pMesogenin1 and 2 function directly downstream of Xtbx6 in Xenopus somitogenesis and myogenesis

Authors


Abstract

T-box transcription factor tbx6 and basic-helix-loop-helix transcription factor pMesogenin1 are reported to be involved in paraxial mesodermal differentiation. To clarify the relationship between these genes in Xenopus laevis, we isolated pMesogenin2, which showed high homology with pMesogenin1. Both pMesogenin1 and 2 appeared to be transcriptional activators and were induced by a hormone-inducible version of Xtbx6 without secondary protein synthesis in animal cap assays. The pMesogenin2 promoter contained three potential T-box binding sites with which Xtbx6 protein was shown to interact, and a reporter gene construct containing these sites was activated by Xtbx6. Xtbx6 knockdown reduced pMesogenin1 and 2 expressions, but not vice versa. Xtbx6 and pMesogenin1 and 2 knockdowns caused similar phenotypic abnormalities including somite malformation and ventral body wall muscle hypoplasia, suggesting that Xtbx6 is a direct regulator of pMesogenin1 and 2, which are both involved in somitogenesis and myogenesis including that of body wall muscle in Xenopus laevis. Developmental Dynamics 237:3749–3761, 2008. © 2008 Wiley-Liss, Inc.

INTRODUCTION

The T-box family of transcription factors has been identified in invertebrates and chordates and is reported to play a crucial role in embryonic development (Kavka and Green,1997; Papaioannou and Silver,1998; Smith,1999; Wilson and Conlon,2002; Showell et al.,2004). A member of this family, tbx6, is expressed in primitive streaks, paraxial mesoderms, muscle-lineage blastomeres, and tailbuds in various species (Chapman et al.,1996; Yasuo et al.,1996; Hug et al.,1997; Griffin et al.,1998; Uchiyama et al.,2001). It is also involved in mesoderm specification and differentiation. In tbx6-null mutant mouse embryos, ectopic neural tube–like structures are formed caudal to the forelimb bud where a somite is normally formed, suggesting that tbx6 is essential for the normal development of posterior paraxial mesoderm (Chapman and Papaioannou,1998). The mouse rib-vertebrae mutation is a hypomorphic tbx6 allele (Watabe-Rudolph et al.,2002; White et al.,2003), which affects somite formation, morphology, and patterning, and leads to malformations of the axial skeleton, such as split vertebrae and neural arches, as well as fusion of adjacent segments (Theiler and Varnum,1985; Beckers et al.,2000; Nacke et al.,2000). In Xenopus laevis, Xtbx6 is a transcriptional activator and the animal cap explants overexpressing Xtbx6 mRNA are reported to differentiate into ventral mesodermal tissues (Uchiyama et al.,2001).

Another group, the basic-helix-loop-helix (bHLH) transcription factors, also plays a key role in embryogenesis. This gene class functions in a variety of transcriptional regulatory networks, including skeletal myogenesis and neurogenesis (Yun and Wold,1996; Kageyama and Nakanishi,1997; Dambly-Chaudière and Vervoort,1998). Mespo/pMesogenin1 was isolated in X. laevis as a bHLH gene expressed in presomitic mesoderm (PSM) and posterior tailbud (Joseph and Cassetta,1999). Subsequently, its ortholog was isolated in mouse and it was revealed that mouse and Xenopus pMesogenins are more similar to each other in the bHLH region than either is to murine MesP1 and 2 or other MesP1/2 orthologs, such as chicken Meso-1 and Xenopus Thylacine1 and 2 (mespb) (Yoon et al.,2000). In X. laevis, new MesP family members, prsm01 and 02 (mespa), have been isolated (Pollet et al.,2005; Moreno et al.,2008). Therefore, pMesogenin1 belongs to an evolutionary branch that diverged from the MesP family. In pMesogenin1-null mutant mouse embryos, there are no identifiable somites nor any segmental patterning in the trunk, posterior to the forelimbs (Yoon and Wold,2000), and knockdown of pMesogenin1 resulted in somite formation defects in X. laevis (Wang et al.,2007). These results suggest that pMesogenin1 plays an essential role in somitogenesis.

It was initially reported that changes in pMesogenin1 expression were not detected in tbx6-null mutant mouse embryos, and that tbx6 expression was either unaffected or little enhanced in pMesogenin1-null mutant mouse embryos, suggesting that tbx6 and pMesogenin1 operate in parallel (Yoon and Wold,2000). However, it was recently reported that the expression of pMesogenin1 is controlled by Wnt signaling and tbx6 synergism (Wittler et al.,2007). Thus, the relationship between tbx6 and pMesogenin1 remains unclear.

In this study, we demonstrate that Xtbx6 regulates pMesogenin1 expression in X. laevis. In addition, a novel bHLH gene, pMesogenin2, whose expression is also regulated by Xtbx6, was isolated. Furthermore, knockdowns of Xtbx6, pMesogenin1, or pMesogenin2 resulted in somite formation defects and ventral body wall muscle hypoplasia. These results suggest that Xtbx6 plays a role in somitogenesis and myogenesis by directly regulating pMesogenin1 and 2 expressions in X. laevis.

RESULTS

Comparison Between Xtbx6 and pMesogenin1 Expression Patterns

The expression of Xtbx6 starts in the early gastrula stage in paraxial mesoderm, continues in the posterior paraxial and lateral plate mesoderms, except the axial mesoderm, and is finally restricted to the tailbud tip (Supp. Fig. 1A–D, which is available online; Uchiyama et al.,2001; Li et al.,2006). This expression pattern is similar to that of pMesogenin1, which is expressed in posterior ventrolateral mesoderm (Supp. Fig. 1E–H; Joseph and Cassetta,1999; Yoon et al.,2000). To compare the areas of expression of Xtbx6 and pMesogenin1, we performed double in situ hybridization. The two expressions overlapped at stage 17, but the anterior border of the Xtbx6 expression spread further anteriorly than the pMesogenin1 expression (Supp. Fig. 1M). The similarity of expression patterns implies a relationship between Xtbx6 and pMesogenin1.

pMesogenin2/Mespo2 Is a Novel bHLH Gene

Two phage clones were obtained when a cDNA fragment containing complete coding sequence (CDS) of pMesogenin1 was used to screen the X. laevis genomic library. One clone had an ORF encoding a protein sharing 91% amino acid identity with pMesogenin1 (Supp. Fig. 1N), and the other had a 3′ portion of the same gene. We named this gene pMesogenin2/Mespo2. Both pMesogenin1 and 2 were detected by PCR from cDNA obtained from embryos of the same fertilization batch. Since X. laevis is a pseudotetraploid species (Bisbee et al.,1977), pMesogenin1 and 2 are probably pseudoalleles. Full-length pMesogenin2 containing putative untranslated regions (UTRs) had 87% nucleotide identity with pMesogenin1, while in the ORF, the identity was 93%.

The spatial expression pattern of pMesogenin2 was similar to that of pMesogenin1, and its area of expression overlapped with that of Xtbx6 (Supp. Fig. 1I–M). Weak pMesogenin2 expression was detected in the head region by whole-mount in situ hybridization. RT-PCR from cDNAs obtained by dissection of stage 33 embryos into four fragments (head-1, head-2, trunk, and tail; Yabe et al.,2006) revealed that pMesogenin2 was expressed only in the tail region (data not shown) suggesting that the weak signal observed in the head region was non-specific.

The temporal expression pattern of pMesogenin2 was similar to that of pMesogenin1. RT-PCR analysis, using specific primers for each gene, revealed that the expression reached a peak at the neurula stage and decreased until stage 45 (Fig. 1A). RT-PCR analysis revealed that Xtbx6 expression started first, followed by pMesogenin1, and finally pMesogenin2 (stages 9, 10, and 10.25, respectively; Fig. 1B).

Figure 1.

Comparison of temporal expression patterns and biological activities of pMesogenin1 and 2. A, B: RT-PCR analysis of pMesogenin1, pMesogenin2, and Xtbx6 in embryos at indicated stages. CE: Overexpressed pMesogenin1 (200 pg) or pMesogenin2 (200 pg) mRNA induced ventrolateral phenotypes in the animal cap explants cultured until stage 33. F, G: RT-PCR analysis of mesodermal marker gene expressions in the animal cap explants injected with pMesogenin1, pMesogenin2, pMesogenin1-VP16, pMesogenin1-EnR, pMesogenin2-VP16, or pMesogenin2-EnR mRNA. These explants were cultured until stage 12.5. Injected mRNAs dosages were pMesogenin1 or pMesogenin2 (500 pg each) and pMesogenin1-VP16, pMesogenin1-EnR, pMesogenin2-VP16, or pMesogenin2-EnR (100 pg each). Ornithine decarboxylase (ODC) was used as the loading control (A, B, F, G).

The biological activities of pMesogenin2 were examined in animal cap explants. Explants overexpressing pMesogenin2 developed a cell-containing cavity, which is similar to the phenotype for ventrolateral mesoderm induction (Fig. 1E). RT-PCR analysis revealed that expressions of ventrolateral mesoderm or panmesodermal markers, such as Xwnt8, XmyoD, Xbra, and Xtbx6, were detected, but that of chordin, a marker of dorsal mesoderm, was not detected in pMesogenin2-injected caps (see Fig. 4F). These results are similar to those reported for pMesogenin1-injected caps (Fig. 1D,F; Yoon et al.,2000).

Figure 4.

The pMesogenin2 5′ regulatory region contains three potential T-box binding sites, and responds to Xtbx6. A: Schematic diagrams of the reporter constructs used in luciferase assays. Bent arrow shows EST-predicted transcriptional start site and black spot displays the putative TATA box. Orange boxes show putative T-box binding sites, and × indicates mutations described in B. Blue boxes display putative LEF/TCF binding sites (see Discussion section). B: Sequences of T1, T2, and T3 conserved in pMesogenin1 5′ regulatory regions of X. laevis and X. tropicalis. Red letters indicate nucleotides that differ from T consensus and lowercase letters indicate the mutations introduced in the putative T-box binding sites. C: Band shift assay. Arrow indicates the position of high-mobility complexes. The first lane for each probe demonstrates its binding to the Xtbx6 protein. In the second lane, unlabeled competitors inhibited the binding. In the third lane, unlabeled competitors containing the mutated binding sites failed to inhibit the binding. D: The pMesogenin2 5′ regulatory region responds to Xtbx6 mRNA (20 pg), but not to β-gal mRNA (20 pg). The luciferase reporter gene induction is dependent on T-box binding sites. E: The pMesogenin2 5′ regulatory region responds more to Xtbx6 than to Xbra. mRNA doses were Xtbx6 mRNA (10, 20, or 40 pg) and Xbra mRNA (20 or 100 pg).

To determine whether pMesogenin1 and/or 2induce these genes as transcriptional activators or repressors, we separately fused pMesogenin1 and 2 to the VP16 activation domain (Friedman et al.,1988) and to the repressor domain of Drosophila engrailed (EnR; Han and Manley,1993). RT-PCR analysis revealed that overexpression of pMesogenin1-VP16 or pMesogenin2-VP16 in animal cap explants induced mesodermal genes, whereas overexpression of pMesogenin1-EnR did not. pMesogenin2-EnR did not induce Xwnt8 and XmyoD, but weakly induced Xbra and Xtbx6 (Fig. 1G; see Discussion section). The observations indicate that both pMesogenin1 and 2 induce mesodermal markers as transcriptional activators.

pMesogenin1 and 2 Are Direct Targets of Xtbx6

To examine whether Xtbx6 acts upstream of pMesogenin1 and 2 in X. laevis, a hormone-inducible version of Xtbx6 (Xtbx6-GR), in which the ligand-binding domain of the human glucocorticoid receptor fused to the Xtbx6 C-terminus, was constructed. This approach has been used to identify T-box target genes (Tada et al.,1998). Synthetic mRNA encoding Xtbx6-GR was injected into the animal pole region at the 2-cell stage and animal cap explants were excised at the blastula stage. Explants were treated with dexamethasone (DEX) for 3 hr with or without the protein synthesis inhibitor emetine (eme). Prior to this, we intended to examine to what degree protein synthesis is inhibited by eme treatment. pGL3-Control vector (driven by SV40 promoter and enhancer) was injected into both blastomeres in the animal pole region at the 2-cell stage and animal cap explants were treated with eme. Translation was evaluated at 0–4 hr during treatment by measuring the luciferase activity. At 3 and 4 hr, 100 μg/ml of eme inhibited the increase of activity to approximately 15 and 4%, respectively (Fig. 2A). Therefore, we set the concentration of eme to 100 μg/ml.

Figure 2.

pMesogenin1 and 2 are direct targets of Xtbx6. A: Luciferase assay in animal caps allowed estimation of the efficiency of protein synthesis inhibition by eme. A broken line indicates a basal activity obtained by averaging values at 0 hour. B: Xtbx6-GR-injected animal cap explants were incubated in DEX, eme, or both. After incubation, they were analyzed for expressions of pMesogenin1, 2, Xvent1 and 2 by RT-PCR. ODC was used as the loading control.

pMesogenin1 and 2 were activated by DEX when eme was present, whereas Xvent1 and 2 were activated by DEX only in the absence of eme (Fig. 2B), suggesting that pMesogenin1 and 2 are direct targets of Xtbx6.

pMesogenin2 5′ Regulatory Region and Expression of an EGFP Reporter Gene

A genomic fragment of pMesogenin2, which included 1,501 nucleotides 5′ from the transcriptional start site, was obtained. To determine if this upstream region can produce correct reporter gene expression, we placed this region together with the whole 5′ UTR upstream of EGFP (pMsgn2-(-1501)-EGFP), and produced transgenic Xenopus embryos using I-Sce I meganuclease (Fig. 3A). Similar to the endogenous gene (Fig. 3C, F), the reporter gene was expressed in posterior PSM, as well as in anterior somites and the head region (Fig. 3D, G, H). This may reflect the stability and perdurance of EGFP mRNA, especially in somites. Alternatively, our construct might have lacked elements responsible for expression suppression in somites and the head region. To examine the first possibility, we constructed pMsgn2-(-1501)-EGFP-p3U in which the pMesogenin2 3′ UTR was fused to 3′ downstream of EGFP (Fig. 3A). In transgenic embryos with this construct, the expression of EGFP could not be detected by whole-mount in situ hybridization (Table 1). To investigate the second possibility, because the 4.3-kbp pMesogenin1 5′ regulatory region was shown to recapitulate endogenous pMesogenin1 expression (Wang and Ding,2006), we constructed pMsgn2-(-1501+)-EGFP in which an X. laevis pMesogenin1 5′ regulatory region sharing a high degree of sequence homology with the X. tropicalis pMesogenin1 gene, was added (Fig. 3A, B). In transgenic embryos with this construct, EGFP expression in the head region was reduced, but expression in somites persisted (Fig. 3E; Table 1). This result suggests that the pMesogenin1 5′ regulatory region contains an element that can suppress pMesogenin2 expression in the head region.

Figure 3.

The pMesogenin2 5′ regulatory region drives expression of the EGFP reporter gene in PSM. A: Schematic diagrams of reporter constructs used in transgenesis. In pMsgn2-(-1501)-EGFP, the pMesognein2 5′ regulatory region-EGFP cassette is flanked at both ends by two I-Sce I recognition sites. In pMsgn2-(-1501)-EGFP-p3U, the pMesogenin2 3′ UTR is inserted between EGFP and the poly(A) signal (pA). In pMsgn2-(-1501+)-EGFP, the pMesogenin1 5′ regulatory region is inserted between pMesogenin2 5′ regulatory region and the I-Sce I recognition site. B: X. laevis pMesogenin1 and X. tropicalis pMesgenin1 5′ regulatory regions compared using mVISTA (http://genome.lbl.gov/vista/mvista/submit.shtml) with the X. laevis sequence as the baseline. The downward bracket indicates the region added in pMsgn2-(-1501+)-EGFP. C–E: Whole-mount in situ hybridization for pMesogenin2 or EGFP. C: The endogenous pMesogenin2 expression at stage 27. D: In transgenic embryos with pMsgn2-(-1501)-EGFP, EGFP expression was observed in PSM, somites, and the head region. E: In transgenic embryos with pMsgn2-(1501+)-EGFP, EGFP expression was observed in PSM and somites, but was reduced in the head. FH: Transverse sections at the positions indicated in C and D.

Table 1. Expression Patterns of EGFP in Transgenic Embryos Using a Series of Reporter Constructsa
Fertilization batch numberConstructThe amount of DNA injected per embryo (pg)Total embryosEmbryos with expression only in PSM and anterior somilesEmbryos with expression only in head regionEmbryos with expression in PSM anterior somites, and head region
++++++++++++++++++
  • a

    +, ++, +++, relative level of EGFP expression assessed from in situ hybridization signals.

1pMsgn2-(−1501)-EGFP10019910003129128
 pMsgn2-(−1501)-EGFP-p3U100168000000000
2pMsgn2-(−1501)-EGFP200182100030431
 pMsgn2-(−1501)-EGFP-p3U200127000000000
3pMsgn2-(−1501)-EGFP10019800008035108
 pMsgn2-(−1501)-Tmut123-EGFP1002920000702330
4pMsgn2-(−1501+)-EGFP200271901201600

pMesogenin2 5′ Regulatory Region Contains Potential T-Box Binding Sites Responsive to Xtbx6

We found three sites resembling a T-box half site (Kispert and Herrmann,1993; Conlon et al.,2001; White and Chapman,2005) positioned at 113 (T1), 141 (T2), and 201 (T3) nucleotides upstream from the pMesogenin2 transcriptional start site (Fig. 4A, B). In a search of both GenBank (Bethesda, MD) and JGI (Walnut Creek, CA) databases for 5′ regulatory regions of pMesogenin1 in X. laevis and X. tropicalis, these potential T-box binding sites were found to be conserved in all genes (Fig. 4B).

Electrophoretic mobility shift assays (EMSAs) were done to test the possibility that Xtbx6 directly binds to these T-like half sites. Xtbx6 interacted strongly with T1, and weakly with T2, but barely with T3 (Fig. 4C). These observations are consistent with the fact that T1 is a better match to the consensus than is T2 or T3 (Fig. 4B). Results described below, however, suggest that T2 and T3 may also possibly play a role in the regulation of pMesogenin2.

To determine whether the pMesogenin2 5′ upstream region responds to Xtbx6, a luciferase reporter gene was placed downstream (pMsgn2-(-1501)-Luc; Fig. 4A) and a luciferase assay was performed. The construct was co-injected into the animal pole region at the 2-cell stage together with Xtbx6 mRNA or β-galactosidase mRNA as the control. Injected embryos were cultured to the late gastrula stage and assayed for luciferase activity. Xtbx6 embryos showed a 12- to 13-fold increase in luciferase activity over the controls (Fig. 4D). To investigate the roles of the potential T-box binding sites in the reporter construct, the sequences were mutated to disrupt putative T-box binding (Fig. 4A, B). When co-injected with Xtbx6 mRNA, the luciferase activity of the construct with T1 mutation (pMsgn2-(-1501)-Tmut1-Luc) was reduced to less than half of that of pMsgn2-(-1501)-Luc. The luciferase activity of a construct with all mutations (pMsgn2-(-1501)-Tmut123-Luc) was further reduced (Fig. 4D). These results suggest that T1 and other potential T-box binding site(s) are necessary for the activation of pMesogenin2 reporter constructs by Xtbx6.

We also examined whether Xbra, a T-box gene expressed in axial mesoderm, can activate pMsgn2-(-1501)-Luc. Xtbx6 activated luciferase activity in a dose-response manner, whereas the activation by Xbra was much less, even with a high dose of mRNA (Fig. 4E).

Xtbx6 is an Upstream Regulator of pMesogenin1 and 2 In Vivo

For Xtbx6 knockdown, an antisense morpholino oligonucleotide (MO) targeting the exon1-intron1 junction of Xtbx6 gene (Xtbx6 exint-MO) was designed based on Xtbx6 genomic sequence obtained by screening of a genomic library (Fig. 5A). As a control, a MO with 5 bases-mismatch was also constructed (Xtbx6 5mis-MO). To examine the performance of Xtbx6 exint-MO, the MO was injected into the marginal zone at the 4-cell stage and the injected embryos were cultured until stage 20, 32, or 41. In RT-PCR analysis, two splicing variant mRNAs were detected in exint-MO-injected embryos (Fig. 5B). The length of the short one was indistinguishable from that of normal mRNA, whereas the long variant had a similar size with a fragment obtained from genomic PCR, and it indeed contained whole intron1 as revealed by sequencing analysis, and had several in-frame stop codons (data not shown). Thus, we concluded that Xtbx6 exint-MO is useful for knockdown of Xtbx6.

Figure 5.

Knockdown of Xtbx6, or pMesogenin1 and 2. A: Schematic diagram of partial genomic structure of the Xtbx6 gene. Bent line displays intron1. B: RT-PCR analysis of Xtbx6 exint-MO- or 5mis-MO-injected embryos (50 ng each). Arrow indicates the position of the long splicing variant mRNAs. Genome, Xtbx6 genomic DNA. ODC was used as the loading control. C: Xtbx6 exint-MO (25 ng) and β-gal mRNA (200 pg) were co-injected into the VMZ and DLMZ at the 4-cell stage. DG: Expression patterns of pMesogenin1 (D, E) and pMesogenin2 (F, G) in Xtbx6 exint-MO-injected embryos at stage 20, revealed by whole-mount in situ hybridization. D, F: Representative embryo injected with Xtbx6 exint-MO. E, G: Representative embryo rescued by co-injecting Xtbx6 mRNA (2 pg) with Xtbx6 exint-MO. H: pMsgn MO designed to bind to the translational start regions of both pMesogenin1 and 2. I: pMsgn MO suppressed the luciferase activity of pMsgn2-Luc. J: pMsgn MO (50 ng) or pMesogenin1-EnR (100 pg) and/or pMesogenin2-EnR (100 pg) was injected with β-gal mRNA (200 pg) into the VMZ at the 4-cell stage. K, L: The expression pattern of Xtbx6 in pMsgn MO- or dominat-negative pMesogenins-injected embryos at stage 18, revealed by whole-mount in situ hybridization.

Xtbx6 exint-MO and β-gal mRNA were co-injected into both the ventral marginal zone (VMZ) and dorsolateral marginal zone (DLMZ) at the 4-cell stage. The uninjected side served as the control (Fig. 5C). Expressions of pMesogenin1 and 2 in embryos injected with Xtbx6 exint-MO were examined at stage 20. Compared with the uninjected side, expressions of pMesogenin1 and 2 in the MO-injected side were reduced in 81% (n = 31) and 80% (n = 30) of embryos, respectively (Fig. 5D, F). Reduction of pMesogenin1 and 2 expressions was rescued by co-injecting Xtbx6 mRNA together with MO (70% of embryos, n = 33 and 75% of embryos, n = 32, respectively; Fig. 5E, G). These results suggest that Xtbx6 acts upstream of pMesogenin1 and 2 in vivo.

pMesogenin1 and 2 Do Not Appear to Be Upstream Regulators of Xtbx6

To examine whether pMesogenin1 and 2 act upstream of Xtbx6, we designed an MO that binds to the translational start regions of both pMesogenin1 and 2 (pMsgn MO; Fig. 5H). To test the effectiveness of pMsgn MO in blocking the translation of pMesogenin1 or 2 mRNA, the whole pMesogenin2 5′ UTR along with 267 bp N-terminal CDS was fused in frame to firefly luciferase (pMsgn2-Luc). pMsgn2-Luc mRNA (100 pg) was co-injected with pMsgn MO (50 or 100 ng) or Xtbx6 exint-MO as the negative control (50 ng) into both blastomeres in the animal pole region at the 2-cell stage. Translation was evaluated at the late gastrula stage by measuring the luciferase activity. Xtbx6 exint-MO had no effect on the luciferase activity of pMsgn2-Luc, whereas pMsgn MO reduced the activity to approximately 30% (Fig. 5I).

pMsgn MO and β-gal mRNA were co-injected into the VMZ at the 4-cell stage, with the uninjected side serving as the control (Fig. 5J). This revealed that Xtbx6 expression was unaffected in pMsgn MO-injected embryos at stage 18 (100% of embryos, n = 32; Fig. 5K). Similarly, in embryos injected with pMesogenin1-EnR, pMesogenin2-EnR, or both, Xtbx6 expression was not obviously changed (95% of embryos, n = 19; 90% of embryos, n = 20; and 84% of embryos, n = 19, respectively; Fig. 5L). These results might suggest that pMesogenin1 and 2 are not upstream regulators of Xtbx6.

Involvement of Xtbx6 and pMesogenin1 and 2 in Somitogenesis and Ventral Body Wall Muscle Formation

For further analysis of phenotypes resulting from Xtbx6 knockdown, Xtbx6 exint-MO-injected embryos were cultured to stage 41 and examined immunohistochemically using the skeletal muscle-specific 12/101 antibody. As a control, only slight reductions of ventral body wall muscles were observed in Xtbx6 5mis-MO-injected embryos (97% of embryos, n = 39; Fig. 6A, B), whereas in Xtbx6 exint-MO-injected embryos, the stained somite was narrower dorso-ventrally, and the boundaries between somites were no longer apparent in the MO-injected side. Furthermore, severe reduction of ventral body wall muscles was observed in the MO-injected side (100% of embryos, n = 32; Fig. 6C, D). We examined XmyoD expression in the somites, and Xmyf-5 expression in the expanding hypaxial myotome (Martin and Harland,2001; Martin et al.,2007). In the MO-injected side, the segmental pattern of XmyoD expression was fused (97% of embryos, n = 31), and Xmyf-5 expression in hypaxial muscle cells was severely reduced (97% of embryos, n = 30) (Fig. 6E–H).

Figure 6.

Knockdowns of Xtbx6 or pMesogenin1 and 2 cause somite malformation and ventral body wall muscle hypoplasia. Representative embryos injected with Xtbx6 5mis-MO (25 ng) (A, B), Xtbx6 exint-MO (25 ng) (CH), pMsgn MO (50 ng) (I, J), or pMesogenin1-EnR and pMesogenin2-EnR (100 pg each) (K, L). In A–L, the uninjected sides are on the left (A, C, E, G, I, K) and injected sides on the right (B, D, F, H, J, L). In some cases, images are flipped horizontally to facilitate comparison. A–D, I–L: Immunohistochemistry with the 12/101 antibody at stage 41. E–H: Analysis of expression patterns of XmyoD (E, F) and Xmyf-5 (G, H) at stage 32 by whole-mount in situ hybridization. Arrows indicate the expression of Xmyf-5 in hypaxial muscle cells. M, N: The expression of NCAM was not detectably changed in Xtbx6 exint-MO (50 ng)-injected embryos at stage 30, revealed by 4d antibody staining.

In tbx6-null mutant mouse embryos, ectopic neural tube–like structures are formed (Chapman and Papaioannou,1998). Therefore, we examined the expression of NCAM, a marker of neural tissue, in Xtbx6 exint-MO-injected embryos immunohistochemically using anti-NCAM antibody 4d. Xtbx6 exint-MO was injected into the VMZ of two blastomeres at the 4-cell stage and the expression of NCAM at stage 30 was viewed laterally (Fig. 6M, N), dorsally, or in sections (data not shown). Xtbx6 exint-MO-injection caused no detectable change in NCAM expression.

Similarly, in 88% (n = 33) of embryos injected with pMsgn MO, the boundaries between somites were unclear just as described by Wang et al. (2007) and simultaneously, ventral body wall muscle hypoplasia was observed (Fig. 6I, J). In addition, this hypoplasia was observed also in embryos injected with pMesogenin1-EnR, pMesogenin2-EnR, or both (67% of embryos, n = 18; 61% of embryos, n = 18; and 60% of embryos, n = 20, respectively; Fig. 6K, L).

A previous study reported that animal cap explants overexpressing Xtbx6 and noggin differentiate into muscle (Uchiyama et al.,2001). We investigated whether pMesogenin1 and 2 can mimic this phenomenon, and observed that animal cap explants overexpressing pMesogenin1 or 2 and noggin did not differentiate into muscle (data not shown). As described in a previous section, pMesogenins induced XmyoD expression at the gastrula stage (Fig. 1F). Therefore, we examined by RT-PCR if pMesogenins could induce the expressions of muscle marker genes at later stages. The expressions of skeletal muscular markers, such as cardiac actin (Cact) and skeletal β-tropomyosin (TMsk), were detected at stage 34 in animal cap explants injected with pMesogenin1 or 2 (Fig. 7A). We also examined whether pMesogenin1 and 2 are involved in myogenesis mediated by Xtbx6 in animal cap explants. Of the animal cap explants overexpressing Xtbx6 and noggin, 84% differentiated into muscle (n = 64; Fig. 7C), whereas the myogenic rate in explants co-injected with pMsgn MO was 48% (n = 70; Fig. 7D). The control MO (Co MO)-injected explants showed a slight inhibition in elongation, but had high myogenic rate (93%, n = 29; Fig. 7E).

Figure 7.

pMesogenin1 and 2 is involved in myogenesis initiated by Xtbx6. A: RT-PCR analysis of muscle marker gene expressions in the animal cap explants injected with pMesogenin1 (200 pg) or pMesogenin2 (200 pg) mRNA. BE: Animal cap explants were immunostained with the 12/101 antibody at stage 24. B: Uninjected explants. C–E: Doses of injected mRNAs or MOs were Xtbx6 mRNA (50 pg), noggin mRNA (100 pg), pMsgn MO (100 ng), and Co MO (100 ng). Note that muscle differentiation was inhibited by pMsgn MO. F: Regulatory hierarchy in somitogenesis and myogenesis involving Xtbx6, Wnt/β-catenin signaling, pMesogenin1 and 2. Solid arrows indicate known controls and the dashed arrow, the deduced control; for details see Discussion section.

DISCUSSION

Transcription factors govern the specification and determination of cell fate during embryogenesis and, hence, studying the transcription factor cascade can help understand the developmental mechanisms. In this study, we focused on two transcription factors, T-box and bHLH. Our results indicate a regulatory cascade governed by Xtbx6 in Xenopus somitogenesis and myogenesis.

pMesogenin1 and 2

In this study, we isolated a novel bHLH gene, pMesogenin2. In a search of databases, only one pMesogenin gene was found in the mouse and X. tropicalis genomes (MGI, http://www.informatics.jax.org/; JGI, http://genome.jgi-psf.org/Xentr4/Xentr4.home.html). Murine pMesogenin1 has 42% amino acid identity with both X. laevis pMesogenin1 and 2 over the entire ORF, and 81 and 79% identity, respectively, in the bHLH region. X. tropicalis pMesogenin1 has 90 and 88% identity with X. laevis pMesogenin1 and 2, respectively, over the entire ORF, and 100 and 95% identity, respectively, in the bHLH region. The difference in the amount of expression between pMesogenin1 and 2 was not examined; however, in PCR cloning using primers common to both genes, cloning frequencies were similar (data not shown). Therefore, we suggest that the amounts of both genes' expression are almost equal.

Although pMesogenin1 and 2 were similar in expression patterns and biological activities, some differences were observed. In animal cap explants overexpressing each mRNA, pMesogenin2 induced Xbra and Xtbx6 expressions more strongly than pMesogenin1 (Fig. 1F). In addition, pMesogenin2-EnR induced weak Xbra and Xtbx6 expressions, whereas pMesogenin1-EnR failed to induce such expressions (Fig. 1G). Since EnR domains were fused to full-length CDSs of pMesogenins, this may be caused by the remnants of activational activity by pMesogenin2 CDS. These results suggest that pMesogenin2 may be more biologically active than pMesogenin1.

pMesogenin1 and 2 as Target Genes of Xtbx6

Previous reports suggest that tbx6 and pMesogenin1 operate in parallel (Yoon and Wold,2000). In X. laevis, it has been reported that Tbx6VP16-GR does not induce pMesogenin1 expression following incubation in DEX, although a weak expression has been observed in Tbx6VP16-GR-injected explants treated with both DEX and a protein synthesis inhibitor, cycloheximide (Li et al.,2006). In the present study, Xtbx6-GR induced pMesogenin1 expression following incubation in DEX with or without eme. Furthermore, knockdown of Xtbx6 reduced pMesogenin1 and 2 expression. Our results show that Xtbx6 is a direct upstream regulator of both pMesogenin1 and 2 (Fig. 7F), which is consistent with the mouse results of Wittler et al. (2007). Animal cap explants overexpressing Xtbx6 have been reported to develop mesothelium (Uchiyama et al.,2001); however, such development in pMesogenin1 or 2 did not occur in this study. In addition, explants overexpressing Xtbx6 and noggin differentiate into muscle (Uchiyama et al.,2001), whereas pMesogenin1 or 2 and noggin did not. These results suggest that pMesogenin1 and 2 are not covering all functions of Xtbx6.

We detected potential T-box binding sites in the pMesogenin2 5′ regulatory region. Activation of this region by Xbra was weaker than that by Xtbx6. This might indicate that Xbra and Xtbx6 utilize different binding site sequences or chromatin structures in the regulatory region. Although no T-box binding site in the pMesogenin2 5′ regulatory region completely met T consensus, the three potential binding sites were necessary for the activation of pMesogenin2 reporter constructs by Xtbx6. This suggests that Xtbx6 can sufficiently bind to sequences that differ from T consensus. Similarly, target genes of VegT have been shown to possess T-box binding sites that fit a TNNCAC(C/T)(T/C) sequence (Taverner et al.,2005), which differ from the core sequence of TCACACCT selected by VegT in PCR-based binding selection (Conlon et al.,2001). We generated a transgenic construct with mutations of the three potential T-box binding sites (pMsgn2-(-1501)-Tmut123-EGFP). In transgenic embryos with this construct, EGFP expression was slightly reduced, but not absent (Table 1). pMesogenin1 expression has been activated by Wnt signaling via the LEF/TCF binding site in X. laevis (Wang et al.,2007) and mouse (Wittler et al.,2007). We observed that the LEF/TCF binding site is conserved in the pMesogenin2 5′ regulatory region (Fig. 4A), suggesting that pMesogenin2 expression is regulated by both Xtbx6 and Wnt signaling (Fig. 7F).

Xtbx6 and pMesogenin1 and 2 Involvement in Somite Formation and Hypaxial Muscle Cell Differentiation

In X. laevis, knockdown of Xtbx6 was reported to disrupt posterior development (Lou et al.,2006). In this study, a knockdown of Xtbx6, using Xtbx6 exint-MO, caused somite formation defects and ventral body wall muscle hypoplasia. During muscle formation, hypaxial muscle cells originating from the ventral part of each somite migrate to form ventral body wall muscle. It has been shown that XHas2, a catalytic enzyme governing vertebrate hyaluronan (HA) biosynthesis, and XCD44, a HA signaling receptor, are required for hypaxial muscle cell migration (Ori et al.,2006). In Xtbx6 exint-MO-injected embryos, Xmyf-5 expression in hypaxial muscle cells was reduced. Since Xmyf-5 expression is indicative of differentiating myoblasts (Martin et al.,2007), Xtbx6 may be required for hypaxial muscle cell differentiation rather than migration.

The reduction of Xmyf-5 expression in Xtbx6 exint-MO-injected embryos is consistent with previous reports that Xmyf-5 is a target of Xtbx6 (Lin et al.,2003; Li et al.,2006). We observed the reduction of Xmyf-5 expression in anterior trunk somites and in the tail region, but not in posterior trunk somites (Fig. 6H). Cell transplantation studies have demonstrated that zebrafish Danio rerio bodies may be divided into three domains: anterior trunk (somites 1–8); posterior trunk (9–15); and tail (16 and beyond), and that each domain requires distinct signals (Szeto and Kimelman,2006). Similar mechanisms might exist in X. laevis. Xmyf-5 expression was also reported to be regulated by p38 MAP kinase (Keren et al.,2005). Knockdown of p38 MAP kinase caused ventral body wall muscle hypoplasia. The relationship between this signaling and Xtbx6 awaits further analysis.

Knockdowns of pMesogenin1 and 22 also resulted in somite formation defects and ventral body wall muscle hypoplasia. Furthermore, animal cap assays showed that pMesogenin1 and 2 are involved in myogenesis mediated by Xtbx6. These results suggest that Xtbx6 plays a role in somitogenesis and myogenesis by regulating pMesogenin1 and 2 expressions (Fig. 7F). Interestingly, genes expressed in the tailbud tip such as Xtbx6 and both pMesogenin1 and 2, govern not only somitogenesis, but also the development of anterior structures such as ventral body wall muscle.

EXPERIMENTAL PROCEDURES

Genomic Library Screening

To generate templates for the X. laevis muscle genomic library screening (ZL1000j, Clontech, Mountain View, CA), PCR products were amplified from pMesogenin1/pBluescript II with a forward primer (5′-ATGGAGACTCTGCACCATCCC-3′) and a reverse primer (5′-AACCTCACGCTCTCTTGGAGC-3′), or Xtbx6/pBluescript I with a forward primer (5′-GAATTCAGCCATGTACCACTCTGAGCTCTTCC-3′) and a reverse primer (5′-TGTAGCGAGAGTTGTCCACGG3′), and ligated into the T-vector. The Hind III-Sma I (526 bp, pMesogenin1) or EcoR I-Xho I (504 bp, Xtbx6) fragment from these plasmids was labeled with 32P, and subsequent screenings were carried out as described by Yabe et al. (2006). GenBank Accession Numbers for Xtbx6 and pMesogenin2genomic D NA sequences are AB462326 and AB462327, respectively.

Plasmid Constructs

To produce Xtbx6-GR, the human glucocorticoid receptor ligand-binding domain, along with the HA tag (hGR-HA) was PCR-amplified from Xsox2-GR-HA/pSP64T (provided by Y. Sasai, RIKEN CDB, Kobe, Japan) with a forward primer (5′-CACATGGCACTCGAGTCTGAAAATCC-3′) and a reverse primer (5′-TCTAGATCCTCGACTAGTCACGCG-3′). At the same time, PCR products were amplified from Xtbx6/pBluescript II (Uchiyama et al.,2001) with a forward primer (5′-GAATTCAGCCATGTACCACTCTGAGCTCTTCC-3′) and a reverse primer (5′-CTGCCAATCAATACTCGAGCATCCAGCC-3′), and inserted into the pCS2+ vector. Finally, a Xho I-Xba I fragment from hGR-HA was inserted into this preparation, fusing to the C-terminus of Xtbx6. To generate Xtbx6/pCS2+, PCR products were amplified from Xtbx6/pBluescript II with a forward primer (5′-AGCCATGTACCACTCTGAG-3′) and a reverse primer (5′-AGTCTCACATCCAGCCCC-3′), and subcloned into the Cla I and EcoR I sites of the pCS2+ vector. To create pMesogenin1/pCS2+ and pMesogenin2/pCS2+, PCR products were amplified from stage-15 whole embryonic cDNA with a common forward primer (5′-ATGGAGACTCTGCACCATCCC-3′) and a common reverse primer (5′-AACCTCACGCT-CTCTTGGAGC-3′), and subcloned into the Xho I and Xba I (pMesogenin1) or the BamH I and Cla I (pMesogenin2) sites of the pCS2+ vector. To clone the 1,052-bp full-length pMesogenin2 gene containing putative UTRs, PCR products were amplified from pMesogenin2 genomic DNA with a forward primer (5′-CAGCCATCAACTCAGAGGC-3′) and a reverse primer (5′-ATCTGGCACAGGATAAATTAAGC-3′), and ligated into T-vector (pMesogenin2 FL/pBluescript II). To produce pMesogenin1-VP16, pMesogenin2-VP16, pMesogenin1-EnR, and pMesogenin2-EnR, PCR products were amplified from pMesogenin1/pCS2+ or pMesogenin2/pCS2+ with a common forward primer, either (5′-CTCGAGATGGAGACTCTGCACCATCC-3′ or 5′-ATCGATGGAGACTCTGCACCATC-C-3′), and a common reverse primer (5′-AGATCTTCGCTCTCTTGGAGCACTGG-3′), and ligated into T-vector. Subsequently, pMesogenin1-VP16/pCS2+, pMesogenin2-VP16/pCS2+, pMesogenin1-EnR/pCS2+, and pMesogenin2-EnR/pCS2+ were constructed according to Uchiyama et al. (2001). To generate pMsgn2-Luc/pCS2+, the region from +1 to +467 at 3′ downstream from the pMesogenin2 transcriptional start site was excised by BamH I and Nco I from pMesogenin2 FL/pBluescript II and ligated to the 5′ end of the firefly luciferase sequence, maintaining the reading frame. To create Xbra/pCS2+, the EcoR I-Stu I fragment from pXT1 (Smith et al.,1991) was inserted into the pCS2+ vector. To obtain β-gal/pBluescript II, the Hind III-Pst I fragment from pENL (provided by Y. Nabeshima, Kyoto University) was inserted into the pBluescript II KS (-) vector.

For transgenesis and luciferase assays, 1,601 bp 5′ upstream from the pMesogenin2 translation start site was PCR-amplified from λDNA containing pMesogenin2 gene with a forward primer (5′-GCGCAACTCGTGAAAGGTAGG-3′) and a reverse primer (5′-ATGGTGCAGAGTAAGC-TTCAGTAAGACTTTCC-3′), and subcloned into the Sac I and Hind III sites of the EGFP I-Sce I RV vector (provided by N. Mizuno, University of Tokyo) and pGL3-Basic vector (Promega, Madison, WI), respectively. To generate pMsgn2-(1501)-EGFP-p3U, the putative pMesogenin2 3′ UTR was PCR-amplified from pMesogenin2 FL/pBluescript II with a forward primer (5′-GGTTGAGAATGATTCTGCAGG-3′) and a reverse primer (5′-ATCTGGCACAGGATAAATTAAGC-3′), and inserted between the EGFP gene and the poly(A) signal of pMsgn2-(-1501)-EGFP. To construct pMsgn2-(-1501+)-EGFP, the region from −3,469 to −2,994 at 5′ of the pMesogenin1 transcriptional start site (Wang and Ding,2006) was PCR-amplified from X. laevis genomic DNA isolated from tadpoles with a forward primer (5′-CAGGATCCCTTTAAGCAGTTGC-3′) and a reverse primer (5′-GGCCAGTCCAACACTGAGC-3′), and inserted into the 5′ end of the pMesogenin2 promoter in pMsgn2-(-1501)-EGFP. Site-directed mutagenesis was performed according to the manufacturer's protocol using the QuickChange XL Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA). Briefly, PCR using sense and antisense oligonucleotides, followed by Dpn I digestion of the parent plasmid, was performed. Oligonucleotide mutants in the three potential T-box binding sites in the pMesogenin2 5′ regulatory region were as follows (mutant nucleotides underlined): Tmut1 (5′-GCCAGTCCATTTTCCATGTTAGCGTCTTGCTGTCTCCTTTG-3′), Tmut2 (5′-TCTCAAAGTAAAGTCAAGTAGCGTTTGGCCAGTCCATTTTCCATG-3′), and Tmut3 (5′-CCCTCCTTCTTTCCTCTCCCAGACCCGTTTACTGAAAGCC-3′).

All PCR products described above were sequenced and ascertained to have no undesirable mutation.

Whole-Mount In Situ Hybridization and Immunohistochemistry

Whole-mount in situ hybridization was performed according to Sive et al. (2000), except that the RNase treatment was omitted. cRNA probes for Xtbx6, pMesogenin1, pMesogenin2, EGFP, XmyoD, and Xmyf-5 were transcribed with T3 or T7 RNA polymerase from Xtbx6/pCS2+, pMesogenin1 FL/pBluescript II, pMesogenin2 FL/pBluescript II, EGFP/pCS2+, XmyoD/pBluescript II, and Xmyf-5/pBluescript II templates linearized with Hind III, BamH I, or EcoR I. In whole-mount double in situ hybridization, cRNA probes for pMesognin1 and 2 were labeled with digoxigenin-UTP (Roche, Basel, Switzerland) and that for Xtbx6 was labeled with fluorescein-UTP (Roche). The subsequent protocol was described by Yabe et al. (2006).

Embryos were fixed in neutral-buffered 4% formaldehyde for 80 min at room temperature. Antibody staining was performed following the standard procedures for indirect immunohistochemistry (Sive et al.,2000) using the skeletal muscle-specific antibody 12/101 or anti-NCAM monoclonal antibody 4d (Developmental Studies Hybridoma Bank, Iowa City, IA) at 1:50–100 dilution, HRP-conjugated anti-mouse IgG+M (American Qualex, San Clemente, CA), and diaminobenzidine for detection.

mRNA Injections, Animal Cap Assay, and Transgenesis

Preparation of X. laevis eggs and in vitro transcription of mRNA were carried out according to Uchiyama et al. (2001). Embryo stages were determined according to Nieuwkoop and Faber (1956). Microinjections were performed in 0.3× Steinberg's solution containing 2% Ficoll. mRNAs and MOs were injected into both blastomeres in the animal pole region at the 2-cell stage. Animal cap explants were excised at stages 8–9 and cultured in 1× Steinberg's solution. Xtbx6-GR mRNA (100 pg) was injected and explants were incubated in 40 μM of DEX for 3 hr at 20°C with or without of emetine (Edwards and Mahadevan,1992). Explants were fixed in Bouin's fluid, embedded in paraffin, sectioned at 4 μm, and stained with Mayer's hematoxylin and eosin. Transgenic X. laevis embryos were generated as described by Ogino et al. (2006). After whole-mount in situ hybridization, embryos were embedded in paraffin and sectioned at 20 μm.

RT-PCR

RT-PCR was performed as described by Yabe et al. (2006), except that PrimeScript Reverse Transcriptase (Takara Bio Inc., Otsu, Japan) was used for reverse transcription. Forward (F) and reverse (R) PCR primers for pMesogenin1, pMesogenin2, Xtbx6, Xtbx6 intron1, Xbra, XmyoD, Xwnt8, chordin, Xvent1, Xvent2, Cact, TMsk, and ODC were as follows: pMesogenin1 (F: 5′-GGACTTGTGCAGGAAACTCCTAG-3′, R: 5′-AACCTCACGCTCTCTTGGAGC-3′); pMesogenin2 (F: 5′-CTTGGGAAAGCAGCTTCTGG-3′, R: 5′-AACCTCACGCTCTCTTGGAGC-3′); Xtbx6 (F: 5′-TCAGTGTGGAACAGGAACAGG-3′, R: CAGCTGGATACAGAATAGATCAGAGG-3′); Xtbx6 intron1 (F: 5′-AGCCATGTACCACTC-TGAGC-3′, R: 5′-GATGGAGTGGAACTGTTTCC-3′); Xbra (F: 5′-GG-AAGTATGTGAATGGAGAATGG-3′, R: GGTCTTTACTTTAGACTGATGGTGG-3′); XmyoD (F: 5′-CTGTTTCTATGGAGCTGTTGCC-3′, R: 5′-GCCTATAAGACGTGATAGATGGTGC-3′); Xwnt8 (F: 5′-ATCTCGAAACTATTCGTCGATGG-3′, R: 5′-TCTGGAATGCCGTCATCTCC-3′); chordin (F: 5′-AGTAGCCAAGGCTAGATTCAACC-3′, R: 5′-CAGAGATGTTGGCATATAGCTCC-3′); Xvent1 (F: 5′-CTTCAGCATGGTTCAACAGG-3′, R: 5′-CAT-ATACTGAGCCCCAAAGAGTG-3′); Xvent2 (F: 5′-GCACTCCAGCAACATCACC-3′, R: 5′-TTCTTCCTAATAGGCCAGAGG-3′); Cact (F: 5′-AATCACAGCACCAGCTCTCC-3′, R: 5′-TAGAAGCACTTCCTGTGCACA-3′); TMsk (F: 5′-TGGAGGAAGAGCTGAAGAACC-3′, R: 5′-ACCCCATGTCTATACTGTCAGACC-3′); and ODC (F: 5′-TGTGAAGACTCTCTCCATTCTTGG-3′, R: 5′-AAGCCTATACATTGATGCTGGC-3′). Multiple PCR assays were performed and the results were reproducible.

Electrophoretic Mobility Shift Assays (EMSAs)

EMSAs were done with synthetic DNA fragments that contain the T1-, T2-, or T3-binding sites from the pMesogenin2 5′ regulatory region. These fragments were prepared by annealing the following complementary oligonucleotides: T1 (5′-TTTTCCATGTTAACACCTTGCTGT-3′ and 5′-GGA-GACAGCAAGGTGTTAACATGG-3′); T2 (5′-TAAAGTCAAGTAACACTTGGCCAG-3′ and 5′-TGGACTGGCCAAGTGTTACTTGAC-3′); and T3 (5′-TTCCTCTCCCAGGTCTGTTTA-CTG-3′ and 5′-CTTTCAGTAAACAGACCTGGGAGA-3′).

The resulting double-stranded DNA fragments were labeled with [γ-32P] ATP (Perkin Elmer, Boston, MA) using T4 polynucleotide kinase (New England Biolabs, Beverly, MA). Six cold competitor DNAs were used in the EMSAs. These included the wild-type T1, T2, and T3 sequences, and mutant form of T1 (T1 mut), T2 (T2 mut), and T3 (T3 mut). The following oligonucleotides were used to prepare these DNA fragment (mutant nucleotides underlined): T1 mut (5′-TTTTCCATGTTAGCGTCTTGCTGT-3′ and 5′-GGAGACAGCAAGACGCTAAC-ATGG-3′); T2 mut (5′-TAAAGTCAAGTAGCGTTTGGCCAG-3′ and 5′-TGGACTGGCCAAACGCTACTT-GAC-3′); T3 mut (5′-TTCCTCTCCCAGACCCGTTTACTG-3′ and 5′-CTTTCAGTAAACGGGTCTGGGAGA-3′).

Xtbx6 protein used in EMSAs were produced by in vitro translation of synthetic mRNA that includes complete CDS, using the Rabbit Reticulocyte Lysate System, Nuclease Treated (Promega). One-microliter out of 50 μl translation reactions was preincubated with 12 pmol of cold competitor DNAs in a solution (total volume of 12 μl) containing 20 mM HEPES-NaOH (pH 7.9), 50 mM KCl, 1 mM EDTA, 1 mM DTT, 10% glycerol, 100 μg/ml BSA, 50 μg/ml poly(dI-dC) · poly(dI-dC) (Sigma, St Louis, MO), 0.5 mM PMSF, and protease inhibitor cocktail (tablets, EDTA-free; Roche) at room temperature for 10 min. Then, 1.3 μl of kinase reaction containing 0.26 pmol of the labeled T1, T2, and T3 DNAs and 190 kBq [γ-32P] ATP was added and incubated at room temperature for 20 min. Subsequently, the reaction mixtures were electrophoresed on a 5% polyacrylamide gel, exposed to an Imaging Plate, and visualized by BAS-2500 Imaging Analyzer (Fujifilm, Tokyo, Japan).

Luciferase Reporter Assays

Luciferase reporter assays were performed using the Dual-Luciferase Reporter Assay System (E1910, Promega) according to the manufacturer's protocol, but at a quarter scale. pGL3-Basic vector (40 pg) containing the 5′ regulatory fragment, 40 pg of pRL-TK vector (Promega), and mRNA encoding β-gal (20 pg), Xtbx6, or Xbra were co-injected into both blastomeres at the 2-cell stage. To test the performance of pMsgn MO, 40 pg of pRL-TK vector, pMsgn2-Luc mRNA, and/or Xtbx6 exint-MO or pMsgn MO were co-injected into both blastomeres at the 2-cell stage. Embryos were cultured until stage 12 and collected as four pools of three embryos. Each pool was suspended in 60 μl of 1× Passive Lysis Buffer (Promega). Following centrifugation, 5 μl was taken for assay, and the photons were counted with a luminometer (Luminescencer-PSN, ATTO, Tokyo, Japan). Each experiment utilized embryos from the same fertilization batch and was replicated at least once. The mean and standard error of the four pools were calculated. In experiments to estimate the efficiency of protein synthesis inhibition by eme, we did not include an internal control, and took care to isolate the pGL3 vector (40 pg)-loaded animal caps as similarily sized as possible, prior to quantifying the firefly luciferase.

Antisense Morpholino Oligonucleotides

Antisense MOs were obtained from Gene Tools LLC (Philomath, OR). The sequences of Xtbx6 exint-MO, pMsgn MO, Co MO (Standard Control Oligo, Classic, Gene Tools), and Xtbx6 5mis-MO are as follows: Xtbx6 exint-MO (5′-TGCCCCAGTCACATACCTGAGTATC-3′), pMsgn MO (see Fig. 5), Co MO (5′-CCTCTTACCTCAGTTACAATTTATA-3′), and Xtbx6 5mis-MO (5′-TGgCCgAGTCAgATACgTGAcTATC-3′; lowercase indicates mismatch) (Supp. Fig. 1).

Acknowledgements

We are grateful to Dr. N. Mizuno and Dr. M. Asashima for their gift of the EGFP I-Sce I RV vector and their kind teaching of the transgenesis method. We are also grateful to Mr. Y. Kawata (in our lab) for generating pMesogenin1 and 2-VP16 and EnR, Dr. Y. Sasai for the Xsox2-GR-HA/pSP64T plasmid, Dr. Y. Nabeshima for the pENL plasmid, and Dr. J. C. Smith (University of Cambridge) for the pXT1 plasmid. H. U. was supported by a 2007 Strategic Research Project (K19007) grant from Yokohama City University, Japan.

Ancillary