus 0.1% Tween-20 supplemented with protease inhibitor cocktail. The beads were then incubated in 50 mM Tris-HCl (pH 6.8), 2% sodium deoxycholate (SDC), 10% glycerol for 1 h while rocking at room temperature. Proteins were processed for LC-MS/MS analysis according to a standard protocol 53 . LC-MS/MS analysis was performed by an UltiMate 3000 Nano LC system (Thermo Fisher Scientific) coupled to a Q-Exactive hybrid quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific).
Chromatin immunoprecipitation. Embryos were dissociated with a solution containing 0.15% trypsin and 0.7 mM EDTA for 20 min while rocking at 37 °C and then diluted 5-fold in DMEM (Sigma) supplemented with 20% FBS to stop the trypsinization. After washing with PBS, the cells were crosslinked with 1% formaldehyde (Sigma) in PBS for 10 min while rocking at room temperature. Crosslinking was stopped by adding glycine solution to a final concentration of 125 mM. Subsequently, the chromatin preparation, immunoprecipitation, and DNA purification were performed using a SimpleChIP Enzymatic Chromatin IP kit (Cell Signaling). ChIP enrichment was determined by qPCR using major satellite primers 54,55 . The antibodies and primers are presented in the Supplementary Methods.