Mouse ZAR1-like (XM_359149) colocalizes with mRNA processing components and its dominant-negative mutant caused two-cell-stage embryonic arrest

Abstract

Maternal effect genes and encoding proteins are necessary for nuclear reprogramming and zygotic genome activation. However, the mechanisms that mediate these functions are largely unknown. Here we identified XM_359149, a Zar1-like gene that is predominantly expressed in oocytes and zygotes, which we designated Zar1-like (Zar1l). ZAR1L-EGFP formed multiple cytoplasmic foci in late two-cell-stage embryos. Expression of the ZAR1L C-terminus induced two-cell-stage embryonic arrest, accompanied with abnormal methylation of histone H3K4me2/3 and H3K9me2/3, and marked down-regulation of a group of chromatin modification factors including Dppa2, Dppa4, and Piwil2. When ectopically expressed in somatic cells, ZAR1L colocalized with P-body components including EIF2C1(AGO1), EIF2C2(AGO2), DDX6 and LSM14A, and germline-specific chromatoid body components including PIWIL1, PIWIL2, and LIN28. ZAR1L colocalized with ZAR1 and interacted with human LIN28. Our data suggest that ZAR1L and ZAR1 may comprise a novel family of processing-body/chromatoid-body components that potentially function as RNA regulators in early embryos. Developmental Dynamics 239:407–424, 2010. © 2009 Wiley-Liss, Inc.

INTRODUCTION

The maternal factors that accumulate during oogenesis play pivotal roles in nuclear reprogramming, zygotic genome activation, and preimplantation embryonic development (Schultz, 1993; Aoki et al., 1997; Latham, 1999; Latham and Schultz, 2001; Ma et al., 2001; Hamatani et al., 2004; Minami et al., 2007; Stitzel and Seydoux, 2007). Oogenesis involves a number of critical events because a growing mouse oocyte is transcriptionally and translationally active. A large number of mRNAs are synthesized and stored to support oocyte maturation and early preimplantation embryogenesis and are not used for immediate translation (Bachvarova, 1985; Wassarman and Kinloch, 1992). Mature oocytes arrest in metaphase during their second meiotic division (MII stage), which is associated with transcriptional shut-down and reduced translation. Fertilization triggers the completion of meiosis, which is followed by the formation of a one-cell embryo (zygote) containing haploid paternal and maternal pronuclei (Schultz, 1993; Aoki et al., 1997; Latham, 1999; Latham and Schultz, 2001; Ma et al., 2001; Hamatani et al., 2004; Minami et al., 2007; Stitzel and Seydoux, 2007). Each pronucleus undergoes DNA replication before entering the first mitosis to produce a two-cell embryo.

Global expression profiling revealed distinct patterns of maternal RNA degradation and zygotic genome activation, which includes three transient waves of de novo transcription: (1) a minor activation before cleavage (minor ZGA), (2) a major activation at the two-cell-stage (major ZGA), and (3) a major activation preceding the dynamic morphological and functional changes that occur during the transition from morula to blastocyst, which is termed mid-preimplantation gene activation (MGA) (Hamatani et al., 2004). The major ZGA promotes dramatic reprogramming of gene expression, coupled with the generation of novel transcripts that are not expressed in oocytes. Thus, the genetic program governed by maternal transcripts/proteins must be switched to one dominated by transcripts/proteins derived from the newly formed zygotic genome (Schultz, 1993; Aoki et al., 1997; Latham, 1999; Latham and Schultz, 2001; Ma et al., 2001; Hamatani et al., 2004; Minami et al., 2007; Stitzel and Seydoux, 2007).

A great amount of maternal effect genes produce mRNAs or proteins that accumulate in the egg during oogenesis. Nevertheless, a limited number of maternal-effect genes have been identified in mice: Nlrp5 (Mater, maternal antigen that embryos require) (Tong et al., 2000); Hsf1 (heat-shock factor 1) (Christians et al., 2000); Dnmt1 (DNA methyltransferase 1, oocyte isoform) (Howell et al., 2001); Npm2 (nucleoplasmin 2) (Burns et al., 2003; De La Fuente et al., 2004); Dppa3 (Stella) (Payer et al., 2003); Zar1 (zygotic arrest 1) (Wu et al., 2003a); Cdh1 (E-cadherin) (De Vries et al., 2004); Pms2 (Gurtu et al., 2002); Ezh2 (enhancer of zeste 2) (Erhardt et al., 2003); Dnmt3a (DNA methyltransferase 3A) (Kaneda et al., 2004); Ube2a (HRA6A) (Roest et al., 2004); and Smarca4 (Brg1) (Bultman et al., 2006). Apart from Smarca4, Cdh1, Pms2, Ezh2, Dnmt3a, and Ube2a, all of these maternal-effect genes are exclusively expressed in oocytes. In addition, only Nlrp5 and Ube2a mutants have similar phenotypes (two-cell arrest) to that of Smarca4 maternally depleted embryos. The other mutants primarily arrest at the one-cell stage (Npm2, Dppa3, Zar1, Hsf1), later stages of preimplantation (Dppa3, Pms2), or during post-implantation (Dnmt3a, Dnmt1o) development. Ezh2 mutant exhibits a postnatal phenotype. Cdh1 mutant appears phenotypically normal because of rescuing by the wild-type paternal allele. Meiotic maturation triggers the degradation of maternal transcripts. About 90% of the maternal mRNAs have been degraded by the two-cell stage. However, the mechanisms that regulate the translation and degradation of maternal transcripts are largely unknown.

In the present study, we identified XM_359149, a Zar1-like gene that is predominantly expressed in oocytes and early preimplantation embryos, which we named Zar1-like (Zar1l). We characterized its sub-cellular localization and its effect on preimplantation development. Our data showed that ZAR1L formed cytoplasmic foci in late two-cell-stage embryos. Its mutant form ZAR1L Cter-Flag-EGFP induced abnormal epigenetic modifications and gene expression changes in late two-cell-stage embryos, and finally caused two-cell-stage arrest. When ectopically expressed in somatic cells, ZAR1L colocalized with P-body components including EIF2C1(AGO1), EIF2C2(AGO2), DDX6, and LSM14A, and germline-specific chromatoid body components including PIWIL1, PIWIL2, and LIN28. ZAR1L colocalized with ZAR1 and interacted with human LIN28. Our data suggested that ZAR1L and ZAR1 comprise a novel family of P-body/C-body-like structure components in late 2-cell embryos.

ABBREVIATIONS AAs amino acids C-body chromatoid-body DAB 3,3′-Diaminobenzidine DAPI 4′-6-Diamidino-2-phenylindole EGFP enhanced green fluorescent protein FBS fetal bovine serum H3K4me1/2/3 Mono-/Di-/Tri-methylated Histone H3 at Lysine 4 H3K9me2/3 Di/Tri- methylated Histone H3 at Lysine 9 HRPhorseradish peroxidasemRNP mRNA ribonucleoprotein ORF open reading frame P-body mRNA processing body PFA paraformaldehyde PHD plant homeodomain PHDL cells pancreatic PNA-HSA double-low cells Phospho-Rpb1 phosphorylated Rpb1 at C-terminal Ser2/5 repeats RFP red fluorescent protein RT-PCR reverse transcription polymerase chain reaction.

RESULTS

Zar1l Gene and Protein Information and Expression Pattern

The XM_359419 sequence (GeneID: 545824; LOCUS: XM_359419) was identified in a search for genes that are preferentially expressed in oocytes and early embryos. Protein blast analysis showed that the XM_359419 ORF encodes for a ZAR1-like protein, which we have named ZAR1L. By genomic analysis and RT-PCR analysis, we successfully cloned the full-length ORF of Zar1l (Fig. 1A). Its orthologs have been found in other organisms, including humans, dogs, cows, and rats. Mouse ZAR1L exhibits greater homology with the predicted human ZAR1L protein than with the mouse ZAR1 protein (Fig. 1A and B). RT-PCR analysis showed that mouse Zar1l has two transcript isoforms and is specifically expressed in adult ovarian tissue (Fig. 2A). Moreover, it is predominantly expressed in oocytes and zygotes. Cloning and sequencing of the full-length ORF showed that one of the mouse Zar1l isoforms is 876 bp (encodes a 291 AA polypeptide) and the other is 982 bp with out-of-frame reading (Fig. 2A and data not shown). RT-PCR analysis showed that mouse Zar1 gene was also predominantly expressed in oocytes and zygotes. In order to determine the protein levels of ZAR1L, Western blot was performed. The results showed that mouse ZAR1L protein was predominantly expressed in oocytes and zygotes, and was also maintained at a certain level in 2-cell- and 4-cell-stage embryos (Fig. 2B). Only the 876-bp transcript isoform of Zar1l was characterized in this study.

Figure 1. Sequence alignment of mouse ZAR1 and ZAR1L proteins, human ZAR1L, and mouse ZAR1L proteins. A: Sequence alignment of mouse ZAR1 and ZAR1L proteins. The eight well-conserved cysteines (with # above them) form an atypical plant homeodomain (PHD) zinc finger domain. The N-terminal sequences exhibit low homology between mouse ZAR1 and ZAR1L protein (analyzed by DNAMAN). B: Sequence alignment of predicted human and mouse ZAR1L proteins. Sequence analysis showed that the ZAR1L protein contains three domains/motifs: one N-terminal function unknown domain, one middle CSE2-like domain, and one C-terminal atypical FVYE/PHD zinc finger domain. The domains/motifs were analyzed by PreDom (2006.1) program and Superfamily (1.73) program. Based on the domain/motif analysis and the BLASTP alignment results, we designed ZAR1L mutants with domain/motif containing fragments. The ZAR1L-Cter fragment contains the well-conserved atypical PHD zinc finger domain (191–291 AAs in mouse ZAR1L, boxed). The ZAR1L-ΔN fragment contains the middle domain (112–190 AAs) and C-terminal domain (191–291 AAs), but lacks the N-terminal domain (1–111 AAs, underlined).

Figure 2. Expression pattern and sub-cellular localization of mouse Zar1l. A: The mouse Zar1l gene is predominantly expressed in the ovary, oocytes, and zygotes (top). Its mRNA level dramatically decreased from 2-cell-stage embryos. Two transcript isoforms of Zar1l gene have been found. The mouse Zar1 gene is also predominantly expressed in oocytes and zygotes and then dramatically decreased (bottom). B: Western blot analysis of ZAR1L protein. ZAR1L is detectable from GV oocytes to 4-cell-stage embryos. C: Design of constructs. The expression constructs were designed based on the predicted protein sequence, with different tags. D–I: Sub-cellular localization of ZAR1L-EGFP fused protein. ZAR1L-EGFP localized to the perinucleus region in some full-grown GV oocytes (E, arrow) and formed cytoplasmic foci in most of the late two-cell-stage embryos (I, arrows). It is distributed predominantly in the cytoplasm in small GV oocytes, MI and MII oocytes, and zygotes. The red signals in E–G represent the F-actin signal labeled with Rhodamine. J: The ZAR1L-Cter-Flag-EGFP mutant predominantly localizes to the cytoplasm and forms multiple cytoplasmic foci (arrow). K: The ZAR1LΔN-EGFP mutant also forms cytoplasm foci, but with a relative lower level (arrows). L: The ZAR1L-C292S/C295S-EGFP mutant completely loses the cytoplasmic foci capacity (arrow). M: The full-length ZAR1-EGFP also forms multiple cytoplasmic foci (arrows). Scale bars = 20 μm.

Sub-Cellular Localization of Mouse ZAR1L Protein and Its Mutants

In order to predict protein domain/structure of ZAR1L, we performed protein sequence analysis by using the PreDom 2006.1 and Superfamily 1.73 protein domain prediction programs, respectively. The PreDom program analysis showed that mouse ZAR1L protein might have three functional domains, one N-terminal domain (51–103 AAs), one C-terminal domain (193–241 AAs), and one larger domain (130–291 AAs, which might contain the middle region and the C-terminal region). The Superfamily 1.73 program analysis showed that mouse ZAR1L protein might contain one CSE2-like domain (159–190 AAs) in the middle region, and one atypical FYVE/PHD zinc finger domain (227–280 AAs) in the C-terminal. Based on these protein domain prediction results, we supposed that mouse ZAR1L protein should have three fragments that each contained one functional domain: a well-conserved C-terminal region (191–291 AAs, containing the FYVE/PHD zinc finger domain), a conserved N-terminal region (1–111 AAs, containing one functional unknown domain), and a relatively conserved middle region (112–190 AAs, containing the CSE2-like domain). Various constructs were designed to express wild-type ZAR1L and dominant-negative mutants of ZAR1L (Fig. 2C). To study the sub-cellular localization of ZAR1L and its mutant proteins, as well as their roles in preimplantation development, in vitro transcribed mRNAs were microinjected into the cytoplasm of GV oocytes, MII oocytes (followed by ICSI), and zygotes. The results showed that the ZAR1L-EGFP signal formed cloud-like structures in the cytoplasm of some fully grown GV oocytes (7.8%, 4/51, Fig. 2E) and formed cytoplasmic foci in most (94.9%, 111/117) of the late two-cell-stage embryos (Fig. 2I). ZAR1L-EGFP cytoplasmic foci could not be observed in MI, MII oocytes, and zygotes (Fig. 2D, F–H). The C-terminus fragment, however, showed similar cytoplasmic foci localization (99.2%, 118/119; Fig. 2J) with the full-length ZAR1L protein. The N-terminus deleted form, ZAR1L ΔN-EGFP, was observed to form weak cytoplasmic foci in most of the late 2-cell-stage embryos (81.5%, 66/81; Fig. 2K). The C249S/C254S-Flag-EGFP point mutant form completely lost its capacity to form cytoplasmic foci (Fig. 2L). The full-length ZAR1 predominantly localized to cytoplasmic foci in late 2-cell-stage embryos (Fig. 2M).

ZAR1L C-Terminus Expression Induced Embryonic Arrest at the Two-Cell Stage

In order to study the function of ZAR1L and its mutants in preimplantation embryonic development, the mRNAs that encoded full-length and mutant ZAR1L were microinjected into zygotes. The results showed that the zygotes injected with EGFP (Fig. 3A–D), ZAR1L-EGFP (Fig. 3E–H), ZAR1L C249S/C254S-Flag-EGFP (Fig. 3Q–T), as well as the full-length ZAR1-EGFP (Fig. 3U–X), could develop to blastocyst stage in vitro. However, most of the zygotes injected with ZAR1L ΔN-EGFP (80.2%, 89/111; Fig. 3I–L) and ZAR1L Cter-Flag-EGFP (96.9%, 156/161; Fig. 3M–P) arrested at the two-cell stage. When EGFP was injected into one blastomere of a two-cell-stage embryo, the injected blastomere was not affected and developed to blastocyst normally (Fig. 3Y). When ZAR1L Cter-Flag-EGFP was injected into one blastomere of a two-cell-stage embryo, the injected blastomere arrested after one cell division. The other non-injected blastomere, however, grew to form a small blastocyst (Fig. 3Z).

Figure 3. Mutant ZAR1L induced cell cycle arrest when injected into zygotes or one blastomere of the two-cell-stage embryos. A–D: The embryos developed normally when EGFP was injected into the zygotes of the control group. E–H: ZAR1L-EGFP-injected embryos develop normally to the blastocyst stage. I–P: Most of the ZAR1L ‡N-EGFP (I–L) and ZAR1L Cter-EGFP (M–P) -injected embryos arrested at a late 2-cell stage. Q–T: Most of the ZAR1L-C292S/C295S-EGFP-injected embryos can develop to the blastocyst stage. U–X: The ZAR1-EGFP-injected embryos develop to the blastocyst stage. Panels of each group show embryos at 36, 72, 96, and 120 hr after hCG treatment, respectively. Y, Z: The one blastomere-injected embryos of the EGFP control group and ZAR1L Cter-EGFP group 120 hr after hCG injection. About half of the cells were EGFP-positive in the control blastocyst (Y). The ZAR1L Cter-EGFP-injected blastomere arrests after one cell division (arrows) and the other non-injected blastomere develops to form a small blastocyst (Z). The inserted small panel in Z shows the cell cycle of the ZAR1L Cter-EGFP injected blastomere is delayed. Zygote mRNA injection was performed during 20–22 hr after hCG treatment. Samples were photographed at 48, 72, 96, and 120 hr after hCG treatment. Scale bars = (A–L, Q–X) 100 μm; (M–P) 150 μm; (Y, Z) 20 μm.

ZAR1L Cter-Flag-EGFP Mutant Did Not Affect BrdU Incorporation But Affected BrUTP Incorporation

In order to investigate whether DNA replication was affected by the ZAR1L Cter-Flag-EGFP, BrdU incorporation assay was performed after the first cell cleavage. BrdU staining with anti-BrdU antibody showed that the BrdU incorporation in the ZAR1L Cter-Flag-EGFP group was similar to the NLS-EGFP control group (Fig. 4A and B).

Figure 4. ZAR1L Cter-EGFP does not affect BrdU incorporation but affects BrUTP incorporation. A, B: BrdU staining of one typical NLS-EGFP-injected 2-cell embryo (A) and ZAR1L Cter-EGFP-injected 2-cell embryos at 36 hr after hCG treatment. High BrdU staining signal was detected in both groups (arrows). C–J: BrUTP staining of injected embryos from middle to late 2-cell stage. C, E, G, I: Water-injected control embryos. D, F, H, J: ZAR1L Cter-EGFP-injected embryos. C–D, E and F, G and H, I and J show the embryos that were injected with BrUTP at 43, 44.5, 46, and 48 hr, respectively, after hCG treatment and collected 90 min after each injection. The BrUTP signals are significantly down-regulated as compared with the control groups at late 2-cell stage (I and J, arrows). Scale bars = 20 μm.

Further, to investigate whether RNA synthesis was affected by the ZAR1L Cter-Flag-EGFP, BrUTP incorporation assay was performed at late 2-cell stage (major zygotic genome activation). BrUTP staining with the anti-BrdU antibody from middle 2-cell stage to late 2-cell stage showed that BrUTP incorporation was moderately down-regulated by ZAR1L Cter-Flag-EGFP (Fig. 4I and J) at late 2-cell embryos. Statistics analysis showed that BrUTP incorporation levels in Zar1l Cter-EGFP injected embryos were down-regulated significantly from middle to late 2-cell stages, as compared with the control group (Supp. Fig. 1, which is available online).

ZAR1L Cter-Flag-EGFP Reduced H3K4me2 and H3K4me3 Methylation Levels and Active RNA Polymerase II (phosphor-Rpb1) Level

Histones H3K4 and H3K9 methylation play important roles in regulating genome structure and gene transcription. In order to confirm and explain how the BrUTP incorporation was affected by the ZAR1L Cter-Flag-EGFP mutant, the RNA transcription-related histone H3K4 methyl-modifications were determined by immunostaining. The results showed that when ZAR1L Cter-Flag-EGFP mutant was injected into zygotes, H3K4me2 and H3K4me3 were down-regulated significantly (Fig. 5C to F), while H3K4me1 was up-regulated moderately (Fig. 5A and B). It is interesting that in about half of the late 2-cell embryos, the H3K4me2 and H3K4me3 level in one blastomere usually down-regulated earlier/faster than the other one (Fig. 5G and H). Quantification and statistics analysis showed that H3K4me1 levels were low in 81.8% of the EGFP-injected late 2-cell embryos. And H3K4me1 levels were middle or high in 89.9% of the Zar1l Cter-EGFP-injected late 2-cell embryos (Fig. 5I). H3K4me2 and H3K4me3 levels were high in 70.7 and 72.3% of the EGFP injected late 2-cell embryos (Fig. 5J and K). However, H3K4me2 and H3K4me3 levels were high in only 18.2 and 14.1% of the Zar1l Cter-EGFP injected late 2-cell embryos (Fig. 5J and K).

Figure 5. ZAR1L Cter-EGFP induces down-regulation of H3K4 methyl-modification and the largest subunit of active RNA polymerase II (phosphor-Rpb1). A, B: ZAR1L Cter-EGFP moderately induces up-regulation of H3K4me1 level in late two-cell-stage embryos (arrow). C–F: ZAR1L Cter-EGFP induces significant down-regulation of H3K4me2 (C and D) and H3K4me3 (E and F) in late two-cell-stage embryos (arrows). G, H: ZAR1L Cter-EGFP severely affects the largest subunit of active RNA polymerase II (phospho-Rpb1) (arrow). Scale bars = 20 μm. I–K: Quantification and statistics analysis of H3K4me1, H3K4me2, and H3K4me3. H3K4me1 levels were low in 81.8% of the EGFP-injected late 2-cell embryos and its levels were middle or high in 89.9% of the Zar1l Cter-EGFP-injected late 2-cell embryos (I). H3K4me2 and H3K4me3 levels were high in 70.7 and 72.3% of the EGFP-injected late 2-cell embryos (J and K). However, H3K4me2 and H3K4me3 levels were high in only 18.2 and 14.1% of the Zar1l Cter-EGFP-injected late 2-cell embryos (J and K). L: Quantification and statistics analysis of Phospho-Rpb1 levels. Phospho-Rpb1 levels were high in 82.3% of the EGFP-injected late 2-cell embryos. However, its levels were low in 67.3% of the Zar1l Cter-EGFP-injected late 2-cell embryos (L).

Activated RNA polymerase II (phosphorylated Rpb1 at C-terminal Ser2/5 repeats, Phospho-Rpb1) was one of the key factors that represented the mRNA transcription/synthesis level. In order to confirm the down-regulation of RNA synthesis (as indicated by BrUTP incorporation assay), we further performed Phospho-Rpb1 immunostaining. The results showed that the Phospho-Rpb1 signal was also down-regulated significantly (Fig. 5G and H). Quantification and statistics analysis showed that Phospho-Rpb1 levels were high in 82.3% of the EGFP injected late 2-cell embryos. However, Phospho-Rpb1 levels were low in 67.3% of the Zar1l Cter-EGFP-injected late 2-cell embryos (Fig. 5L).

ZAR1L Cter-Flag-EGFP Significantly Down-Regulated H3K9me2 But Up-Regulated H3K9me3 in Late Two-Cell-Stage Embryos

To investigate whether Histone H3K9 methylation levels changed according to the phenotypes observed above, we performed immunostaining of H3K9me2 and H3K9me3 at late 2-cell-stage embryos. The results demonstrated that the H3K9me2 signal was dramatically down-regulated (Fig. 6A–C) in most of the late 2-cell embryos that were injected with ZAR1L Cter-Flag-EGFP. The H3K9me3 signal, however, was dramatically up-regulated and formed a perinuclear ring in the ZAR1L Cter-Flag-EGFP group (Fig. 6D–F). The H3K9me3 signal was dramatically down-regulated for a short period of time in normal late two-cell-stage embryos (Fig. 6D). Statistics analysis demonstrated that about 77% of the EGFP control group embryos showed a high level of H3K9me2. In contrast, 74% of the ZAR1L Cter-Flag-EGFP group embryos showed very weak levels of H3K9me2. Our results showed that about 5% (3/57) of the late two-cell-stage embryos in the control groups exhibited strong staining of anti-H3K9me3. However, more than 80% (83.9 ± 6.0%, n = 4; 49/58) of the late two-cell-stage embryos in the ZAR1L Cter-Flag-EGFP group had strong anti-H3K9me3 staining (Fig. 6E and F). DAPI staining showed that crescent-like nuclei were observed in about a half of the ZAR1L Cter-Flag-EGFP-injected two-cell-stage embryos (Figs. 5B, D, 6D). The crescent-like nucleus and H3K9me3 perinuclear ring had not been observed in the control groups (Figs. 5A,C, 6A,D).

Figure 6. ZAR1L Cter-EGFP down-regulates H3K9me2 but up-regulates H3K9me3 methylation level in late 2-cell embryos. A–C: H3k9me2 is down-regulated significantly by ZAR1L Cter-EGFP mutant in 74.0% of the late 2-cell embryos (arrow). D–F: H3k9me3 is up-regulated significantly by ZAR1L Cter-EGFP mutant in 83.8% of the late 2-cell embryos. Open arrow indicates the down-regulation of H3K9me3 in the control group (D). Solid arrow indicates the high level and perinuclear ring formation of H3K9me3 signal (E). One typical crescent-like nucleus is shown (E, arrow). Scale bars = 20 μm.

ZAR1L Colocalized With P-Body Components and Germline-Specific Chromatoid Body Components in Somatic Cells

In order to explore the functional properties and molecular mechanisms of ZAR1L protein, we over-expressed ZAR1L-Flag-EGFP in 293T cells and then performed co-immunoprecipitation with the anti-Flag monoclonal antibody. A set of widely expressed P-body components was found in the immunoprecipitates with ZAR1L-Flag-EGFP, including EIF2C1, EIF2C2, DDX6, and LSM14A. To test colocalization between ZAR1L and the P-body components, we co-expressed ZAR1L and the P-body components transiently in mouse pancreatic PNA-HSA double-low (PHDL) cells and human 293T cells. The results showed that ZAR1L colocalized well with EIF2C1, EIF2C2, DDX6, and LSM14A in mouse PHDL cells (Fig. 7A–D) and human 293T cells (data not shown).

Figure 7. ZAR1L colocalizes with P-body components and germline-specific chromatoid-body components in mouse pancreatic PHDL epithelial cells. A, B: ZAR1L-EGFP colocalizes well with human EIF2C1-myc (A) and human EIF2C2-myc (B), respectively (arrows). C, D: ZAR1L-RFP colocalizes well with DDX6-EGFP (C) and LSM14A-EGFP (D), respectively (arrows). E–G: ZAR1L colocalizes well with germline-specific mouse Piwil1-Flag (E), Piwil2-Flag (F), and human LIN28-RFP (G), (arrows). H: ZAR1L-RFP colocalizes extensively with ZAR1-EGFP (arrows). Monoclonal antibody against Myc tag and Flag tag, and anti-mouse Alex-594 secondary antibody are used. Scale bars = 10 μm.

The chromatoid body has a structure similar to that of the P-body in the male germline. It was observed in cytoplasm as cytoplasmic foci. Because ZAR1L was a female germline-specific protein, we assessed whether ZAR1L colocalized with germline-specific chromatoid body components. The results showed that ZAR1L colocalized extensively with the germline-specific PIWIL1, PIWIL2, and LIN28 when they were ectopically expressed in PHDL cells (Fig. 7E–G) and 293T cells (data not shown). In addition, ZAR1L also colocalized extensively with the germline-specific ZAR1.

Colocalization of ZAR1L, ZAR1L C-ter-Flag-EGFP, LIN28, LSM14A, and ZAR1 in 2-Cell Embryos

The protein sequence alignment analysis suggested that mouse ZAR1L protein might have three functional domains. In order to characterize the function properties of different domains, series mutants have been designed to analyze their sub-cellular localization and colocalization with germline P-body/C-body components. Our results showed that both ZAR1L-EGFP and ZAR1L Cter-Flag-EGFP colocalized extensively with LIN28-RFP (Fig. 8A and B) and LSM14A-RFP (Fig. 8C and D) in late 2-cell-stage embryos. In order to characterize the colocalization between ZAR1L C-terminus and ZAR1L or ZAR1, we co-expressed them in zygotes and analyzed them in late 2-cell embryos. We found that both ZAR1L-RFP and ZAR1-EGFP colocalized extensively with the C-terminus of ZAR1L in nearly all the 2-cell embryos in the cytoplasmic foci region (Fig. 8E–H). However, we also found that ZAR1L Cter-Flag-EGFP and ZAR1L Cter-RFP localized to the nucleus in about one fourth of the late 2-cell embryos (23.5%, 12/51 and 27.5%, 19/69, respectively, Fig. 8F and H). Further experiment showed that ZAR1L-RFP colocalized extensively with ZAR1-EGFP (Fig. 8I) and ZAR1L Cter-RFP colocalized well with ZAR1 Cter-EGFP (Fig. 8J). Serial observation showed that the ZAR1L C-terminus moved from the nucleus to the cytoplasmic foci in middle to late 2-cell-stage embryos (Fig. 4D–J, left panels). The colocalization study of the Zar1l, Zar1, and Zar1l C-terminus demonstrated that the full-length Zar1 and Zar1l, as well as Zar1l, C-terminus, could shuttle from the nucleus to the cytoplasm. Simultaneously, we found that co-expression of ZAR1L and C-terminus of ZAR1L could partially (48.2%, 41/85) rescue the 2-cell-block phenotype that was induced by the C-terminus of ZAR1L at 72 hr after hCG injection (Fig. 8K and L). However, these embryos still failed to develop to the blastocyst stage at 120 hr after hCG injection (Fig. 8M and N).

Figure 8. Colocalization of ZAR1L, ZAR1L Cter mutant, LIN28, LSM14A, and ZAR1 in 2-cell embryos. A, B: Both ZAR1L-RFP (A) and ZAR1L Cter-EGFP (B) colocalize extensively with LIN28-EGFP in late two-cell-stage embryos (arrows). C, D: Both ZAR1L-RFP (C) and ZAR1L Cter-EGFP (D) colocalize extensively with LSM14A-EGFP in late two-cell-stage embryos (arrows). ZAR1L Cter-EGFP shows higher nucleus localization (open arrows). E: ZAR1L-RFP colocalizes extensively with the ZAR1L Cter-Flag-EGFP in cytoplasmic foci. F: ZAR1L Cter-Flag-EGFP localizes to the nucleus in 23.5% (12/51) of the late 2-cell embryos (open arrows). G: ZAR1-EGFP colocalizes well with ZAR1L Cter-RFP in cytoplasmic foci. H: ZAR1L Cter-RFP localizes to the nucleus in 27.5% (19/69) of the late 2-cell embryos (open arrows). I, J: ZAR1L-RFP colocalizes extensively with ZAR1-EGFP, and ZAR1L Cter-RFP colocalizes well with ZAR1 Cter-EGFP (Fig. 8J, arrows). Open arrows in D, F, and H indicate the relatively high nuclear localization of ZAR1L C-terminus mutant. Scale bars = 20 μm. K–N: Co-injection of Zar1l-EGFP and Zar1l Cter-EGFP could partially rescue the 2-cell-block phenotype that was induced by Zar1l Cter-EGFP. At 72 hr after hCG injection, EGFP and Zar1l Cter-RFP co-injected embryos were still blocked at the 2-cell stage (K, open arrows), while 48.2% (41/85) of Zar1l(full)-EGFP and Zar1l Cter-RFP co-injected embryos were 3-cell or 4-cell embryos (L, solid arrows). At 120 hr after hCG injection, EGFP singly injected embryos developed to blastocyst stage (M). However, 76.1% (64/84) of Zar1l(full)-EGFP and Zar1l Cter-RFP co-injected embryos developed beyond the 2-cell stage, but were still restrained to 3-cell to 8-cell embryos (N, open arrows).

Mouse ZAR1L Protein Interacts With Human LIN28

In order to identify some molecular mechanisms of ZAR1L function, we performed immunoprecipitation and silver staining experiments to try to find out some proteins that might interact with mouse ZAR1L. Mass spectrometry analysis revealed that human LIN28 protein was one of the ZAR1L immunoprecipitated components when mouse ZAR1L and human LIN28 were co-expressed in 293T cells. Moreover, ZAR1L protein was rich in the 60-kD band that was in the lysates generated from LIN28-Flag-EGFP immunoprecipitation (with anti-Flag monoclonal antibody) (Fig. 9A). Western blotting analysis further supported the interaction between mouse ZAR1L and human LIN28 (Fig. 9B). Further analysis showed that both ZAR1L-RFP and ZAR1L Cter-RFP colocalized extensively with LIN28-EGFP in late two-cell-stage embryos (Fig. 8A and B).

Figure 9. Biochemical analysis of interaction between mouse ZAR1L and human LIN28 in 293T cells. A: Silver staining of samples obtained by immunoprecipitation. Left: ZAR1L Cter-EGFP immunoprecipitation and silver staining. Middle: ZAR1L-Flag-EGFP was co-transfected with hLIN28-RFP into 293T cells and immunoprecipitation was performed with anti-Flag monoclonal antibody. Right: ZAR1L-RFP was co-transfected with hLIN28-Flag-EGFP into 293T cells and immunoprecipitation was performed with anti-Flag monoclonal antibody. Arrows indicate the bands of ZAR1L (top) and human LIN28 (bottom). B: Western blotting analysis of samples obtained by immunoprecipitation. ZAR1L-RFP and LIN28-Flag-EGFP are co-transfected into 293T cells. ZAR1L-RFP and LIN28-Flag-EGFP are used as bait proteins. Before elution, samples are treated with or without RNaseA (20 mg/ml) for 30 min, respectively. Anti-ZAR1L polyclonal antibody and anti-Flag monoclonal antibody have been used.

ZAR1L Cter-Flag-EGFP Down-Regulated a Set of Chromatin Modification-Associated Genes in mRNA Level

It was well known that P-body components played important roles in regulating post-transcriptional mRNA stability and translation. In order to characterize whether the ZAR1L Cter-Flag-EGFP mutant resulted in down-regulation of some zygote and 2-cell embryo highly expressed genes, we performed RT-PCR analysis of a group of selected genes. Our data showed that a set of chromatin-modification genes, including Dppa2, Dppa4, and Piwil2 (Mili), were dramatically down-regulated by ZAR1L Cter-Flag-EGFP at an mRNA level in late two-cell-stage embryos (Fig. 10A). However, Cbx1, Dppa3, Oct4, Gata6, and Tbpl1 genes were not affected significantly (Fig. 10A). In order to confirm the effect of ZAR1L Cter-Flag-EGFP expression on DPPA2 in the protein level, we performed immunostaining with DPPA2-specific antibody. Our results showed that DPPA2 was dramatically down-regulated at protein levels in about 90% (90.8 ± 2.2%, n = 4; total 42/54) of the late two-cell-stage embryos that were injected with ZAR1L Cter-Flag-EGFP (Fig. 10B–D). Only about 4% (2/55) of the late two-cell-stage embryos had a high staining of DPPA2 (Fig. 10D). In the EGFP control group, about 95% (94.7 ± 3.8%, n = 4; 50/53) of the injected embryos had strong staining of DPPA2 (Fig. 10D). To preliminarily test whether DPPA2 had important roles in preimplantation development, the mRNA corresponding to the N-terminus deleted DPPA2-ΔN-EGFP mutant (Fig. 10E) was injected into zygotes. The DPPA2-ΔN-EGFP mutant predominantly localized in the nucleus and induced two-cell-stage arrest (Fig. 10F and G). To further investigate whether DPPA2 was mainly responsible for the ZAR1L Cter-Flag-EGFP-induced 2-cell arrest phenotype, we performed a DPPA2 rescue experiment. We found that over-expression of DPPA2 could not rescue the 2-cell arrest phenotype that was induced by ZAR1L Cter-Flag-EGFP (Fig. 10H and I). Simultaneously, we found that, if co-injected with ZAR1L Cter-Flag-EGFP, ectopically expressed DPPA2-RFP protein was also down-regulated significantly at late 2-cell stage (Fig. 10J and K).

Figure 10. ZAR1L Cter-EGFP induces dramatic down-regulation of a set of chromatin modification factors. A: RT-PCR analysis. ZAR1L Cter-EGFP induces dramatic down-regulation of Dppa2, Dppa4, and Piwil2 at the mRNA level. B, C: Immunofluorescent staining. ZAR1L Cter-EGFP induces dramatic down-regulation of DPPA2 at the protein level (arrow). B is the NLS-EGFP control group. C: The ZAR1L Cter-EGFP group. D: Statistical analysis of the DPPA2 staining intensity (analyzed by Adobe Photoshop). The DPPA2 staining signal is dramatically down-regulated in 90.8% of late two-cell-stage embryos in the ZAR1L Cter-EGFP group. E: The construct design for DPPA2 and DPPA2 ΔN-EGFP. The N-terminus SAP domain is deleted and the nuclear localization signal remains intact in the mutant construct. F, G: Development of zygotes that were injected with DPPA2 ΔN-EGFP, 60 hr after hCG injection. The N-terminus deleted DPPA2 localizes to the nucleus and induces arrest at the 2-cell stage. H–K: Co-injection of ZAR1L Cter-EGFP and DPPA2-RFP into zygotes. H and J show DPPA2-RFP signals and the inset panels show the bright field, at 42 and 72 hr after hCG treatment, respectively. Over-expression of DPPA2 does not rescue the 2-cell arrest phenotype that was induced by ZAR1L Cter-EGFP. I and K show that DPPA2 protein is down-regulated dramatically in co-injected late 2-cell embryos (arrows). Some embryos occasionally develop beyond the 2-cell stage (arrows in the inset bright field panel in I). Scale bars = (B, C, J, and K) 20 μm; (F–I) 100 μm.

DISCUSSION

Wu et al. (2003a) have reported that Zar1 (−/−) mice are viable and grossly normal, but the Zar1(−/−) females are infertile. They found that ovarian development, oogenesis, and fertilization are unimpaired in Zar1 (−/−) mice. Most of the embryos generated from Zar1(−/−) females, however, arrest at the one-cell and two-cell stages (Wu et al., 2003a). They observed that pronuclei formation and DNA replication occur, but the maternal and paternal genomes remain separate entities in arrested zygotes. The mechanism by which null Zar1 expression caused zygote and two-cell-stage arrest is unknown. In this study, we characterized a Zar1-like gene, XM_359149, which encoded a ZAR1-like protein and we have designated it as Zar1l. Zar1l ORF sequences have a poor similarity to the Zar1 ORF. However, they exhibit a high similarity in protein sequence in their C-terminus. When our Zar1l project was ongoing, Sangiorgio et al. (2008) reported the expression and preliminary sequence analysis of Bovine ZAR1-like gene and its orthologs in vertebrates. Unlike bovine ZAR1-like gene, we found that mouse Zar1-like gene was predominantly expressed in ovaries, oocytes, and early preimplantation embryos. The differences in expression patterns suggest that ZAR1L may have species-variant functions. The ZAR1-like orthologous proteins are conserved in their predicted N-terminus, middle-region, and C-terminus domains. The C-terminus zinc-finger-containing domains are well conserved from zebrafish to human (data not shown). These results suggested that the mouse Zar1l gene may play some important roles in female germline and/or embryonic development.

Mutant ZAR1L Induced Cell Cycle Arrest in Preimplantation Embryos

Protein sequence analysis suggested that mouse ZAR1L might contain three functional domains: an N-terminus unknown domain, a CSE2-like middle domain, and a C-terminus atypical PHD zinc finger domain (Wu et al., 2003b). Series mutations were designed for functional analysis of ZAR1L protein in preimplantation embryos. Our data showed that both the C-terminus ZAR1L (ZAR1L Cter-Flag-EGFP) and the N-terminus deleted ZAR1L (ZAR1L ΔN-EGFP) caused two-cell-stage arrest. The ZAR1L Cter-Flag-EGFP induced a more severe phenotype. When ZAR1L Cter-Flag-EGFP was injected into one blastomere of the two-cell embryos, it resulted in arrest of the injected blastomere after one time cell division, whereas the other blastomere grew to form a small blastocyst. These data suggested that mouse Zar1l might play important roles in preimplantation development in mice.

Mutant ZAR1L Did Not Affect DNA Replication But Affected RNA Synthesis

To find out whether DNA replication and/or RNA synthesis was affected by the ZAR1L Cter-Flag-EGFP mutant, we performed BrdU incorporation assay (Vitale et al., 1989) and BrUTP incorporation assay (Aoki et al., 1997). Our results demonstrated that BrdU incorporation was not affected, but the BrUTP incorporation was significantly down-regulated in late 2-cell embryos. These data indicated that RNA synthesis during late 2-cell embryos was affected by ZAR1L Cter-Flag-EGFP mutant. RNA synthesis is one of the key events during major zygotic gene activation. To further confirm the influence of ZAR1L Cter-Flag-EGFP mutant on zygotic gene activation, we detected the largest active RNA polymerase II subunit, Phospho-Rpb1(Ser2/5) (Svejstrup, 2002; Shilatifard et al., 2003). Dramatic down-regulation of phosphor-Rpb1 indicated the RNA synthesis activity was affected by ZAR1L Cter-Flag-EGFP.

ZAR1L Cter-Flag-EGFP Induced Abnormalities in H3K4me1/2/3 and H3K9me2/3 Modifications in Two-Cell-Stage Embryos

It was reported that Histone H3K4 methylation levels correlated with the genome activation (Lepikhov and Walter, 2004; Lepikhov et al., 2008; VerMilyea et al., 2009), while H3K9me3 methylation (Lachner and Jenuwein, 2002; Horn and Peterson, 2006; Grewal and Jia, 2007) correlated with genome inactivation. So we further determined the H3K4me1/2/3 levels and H3K9me2/3 levels. Our results showed that H3K4me2/3 was significantly down-regulated, while H3K4me1 was moderately up-regulated. These data consisted with the transcription activity as indicated by a BrUTP incorporation test and phospho-Rpb1 staining. H3K9me2/3 methylation staining results showed that the H3K9me3 demethylation during the late 2-cell stage was severely affected by the ZAR1L Cter-Flag-EGFP mutant and, correspondingly, the H3K9me2 was at low levels. The time point of down-regulation of H3K9me3 at the late two-cell stage in the EGFP control group is in accordance with the wave of major zygotic genome activation (embryonic gene transcription) (Nakayama et al., 2001; Lachner and Jenuwein, 2002; Hamatani et al., 2004; Yeo et al., 2005; Horn and Peterson, 2006; Grewal and Jia, 2007). The perinuclear ring formation of the H3K9me3 signal, as well as the crescent-like nucleus formation, indicated that the nuclear morphology was changed by the ZAR1L Cter-Flag-EGFP mutant, and the chromatin might be in a heterochromatin state. These data are also consistent with the results that significant down-regulation of BrUTP incorporation and phosphor-Rpb1 levels in the nucleus have been induced by Zar1l Cter-EGFP.

ZAR1L Colocalized With the P-body and C-body Components in Both Somatic Cells and Late 2-Cell-Stage Embryos

Many maternal mRNAs accumulate in growing oocytes and are stored in MII oocytes. They are translationally repressed until fertilization. Most of the maternal mRNAs have been degraded by the end of the two-cell stage. Correspondingly, major zygotic genome activation occurs at the late two-cell stage. How the maternal mRNAs are tightly controlled in terms of their stability, translational repression, and/or initiation, and degradation is largely unknown. Recent studies have revealed the existence of specific mRNA processing bodies (P-bodies) as multiple cytoplasmic foci in somatic cells (Hannon, 2002; van Dijk et al., 2002; Sheth and Parker, 2003; Cougot et al., 2004; Brengues et al., 2005; Liu et al., 2005). P-bodies contain untranslated mRNAs and can serve as sites of mRNA translational repression and degradation. P-bodies are highly dynamic structures, and the components are altered depending on the cell state. Many proteins have been reported to be localized to P-body structures (Hannon, 2002; van Dijk et al., 2002; Sheth and Parker, 2003; Cougot et al., 2004; Brengues et al., 2005; Fillman and Lykke-Andersen, 2005; Liu et al., 2005; Yang et al., 2006; Eulalio et al., 2007; Parker and Sheth, 2007; Pressman et al., 2007). In the male germline, chromatoid-body (C-body) structures have been found to be similar to the P-body structures (Matsumoto et al., 2005; Kotaja et al., 2006; Kotaja and Sassone-Corsi, 2007).

The male germline-specific cytoplasmic foci (chromatoid body) share components found in somatic cell P-bodies, such as the Agonaute proteins, some of the RNA enzymes, and ribosomal proteins. We found that the maternal effect gene Zar1l encoded a female germline-specific protein ZAR1L, which localized to the cytoplasmic foci structures in late two-cell-stage embryos. Based on the knowledge that mRNA processing mechanisms are evolutionarily conserved from germ cells to somatic cells, we speculated that ZAR1L might be involved in P-body- or C-body-like structures. Indeed, we found that mouse ZAR1L was colocalized extensively with widely expressed P-body components, including EIF2C1, EIF2C2, DDX6, and LSM14A and with germline-specific chromatoid-body components including PIWIL1, PIWIL2, and LIN28 in somatic cells. We also confirmed that mouse ZAR1L colocalized extensively with LIN28 and LSM14A in late 2-cell embryos. Our data further indicated co-expression of ZAR1L could partially rescue the 2-cell-block phenotype that was caused by ZAR1L C-terminus. The colocalization study of the Zar1l, Zar1, and Zar1l C-terminus demonstrated that the full-length Zar1 and Zar1l, as well as Zar1l C-terminus, could shuttle from the nucleus to the cytoplasm. These results indicate that the C-terminus of ZAR1L may induce 2-cell blocks through a dominant-negative effect. Because of the lack of commercialized ZAR1L antibody, we customized peptide antibody against mouse ZAR1L. It works well for Western blotting but not for immunostaining. We confirmed the interaction between mouse ZAR1L and human LIN28 in somatic cells using immunoprecipitation and Western blotting. Our results also showed that ZAR1L extensively colocalized with ZAR1. These data demonstrated that ZAR1L, as well as ZAR1, may play some roles in P-body and/or C-body structures and might have functions in regulating oocyte-to-embryo transition.

The ZAR1L Cter-EGFP Down-Regulated a Set of Chromatin Modification Factors, Including Dppa2, Dppa4, and Piwil2 at the mRNA Level

Nuclear reprogramming is a critical event that occurs during zygotic genome activation (Schultz, 1993; Aoki et al., 1997; Latham, 1999; Latham and Schultz, 2001; Ma et al., 2001; Hamatani et al., 2004; Minami et al., 2007; Stitzel and Seydoux, 2007), through which the transcriptionally inactive genome changes into an active genome. Several maternal factors are associated with nuclear reprogramming. Depletion of Smarca4 (Brg1), a chromatin-remodeling factor, caused arrest at the two-cell stage, which was accompanied by down-regulation of a multitude of mRNAs (Bultman et al., 2006). Dppa3(Stella) is required for protection of the maternal genome from DNA-demethylation during early embryonic development (Nakamura et al., 2007). Similar to DPPA3, DPPA2 and DPPA4 have one DNA-binding SAP domain and one uncharacterized C-terminal domain, and are associated with chromatin (Aravind and Koonin, 2000; Maldonado-Saldivia et al., 2007; Masaki et al., 2007). Recent studies have indicated that mouse DPPA4 protein associated with transcriptionally active chromatin in ES cells (Masaki et al., 2007). Comprehensive ChIP-on-chip analysis demonstrated that POU5F1 (OCT-3/4), SOX-2, and NANOG each bind to the Dppa4 promoter region in human ES cells (Boyer et al., 2005). However, the roles of DPPA2 and DPPA4 in early embryonic development remain largely unknown. Our results showed that Dppa2 and Dppa4 were dramatically down-regulated by the mutant ZAR1L Cter-EGFP in late two-cell-stage embryos. In order to test whether DPPA2 protein plays important roles in preimplantation development, we designed a dominant-negative mutant of mouse DPPA2 and injected it to the zygotes. Our data showed that deletion of the N-terminal SAP domain of mouse DPPA2 caused arrest at the two-cell stage in vitro. Our data suggested that DPPA2 may play an important role in embryonic development. We tried to rescue the 2-cell arrest phenotype through over-expression of DPPA2 but failed. These data indicated that DPPA2 is one of the important but not the dominant factors affected by ZAR1L Cter-Flag-EGFP.

Piwi family members regulate chromatin structure, transposon control, mRNA transcription and translation, and mRNA degradation through interactions with piRNAs and associated complexes (Kuramochi-Miyagawa et al., 2004; Parker et al., 2004; Kavi et al., 2006; Lau et al., 2006; Aravin et al., 2007; Brower-Toland et al., 2007; Carmell et al., 2007; Hartig et al., 2007; Houwing et al., 2007; Lin, 2007; Klattenhoff and Theurkauf, 2008). Recently, piRNAs have been isolated from murine mature oocytes (Brennecke et al., 2008; Tam et al., 2008; Watanabe et al., 2008). Piwil2 (Mili), but not Piwil1 (Miwi), or Piwil4 (Miwi2) is specifically expressed in mature oocytes (Watanabe et al., 2008). Piwi family proteins and piRNAs play important roles in chromatin modification and genome stability (Aravin et al., 2007, 2008; O'Donnell and Boeke, 2007). Our data showed that ZAR1L Cter-EGFP dramatically induced down-regulation of Piwil2 mRNAs in two-cell embryos. Down-regulation of Dppa2, Dppa4, and Piwil2 mRNA by ZAR1L Cter-Flag-EGFP in vitro in late two-cell-stage embryos indicated that ZAR1L may correlate with a set of mRNAs' stability or degradation.

In summary, our data demonstrated that ZAR1L plays important roles in regulating oocyte-to-embryo transition and preimplantation development. The ZAR1L Cter-Flag-EGFP mutant induced epigenetic abnormalities and down-regulation of a group of chromatin modification factors in late two-cell-stage embryos and finally caused arrest at the two-cell stage. ZAR1L colocalized with multiple mRNA chromatoid-body/processing-body components in somatic cells and late two-cell-stage embryos, and it interacted with LIN28. ZAR1L could be the first tissue- and stage-specific chromatoid-body/processing-body component that has been identified in 2-cell-stage mouse embryos. Zar1l knockout mice will be generated to analyze the functional role of ZAR1L in the female germline and during embryonic development. The biochemical nature of the domains of ZAR1L, as well as ZAR1, remains to be further characterized.

EXPERIMENTAL PROCEDURES

Animals, Collection of Oocytes and Embryos

B6D2F1 (C57BL/6JxDBA2) female mice (8–10 weeks old) were used for the collection of fully grown germinal vesicles (GV) and MII oocytes. GV oocytes were collected according to a previous study (Wang et al., 2008). Zygotes were collected from successfully mated B6D2F1 females. All studies adhered to procedures consistent with the National Institute of Biological Sciences Guide for the care and use of laboratory animals.

In Vitro Transcription, Microinjection, and Preimplantation Embryo Incubation

The predicted Zar1l ORF was cloned into pBS-RN3 (Lemaire et al., 1995), a modified in vitro transcription vector, in which the wild type and mutant ZAR1L were expressed with Flag and/or EGFP as fused proteins. In brief, zygotes were collected from adult male B6D2F1 mice at 18 hr after hCG treatment. mRNA injection was performed during 20–22 hr after hCG treatment. The surviving zygotes were cultured in KSOM medium. Samples were photographed at 48, 72, 96, and 120 hr after hCG treatment for developmental recording. Late 2-cell embryos were collected during 48–50 hr after hCG treatment. Capped RNAs were transcribed under the control of a T3 promoter with mMessage mMachine (Ambion, Austin, TX), as the protocol dictated. In vitro transcribed mRNAs were injected into GV, MII oocytes, and zygotes using a PIEZO micro-injector.

Cell Culture

293T cells were cultured in DMEM-based medium, which contained 10% FBS (Hyclone, Logan, UT), 2 mM glutamine, 1× nucleosides (Gibco, Gaithersburg, MD), 1× nonessential amino acids (Gibco), 1× beta-mercaptoethanol (Gibco), 2 mM glutamine (Gibco), 100 IU/ml penicilLin, and 100 μg/ml streptomycin (Gibco). Pancreatic PNA-HSA double-low cells were incubated under similar conditions as described above, with 3% FBS added.

Construction of Transient Expression Vectors and Transfection

The mouse Zar1l ORF was obtained by performing RT-PCR from adult ovarian tissue. The sequences were confirmed by sequencing. The DNA sequences of Zar1l ORF reported in this study have been deposited in the Genbank database (www.ncbi.nlm.nih.gov/Genbank, accession no. FJ858201). The ORF and mutants were cloned into the pEGFP-N1 vector for transient expression. Eukaryotic expression vectors inserted with human EIF2C1, human EIF2C2, mouse Ddx6, mouse Lsm14a, mouse Piwil1 (Miwi), mouse Piwil2 (Mili), and human LIN28 were co-transfected with Zar1l into PHDL cells or 293T cells to analyze their colocalization with mouse ZAR1L. Vigofect reagent was used according to the manufacturer's protocol (Vigorous). Cells were collected 24 or 36 hr after transfection. Please see Supp. Table S1 to view the primer sequences used.

RT-PCR

Total RNA samples were prepared from adult ovary, testis, liver, spleen, oocytes, and preimplantation embryos. The RNA was extracted using conventional methods for adult tissues. The PicoPure RNA isolation kit (Arcturus) was used to extract RNA from collected oocytes and preimplantation embryos. Reverse transcription and PCR were performed by conventional methods using MMLV reverse transcriptase (Promega, Madison, WI). Genomic DNA was extracted by conventional methods. The RNA was reverse-transcribed by MMLV reverse transcriptase (Promega) and amplified by PCR for 25 or 30 cycles. Primers were selected that encompassed the intronic sequences. PCR cycling was performed at 98°C for 2 min followed by 98°C for 15 sec, 54–60°C for 15 sec, 72°C for 40 sec, and finally 72°C for 8 min, using the PrimeSTAR HS DNA polymerase (Takara). Please see Supp. Table S2 to view the primer sequences.

Immunofluorescent Staining and Confocal Microscopy

Conventional immunostaining were performed for H3K4me1, H3K4me2, H3K4me3, H3K9me2, H3K9me3, Phospho-Rpb1, and Dppa2 antibodies. In brief, samples were fixed by 4% paraformeldehyde for 20 min. Then, the samples were permeabilized with 0.5% triton X-100 and blocked with 5% normal horse serum for 2 hr. The primary antibodies were incubated overnight at 4°C. The antibodies against H3K4me1, H3K4me2, H3K4me3, H3K9me2, H3K9me3 (all from Upstate, Billerica, MA), and Dppa2 (gifted by Dr. Western at the ARC Centre, Australia) were diluted at 1:300. Phospho-Rpb1 antibody (Cell Signaling Technology, Danvers, MA) was diluted at 1:100. For BrdU staining, BrdU was added to the medium at 28 hr after hCG treatment. The BrdU-labeled samples were collected 36 hr after hCG treatment. Samples were treated with 2 M HCl for 10 min after permeabilization and washing. Then, samples were washed once with pH 8.0 Tris-HCl and then with pH 7.4 PBS four times after HCl treatment. Conventional methods were used for the other staining steps. For BrUTP staining, a similar amount of 100 mM BrUTP (Sigma, St. Louis, MO) was injected into ZAR1L Cter-EGFP or EGFP pre-injected (at zygote stage) 2-cell embryos at 43, 44.5, 46, and 48 hr after hCG treatment. Samples were collected 90 min after each injection of BrUTP (Sigma). Both the incorporated BrdU and BrUTP were stained with anti-BrdU monoclonal antibody. The Alexa-594-conjugated secondary antibodies were incubated for 1 hr at room temperature. Samples were further counterstained with 100 ng/ml of DAPI. Images were obtained with an Olympus IX 71 microscope equipped with a CCD camera (DVC, Austin, TX), or LSM510 Meta confocal microscope (Zeiss, Oberkochen, Germany). The staining intensity was analyzed by Adobe Photoshop.

Immunoprecipitation and Western Blotting

293T cells were transfected to transiently express ZAR1L Cter-Flag-EGFP, ZAR1L-Flag-EGFP, ZAR1L-RFP, human LIN28-Flag-EGFP, and LIN28-RFP, either independently or in combination. Samples were collected 36 hr after transfection. Immunoprecipitations were performed according to the manufacturer's protocol (FLAGIPT-1; Sigma) or with anti-ZAR1L polyclonal antibody prepared using a specific polypeptide (134-RRPQDGE DEESQEE-147). The final concentration of the ZAR1L antibody was 1 μg/ml. Proteins were separated in 10% SDS-polyacrylamide gels and transferred to PVDF membrane. Immunoblotting analysis was performed using anti-Flag monoclonal and anti-ZAR1L polyclonal antibodies. Blots were detected using ECL (Amersham Biosciences, Pittsburgh, PA) according to the manufacturer's protocol. Blots were detected using ECL (GE Healthcare, London, UK).

Silver Staining and Mass Spectrometry Analysis

Proteins were separated in 10% SDS-polyacrylamide gels. Silver staining was performed according to the manufacturer's protocol (PROTSIL2-1KT, Sigma). The bands of interest were analyzed by a LTQ linear ion trap mass spectrometer (Thermo Electron, Waltham, MA).

Acknowledgements

We greatly appreciate Dr. Gurdon at the Gurdon Institute for providing the pBS-RN3 vector. We thank Dr. Western at the ARC Centre (Australia) for providing the DPPA2 antibody. We thank Dr. Satomi at Osaka University for providing the Piwil1 and Piwil2 cDNA. We thank the lab members and the two anonymous reviewers for helpful comments on the manuscript.