MOLECULAR AND CELLULAR BIOLOGY, May 2008, p. 32363244 0270-7306/08/$08.000 doi:10.1128/MCB.01754-07 Copyright 2008, American Society for Microbiology. All Rights Reserved.

Vol. 28, No. 10

Activator-to-Repressor Conversion of T-Box Transcription Factors by the Ripply Family of Groucho/TLE-Associated Mediators
Akinori Kawamura,1,2,3 Sumito Koshida,1,4 and Shinji Takada1,4*
Okazaki Institute for Integrative Biosciences, National Institutes of Natural Sciences, Okazaki, Aichi 444-8787, Japan1; Graduate School of Science and Engineering, Bioscience and Biomedical Engineering, Waseda University, Shinjuku-ku, Tokyo 169-8050, Japan2; Department of Life Science, Graduate School of Science and Engineering, Saitama University, Sakura-ku, Saitama 338-8570, Japan3; and Department of Basic Biology, Graduate University for Advanced Studies (SOKENDAI), Okazaki, Aichi 444-8787, Japan4
Received 25 September 2007/Returned for modication 26 November 2007/Accepted 28 February 2008
The T-box family of transcription factors, dened by a conserved DNA binding domain called the T-box, regulate various aspects of embryogenesis by activating and/or repressing downstream genes. In spite of the biological signicance of the T-box proteins, how they regulate transcription remains to be elucidated. Here we show that the Groucho/TLE-associated protein Ripply converts T-box proteins from activators to repressors. In cultured cells, zebrash Ripply1, an essential component in somite segmentation, and its structural relatives, Ripply2 and -3, suppress the transcriptional activation mediated by the T-box protein Tbx24, which is coexpressed with ripply1 during segmentation. Ripply1 associates with Tbx24 and converts it to a repressor. Ripply1 also antagonizes the transcriptional activation of another T-box protein, No tail (Ntl), the zebrash ortholog of Brachyury. Furthermore, injection of a high dosage of ripply1 mRNA into zebrash eggs causes defective development of the posterior trunk, similar to the phenotype observed in homozygous mutants of ntl. A mutant form of Ripply1 defective in association with Tbx24 also lacks activity in zebrash embryos. These results indicate that the intrinsic transcriptional property of T-box proteins is controlled by Ripply family proteins, which act as specic adaptors that recruit the global corepressor Groucho/TLE to T-box proteins. T-box proteins (Tbx) play many crucial roles in development by activating and/or repressing the transcription of their target genes (23, 32). For instance, reduced function of T/Brachyury, the founder member of this family, causes a truncated tail in the mouse (10). The importance of T-box transcription factors is also shown by the fact that mutations in human T-box genes cause severe congenital disorders, such as DiGeorge, ulnarmammary, and Holt-Oram syndromes (26). Furthermore, the Tbx family is evolutionally conserved from Caenorhabditis elegans to insects and vertebrates, showing remarkable functional conservation across species. However, in spite of the biological signicance of the T-box proteins, how they regulate transcription remains to be elucidated. One of the processes in which T-box proteins play crucial roles is somite segmentation in vertebrates (3, 4, 24). Somitogenesis is the sequential subdivision of segmented precursors of the vertebral column and musculature from the presomitic mesoderm (PSM) (27, 29). Prior to morphological subdivision, a segmental prepattern is established in the anterior PSM. Genetic analyses of the mouse and the zebrash indicate that several Tbx genes are required for distinct processes of this prepattern formation (11, 24, 36), which also involves Notch signaling and Mesp2, a basic helix-loop-helix transcription factor (21, 33). For instance, Tbx6 induces the expression of Mesp2 and is required for proper patterning of somites of the mouse (36). Similarly, tbx24/fused somites, structurally related to mouse Tbx6, is expressed in the anterior PSM of the zebrash embryo and required for the formation of somite boundaries and the expression of mesp-b, a zebrash homolog of Mesp2 (24, 30), suggesting that Tbx24 is involved in the transcriptional regulation of the mesp-b locus in zebrash. Because Mesp2, as well as mesp-b, is expressed in only a portion of Tbx-expressing cells (24, 28, 36), however, another molecule also seems to be involved in Tbx-mediated transcriptional regulation of mesp genes in the anterior PSM. In a previous study, we showed that Ripply1, which is associated with the global transcriptional corepressor Groucho/ TLE, terminates the expression of segmentation genes in the anterior PSM of zebrash embryos (16). In ripply1-decient embryos, the expression of mesp-b is up-regulated in a cellautonomous manner, whereas in embryos injected with ripply1 mRNA, the expression of mesp-b is highly suppressed in the anterior PSM (16). These results suggest that Ripply1 regulates the proper expression of mesp-b in the anterior PSM. Taking into account that the expression of mesp-b could be induced by T-box proteins, we can speculate that Ripply1 may antagonize the function of T-box proteins in the transcription of mesp genes. Therefore, we examined precisely the relationship between Tbx24 and Ripply1 in the transcription of mesp-b. We showed that Ripply1 is coprecipitated with Tbx24 and converts it from an activator to a repressor in cultured cells. The other members of the Ripply family, Ripply2 and Ripply3, can also antagonize the transcriptional activation mediated by Tbx24. On the other hand, Ripply1 also antagonizes the transcriptional activation of another T-box protein, No tail (Ntl),

* Corresponding author. Mailing address: Okazaki Institute for Integrative Biosciences, National Institutes of Natural Sciences, Okazaki, Aichi 444-8787, Japan. Phone: 81-564-59-5241. Fax: 81-564-59-5240. E-mail: stakada@nibb.ac.jp. Supplemental material for this article may be found at http://mcb.asm.org/. Present address: Department of Biological Sciences, Graduate School of Science, University of Tokyo, Bunkyo-ku, Tokyo, Japan. Published ahead of print on 10 March 2008.

VOL. 28, 2008

MEDIATOR IN REPRESSION BY T-BOX PROTEINS
both in vitro and in vivo. These results reveal a molecular mechanism underlying dynamic conversion of the regulatory property of T-box proteins in the transcription of their targets.
MATERIALS AND METHODS Plasmid construction. The mesp-b 2.5-kb luciferase reporter plasmid was constructed by inserting a 2.5-kb genomic fragment upstream of the translational initiation site of zebrash mesp-b fused to the rey luciferase gene in the pGL3 Basic vector (Promega). For pBP-TbxX2 and pBP-mutTbxX2, a T-box site (5AATTCACACCT-3) and a mutT-box sequence (5-AATTCAgACCT-3), respectively, were inserted into the beta interferon basal promoter fused to the luciferase gene. Luciferase assay and TSA treatment. Human embryonic kidney 293T cells were cultured at 37C in Dulbeccos modied Eagles medium (DMEM) supplemented with 10% fetal calf serum and antibiotics. For luciferase assay, 293T cells (1.0 104/well) were seeded into 24-well collagen-coated plates (Iwaki). The cells were transiently transfected with the reporter gene plasmids along with the indicated expression vector, using Fugene 6 reagent (Roche). The total amount of DNA in each transfection mix was kept the same by supplementation with the empty vector pCS2. After incubation for 24 h, the cells were washed with phosphate-buffered saline (Ca2 and Mg2 free) and lysed, and luciferase activity was assayed by using a dual-luciferase reporter assay system (Promega). For normalization of the transfection efciencies, pRL-tk-Luc (Promega), carrying a Renilla luciferase reporter gene under the control of the thymidine kinase promoter, was cotransfected. The average normalized rey luciferase level with mock transfection was set at 100%, and error bars represent standard deviations. All transfections were performed in triplicate, and similar results were obtained in at least duplicate experiments. For trichostatin A (TSA) treatment, the cells were rst incubated with DMEM for 12 h posttransfection and with DMEM containing 300 nM TSA (Sigma). After incubation for another 24 h, the cells were lysed for analysis of luciferase activity. Coimmunoprecipitation. Whole-cell extracts were prepared from 293T cells transfected with the various expression vectors and incubated with either antiMyc (4A6; Upstate Biotechnology) or anti-FLAG (D-8; Santa Cruz Biotechnology) antibody for 1 h at 4C. The antigen-antibody complexes were collected using protein G-Sepharose (GE Healthcare) and washed several times. Precipitated proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electrophoretically transferred to polyvinylidene diuoride membranes. Western blotting was carried out according to a standard procedure by using rabbit polyclonal anti-FLAG or mouse monoclonal anti-Myc antibodies. ChIP experiments. Chromatin immunoprecipitation (ChIP) experiments were performed by using some modications of a previously described method (20). Briey, 293T cells (2.0 106/10-cm dish) were transfected with a linearized mesp-b 2.5-kb luciferase gene and each expression vector. After a 40-h incubation, the cells were cross-linked with formaldehyde, and the genomic DNA was sheared to an average size of 400 bp by sonication. Sonicated samples were incubated with anti-Myc antibody (9E10; Santa Cruz Biotechnology) or with anti-FLAG antibody (D9; SantaCruz), collected with protein G-Sepharose (GE Healthcare), and washed several times. After reversal of the cross-link, the DNA isolated by ChIP was amplied by PCR using the following primers: T-site containing region-s, 5-CAACAAACACAAAAAGCACACGTT-3; T-site containing region-as, 5-GGTGAAAGGAGGATGGAGGTTTAT-3; Proximal-region-s, 5-GCATCTCTATTGACGATATC-3; Proximal-region-as, 5-CATT TCATTGCTCTAGCTCA-3; Distal-region-s, 5-TTTCAGAGCACAACAA AAGCAGTG-3; and Distal-region-as, 5-TTATGTATGCATGAACGCAG ATGG-3. Fish and microinjection. The zebrash TL2 inbred line was used for sh experiments (18). Capped mRNA encoding the entire open reading frame of ripply1 was synthesized from linearized pCS2ripply1 (8) by using an mMessage mMachine Sp6 kit (Ambion). The sequences of morpholino antisense oligonucleotides (MOs) used in the study were as follows: ripply1 MO, 5-CATCGTC ACTGTGTTTTTCGTTTTG-3; tbx24 MO, 5-CATTTCCACACCCAGCATG TCTCGG-3; and deltaD MO, 5-AACAGCTATCATTAGTCGTCCCATG-3. Synthesized mRNA or MO was injected into the yolks of fertilized embryos at the one-cell stage. Whole-mount in situ hybridization. Whole-mount in situ hybridization was performed as described previously (17). An antisense RNA probe was transcribed according to a standard procedure.

FIG. 1. Tbx24-dependent transcriptional suppression of mesp-b mRNA by Ripply1 in zebrash embryos, shown by whole-mount in situ hybridization. The stained embryos were at mounted by removing the yolk (dorsal view and anterior to the top). The MOs were injected into the fertilized embryos at the appropriate concentration as previous described. (A) Ectopic expression of mesp-b mRNA is not detectable in tbx24 and ripply1 MO-injected embryos. The embryos were simultaneously stained with the probes for no tail (notochord and tailbud) and mesp-b (segmental expression in the anterior PSM). Expression of mesp-b, which is normally restricted to the anterior PSM, was signicantly decreased in tbx24 MO-injected embryos. In ripply1 MO-injected embryos, mesp-b mRNA was ectopically expressed in the presumptive somitic regions. In both the tbx24 and ripply1 MO-injected embryos, mesp-b mRNA was not ectopically expressed in the somitic regions, which is totally different from what was observed in ripply1 MO-injected embryos. Staining of no tail appeared to be comparable in either case. (B) Ectopic expression of mesp-b mRNA was detected in the deltaD and ripply1 MO-injected embryos. The embryos were stained with the probe for mesp-b. Expression of mesp-b mRNA was scattered in the anterior PSM of deltaD MO-injected embryos. In deltaD and ripply1 MO-injected embryos, expression of mesp-b mRNA was not properly terminated in the anterior PSM, in a similar manner to that in ripply1 MO-injected embryos.
RESULTS Up-regulation of mesp-b mRNA in ripply1-decient embryos is dependent on Tbx24. In ripply1-decient embryos, the expression of mesp-b was earlier shown to be up-regulated in the anterior PSM and extended anteriorly (16). The expression pattern of mesp-b in the anterior PSM is also known to be perturbed in zebrash embryos defective in either Notch signaling or the transcription factor Tbx24 (12, 24, 30). Thus, as a rst approach to address how Ripply1 regulates mesp-b expression, we investigated the epistatic relationship between ripply1 and tbx24 or deltaD, encoding a Notch ligand. As previously

KAWAMURA ET AL.

MOL. CELL. BIOL.
FIG. 2. Ripply1 antagonizes luciferase reporter gene activation mediated by Tbx24. Luciferase assays with 293T cells were conducted to examine the transcriptional relationship between Ripply1 and Tbx24. The amounts of plasmids used for transfection are shown under each bar. The average normalized rey luciferase activity with pCS2 alone was set at 100%, and error bars represent the standard deviations. (A) Schematic representation of the mesp-b 2.5-kb luciferase reporter construct. Two T sites and RBP-J-binding sequences are emphasized with boxes, and numbers show their locations relative to the mesp-b initiation codon. (B) Tbx24 signicantly activates reporter gene activity in 293T cells. (C) Ripply1 completely suppresses the enhancement of Tbx24 reporter activity. (D) Ripply1 does not suppress the transcriptional up-regulation of the reporter gene by the Notch intracellular domain (NICD).

described, the expression of mesp-b was signicantly reduced in tbx24 MO-injected embryos (Fig. 1A) (n 12; 100% were affected), suggesting the activator function of Tbx24 (24, 30), but was up-regulated in ripply1 MO-injected embryos (n 15; 100% were affected) (16). In embryos injected with both tbx24 MO and ripply1 MO, the mesp-b mRNA was almost completely suppressed, as in the tbx24 MO-injected embryos (Fig. 1A) (n 7; 100% were affected). In contrast, the anteriorly expanded expression of mesp-b caused by the ripply1 MO injection was not reduced in the embryos injected with deltaD MO and ripply1 MO, although its expression in the PSM was perturbed, showing a scattered salt-and-pepper pattern (Fig. 1B) (n 9; 100% were affected). These results suggest that the transcriptional suppression of mesp-b mRNA by Ripply1 depends on the function of Tbx24 in zebrash embryos. Ripply1, as well as Ripply2 and -3, antagonizes the transcriptional activation of Tbx24 through the T-box binding site. To investigate more precisely the relationship between Tbx24 and Ripply1 in the transcription of mesp-b, we next carried out in vitro luciferase reporter assays in which the reporter gene was under the control of the 2.5-kb genomic fragment upstream of the mesp-b initiation codon (Fig. 2A). We found that Tbx24 signicantly activated the luciferase activity in a dosedependent manner (Fig. 2B). This Tbx24-mediated activation of the reporter gene, however, was completely prevented by cotransfection with Ripply1 (Fig. 2C). In contrast, Ripply1 did not signicantly affect the enhancement of reporter activity by the Notch intracellular domain of notch1a (Fig. 2D), which is a constitutively active form of the Notch receptor (34). These results indicate that Ripply1 specically suppresses transcriptional activation by Tbx24.
To identify the regulatory element targeted by Ripply1, we next constructed a series of reporter plasmids containing progressive 5 deletions of the 2.5-kb mesp-b fragment fused to the luciferase gene. As summarized in Fig. 3A, reporter genes containing at least 82 bp upstream of mesp-b could be activated by Tbx24 and suppressed by Ripply1, but those with 56 bp upstream of mesp-b were not affected by either Tbx24 or Ripply1. Because a putative T-box binding site (T site) exists between 82 and 56 bp upstream of mesp-b, we next determined whether the T site (5-AATTCACACCT-3) is sufcient for the suppression of Tbx24 activity by Ripply1. The suppression by Ripply1 was also recapitulated by pBP-TbxX2, which contains two tandem T sites upstream of the beta interferon basal promoter and the rey luciferase gene (Fig. 3B). This suppression by Ripply1 was abolished by substitution of a single nucleotide within the T site that is crucial for T-box binding (mutT) (8, 22), suggesting that the regulatory element targeted by Ripply1 cannot be separated from the T site targeted by Tbx24. In addition to Ripply1, zebrash Ripply2 and Ripply3 also antagonized Tbx24 activation in a dose-dependent manner (Fig. 3C). Thus, the Ripply family proteins can commonly antagonize transcriptional activation by Tbx24. Ripply1-Tbx24 complex associates with T-box binding site. The inability to separate the regulatory element for Ripply from the Tbx-binding element indicated that Ripply1 might inhibit Tbx24 transcriptional activation via a physical interaction. To examine this possibility, we generated Myc-tagged Ripply1 and FLAG-tagged Tbx24 proteins. We conrmed that both proteins were functional in the luciferase assay (see Fig. S1 in the supplemental material). The two proteins were pre-

FIG. 3. Ripply1 antagonizes the transcriptional activation of Tbx24 through the Tbx binding site. (A) Deletion constructs of the mesp-b luciferase reporter gene. The effects of Tbx24 and Ripply1 on each reporter gene are summarized on the right side of the panel. (B) T site is sufcient for the transcriptional suppression of Tbx24 by Ripply1. The sequence of each T site or mutated T site inserted upstream of the beta interferon basal promoter of the luciferase reporter gene is shown. (C) Using the pBP-TbxX2 reporter gene, zebrash Ripply1, as well as Ripply2 and -3, antagonizes the transcriptional activation of Tbx24 through the T site in a dose-dependent manner.
dominantly colocalized in the nucleus (Fig. 4A), suggesting that the suppression by Ripply1 is not due to removal of Tbx24 from the nucleus. Rather, coimmunoprecipitation assays showed preferential association of Tbx24-FLAG with Myc-Ripply1, and vice versa, in 293T cell lysates (Fig. 4B). To examine whether this Tbx24-Ripply1 complex is associated with DNA at the T site, a ChIP assay was performed with 293T cells. Myc-Ripply1 was specically associated with the DNA region containing T sites (T region), but not with another two regions in the 5 region upstream of mesp-b, only in the presence of Tbx24-FLAG (Fig. 4C). On the other hand, Tbx24-FLAG was consistently associated with the T region independently of Ripply1 (Fig. 4C). These results strongly suggest that Ripply1 represses the function of Tbx24 by forming a complex with Tbx24 that binds to the T site. T domain and Ripply homology domain are indispensable for interactions between Tbx24 and Ripply1. To gain further mechanistic insights into the interactions between Tbx24 and Ripply1, we next prepared several deletion or amino acid substitution constructs of Tbx24 and Ripply1 (Fig. 5A). By conducting immunoprecipitation assays, we found that the T domain alone, which is highly conserved among the Tbx family
proteins, is sufcient for the association with Ripply1 but that other regions of Tbx24 are dispensable for the interactions (Fig. 5D). On the other hand, an 50-amino-acid sequence at the carboxyl terminus, called the Ripply homology domain, is conserved among the Ripply family proteins. Especially highly conserved sequences exist in this domain (Fig. 5B). Replacement of some of these sequences with an alanine stretch (Ripply1-mutFPVQ) resulted in a signicant reduction in the transcriptional repression of the pBP-TbxX2 reporter construct (Fig. 5C), indicating that this conserved sequence is required for Ripply-mediated transcriptional repression. This is at least caused by a signicantly decreased afnity of Ripply1-mutFPVQ for Tbx24, because the immunoprecipitation assay revealed that Ripply1-mutFPVQ was not efciently precipitated with Tbx24 compared with the coprecipitation with wild-type Ripply1 (Fig. 5E). Ripply1 converts the transcriptional property of Tbx24 from activator to repressor by recruiting the transcriptional corepressor Groucho/TLE. Ripply1 interacts with the transcriptional corepressor Groucho/TLE through the WRPW motif (7,

FIG. 4. The Tbx24-Ripply1 protein complex associates with the T site. (A) Colocalization of Myc-Ripply1 and Tbx24-FLAG proteins in the nuclei of 293T cells. (B) Myc-Ripply1 and Tbx24-FLAG proteins preferentially coimmunoprecipitate from lysates of 293T cells. (C) ChIP analysis reveals that Myc-Ripply1 and Tbx24-FLAG complexes, but not Myc-Ripply1 alone, preferentially associate with the DNA sequence containing the T sites. The DNA region containing the T sites (T region) was coprecipitated with Myc-Ripply1 in the presence of Tbx24, while a DNA region close to the T sites (proximal region), which includes potential binding sequences for the Pax, Hox, TCF/Lef1, and E47 transcription factors, as well as another region 2.0 kb upstream from the T sites (distal region), was not immunoprecipitated. The length of the amplied band was 250 bp in each case. On the other hand, Tbx24-FLAG was consistently associated with the T region in the presence or absence of Ripply1.
16). To determine whether Groucho/TLE is involved in the suppression of Tbx24 by Ripply, we next examined the effect of removing the WRPW motif. Compared with the wild-type Ripply1 protein, Ripply1 lacking the WRPW tetrapeptide did not efciently suppress Tbx24-mediated enhancement of luciferase activity (Fig. 6A). Because multimeric Groucho/TLE complexes contain histone deacetylases (HDACs) in Drosophila and mammals (5, 7), we next examined the effect of TSA, a chemical inhibitor of HDACs (37). We found that TSA partially rescued the suppression of luciferase activity by Ripply1 (Fig. 6B). These results suggest that Ripply1 counteracts the
transcriptional activity of Tbx24 by recruiting the Groucho/ TLE complex via the WRPW motif. Specically, Ripply acts as an adaptor that converts Tbx24 from an activator to a repressor. In fact, whereas Tbx24 activates the reporter genes in the absence of Ripply1 in 293T cells, it acts as a transcriptional repressor when there is an excess of Ripply1 (Fig. 6C). Thus, Ripply1 controls the characteristics of transcriptional regulation by Tbx24 from activator to repressor. Ripply1 can antagonize the transcriptional activation of another T-box gene, no tail, in vivo and in vitro. The result showing that the T domain, which is conserved among all of the
FIG. 5. Ripply homology domain is required for interactions with T domain of Tbx24 and suppression of mesp-b mRNA in zebrash embryos. (A) Schematic representation of modied FLAG-tagged Tbx24 and Myc-tagged Ripply1 variants used in this study. Conserved domains of Tbx24 and Ripply1 are emphasized with different colors. (B) Comparisons of Ripply homology domains in zebrash Ripply proteins. Identical amino acid residues are shown on a black background, and similar residues are shown on a gray background. A four-amino-acid stretch (the 97th to 100th amino acids) was replaced with an alanine stretch (emphasized with asterisks). This variant was designated ripply1-mutFPVQ. (C) Luciferase reporter assay. The assay employed a luciferase reporter in which two T sites were placed upstream of the basal promoter used for the experiment shown in Fig. 3B. In contrast to wild-type Ripply1, replacement of the highly conserved sequences in the Ripply homology domain with a stretch of four alanines resulted in a signicant reduction of the transcriptional repression of the pBP-TbxX2 reporter construct. (D and E) Immunoprecipitation assays. The T domain alone (173 amino acids) preferentially coimmunoprecipitated with Myc-Ripply1 (D). Substitution of four amino acid residues highly conserved in the Ripply homology domain results in a signicant reduction of afnity for Tbx24 (E). (F) Whole-mount in situ staining with mesp-b and mesp-a probes. Injection of ripply1 mRNA resulted in a signicant reduction of mesp-b, but not mesp-a, mRNA in the anterior PSM (arrowheads) at the three-somite stage. However, the same amount of substituted ripply1-mutFPVQ mRNA did not cause such suppression of mesp-b mRNA. Stained embryos are viewed from the dorsal side and anterior to the top.

FIG. 6. Ripply1 exerts Tbx24-dependent transcriptional repression by recruiting the Groucho/TLE-HDAC complex. (A) Suppression of Tbx24 transcriptional activity depends on the Groucho/TLE-binding motif WRPW of Ripply1. (B) Transcriptional suppression of Tbx24 by Ripply1 involves HDAC activity. (C) Transcriptional property of Tbx24 proteins is converted from activation to repression by Ripply1. The assay employed a luciferase reporter in which two T sites were placed upstream of the thymidine kinase promoter and the luciferase gene of pST0/TLN (25). Although Tbx24 normally acts as a transcriptional activator, Tbx24 acts as a repressor in the presence of Ripply1 (100 ng).
Tbx family proteins, is necessary and sufcient for the association between Ripply1 and Tbx24 suggests that Ripply acts on other T-box transcription factors (Fig. 3C). Indeed, Ripply1 suppressed transcriptional activation by Ntl, a zebrash ortholog of the mouse T-box protein T/Brachyury (9, 31) (Fig. 7A). Thus, Ripply proteins appear to suppress the function of various T-box transactivators in vitro. In addition, we investigated the ability of Ripply1 to antagonize T-box proteins in vivo by injecting ripply1 mRNA into eggs at the one-cell stage. As previously reported, injection of wild-type ripply1 mRNA at a low dosage caused segmental disruption of somites and suppression of mesp-b expression in the PSM. On the other hand, misexpression of ripply1 mRNA at a higher dosage resulted in truncation of the posterior body and the absence of a notochord and horizontal myoseptum (Fig. 7B). Although the altered somite patterning is similar to the phenotype observed for fused somites/tbx24 mutants (24), the posterior truncation and the absence of a differentiated notochord and horizontal myoseptum, regardless of the presence of a neural tube and somites (Fig. 7C) (n 3; 100% were affected), are strikingly similar to the phenotype of ntl homozygous mutants (9). Furthermore, as observed for ntl homozygous mutants (1, 9, 35), in the ripply1 mRNA-injected embryos, the expression of myod in the adaxial cells, which is induced by notochord, was missing at the one-somite stage (Fig. 7D). In a later stage, bilateral expression of myod was fused in the posterior trunk, as observed in ntl homozygous mutants (Fig. 7E). In contrast, the expression of mesp-a was normal in the ripply1 mRNA-injected embryos (Fig. 7D). Of note, the expression of ntl itself was also unchanged in the ripply1 mRNA-injected embryos (Fig. 7F). Interestingly, conversion of the Xenopus homolog of T/Brachyury into a transcriptional repressor by fusion with the En1 repressor domain also produces a similar phenotype to that of ntl homozygous mutants (6). Therefore, the morphological defect observed in ripply-injected embryos could be consistent with the expected defect when Ntl is converted into a repressor. Collectively, our results suggest that

Ripply proteins are regulators of a wide range of T-box proteins. Association with T-box protein is required for the function of Ripply in embryos. Finally, to see whether interaction with T-box proteins is required for Ripply activities in vivo, we injected mRNA encoding either ripply1 or ripply1-mutFPVQ into zebrash embryos. As previously described, injection of the wild-type ripply1 mRNA resulted in reduced expression of mesp-b but not mesp-a (Fig. 5F). In contrast, the injection of the mutated ripply1 mRNA did not cause an apparent decrease in mesp-b expression in the anterior PSM (Fig. 5F). Thus, the interaction between Ripply and T-box protein is required for the activity of Ripply1, even in embryos. DISCUSSION Previously, we showed that zebrash ripply1 plays roles in the repression of gene expression in the anterior PSM. Although we suggested that the Groucho/TLE transcriptional corepressor is involved in this suppression, the molecular mechanism underlying this suppression has mostly remained. Here we have provided evidence to support a model by which Ripply proteins convert the transcriptional property of T-box proteins from activators to repressors. During the preparation of this manuscript, Kondow et al. reported that Bowline, a Xenopus counterpart of Ripply, suppresses the activity of an activated form of Tbx6 fused with VP16 in cultured cells (19). Although this report implies potential interaction between Ripply/Bowline and a T-box protein, it still remains unclear whether Ripply/Bowline can actually suppress the activity of intact T-box proteins. In contrast, our results clearly indicate that the activity of the intact Tbx24 molecule is modulated by zebrash Ripply1 and that Tbx24 can actually be converted from an activator to a repressor by Ripply1. Furthermore, treatment with an inhibitor specic for HDAC activity strongly suggests that HDAC, which is known to be associated with Groucho/TLE, is actually involved in the Ripply-mediated sup-
FIG. 7. Transcriptional activation of Ntl, a zebrash ortholog of T/Brachyury, is antagonized by Ripply1. (A) Luciferase assays with 293T cells. Zebrash Ripply1 suppresses the transcriptional activation of Ntl proteins. (B) Injection of ripply1 mRNA causes truncation of the posterior body, similar to the phenotype of ntl homozygous mutants (9). Synthetic ripply1 mRNA was injected into fertilized embryos at a concentration of 10 ng/l. The image shown is a lateral view at 36 h postfertilization (hpf). (C) The notochord (N) and horizontal myoseptum (arrowhead) are absent in ripply1 mRNA-injected embryos. Transverse sections (7 m) at the trunk region above the yolk tube of embryos at 36 h postfertilization are shown. The sections were briey stained with hematoxylin. DA, dorsal aorta. (D) Expression of myod in adaxial cells, induced by the presence of the notochord, was absent from ripply1 mRNA-injected embryos at the one-somite stage. In contrast, expression of mesp-a in the anterior PSM appears indistinguishable between the control and ripply1 mRNA-injected embryos. (E) Bilateral expression of myod is fused in the posterior trunk (*), as observed in ntl homozygous mutants (1, 35). (F) Expression of ntl appears indistinguishable between control and ripply1 mRNA-injected embryos. The images show the animal pole view at the shield stage.

pression. We also show that Ripply1 is actually recruited to the T-box protein-binding DNA sequence through Tbx24. In addition, the T domain and a highly conserved amino acid stretch in the Ripply homology domain are indispensable for the interactions between Tbx24 and Ripply1, and as anticipated from this result, other members of either the Ripply or Tbx family are also interactive with Tbx24 or Ripply1, respectively. Finally, we show that this interaction is required for the activity of Ripply in embryos. Based on these results, we conclude that T-box proteins are converted from activators to repressors through specic interactions with Ripply proteins, which recruit the global corepressor Groucho/TLE. The spatiotemporal regulation of gene expression, which is critical for proper development, is managed by a number of molecular processes, including the replacement of coactivator and corepressor complexes. For instance, external signals, such as Wnt and Notch, displace corepressor complexes, which maintain default repression of these signals targets, from
DNA-binding transcriptional regulators (2, 1315). Here we show that spatiotemporal gene regulation is also accomplished by another fashion of replacement of corepressor complexes. Our study provides evidence that T-box proteins are converted from activators to repressors through specic interactions with Ripply proteins, which recruit the global corepressor Groucho/ TLE-HDAC complex. During somite segmentation in zebrash, the expression of mesp-b, a target of Tbx24, is dynamically regulated to yield the characteristic stripe pattern in the anterior PSM (30), even though Tbx24 is constantly expressed in this region (24). Considering that ripply1 is expressed in a stripe fashion in the anterior PSM (16), Ripply1 appears to provide conversion of the transcriptional property of Tbx24 in a stripe fashion in this region, and this conversion regulates the expression of mesp-b. Because defective function of Ripply1 or Mesp-b results in abnormal anteroposterior compartmentation within a single somite unit in zebrash embryos, the Ripply1mediated transcriptional control of mesp-b appears to be cru-
12. Holley, S. A., R. Geisler, and C. Nusslein-Volhard. 2000. Control of her1 expression during zebrash somitogenesis by a delta-dependent oscillator and an independent wave-front activity. Genes Dev. 14:16781690. 13. Hsieh, J. J., S. Zhou, L. Chen, D. B. Young, and S. D. Hayward. 1999. CIR, a corepressor linking the DNA binding factor CBF1 to the histone deacetylase complex. Proc. Natl. Acad. Sci. USA 96:2328. 14. Hurlstone, A., and H. Clevers. 2002. T-cell factors: turn-ons and turn-offs. EMBO J. 21:23032311. 15. Kao, H. Y., P. Ordentlich, N. Koyano-Nakagawa, Z. Tang, M. Downes, C. R. Kintner, R. M. Evans, and T. Kadesch. 1998. A histone deacetylase corepressor complex regulates the Notch signal transduction pathway. Genes Dev. 12:22692277. 16. Kawamura, A., S. Koshida, H. Hijikata, A. Ohbayashi, H. Kondoh, and S. Takada. 2005. Groucho-associated transcriptional repressor ripply1 is required for proper transition from the presomitic mesoderm to somites. Dev. Cell 9:735744. 17. Kawamura, A., S. Koshida, H. Hijikata, T. Sakaguchi, H. Kondoh, and S. Takada. 2005. Zebrash hairy/enhancer of split protein links FGF signaling to cyclic gene expression in the periodic segmentation of somites. Genes Dev. 19:11561161. 18. Kishimoto, Y., S. Koshida, M. Furutani-Seiki, and H. Kondoh. 2004. Zebrash maternal-effect mutations causing cytokinesis defect without affecting mitosis or equatorial vasa deposition. Mech. Dev. 121:7989. 19. Kondow, A., K. Hitachi, K. Okabayashi, N. Hayashi, and M. Asashima. 2007. Bowline mediates association of the transcriptional corepressor XGrg-4 with Tbx6 during somitogenesis in Xenopus. Biochem. Biophys. Res. Commun. 359:959964. 20. Lavrrar, J. L., and P. J. Farnham. 2004. The use of transient chromatin immunoprecipitation assays to test models for E2F1-specic transcriptional activation. J. Biol. Chem. 279:4634346349. 21. Morimoto, M., Y. Takahashi, M. Endo, and Y. Saga. 2005. The Mesp2 transcription factor establishes segmental borders by suppressing Notch activity. Nature 435:354359. 22. Muller, C. W., and B. G. Herrmann. 1997. Crystallographic structure of the T domain-DNA complex of the Brachyury transcription factor. Nature 389: 884888. 23. Naiche, L. A., Z. Harrelson, R. G. Kelly, and V. E. Papaioannou. 2005. T-box genes in vertebrate development. Annu. Rev. Genet. 39:219239. 24. Nikaido, M., A. Kawakami, A. Sawada, M. Furutani-Seiki, H. Takeda, and K. Araki. 2002. Tbx24, encoding a T-box protein, is mutated in the zebrash somite-segmentation mutant fused somites. Nat. Genet. 31:195199. 25. Nishikawa, J., M. Amano, Y. Fukue, S. Tanaka, H. Kishi, Y. Hirota, K. Yoda, and T. Ohyama. 2003. Left-handedly curved DNA regulates accessibility to cis-DNA elements in chromatin. Nucleic Acids Res. 31:66516662. 26. Packham, E. A., and J. D. Brook. 2003. T-box genes in human disorders. Hum. Mol. Genet. 12(Spec. No. 1):R37R44. 27. Pourquie, O. 2000. Segmentation of the paraxial mesoderm and vertebrate somitogenesis. Curr. Top. Dev. Biol. 47:81105. 28. Saga, Y., N. Hata, H. Koseki, and M. M. Taketo. 1997. Mesp2: a novel mouse gene expressed in the presegmented mesoderm and essential for segmentation initiation. Genes Dev. 11:18271839. 29. Saga, Y., and H. Takeda. 2001. The making of the somite: molecular events in vertebrate segmentation. Nat. Rev. Genet. 2:835845. 30. Sawada, A., A. Fritz, Y. J. Jiang, A. Yamamoto, K. Yamasu, A. Kuroiwa, Y. Saga, and H. Takeda. 2000. Zebrash Mesp family genes, mesp-a and mesp-b are segmentally expressed in the presomitic mesoderm, and Mesp-b confers the anterior identity to the developing somites. Development 127: 16911702. 31. Schulte-Merker, S., F. J. van Eeden, M. E. Halpern, C. B. Kimmel, and C. Nusslein-Volhard. 1994. no tail (ntl) is the zebrash homologue of the mouse T (Brachyury) gene. Development 120:10091015. 32. Showell, C., O. Binder, and F. L. Conlon. 2004. T-box genes in early embryogenesis. Dev. Dyn. 229:201218. 33. Takahashi, Y., K. Koizumi, A. Takagi, S. Kitajima, T. Inoue, H. Koseki, and Y. Saga. 2000. Mesp2 initiates somite segmentation through the Notch signalling pathway. Nat. Genet. 25:390396. 34. Takke, C., P. Dornseifer, E. von Weizsacker, and J. A. Campos-Ortega. 1999. her4, a zebrash homologue of the Drosophila neurogenic gene E(spl), is a target of NOTCH signalling. Development 126:18111821. 35. Weinberg, E. S., M. L. Allende, C. S. Kelly, A. Abdelhamid, T. Murakami, P. Andermann, O. G. Doerre, D. J. Grunwald, and B. Riggleman. 1996. Developmental regulation of zebrash MyoD in wild-type, no tail and spadetail embryos. Development 122:271280. 36. Yasuhiko, Y., S. Haraguchi, S. Kitajima, Y. Takahashi, J. Kanno, and Y. Saga. 2006. Tbx6-mediated Notch signaling controls somite-specic Mesp2 expression. Proc. Natl. Acad. Sci. USA 103:36513656. 37. Yoshida, M., M. Kijima, M. Akita, and T. Beppu. 1990. Potent and specic inhibition of mammalian histone deacetylase both in vivo and in vitro by trichostatin A. J. Biol. Chem. 265:1717417179.

cial for proper segmentation of somites. Thus, the regulatory recruitment of corepressor complexes to T-box proteins, established by the dynamic expression of ripply1, is likely to be a key process to establish segmental gene expression during somite development. The identication of Ripply1 as a switching molecule for T-box genes suggests a novel transcriptional mechanism that could participate in other aspects of development, although the relationship between the Ripply family and T-box factors besides Tbx24 and Ntl remains to be elucidated. The overlapping expression of ripply and T-box genes observed in various developing tissues and organs (A. Kawamura and S. Takada, unpublished data) supports the idea that the two proteins cooperate in development. Further understanding of the cooperative transcriptional regulation by Ripply and T-box proteins could help to elucidate the mechanisms underlying the disproportionate activation or repression found in human genetic disorders associated with mutations in T-box genes.
ACKNOWLEDGMENTS We thank T. Ohyama, J. Campos-Ortega, K. Hoshijima, and M. Nikaido for reagents and technical advice. We are also grateful to T. Higashinakagawa, H. Kondoh, K. Inoue, and K. Yamasu for support and to the members of the Takada and Higashinakagawa laboratories for critical discussions. This work was supported by a grant-in-aid for scientic research from the Ministry of Education, Science, Culture, and Sports of Japan to S.T. and S.K. A.K. was supported by a research fellowship from the Japan Society for the Promotion of Science.
REFERENCES 1. Amacher, S. L., B. W. Draper, B. R. Summers, and C. B. Kimmel. 2002. The zebrash T-box genes no tail and spadetail are required for development of trunk and tail mesoderm and medial oor plate. Development 129:3311 3323. 2. Barolo, S., T. Stone, A. G. Bang, and J. W. Posakony. 2002. Default repression and Notch signaling: Hairless acts as an adaptor to recruit the corepressors Groucho and dCtBP to Suppressor of Hairless. Genes Dev. 16: 19641976. 3. Bussen, M., M. Petry, K. Schuster-Gossler, M. Leitges, A. Gossler, and A. Kispert. 2004. The T-box transcription factor Tbx18 maintains the separation of anterior and posterior somite compartments. Genes Dev. 18:1209 1221. 4. Chapman, D. L., and V. E. Papaioannou. 1998. Three neural tubes in mouse embryos with mutations in the T-box gene Tbx6. Nature 391:695697. 5. Chen, G., J. Fernandez, S. Mische, and A. J. Courey. 1999. A functional interaction between the histone deacetylase Rpd3 and the corepressor Groucho in Drosophila development. Genes Dev. 13:22182230. 6. Conlon, F. L., S. G. Sedgwick, K. M. Weston, and J. C. Smith. 1996. Inhibition of Xbra transcription activation causes defects in mesodermal patterning and reveals autoregulation of Xbra in dorsal mesoderm. Development 122:2427 2435. 7. Fisher, A. L., and M. Caudy. 1998. Groucho proteins: transcriptional corepressors for specic subsets of DNA-binding transcription factors in vertebrates and invertebrates. Genes Dev. 12:19311940. 8. Goering, L. M., K. Hoshijima, B. Hug, B. Bisgrove, A. Kispert, and D. J. Grunwald. 2003. An interacting network of T-box genes directs gene expression and fate in the zebrash mesoderm. Proc. Natl. Acad. Sci. USA 100: 94109415. 9. Halpern, M. E., R. K. Ho, C. Walker, and C. B. Kimmel. 1993. Induction of muscle pioneers and oor plate is distinguished by the zebrash no tail mutation. Cell 75:99111. 10. Herrmann, B. G., S. Labeit, A. Poustka, T. R. King, and H. Lehrach. 1990. Cloning of the T gene required in mesoderm formation in the mouse. Nature 343:617622. 11. Hofmann, M., K. Schuster-Gossler, M. Watabe-Rudolph, A. Aulehla, B. G. Herrmann, and A. Gossler. 2004. WNT signaling, in synergy with T/TBX6, controls Notch signaling by regulating Dll1 expression in the presomitic mesoderm of mouse embryos. Genes Dev. 18:27122717.

MCB Accepts, published online ahead of print on 10 March 2008 Mol. Cell. Biol. doi:10.1128/MCB.01754-07 Copyright 2008, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.

Kawamura et al

Activator-to-repressor conversion of T-box transcription factors by the Ripply family of Groucho/TLE-associated mediators
Running title: Mediator in the repression by T-box proteins Akinori Kawamura1,2,3, Sumito Koshida1,4, #, and Shinji Takada1,4,* Okazaki Institute for Integrative Biosciences, National Institutes of Natural Sciences, Okazaki, Aichi 444-8787, Japan 1 Graduate School of Science and Engineering, Bioscience and Biomedical Engineering Waseda University, Shinjuku, Tokyo 169-8050, Japan 2 Department of Life Science, Graduate School of Science and Engineering, Saitama University, Sakura-ku, Saitama 338-8570, Japan 3 Department of Basic Biology, Graduate University for Advanced Studies (SOKENDAI), Okazaki, Aichi 444-8787, Japan 4
*Author to whom correspondence should be addressed Shinji Takada, Ph. D. Okazaki Institute for Integrative Biosciences, National Institutes of Natural Sciences, Okazaki, Aichi 444-8787, Japan, Phone: +81-564-59-5241, Fax: +81-564-59-5240, E-mail: stakada@nibb.ac.jp
Present address: Department of Biological Sciences, Graduate School of Science, University of Tokyo, Bunkyo-ku, Tokyo, Japan. The word count for the Materials and Methods section is 559. The combined word count for the Introduction, Results, and Discussion sections is

2,794.

ABSTRACT The T-box family of transcription factors, defined by a conserved DNA binding domain called the T-box, regulates various aspects of embryogenesis by activating and/or repressing downstream genes. In spite of the biological significance of the
T-box proteins, how they regulate transcription remains to be elucidated. Here, we show that the Groucho/TLE-associated protein, Ripply, converts T-box proteins from activators to repressors. In cultured cells, zebrafish Ripply1, an essential
component in somite segmentation, and its structural relatives, Ripply2 and 3, suppress the transcriptional activation mediated by the T-box protein Tbx24,
which is coexpressed with ripply1 during segmentation. Ripply1 associates with Tbx24 and converts it to a repressor. Ripply1 also antagonizes the transcriptional
activation of another T-box protein, No tail (Ntl), the zebrafish ortholog of Brachyury. Furthermore, injection of high dosage of ripply1 mRNA into zebrafish eggs causes defective development of the posterior trunk, similar to the phenotype observed in homozygous mutants of ntl. A mutant form of Ripply1, defective in association with Tbx24, also lacks the activity in zebrafish embryos. These results indicate that the intrinsic transcriptional property of T-box proteins is controlled by Ripply family proteins, which act as specific adaptors that recruit the global corepressor Groucho/TLE to T-box proteins.
INTRODUCTION T-box proteins play many crucial roles in development by activating and/or repressing the transcription of their target genes (23, 32). For instance, reduced function of
T/Brachyury, the founder member of this family, causes truncated tail in the mouse (10). The importance of T-box transcription factors is also shown by the facts that mutations in human T-box genes cause severe congenital disorders, such as DiGeorge,
Ulnar-mammary, and Holt-Oram syndromes (26). Furthermore, the T-box family is evolutionally conserved from C. elegans to insects and vertebrates showing remarkable functional conservation across species. However, in spite of the biological significance of the T-box proteins, how they regulate transcription remains to be elucidated. One of the processes in which T-box proteins play crucial roles is the somite
segmentation in vertebrates (3, 4, 24). Somitogenesis is the sequential subdivision of segmented precursors of the vertebral column and musculature from the presomitic mesoderm (PSM) (27, 29). Prior to morphological subdivision, a segmental prepattern is established in the anterior PSM. Genetic analyses of the mouse and the zebrafish indicate that several Tbx genes are required for distinct processes of this prepattern formation (11, 24, 36), which also involves Notch signaling and Mesp2, a basic helix-loop-helix transcription factor (21, 33). For instance, Tbx6 induces the expression

of Mesp2 and is required for proper patterning of somites of the mouse (36). Similarly, tbx24/fused somites, structurally related to mouse Tbx6, is expressed in the anterior PSM of the zebrafish embryo, and required for the formation of somite boundaries and
the expression of mesp-b, a zebrafish homolog of Mesp2 (24, 30), suggesting that Tbx24 is involved in the transcriptional regulation of the mesp-b locus in zebrafish. Because
Mesp2, as well as mesp-b, is expressed in only a portion of Tbx-expressing cells (24, 28, 36), however, another molecule also seems to be involved in Tbx-mediated transcriptional regulation of mesp genes in the anterior PSM.
In a previous study, we showed that Ripply1, which is associated with the global
transcriptional corepressor Groucho/TLE, terminates the expression of segmentation genes in the anterior PSM of zebrafish embryos (16). In ripply1-deficient embryos, the expression of mesp-b, is upregulated in a cell-autonomous manner, whereas, in embryos injected with ripply1mRNA, the expression of mesp-b is highly suppressed in the anterior PSM (16). These results suggest that Ripply1 regulates the proper expression of mesp-b in the anterior PSM. Taking it into account that the expression of mesp-b could be induced by Tbx, we can speculate that Ripply1 may antagonize the function of Tbx in the transcription of mesp genes. Therefore, we examined precisely the relationship between Tbx24 and Ripply1 in the transcription of mesp-b. We showed that Ripply1 is
coprecipitated with Tbx24 and converts it from an activator to a repressor in culture cells. The other members of the Ripply family, Ripply2 and Ripply3, can also antagonize the transcriptional activation mediated by Tbx24. On the other hand,
Ripply1 also antagonizes the transcriptional activation of another T-box protein, No tail (Ntl), both in vitro and in vivo. These results reveal a molecular mechanism underlying dynamic conversion of the regulatory property of Tbx proteins in the transcription of their targets.

MATERIALS AND METHODS

Plasmid Constructions. The mesp-b 2.5-kb luciferase reporter plasmid was

function of Tbx24 in zebrafish embryos.
Ripply1, as well as Ripply 2 and 3, antagonizes the transcriptional activation of Tbx24 through the T-box binding site. To investigate the relationship between Tbx24 and Ripply1 in the transcription of mesp-b more precisely, we next carried out in vitro
luciferase reporter assays in which the reporter gene was under the control of the 2.5-kb genomic fragment upstream of the mesp-b initiation codon (Fig. 2A). We found that Tbx24 significantly activated the luciferase activity in a dose-dependent manner (Fig. 2B). This Tbx24-mediated activation of the reporter gene, however, was completely prevented by cotransfection with Ripply1 (Fig. 2C). In contrast, Ripply1 did not
significantly affect the enhancement of reporter activity by the Notch intracellular
domain of notch1a (Fig. 2D), which is a constitutively active form of the Notch receptor (34). These results indicate that Ripply1 specifically suppresses transcriptional activation by Tbx24.
To identify the regulatory element targeted by Ripply1, we next constructed a series
of reporter plasmids containing progressive 5-deletions of the 2.5 kb mesp-b fragment fused to the luciferase gene. As summarized in Fig. 3A, reporter genes containing at least 82 bp upstream of mesp-b could be activated by Tbx24 and suppressed by Ripply1, but those with 56 bp upstream of mesp-b were not affected by either Tbx24 or Ripply1. Because a putative T-box binding site (T-site) exists between 82 and 56 bp upstream of mesp-b, we next determined whether the T-site (5-AATTCACACCT-3) is sufficient for the suppression of Tbx24 activity by Ripply1. The suppression by Ripply1 was also recapitulated by pBP-TbxX2, which contains two tandem T-sites upstream of the
interferon- basal promoter and the firefly luciferase gene (Fig. 3B). This suppression by Ripply1 was abolished by substitution of a single nucleotide within the T-site that is crucial for T-box binding (mutT) (8, 22), suggesting that the regulatory element targeted by Ripply1 cannot be separated from the T-site targeted by Tbx24.
In addition to Ripply1, zebrafish Ripply2 and Ripply3 also antagonized Tbx24
activation in dose-response manners (Fig. 3C). Thus, the Ripply family proteins can commonly antagonize the transcriptional activation by Tbx24.
Ripply1-Tbx24 complex associates with T-box binding site. The inability to
separate the regulatory element for Ripply from the Tbx-binding element indicated that Ripply1 might inhibit Tbx24 transcriptional activation via a physical interaction. To examine this possibility, we generated Myc-tagged Rippy1 and FLAG-tagged Tbx24 proteins. We confirmed that both proteins were functional in the luciferase assay (See Fig. S1 in the supplemental material). The two proteins were predominantly co-localized in the nuclei (Fig. 4A), suggesting that the suppression by Ripply1 is not due to removal of Tbx24 from the nucleus. Rather, co-immunoprecipitation assays showed preferential association of Tbx24-FLAG with Myc-Ripply1 and vice versa in 293T cell lysates (Fig. 4B). To examine whether this Tbx24-Ripply1 complex is

associated with DNA at the T-site, a chromatin immunoprecipitation (ChIP) assay was performed in 293T cells. Myc-Ripply1 is specifically associated with the DNA region containing T-sites (T-region), but not with another two regions in the 5 upstream of the mesp-b, only in the presence of Tbx24-FLAG (Fig. 4C). On the other hand,
Tbx-24-FLAG is consistently associated with the T-region independently of Ripply1
(Fig. 4C). These results strongly suggest that Ripply1 represses the function of Tbx24 by forming a complex with Tbx24 that binds to the T-site.
T-domain and Ripply homology domain are indispensable for the interactions
between Tbx24 and Ripply1. To get further mechanistic insights into the interactions between Tbx24 and Ripply1, we next prepared several deletion or amino acid-substituted constructs of Tbx24 and Ripply1 (Fig 5A). By conducting immunoprecicipitation assays, we found that the T-domain alone, which is highly conserved among the Tbx family proteins, is sufficient for the association with Ripply1 but that other region of Tbx24 is dispensable for the interactions (Fig. 5D). On the other hand, a ~50 amino acid length sequence at the carboxyl terminus, called the Ripply homology domain, is conserved among the Ripply family proteins. Especially, highly conserved sequences exist in this domain (Fig. 5B). Substitution of
some of these sequences to an alanine stretch (Ripply1-mutFPVQ) results in a significant reduction in the transcriptional repression of the pBP-TbxX2 reporter construct (Fig. 5C), indicating that this conserved sequence is required for
Ripply-mediated transcriptional repression. This is at least caused by a significantly
decreased affinity of Ripply1-mutFPVQ for Tbx24, because the immunoprecipitation assay revealed that Ripply1-mutFPVQ was not efficiently precipitated with Tbx24 compared with the co-precipitation with wild-type Ripply1 (Fig 5E).
Ripply1 converts the transcriptional property of Tbx24 from activator to
repressor by recruiting the transcriptional corepressor Groucho/TLE. Ripply1 interacts with the transcriptional corepressor, Groucho/TLE, through the WRPW motif (7, 16). To determine whether Groucho/TLE is involved in the suppression of Tbx24 by Ripply, we next examined the effect of removing the WRPW motif. Compared with the wild-type Ripply1 protein, Ripply1 lacking the WRPW tetrapeptide did not efficiently suppress Tbx24-mediated enhancement of luciferase activity (Fig. 6A). Because multimeric Groucho/TLE complexes contain histone deacetylases (HDACs) in Drosophila and mammals (5, 7), we next examined the effect of trichostatin A (TSA), a chemical inhibitor of HDACs (37). We found that TSA partially rescued the suppression

of luciferase activity by Ripply1 (Fig. 6B). These results suggest that Ripply1 counteracts the transcriptional activity of Tbx24 by recruiting the Groucho/TLE complex via the WRPW motif. Specifically, Ripply acts as an adaptor that converts
Tbx24 from an activator to a repressor. In fact, whereas Tbx24 activates the reporter
genes in the absence of Ripply1 in 293T cells, it acts as a transcriptional repressor when there is an excess of Ripply1 (Fig. 6C). Thus, Ripply1 controls the characteristics of transcriptional regulation by Tbx24 from activator to repressor.
Ripply1 can antagonize the transcriptional activation of another T-box gene, no
tail, in vivo and in vitro. The result that the T-domain, which is conserved among the all of Tbx family proteins, is necessary and sufficient for the association between Ripply1 and Tbx24 suggests that Ripply acts on other T-box transcription factors (Fig. 3C). Indeed, Ripply1 suppressed transcriptional activation by No tail (Ntl), a zebrafish ortholog of the mouse T-box protein T/Brachyury (9, 31) (Fig. 7A). Thus, Ripply proteins appear to suppress the function of various T-box trans-activators in vitro. In addition, we investigated the ability of Ripply1 to antagonize T-box proteins in vivo by injecting ripply1 mRNA into one-cell stage eggs. As previously reported, injection of wild type ripply1 mRNA at low dosage caused segmental disruption of somites and
suppression of mesp-b expression in the PSM. On the other hand, misexpression of ripply1 mRNA at higher dosage resulted in truncation of the posterior body and the absence of a notochord and horizontal myoseptum (Fig. 7B). Although the altered
somite patterning is similar to the phenotype observed in fuses somites/tbx24 mutants (24), the posterior truncation and the absence of a differentiated notochord and
horizontal myoseptum, regardless of the presence of a neural tube and somites (Fig. 7C, n=3; 100% affected), is strikingly similar to ntl homozygous mutants (9). Furthermore, as observed in ntl homozygous mutants (1, 9, 35), in the ripply1 mRNA-injected embryos, the expression of myod in the adaxial cells, which is induced by notochord, was missing at the one-somite stage (Fig. 7D). In a later stage, bilateral expression of myod is fused in the posterior trunk, as observed in ntl homozygous mutants (Fig. 7E).

In contrast, the expression of mesp-a was normal in the ripply1 mRNA-injected embryos (Fig. 7D). Of note, the expression of ntl itself was also unchanged in the ripply1 mRNA-injected embryos (Fig. 7F). Interestingly, conversion of the Xenopus homolog of T/Brachyury into a transcriptional repressor by fusion with the En1 repressor domain also produces a similar phenotype as ntl homozygous mutants (6). Therefore, the morphological defect observed in ripply-injected embryos could be consistent with the expected defect when Ntl is converted into a repressor. Collectively,
our results suggest that Ripply proteins are regulators of a wide range of T-box proteins.
Association with T-box protein is required for the function of Ripply in embryos.
Finally, to see whether interaction with T-box proteins is required for Ripply activities in vivo, we injected mRNA encoding either ripply1 or ripply1-mutFPVQ into zebrafish
embryos. As previously described, injection of the wild type ripply1 mRNA results in
reduced expression of mesp-b, but not mesp-a (Fig. 5F). In contrast, the injection of this mutated ripply1 mRNA did not cause an apparent decrease in mesp-b expression in the anterior PSM (Fig. 5F). Thus, the interaction between Ripply and T-box protein is required for the activity of Ripply1 even in embryos.

DISCUSSION

Previously, we showed that zebrafish ripply1 plays roles in the repression of gene
expression in the anterior PSM. Although we suggested that the Groucho/TLE transcriptional co-repressor is involved in this suppression, the molecular mechanism underlying this suppression has mostly remained. Here, we have provided evidence to support a model by which Ripply proteins convert the transcriptional property of T-box proteins from activators to repressors. During the preparation of this manuscript,
Kondow et al. reported that Bowline, a Xenopus counterpart of Ripply, suppresses the activity of an activated form of Tbx6 fused with VP16 in cultured cells (19). Although this report implies potential significance between Ripply/Bowline and a T-box protein, it still remains unclear whether Ripply/Bowline can actually suppress the activity of intact T-box proteins. In contrast, our results clearly indicate that the activity of the intact Tbx24 molecule is modulated by zebrafish Ripply1 and that the property of
Tbx24 can be actually converted from activator to repressor by Ripply1. Furthermore, treatment with an inhibitor specific for HDAC activity strongly suggests that HDAC, which is known to be associated with Groucho/TLE, is actually involved in the Ripply-mediated suppression. We also show that Ripply1 is actually recruited to the T-box protein-binding DNA sequence through Tbx24. In addition, the T-domain and a highly conserved amino acid stretch in the Ripply homology domain are indispensable for the interactions between Tbx24 and Ripply1; and, as anticipated from this result, other members of either the Ripply or Tbx family are also interactive with Tbx24 or Ripply1, respectively. Finally, we show that this interaction is required for the activity of Ripply in embryos. Based on these results, we conclude that T-box proteins are converted from activators to repressors through specific interactions with Ripply proteins, which recruit the global corepressor Groucho/TLE.

The spatiotemporal regulation of gene expression, which is critical for proper development, is managed by a number of molecular processes, including the replacement of coactivator and corepressor complexes. For instance, external signals, such as Wnt and Notch, displace corepressor complexes, which maintain default
repression of these signals targets, from DNA-binding transcriptional regulators (2,
13-15). Here, we show that the spatiotemporal gene regulation is also accomplished by
another fashion of replacement of corepressor complexes. Our study provides evidence that T-box proteins are converted from activators to repressors through specific interactions with Ripply proteins, which recruit the global corepressor Groucho/TLE-HDAC complex. During somite segmentation in zebrafish, the expression of mesp-b, a target of Tbx24, is dynamically regulated to yield the characteristic stripe pattern in the anterior PSM (30), even though Tbx24 is constantly expressed in this region (24). Considering that ripply1 is expressed in a stripe fashion in the anterior PSM (16), Ripply1 appears to provide conversion of the transcriptional property of Tbx24 in a stripe fashion in this region, and this conversion generates the characteristic expression pattern of mesp-b. Because defective function of Ripply1 or Mesp-b results in abnormal antero-posteror compartmentation within a single somite unit in zebrafish embryos, the Ripply1-mediated transcriptional control of mesp-b
appears to be crucial for proper segmentation of somites. Thus, the regulatory recruitment of corepressor complexes to T-box proteins, established by the dynamic expression of ripply1, is likely to be a key process to establish the segmental gene expression during somite development.
The identification of Ripply1 as a switching molecule for T-box genes suggests a novel transcriptional mechanism that could participate in other aspects of development,
although the relationship between the Ripply family and T-box factors besides Tbx24 and Ntl remains to be elucidated. The overlapping expression of ripply and T-box genes, observed in various developing tissues and organs (A.K and S.T, unpublished data), supports the idea that the two proteins cooperate in development. Further understanding of the cooperative transcriptional regulation by Ripply and T-box proteins could help elucidate the mechanisms underlying the disproportionate activation or repression found in human genetic disorders associated with mutations in T-box genes.

ACKNOWLEDGEMENTS We thank T. Ohyama, J. Campos-Ortega, K. Hoshijima, and M. Nikaido for reagents
and technical advice. We are also grateful to T. Higashinakagawa, H. Kondoh, K. Inoue, and K Yamasu for support and the members of the Takada and Higashinakagawa laboratories for critical discussions. This work was supported by a grant-in-aid for
scientific research from the Ministry of Education, Science, Culture, and Sports of
Japan to S.T. and S.K. A.K. was supported by a research fellowship from the Japan Society for the Promotion of Science.
REFERENCES 1. Amacher, S. L., B. W. Draper, B. R. Summers, and C. B. Kimmel. 2002. The zebrafish T-box genes no tail and spadetail are required for development of trunk and tail mesoderm and medial floor plate. Development 129:3311-23. Barolo, S., T. Stone, A. G. Bang, and J. W. Posakony. 2002. Default repression and Notch signaling: Hairless acts as an adaptor to recruit the corepressors
Groucho and dCtBP to Suppressor of Hairless. Genes Dev 16:1964-76. Bussen, M., M. Petry, K. Schuster-Gossler, M. Leitges, A. Gossler, and A. Kispert. 2004. The T-box transcription factor Tbx18 maintains the separation of anterior and posterior somite compartments. Genes Dev 18:1209-21. Chapman, D. L., and V. E. Papaioannou. 1998. Three neural tubes in mouse embryos with mutations in the T-box gene Tbx6. Nature 391:695-7. Chen, G., J. Fernandez, S. Mische, and A. J. Courey. 1999. A functional interaction between the histone deacetylase Rpd3 and the corepressor groucho in Drosophila development. Genes Dev 13:2218-30. Conlon, F. L., S. G. Sedgwick, K. M. Weston, and J. C. Smith. 1996. Inhibition of Xbra transcription activation causes defects in mesodermal patterning and reveals autoregulation of Xbra in dorsal mesoderm. Development

7. 8. 9. 10. 11.

122:2427-35. Fisher, A. L., and M. Caudy. 1998. Groucho proteins: transcriptional corepressors for specific subsets of DNA-binding transcription factors in vertebrates and invertebrates. Genes Dev 12:1931-40. Goering, L. M., K. Hoshijima, B. Hug, B. Bisgrove, A. Kispert, and D. J. Grunwald. 2003. An interacting network of T-box genes directs gene expression and fate in the zebrafish mesoderm. Proc Natl Acad Sci U S A 100:9410-5. Halpern, M. E., R. K. Ho, C. Walker, and C. B. Kimmel. 1993. Induction of muscle pioneers and floor plate is distinguished by the zebrafish no tail mutation. Cell 75:99-111. Herrmann, B. G., S. Labeit, A. Poustka, T. R. King, and H. Lehrach. 1990. Cloning of the T gene required in mesoderm formation in the mouse. Nature 343:617-22. Hofmann, M., K. Schuster-Gossler, M. Watabe-Rudolph, A. Aulehla, B. G. Herrmann, and A. Gossler. 2004. WNT signaling, in synergy with T/TBX6, controls Notch signaling by regulating Dll1 expression in the presomitic mesoderm of mouse embryos. Genes Dev 18:2712-7.

Holley, S. A., R. Geisler, and C. Nusslein-Volhard. 2000. Control of her1 expression during zebrafish somitogenesis by a delta-dependent oscillator and an independent wave-front activity. Genes Dev 14:1678-90. Hsieh, J. J., S. Zhou, L. Chen, D. B. Young, and S. D. Hayward. 1999. CIR, a corepressor linking the DNA binding factor CBF1 to the histone deacetylase complex. Proc Natl Acad Sci U S A 96:23-8. Hurlstone, A., and H. Clevers. 2002. T-cell factors: turn-ons and turn-offs. Embo J 21:2303-11. Kao, H. Y., P. Ordentlich, N. Koyano-Nakagawa, Z. Tang, M. Downes, C. R. Kintner, R. M. Evans, and T. Kadesch. 1998. A histone deacetylase corepressor complex regulates the Notch signal transduction pathway. Genes Dev 12:2269-77. Kawamura, A., S. Koshida, H. Hijikata, A. Ohbayashi, H. Kondoh, and S. Takada. 2005. Groucho-associated transcriptional repressor ripply1 is required for proper transition from the presomitic mesoderm to somites. Dev Cell 9:735-44. Kawamura, A., S. Koshida, H. Hijikata, T. Sakaguchi, H. Kondoh, and S. Takada. 2005. Zebrafish hairy/enhancer of split protein links FGF signaling to cyclic gene expression in the periodic segmentation of somites. Genes Dev 19:1156-61. Kishimoto, Y., S. Koshida, M. Furutani-Seiki, and H. Kondoh. 2004. Zebrafish maternal-effect mutations causing cytokinesis defect without affecting

14. 15.

18. 19. 20. 21. 22. 23.
mitosis or equatorial vasa deposition. Mech Dev 121:79-89. Kondow, A., K. Hitachi, K. Okabayashi, N. Hayashi, and M. Asashima. 2007. Bowline mediates association of the transcriptional corepressor XGrg-4 with Tbx6 during somitogenesis in Xenopus. Biochem Biophys Res Commun
359:959-64. Lavrrar, J. L., and P. J. Farnham. 2004. The use of transient chromatin immunoprecipitation assays to test models for E2F1-specific transcriptional activation. J Biol Chem 279:46343-9. Morimoto, M., Y. Takahashi, M. Endo, and Y. Saga. 2005. The Mesp2 transcription factor establishes segmental borders by suppressing Notch activity. Nature 435:354-9. Muller, C. W., and B. G. Herrmann. 1997. Crystallographic structure of the T domain-DNA complex of the Brachyury transcription factor. Nature 389:884-8. Naiche, L. A., Z. Harrelson, R. G. Kelly, and V. E. Papaioannou. 2005. T-box

26. 27. 28.

genes in vertebrate development. Annu Rev Genet 39:219-39. Nikaido, M., A. Kawakami, A. Sawada, M. Furutani-Seiki, H. Takeda, and K. Araki. 2002. Tbx24, encoding a T-box protein, is mutated in the zebrafish somite-segmentation mutant fused somites. Nat Genet 31:195-9. Nishikawa, J., M. Amano, Y. Fukue, S. Tanaka, H. Kishi, Y. Hirota, K. Yoda, and T. Ohyama. 2003. Left-handedly curved DNA regulates accessibility to cis-DNA elements in chromatin. Nucleic Acids Res 31:6651-62. Packham, E. A., and J. D. Brook. 2003. T-box genes in human disorders. Hum Mol Genet 12 Spec No 1:R37-44. Pourquie, O. 2000. Segmentation of the paraxial mesoderm and vertebrate somitogenesis. Curr Top Dev Biol 47:81-105. Saga, Y., N. Hata, H. Koseki, and M. M. Taketo. 1997. Mesp2: a novel mouse gene expressed in the presegmented mesoderm and essential for segmentation initiation. Genes Dev 11:1827-39. Saga, Y., and H. Takeda. 2001. The making of the somite: molecular events in vertebrate segmentation. Nat Rev Genet 2:835-45. Sawada, A., A. Fritz, Y. J. Jiang, A. Yamamoto, K. Yamasu, A. Kuroiwa, Y. Saga, and H. Takeda. 2000. Zebrafish Mesp family genes, mesp-a and mesp-b are segmentally expressed in the presomitic mesoderm, and Mesp-b confers the

29. 30.

31. 32. 33. 34. 35. 36.
anterior identity to the developing somites. Development 127:1691-702. Schulte-Merker, S., F. J. van Eeden, M. E. Halpern, C. B. Kimmel, and C. Nusslein-Volhard. 1994. no tail (ntl) is the zebrafish homologue of the mouse T (Brachyury) gene. Development 120:1009-15. Showell, C., O. Binder, and F. L. Conlon. 2004. T-box genes in early embryogenesis. Dev Dyn 229:201-18. Takahashi, Y., K. Koizumi, A. Takagi, S. Kitajima, T. Inoue, H. Koseki, and Y. Saga. 2000. Mesp2 initiates somite segmentation through the Notch signalling pathway. Nat Genet 25:390-6. Takke, C., P. Dornseifer, E. v Weizsacker, and J. A. Campos-Ortega. 1999. her4, a zebrafish homologue of the Drosophila neurogenic gene E(spl), is a target of NOTCH signalling. Development 126:1811-21. Weinberg, E. S., M. L. Allende, C. S. Kelly, A. Abdelhamid, T. Murakami, P. Andermann, O. G. Doerre, D. J. Grunwald, and B. Riggleman. 1996. Developmental regulation of zebrafish MyoD in wild-type, no tail and spadetail embryos. Development 122:271-80. Yasuhiko, Y., S. Haraguchi, S. Kitajima, Y. Takahashi, J. Kanno, and Y.
Saga. 2006. Tbx6-mediated Notch signaling controls somite-specific Mesp2 expression. Proc Natl Acad Sci U S A 103:3651-6. Yoshida, M., M. Kijima, M. Akita, and T. Beppu. 1990. Potent and specific inhibition of mammalian histone deacetylase both in vivo and in vitro by trichostatin A. J Biol Chem 265:17174-9.
Legends for figures FIG. 1. Tbx24-dependent transcriptional suppression of mesp-b mRNA by Ripply1 in zebrafish embryos. Whole-mount in situ hybridization. The stained embryos are flat-mounted by removing the yolk (dorsal view and anterior to the top). The MOs were injected into the fertilized embryos at the appropriate concentration as previous
described. (A) Ectopic expression of mesp-b mRNA is not detectable in the tbx24 and ripply1 MO-injected embryos. The embryos were simultaneously stained with the
probes for no tail (notochord and tailbud) and mesp-b (segmental expression in the
anterior PSM). Expression of mesp-b, which is normally restricted to the anterior PSM, is significantly decreased in tbx24 MO-injected embryos. In ripply1 MO-injected

embryos, mesp-b mRNA is ectopically expressed in the presumptive somatic regions. In
both the tbx24 and ripply1 MO-injected embryos, mesp-b mRNA is not ectopically expressed in the somatic regions, which is totally different from those observed in ripply1 MO-injected embryos. Staining of no tail appears comparable in either case. (B) Ectopic expression of mesp-b mRNA is detected in the deltaD and ripply1 MO-injected embryos. The embryos was stained with the probe for mesp-b. Expression of mesp-b mRNA is scatted in the anterior PSM of deltaD MO-injected embryos. In deltaD and ripply1 MO-injected embryos, expression of mesp-b mRNA is not properly terminated in the anterior PSM in a similar manner to those in ripply1 MO-injected embryos.
FIG. 2. Ripply1 antagonizes the luciferase reporter gene activation mediated by Tbx24. Luciferase assays in 293T cells were conducted to examine the transcriptional relationship between Ripply1 and Tbx24. The amounts of plasmids used for transfection are shown under each bar. The average normalized firefly luciferase activity with
pCS2+ alone was set at 100%, and error bars represent the standard deviations. (A) Schematic representation of the mesp-b 2.5-kb luciferase reporter construct. Two T-sites and RBP-J-binding sequences are emphasized with boxes, and the number shows their location relative to the mesp-b initiation codon. (B) Tbx24 significantly activates reporter gene activity in 293T cells. (C) Ripply1 completely suppresses the enhancement of Tbx24 reporter activity. (D) Ripply1 does not suppress the
transcriptional up-regulation of the reporter gene by the Notch intracellular domain (NICD).
FIG. 3. Ripply1 antagonizes the transcriptional activation of Tbx24 through the Tbx binding site. (A) Deletion constructs of the mesp-b luciferase reporter gene. The effects
of Tbx24 and Ripply1 on each reporter gene are summarized on the right side of the panel. (B) The T-site is sufficient for the transcriptional suppression of Tbx24 by Ripply1. Sequence of each T-site or mutT-site inserted upstream of interferon- basal
promoter of luciferase reporter gene is shown. (C) Using the pBP-TbxX2 reporter gene, zebrafish Ripply1, as well as Ripply2 and 3, antagonize the transcriptional activation of Tbx24 through the T-site in dose-dependent manners.
FIG. 4. The Tbx24-Ripply1 protein complex associates with the T-site. (A) Colocalization of Myc-Ripply1 and Tbx24-FLAG proteins in the nucleus of 293T cells. (B) Myc-Ripply1 and Tbx24-FLAG proteins preferentially coimmunoprecipitate from lysates of 293T cells. (C) ChIP analysis reveals that Myc-Ripply1 and Tbx24-FLAG complex but not Myc-Ripply1 alone preferentially associate with the DNA sequence containing the T-sites. The DNA region containing the T sites (T-region) is

Ripply1 (100ng).

property of Tbx24 proteins is converted from activator to repressor by Ripply1. The assay employed a luciferase reporter in which two T-sites were placed upstream of the thymidine kinase promoter and the luciferase gene of pST0/TLN (25). Although Tbx24 normally acts as a transcriptional activator, Tbx24 acts as a repressor in the presence of
FIG. 7. Transcriptional activation of Ntl, a zebrafish ortholog of T/Brachyury, is antagonized by Ripply1. (A) Luciferase assays in 293T cells. Zebrafish Ripply1 suppresses the transcriptional activation of Ntl proteins. (B) Injection of ripply1 mRNA causes the truncation of the posterior body, similar to ntl homozygous mutants (9). Synthetic ripply1 mRNA was injected into fertilized embryos at a concentration of 10 ng/l. Shown is the lateral view 36 hour post-fertilization (hpf). (C) The notochord (N) and horizontal myoseptum (arrowhead) are absent in the ripply1 mRNA-injected
embryos. Transverse sections (7m) at the trunk regions above the yolk tube of 36-hpf embryos. The sections were briefly stained with hematoxylin. DA, dorsal aorta. (D) Expression of myod in the adaxial cells, induced by the presence of the notochord, is absent from the ripply1 mRNA-injected embryos at the one-somite stage. In contrast,
expression of mesp-a in the anterior PSM appears indistinguishable between the control and ripply1 mRNA-injected embryos. (E) Bilateral expression of myod is fused in the
posterior trunk (*), as observed in ntl homozygous mutants (1, 35). (F) Expression of ntl appears indistinguishable between the control and ripply1 mRNA-injected embryos. Animal pole view at the shield stage.