SEARCH

SEARCH BY CITATION

Keywords:

  • enhancer;
  • medaka;
  • mesp;
  • somitogenesis;
  • transgenic

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Somitogenesis is a critical step during the formation of metameric structures in vertebrates. Recent studies in mouse, chick, zebrafish and Xenopus have revealed that several factors, such as T-box genes, Notch/Delta, Wnt, retinoic acid and FGF signaling, are involved in the specification of nascent somites. By interacting with these pathways, the Mesp2-like bHLH transcription factors are transiently expressed in the anterior presomitic mesoderm and play a crucial role in somite formation. The regulatory mechanisms of Mesp2 and its related genes during somitogenesis have been studied in mouse and Xenopus. However, the precise mechanism that regulates the transcriptional activity of Mesp2 has yet to be determined. In our current report, we identify the essential enhancer element of medaka mesp-b, an orthologue of mouse Mesp2, using transgenic techniques and embryo manipulation. Our results demonstrate that a region of approximately 2.8 kb, upstream of the mesp-b gene, is responsible for both the initiation and anterior localization of mesp-b transcription within a somite primordium. Furthermore, putative motifs for both T-box transcription factors and Notch/Delta signaling are present in this enhancer region and are essential for activity.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Somites constitute the first visible metameric units in vertebrate embryos and arise from the segmental subdivision of the paraxial mesoderm called presomitic mesoderm (PSM). The process of somitogenesis begins in the cranial region of the embryo and successively proceeds caudally on both sides of the notochord and neural tube. The strict periodicity of somitogenesis is thought to be governed by a ‘clock-and-wavefront’ mechanism (Cooke & Zeeman 1976), whereby a biochemical oscillator (the segmentation clock) operates in PSM cells and a wavefront of maturation sweeps back through this tissue, arresting oscillation and initiating somite differentiation. Recent genetic and molecular studies have supported this model and identified a number of genes that are critical to this process (reviewed in Aulehla & Herrmann 2004; Rida et al. 2004; Pourque 2003). The periodicity is generated by the oscillation of Notch/Delta signaling components (the clock) in the posterior PSM, and this temporal periodicity is then translated into the segmental units in the ‘wavefront’. The wavefront is thought to exist in the anterior PSM and progress backwards at a constant rate. The progression of this wavefront was found to be regulated by antagonistic interaction of two signaling molecules, FGF and retinoic acid (RA). FGF are highly expressed in the tail region and antagonize the function of RA which is produced in the segmented somites (Dubrulle & Pourquie 2004).

In the wavefront or anteriormost PSM, mouse Mesp2 and its related proteins, such as Mesp-b (zebrafish) and Thylacine1 (Thy1; Xenopus), play an essential role in arresting both the clock oscillation and rostro-caudal patterning of somites (Saga et al. 1997; Sparrow et al. 1998; Sawada et al. 2000; Moreno & Kintner 2004; Morimoto et al. 2005). The Mesp genes are first activated as a broad stripe, probably corresponding to one presumptive somite at the S-II position (we have adopted the recently proposed presomite nomenclature system, in which the forming somite is referred to as S0, the newly formed somite as SI, and S-I and S-II designate the next two presumptive somites; Pourquie & Tam 2001), and this expression becomes localized to the anterior half of the presumptive somite at S-I prior to somite formation (Takahashi et al. 2000). Interestingly, the expressions of Notch-related bHLH transcriptional repressors such as c-hairy1 (chick; Palmeirim et al. 1997), her1 and her7 (zebrafish; Sawada et al. 2000; Oates & Ho 2002), Hes1 and Hes7 (mouse; Jouve et al. 2000; Bessho et al. 2001), and her7 and hey1 (medaka; Elmasri et al. 2004a), which oscillate and move anteriorly through the posterior-to-intermediate PSM, also take on a fixed segmental pattern resembling that of the Mesp2-like genes in the wavefront. Recent genetic analysis indicates that these genes contribute to segmental patterning, in part by regulating the segmental expression of the Mesp2-like bHLH proteins (Jen et al. 1999; Takahashi et al. 2000, 2003). Conversely, Mesp2 arrests the oscillation of Notch-related genes by suppressing Notch signaling (Morimoto et al. 2005). Moreover, zebrafish mesp-b is placed genetically downstream of Tbx24, which is broadly expressed in the PSM, as mesp-b expression is completely lost in the fused somites mutant in which tbx24 is mutated (Nikaido et al. 2002). Thus, a sophisticated genetic network that is centered by Mesp2 operates in the anterior PSM prior to segment boundary formation (Saga & Takeda 2001).

Given the essential role of the Mesp genes and their strictly regulated expression pattern in somitogenesis, it will be of great interest to determine the precise mechanisms underlying their regulation. This will also help us to further understand somitogenesis at the molecular level. The regulation of Mesp2 and its related genes during somitogenesis have been studied in mouse (Haraguchi et al. 2001) and Xenopus (Moreno & Kintner 2004), and the core PSM enhancer of mouse Mesp2 was found to be located in the region 300 bp upstream of the ATG translation initiation site. On the other hand, an approximately 3.5 kb upstream region of Xenopus Thy1 has been shown to recapitulate endogenous Thy1 expression in the PSM (Moreno & Kintner 2004). In any case, the upstream factors that directly regulate the transcriptional activity of Mesp2 and its associated genes remain unknown.

In our current study, we report the elucidation of the medaka mesp-b enhancer using transgenic techniques and embryo manipulation. We chose the medaka embryo as our experimental model for the analysis of the mesp gene enhancer as this is an emerging vertebrate model system. Moreover, medaka studies have been recently boosted by a high-quality draft genome (Naruse et al. 2004; Medaka Genome Sequencing Project, 2005) and a variety of developmental mutants (Furutani-Seiki et al. 2004). Intriguingly, some of the medaka somite mutants seem to be unique and are not covered by the zebrafish mutagenesis screening studies that have been conducted so far (Elmasri et al. 2004b). We demonstrate that an approximately 2.8 kb sequence upstream of the translation initiation site can recapitulate endogenous mesp-b expression. We further show that this region contains several putative motifs which can mediate Tbx24, Notch/Delta and RA signaling, and provide experimental evidence that some of these motifs function cooperatively in vivo during the establishment of the mesp-b expression domain. We further compare these results with reports of the mouse Mesp2 enhancer, and discuss the significance of both the conserved and divergent enhancer structures in vertebrate mesp loci.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Fish strains

We used the southern strain of medaka (Olyzias latipes), cab. Embryos were staged according to Iwamatsu (1994).

Isolation of the medaka mesp-a and mesp-b gene loci

Medaka mesp-a cDNA was partially isolated as an expressed sequence tag (EST) clone MF01SSA178C10 (Kimura et al. 2004). A medaka BAC library was screened with this clone using the AlkPhos Direct Labeling and Detection System (Amersham Biosciences, Piscataway, NJ, USA). Restriction mapping revealed the relative location of mesp-a within five positive BAC clones. The mesp-a gene locus was most centrally located in BAC clone 180H12, compared with the other four clones. Shotgun sequencing of BAC clone 180H12 was performed at the DNA Sequencing Center of the National Institute of Genetics. Sequence analysis revealed the presence of mesp-b, which encodes a Mesp-type bHLH domain.

Total RNA isolates from somite-stage medaka embryos were prepared using ISOGEN (Nippongene, Tokyo, Japan). First strand cDNA were synthesized from 2 µg of total RNA using SuperscriptII Reverse Transcriptase with 3′ rapid amplification of cDNA ends (RACE) oligo primer (Invitrogen, Carlsbad, CA, USA). Partial mesp-b cDNA fragments were amplified with an Abridged Universal Amplification Primer (Invitrogen) and gene specific primers for mesp-b (5′-TACCTCCTCGGTCCCTTTT-3′ and 5′- GACCTCTCCAGCACGTCTTC -3′) then separated by gel electrophoresis. These fragments was cloned into the pCRII-TOPO vector (Invitrogen) and sequenced.

Whole-mount in situ hybridization

Embryos were fixed with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS), dechorionated manually with forceps and stored in methanol at −20°C. After rehydration, embryos were treated with proteinase K and re-fixed with 4% PFA/PBS. Hybridization was performed at 65°C with digoxigenin-labeled probes overnight. Hybridized embryos were washed with 50% formamide/2× SSC + 0.1% Tween-20 (SSCT), 2× SSCT and 0.2× SSCT, then treated with antidigoxigenin antibodies labeled with alkaline phosphatase (AP) at 1/7000 dilution. Signals were developed using 4-nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl-phosphate.

mesp-a cDNA and the mesp-b genomic fragments were used as templates for the probes. The mesp-a cDNA and mesp-b genomic fragments were isolated from cab strain of medaka by either reverse transcription–polymerase chain reaction or polymerase chain reaction (PCR) with specific primers designed according to medaka whole-genome shotgun data (for mesp-a, 5′-ATGGAGATGTCCTTCTGCTTCCCCCTTCAG-3′ and 5′-TCAACACTGTCCTGAGAAGATCACAGCTGG-3′; for mesp-b, 5′-ATGGATACCTCCTCGGTCCCTTTTCTTGAC-3′ and 5′-ACCCCAGTACTCTCTTGGAACCAGAGGAGT-3′).

Additional cDNA used as probe templates were as follows: tbx24 (EST ID; MF01SSA103E01); her7 (Elmasri et al. 2004a); deltaD (EST ID; MF01SSA044D02).

Plasmid construction and mutagenesis

The 5 kb genomic fragments upstream of medaka mesp-b were amplified by PCR from the BAC clone containing the mesp genes, using the following primers with introduced XhoI or BamHI restriction sites (restriction sites underlined; the bold letter indicates the ±1 translational initiation site of mesp-b): 5′-TACTCGAGATAGTTCACATCAGAACCTCCC-3′ and 5′-ATGGATCCATGTTGTGAGGAGGAGCTCCAC-3′. These fragments were subsequently cloned into the pMEPI vector (Kakihara et al. unpubl. data, 2006). The 2.8 kb EGFP vector was generated by self-ligation of a fragment amplified by PCR from the 5 kb EGFP vector, using the following primers with introduced BglII restriction sites (restriction sites underlined): 5′-GAAGATCTAACTATTATTAATATGCTGACA-3′ and 5′-GAAGATCTCTCGACCTGCAGGCATGCAAGC-3′. The 1.4 kb EGFP vector was generated by self-ligation of a SphI–NsiI fragment from the 5 kb EGFP vector. The sequence upstream of EGFP translation initiation site is as follows (small letters, upstream sequence of mesp-b; +1, the translation initiation site of mesp-b; capital bold letters, coding sequence of EGFP; and underlined, BamHI site); 5′-ccagaacagaccgtggagctcctcctcacaacatgGATCCCCGGGTACCGGTCGCCACCATGGTGAGCAA-3′.

The pMEPI vector was generated as follows: an I-SceI site was inserted into the pEGFP vector (Clontech, Palo Alto, CA, USA) at a position just downstream of the 3′ MCS by self-ligation of the PCR product with the following primers (I-SceI site underlined): 5′-AATTACCCTGTTATCCCTAGGGCCCGTACGGCCGA-3′ and 5′-TTCGTCTCGCGCGTTTCGGTGATGACGGTG-3′. An SV40 polyadenylation sequence was then cloned into the SpeI–BsiWI site of this vector. This was isolated from pEGFP-1 (Clontech) by PCR using primers with an introduced SpeI–BsiWI site. Finally, this MCS–EGFP–pA–I-SceI fragment was amplified by PCR and subsequently cloned into the pSMART LC Amp vector (Lucigen, Middleton, WI, USA).

Further deletion constructs of putative binding motifs in the 2.8 kb mesp-b enhancer fragment were generated by PCR using the following primers: for dT1, 5′-CTTTATTTGTTTCCATACGTGATCATAT-3′ and 5′-ATGCCAAGACCAGCACTTGAGTCAGGCGG-3′; for dT2, 5′-CTGTGGCCTGGCTTGGATTTATAAACCAAG-3′ and 5′-AAGTGACACCTCTGCAGATTCTGCACTGTC-3′; for dN1, 5′-TTATGTTTTGGCTTAGGTCCACGAAATAA-3′ and 5′-AAGTAAATGTTTCAGAAAACAAAAGCAAT-3′; for dN2, 5′-ACGTTGTATCAGGATGAGTTCAGGCTCATC-3′ and 5′-AAAACAGGGAATCTTTACGGTTAATCCAAC-3′; for dR, 5′-GCCCGCCTGACTCAAGTGCTGGTCTTGGCA-3′ and 5′-CACCGCTCGACCCTGCTGAGGTGAGGCTGT-3′; for dW1+2, 5′-TGGCGGAAGCCTTGCAAGCAAAAGTGGCGT-3′ and 5′-GTCAACAAAACTTTGCCTGTGGT-3′; for dW3+4, 5′-CATCTATTTGTATTTTTCAGTTAGCTTTAG-3′ and 5′-CTGTGAGCGCCGTTTTGTAGTCCAAATTCA-3′ (see also Fig. 4A).

image

Figure 4. Deletion mapping of the medaka mesp-b enhancer elements. (A) The region between −2.8 kb and −1.4 kb upstream of the mesp-b gene contains a putative T-box binding site (T1), RBPJκ binding site (N1), TCF binding sites (W1–4) and a retinoic acid response element (RARE, R). Gray letters indicate mismatches to the consensus. Color codes for each of the transcription factors and the nucleotide sequence of their binding sites are indicated at the bottom. Arrows indicate the positions of primers used to make each deletion construct in (B). (B) Removal of T1, T2, N1, N2, W1+2 resulted in the reduced expression of the reporter green fluorescent protein (GFP). A double deletion of either T1 and T2 (dT1+2), or N1 and N2 (dN1+2) further reduces the levels of reporter expression. Note that the 1.4 kb upstream sequence of mesp-b also contains putative T-box binding (T2) and RBPJκ binding sites (N2), which were also tested here. In all cases examined, except for the 1.4 kb fragment, the anterior localized pattern of GFP expression was maintained (pattern is +). For reporter activity, ++ denotes normal levels of expression; + indicates reduced expression; ± equals very weak levels of expression.

Download figure to PowerPoint

Microinjection

Microinjection and generation of transgenic medaka was performed according to Thermes et al. (2002). For microinjections, 1-cell stage embryos of cab inbred strain were used. DNA was prepared using a Qiagen Midiprep kit (Qiagen, Tokyo, Japan) and was co-injected with the restriction enzyme I-SceI to facilitate incorporation of the constructs into the genome (Thermes et al. 2002), through the chorion into the cytoplasm of the 1-cell stage embryos. Embryos were raised to sexual maturity and transgenic carriers were identified by crossing inter se.

Pharmacological treatments

N-(N-(3,5-Difluorophenacetyl-L-alanyl))-S-phenylglycine t-butyl ester (DAPT; γ-secretase inhibitor IX; Calbiochem, San Diego, CA, USA) treatments were performed according to the method of Geling et al. (2002). Briefly, a 10 mm stock of DAPT in dimethylsulfoxide (DMSO) was diluted to 100 µm in Ringer solution (Yamamoto 1975) and applied to dechorionated medaka embryos. Disulfiram (Sigma, St Louis, MO, USA) treatments were performed according to Vermot et al. (2005). A 400 mm stock of disulfiram in DMSO was diluted to 800 µm in Ringer solution and applied to dechorionated medaka embryos. Embryos were incubated with DAPT or disulfiram at 28.5°C from stage 15 until stage 25. Control embryos for each experiment were incubated in medium containing an equivalent concentration of DMSO.

Morpholino knockdown

The antisense morpholino oligonucleotide (MO) used against medaka tbx24 was 5′-TTCAAAGTATTCTCACCTGCCCTTT-3′ and the control MO was 5′-CCTCTTACCTCAGTTACAATTTATA-3′ (GeneTools LLC, Philomath, OR, USA). MO were resuspended in sterile water as a 3 mm stock solution and diluted to 1.5 mm in Ringer solution.

Image recording and processing

Transgenic F1 embryos were selected for reporter green fluorescent protein (GFP) expression under a MZFLIII Leica dissecting microscope with a GFP filter (Leica Microsystems, Wetzlar, Germany). Images were recorded with Zeiss AxioCam HRc (Zeiss, Oberkochen, Germany) and processed and mounted using Photoshop CS (Adobe, San Jose, CA, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Isolation of two medaka Mesp-related genes

Partial cDNA fragments of a medaka mesp-related gene were identified as an EST clone (MF01SSA178C10). Employing the RACE strategy, we cloned the full-length cDNA of this mesp gene. Within the BAC clone harboring this mesp gene, we identified an additional mesp-related gene. Both of these mesp genes were found to contain two exons and to be positioned in a head-to-tail orientation, separated by an approximately 6 kb intergenic sequence (Fig. 1A). Based on their expression pattern (see below; Fig. 2A–H), the first of these factors was designated as medaka mesp-a, and is expressed in both the blastoderm margin (nascent mesoderm) and PSM. The second of the two genes was designated as mesp-b and was found to be expressed only in the PSM. The cluster mesp-a/b is the co-orthologue of the mouse Mesp-1/2 and has a similar genomic organization as the zebrafish mesp region, which has an approximately 14 kb intergenic sequence between the two genes.

image

Figure 1. Genomic organization of the medaka mesp region and medaka Mesp protein sequences among different vertebrates. (A) Schematic representation of the medaka mesp-a and mesp-b gene locus. (B) Complete peptide sequences of the medaka Mesp-a and Mesp-b aligned with zebrafish Mesp-a and Mesp-b, and mouse MesP1 and MesP2. Identical amino acids are indicated by dots, and dashes represent gaps introduced to maximize the alignment. (C) Comparison of the medaka Mesp-a and Mesp-b bHLH domains with those of other closely related transcription factors. (D) A phylogenetic tree of the bHLH transcription factors, constructed using the neighbor-joining method (Saitou & Nei 1987) with cDNA sequences of the bHLH domains.

Download figure to PowerPoint

image

Figure 2. Normal expression patterns of the mesp-a and mesp-b genes in the medaka embryo. In each panel, the probe used is indicated at the bottom. The embryos are viewed from the dorsal side, and are oriented with the anterior side (animal pole) towards the top. (A–H) mesp expression in wild-type embryos at stage 15 (A,E) and stage 22 (B–D,F–H) showing a dynamic on–off pattern. (I–M) Comparison of the expression domains of mesp-a (I), mesp-b (J), her7 (K), deltaD (L) and tbx24 (M) in the presomitic mesoderm. SI, newly formed somite; S0, forming somite. Bars, (A,E) 0.25 mm; (B–D,F–H) 0.1 mm; (I–M) 0.1 mm. Arrowhead, mesp-a expression in (A).

Download figure to PowerPoint

The mesp-a cDNA consists of 899 bp and encodes a protein of 245 amino acids from the first possible initiation site at the 23rd nucleotide (nt). The mesp-b cDNA yields a protein of 262 amino acids. The bHLH domains of these two proteins share more than 90% amino acid identity (49/51), but no significant homology between them exists outside of these motifs (Fig. 1B). Homology searches using the bHLH domains indicated that among all members of the Mesp-related family, isolated from vertebrate model organisms, medaka mesp-a and mesp-b are most closely related to each other. Less sequence similarity was found between the bHLH domains of these two genes and other bHLH transcription factors, such as medaka Her7 and Myf5 (Fig. 1C). A phylogenic tree (Fig. 1D) with the cDNA sequences of bHLH domains confirms that medaka mesp-a and mesp-b belong to a Mesp-related subfamily of bHLH transcription factors.

Expression of the medaka mesp-a and mesp-b genes during embryonic development

Whole-mount in situ hybridization with digoxigenin-labeled probes was performed to investigate the embryonic expression pattern of the two medaka mesp genes. mesp-a transcripts were detectable at the epiboly stage in the blastoderm margin (Fig. 2A) and during the segmentation period in stripes of cells located on either side of the neural tube (Fig. 2B–D). In contrast, mesp-b is not expressed at the epiboly stage (Fig. 2E), but does show an expression pattern that is similar to mesp-a during the segmentation period (Fig. 2F–H). It is noteworthy that the expression patterns of medaka mesp-a and mesp-b are nearly identical to zebrafish mesp-a and mesp-b, respectively. Since mesp-b, but not mesp-a, has been shown to play an essential role in somitogenesis (Sawada et al. 2000), we focused on the medaka mesp-b gene in our subsequent experiments.

Before attempting to elucidate the mesp-b enhancer, we examined the expression pattern of the key segmentation genes, her7, deltaD and tbx24, because their expression profiles have not been well documented in the medaka embryo (Fig. 2I–M). Medaka her7 shows a dynamic expression pattern and its expression domain sweeps from posterior to anterior through the PSM. The anterior her7 stripe overlaps that of the medaka mesp genes. deltaD is broadly expressed in the PSM with high expression in the posterior. The anterior part of the segmented somite also expresses deltaD. The expression domain of tbx24 expands from the intermediate to anterior PSM, which overlaps that of the mesp genes. Taken together, we found that the expression domains of these genes partially overlap with mesp-b in the anterior PSM, suggesting their possible involvement in the regulation of mesp-b. Under our culture conditions, a pair of somites is formed in the medaka embryo every 60 min.

Identification of a presomitic mesoderm enhancer of mesp-b

To map the enhancer element that mediates PSM-specific mesp-b transcription, we first analyzed the region upstream of the translation initiation site in the mesp-b gene fragment. A 5 kb upstream fragment of the mesp-b gene was fused to the EGFP reporter cassette (Fig. 3A), and this construct was then injected into fertilized eggs. The animals that resulted from these injected eggs were raised to adulthood and then crossed inter se. The progeny from these crosses were then screened for EGFP-mediated fluorescence. In our subsequent enhancer analyses, we used only stable transgenic fish, which were the offspring of founder fish (G0), because transient expression patterns in these medaka embryos, when injected with enhancer constructs, tended to be mosaic and variable. The number of G0 animals used in each assay is shown in Figure 3E.

image

Figure 3. Green fluorescent protein (GFP) reporter expression under the control of a 5 kb medaka mesp-b upstream sequence recapitulates the endogenous mesp-b expression pattern. (A) Genomic locus of the medaka mesp genes and the construct used to generate the mesp-b transgenic lines. A 5 kb fragment upstream of the mesp-b gene was ligated to the GFP reporter gene. (B–D) Transmitted light images (B), GFP fluorescence (C) and GFP mRNA expression (D) of the 5 kb mesp-b enhancer transgenic medaka embryo at stage 22. The anterior is to the left. (E) Comparable enhancer activity was detected in reporter assays using both the 5 kb and 2.8 kb upstream fragments of the mesp-b gene. A more truncated 1.4 kb species however, greatly reduces this enhancer activity. GFP expression in the offspring derived from positive G0 are shown. The expression pattern and strength in the offspring was found to be invariable among G0 founders with the same transgene (see also Table 1). The number of G0 founders that produced offspring showing reporter gene expression is shown on the right. Bars, (BC) 0.1 mm; (D) 0.05 mm.

Download figure to PowerPoint

The embryos carrying the 5 kb reporter cassette showed a segmental GFP expression pattern in the PSM and also in segmented somites (Fig. 3C). The GFP reporter protein became detectable in the somite primordia, posterior to the forming somite (S-I), and was observed to be present in segmented somites. GFP fluorescence was first detected in the anterior half of the presumptive somite and formed somites, and then gradually expanded into the posterior half. Significantly, the timing and the initial profile of the GFP reporter expression in these transgenic embryos is consistent with the endogenous mesp-b expression pattern. Mesp-b expression is first detectable at the level of S-II or S-I in a relatively broad stripe and then becomes restricted to the anterior half of the presumptive somite (Fig. 2J). However, the GFP reporter expression pattern in our experiments was found to persist for a longer period after the completion of boundary formation, and even became enhanced as somatic cells differentiated into the muscle lineage. Endogenous mesp-b expression, however, is shut off as the morphological boundary becomes visible. The longer retention of reporter expression could be due to the high stability of the GFP protein, and the precise timing of transcription could also be masked by the time required for the maturation of GFP. We thus performed in situ hybridization staining for GFP in our transgenic embryos. As shown in Figure 3D, transcription of the GFP reporter gene initiates in a broad stripe or as an anteriorly localized stripe in the presumptive S-I and/or S-II somites. This is also observed for endogenous mesp-b, but once again the reporter transcripts are detectable over four to five somite levels after the completion of boundary formation, leaving six to seven stripes of GFP mRNA in the transgenic embryos. Hence, the mesp-b 5 kb upstream sequence does contain a cis-regulatory element that is necessary for transcriptional initiation and anterior localization, but may lack the region required for the downregulation of gene expression following segmentation.

To further elucidate the regulatory regions that harbor mesp-b enhancer activity, we generated two deletion reporter constructs containing a 2.8 kb or 1.4 kb segment of the upstream mesp-b gene sequence (Fig. 3E). Transgenic medaka that were generated using the 2.8 kb species exhibited a GFP expression pattern that was similar to the transgenic animals containing the 5 kb enhancer. We also observed little variation in the expression pattern among the founder fish harboring the 2.8 kb enhancer as 5/10 of these produced transgenic offspring showing nearly identical GFP expression patterns (Table 1). In contrast, the 1.4 kb upstream sequence reporter exhibited reduced activity in transgenic progeny. Of the 14 founder animals used in the set of experiments with this deletion construct, one produced offspring showing low levels of ubiquitous expression, and another two transmitted a weak segmental expression pattern (Fig. 3E). This suggested that an important regulatory element resides within the region between −2.8 and −1.4 kb upstream of the mesp-b translation initiation site.

Table 1.  GFP expression driven by 5 kb, 2.8 kb and 1.4 kb upstream sequences
G0No. eggs examinedGFP expression level
Enhancer sequenceID+++±
  • GFP expression driven by 5 kb, 2.8 kb and 1.4 kb upstream sequences. Three out of five G0 producing GFP-positive offspring for the 5 kb sequence, two out of five for the 2.8 kb sequence and two out of three for the 1.4 kb sequence were examined. The number of offspring assigned to each GFP expression level is shown. Reporter activity was rated as follows: ++, normal levels of expression; +, reduced; ±, very weak.

  • The GFP expression pattern driven by the 1.4 kb fragment was also lost or disturbed. GFP, green fluorescent protein.

5 kbw2172800 0
s1116170 0
s7 37 60 0
2.8 kbt7 50182 0
t9 67 60 0
1.4 kbt6 33 0012
t7 12 0012

T-box binding sites and RBPJκ binding sites cooperatively participate in mesp-b enhancer activity

The 2.8 kb sequence upstream of mesp-b gene was assessed for putative transcription factor-binding motifs and revealed a number of putative sites including T-box core binding sequences (designated T1 and T2; Conlon et al. 2001), RBPJκ binding-like sequences (N1 and N2; Tun et al. 1994), a retinoic acid (RA) response element (RARE; reviewed in Bastie & Rochette-Egly 2004) and Tcf binding sequences (W1–4; Travis et al. 1991; Brannon et al. 1997; Fig. 4A). Significantly, these sites were found to be nearly identical to the corresponding published consensus sequences. We also found several putative binding sites with less conserved motifs distributed along the entire 2.8 kb mesp-b enhancer region (Fig. 7). However, in this present study we focused upon examining the function of the following highly conserved binding sites: T1, T2, N1, N2, RARE and W1–4. Based upon previous reports that we cited earlier, we reasoned that these motifs would be indispensable for enhancer activity.

image

Figure 7. Diagrams showing conserved blocks of sequence from the 5′ flanking regions of vertebrate mesp genes. Comparisons are shown between the mesp upstream regions of medaka, fugu, zebrafish and mouse. The sequence motifs that we used to search for putative binding sites are shown at the bottom. Gray letters indicate sequences that can be mismatched. Thus, the standard we applied in this figure is less stringent compared with Figure 4. The critical enhancer region identified in the present study is indicated by the boxed area in red, and the mouse core enhancer (300 bp) previously identified by Haraguchi et al. (2001) is shown in black.

Download figure to PowerPoint

To test for the requirement of these binding motifs during mesp-b expression, we generated deletion reporters for each site by PCR and examined the subsequent effects upon enhancer activity in transgenic fish (Figs 4B and 5). Single deletion mutants of T1 (dT1), T2 (dT2), N1 (dN1) and N2 (dN2) were found to reduce the levels of EGFP reporter expression (Figs 5A–C,E,F). In addition, a double deletion of W1+W2 (dW1+2) also lowered GFP expression levels (Fig. 5I). In contrast, deletion of either the RARE (dR) or W3+W4 motifs (dW3+4) did not alter either the expression pattern or the reporter protein levels (Figs 4B, 5H,J and 6Q). Importantly, the GFP expression profiles for all of these deletion mutants still exhibited a segmental or anterior restricted pattern, even in cases where the expression levels were greatly reduced, suggesting that these motifs are mainly required for the activation and maintenance of mesp-b expression, but not for anterior restriction of mesp-b within a presumptive somite. To examine whether T1 and T2, or N1 and N2, function independently of each other, we generated double deletions for both sets of motifs (dT1+2 and dN1+2). The resulting mutant enhancers displayed a further reduction in GFP expression levels when compared with the corresponding single deletion constructs (Fig. 5D,G). This indicated that the T1/T2 and N1/N2 sites have cooperative roles in directing mesp-b gene expression in the PSM and somites. A longer exposure time also revealed that a stripe expression pattern was maintained in the dN1+2 transgenic embryos (see Fig. 6K′).

image

Figure 5. The activity of both the wild-type and 2.8 kb mutant mesp-b enhancers. Transmitted light (A–J), or fluorescent (A′–J′) images of transgenic medaka embryos at stage 22 are shown. The reporter construct used is shown on the left. For details, see the main body of the text. Bar, 0.05 mm.

Download figure to PowerPoint

image

Figure 6. The involvement of Tbx24 (A–F), Notch signaling (G–L), and retinoic acid (RA) signaling (M–R) in endogenous medaka mesp-b expression and in transgene expression driven by both wild-type and mutant 2.8 kb mesp-b enhancer fragments. Dorsal views of the medaka embryos at stage 22 are shown with the anterior towards the top. In the case of the transgenic animals, the enhancer that drives EGFP reporter expression is indicated at the top right corner. (A–F) The effect of tbx24 gene knockdown. Control embryos injected with control morpholino oligonucleotide (MO) (A,C,E) and embryos injected with MO-tbx24 (B,D,F) are shown. Injection of MO-tbx24 abolishes both endogenous mesp-b expression and transgene expression. Morphologically, MO-tbx24 also completely inhibits somite segmentation (D,F). (G–L) The effects of N-(N-(3,5-difluorophenacetyl-L-alanyl))-S-phenylglycine t-butyl ester (DAPT; an inhibitor of Notch signaling) treatment upon mesp-b expression. Embryos treated with dimethylsulfoxide (DMSO; control; G,I,K) and with DAPT (H,J,L) are shown. DAPT treatment reduces the expression levels of both endogenous and transgene expression of mesp-b. Note that DAPT treatment also causes the formation of a salt-and-pepper pattern for both endogenous and transgene expression, driven by both the 2.8 kb wild-type and dN1+2 mesp-b enhancer fragment lacking both RBPJκ binding sites. In contrast, GFP expression driven by the dN1+2 enhancer maintains a segmental pattern (compare K′ with L′). DAPT treatment disturbs somite boundary formation, particularly in the posterior trunk region (J,L). (J′,K′,L′) show enlarged images taken after longer exposures. (M–R) The effect of disulfiram (an inhibitor of RA synthesis) treatment on mesp-b expression in the medaka embryo. Embryos treated with DMSO (control; M,O,Q) and with disulfiram (N,P,R) are shown. Disulfiram treatment was found to reduce both endogenous and transgene mesp-b expression. Note that the dR enhancer lacking the RARE motif still responds to disulfiram treatment (R′). Disulfiram treatment also occasionally induces fusion of the somites (R) in addition to left–right asymmetric somitogenesis (P). 2.8 indicates the embryo carrying the 2.8 kb mesp-b enhancer construct, T1±2 indicates the embryo with dT1+2 enhancer which lack both the two T-box binding sites and N1+2 indicates the embryo carrying the dN1+2 enhancer which lacks both the two RBPJκ binding sites. Bars, (A,B,G,H,M,N) 0.1 mm; (C–F,I–L,O–R) 0.1 mm; (J′,K′,L′) 0.05 mm.

Download figure to PowerPoint

Taken together, these data demonstrate that the putative T-box and RBPJκ binding-like sites, and some of Tcf-binding sites in the 2.8 kb upstream region of the mesp-b gene are involved in the activation of mesp-b, but that the RARE motif does not play an essential role in this upregulation.

The Tbx24, Delta/Notch and retinoic acid signaling pathways regulate mesp-b expression via its 2.8 kb enhancer region

We next examined how both the wild type or mutant 2.8 kb mesp-b enhancers responded to alterations in the levels of key signaling molecules such as Tbx24, Notch and RA. We first tested whether the endogenous medaka mesp-b responded in a similar way to its zebrafish and Xenopus counterparts, and, as expected, the morpholino knockdown of tbx24 (MO-tbx24) completely abolished endogenous mesp-b expression and inhibited morphological boundary formation (Fig. 6A,B,D,F; n = 5). Similarly, treatment of developing medaka embryos with DAPT, an inhibitor of Notch signaling (Dovey et al. 2001), resulted in a salt-and-pepper mesp-b expression pattern with reduced expression levels (Fig. 6G,H; n = 9), whereas treatment with disulfiram, an inhibitor to RA synthesis (Vermot et al. 2005), gave rise to a weak and diffuse pattern of mesp-b expression (Fig. 6M,N; n = 12). In both cases, morphological somite formation was also disrupted; DAPT treatment resulted in defects in boundary formation of the posterior somites (Fig. 6J,L), as observed in zebrafish (Geling et al. 2002), whereas disulfiram treatment occasionally caused a disturbance in boundary formation and left–right asymmetric somitogenesis. The latter phenotype has also been reported in chick embryos treated with disulfiram (Vermot et al. 2005).

Similar to endogenous mesp-b expression, GFP reporter expression in the 2.8 kb transgenic medaka embryos was almost completely lost following injection by MO-tbx24 (Fig. 6C,D; n = 3). MO injection, however, did not affect the low levels of GFP expression in the dT1+2 transgenic fish (Fig. 6E,F; n = 5). In a similar fashion, the perturbation of Notch signaling by DAPT treatment caused a reduction in GFP reporter expression, driven by the 2.8 kb enhancer (Fig. 6I′,J′; n = 6 for DAPT and 11 for control). Importantly, the 2.8 kb transgenic embryos in this experiment exhibited a salt-and-pepper GFP expression pattern, although the overall expression levels remained low (Fig. 6J′). A similar salt-and-pepper pattern was detectable in medaka embryos carrying the dN1+2 enhancer, and the initial low levels of transgene expression were maintained (Fig. 6K′,L′). We next treated transgenic embryos with disulfiram and unexpectedly found that the inhibition of RA signaling reduced the activity of both the 2.8 kb (Fig. 6O,P; n = 5 for disulfiram and 4 for control) and dR enhancers (Fig. 6Q,R; n = 12 for disulfiram and 8 for control).

Taken together, these results strongly suggest that the Tbx24 transcription factor and the Delta/Notch signaling pathway participate in medaka mesp-b transcription via the T1/T2 and N1/N2 sites on the upstream mesp-b enhancer, respectively. In contrast, our current findings show that RA signaling regulates mesp-b transcription via an unidentified element that is distinct from the canonical RARE sequence in this enhancer region. In addition, the N1 and N2 motifs were found to be dispensable for the Notch-dependent anterior localization of mesp-b transcription.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The Mesp2-like bHLH transcription factors are key determinants in the establishment of both the segmental boundaries and the antero-posterior polarity within somites. Thus, the cells of the PSM need to express these genes at the correct time and location prior to the formation of a segmental unit (Saga et al. 1997; Sparrow et al. 1998; Sawada et al. 2000). Several signaling pathways, including Notch, FGF, RA and Wnts, and the T-box transcription factors are likely to be critical for the transcriptional regulation of the Mesp2-like genes during segmentation (Aulehla et al. 2003; Jen et al. 1999; Takahashi et al. 2000). In the present study, we have isolated the medaka mesp genes and identified the PSM enhancer element of mesp-b using stable transgenic medaka embryos. The PSM enhancer contains several putative binding sites for key signaling molecules, and the roles of these sites during mesp-b upregulation were tested in vivo.

The cloning and genomic organization of the medaka mesp genes

We first isolated two medaka mesp genes and characterized their expression patterns in the developing embryo. Their genomic organization was also analyzed using the medaka BAC clone sequence (180H12, an approximately 155 kb fragment mapped on Medaka LG3). Mesp and its related genes have been isolated in mouse (MesP1 and 2; Saga et al. 1996, 1997), chick (cMeso-1 and 2; Buchberger et al. 1998, 2002), zebrafish (mesp-a and b; Sawada et al. 2000) and Xenopus (Thylacine1, Sparrow et al. 1998; Mespo, Joseph & Cassetta 1999). Their bHLH domains are highly homologous among each other and form the Mesp gene family (Fig. 1B). The Mesp phylogenetic tree (Fig. 1D) suggests that independent duplication of the mesp gene occurred in each vertebrate during evolution. This is further supported by the differences in genomic organization that can be observed, at least between the fish and tetrapod lineages. The medaka mesp-a and mesp-b genes are located head-to-tail in the same BAC clone and are separated by a 6 kb intergenic sequence (Fig. 1A). Similarly, zebrafish mesp-a and mesp-b and the fugu Mesp-related genes are aligned head-to-tail, 14 kb and 5 kb apart, respectively (on LG7 in zebrafish and on scaffold 425 of JGI in fugu). In contrast, mouse Mesp1 and Mesp2 localize to chromosome 7 and exhibit a head-to-head orientation with a 23 kb intergenic sequence (Saga et al. 1997). This is also true for human and chick Mesp-related genes, which are oriented head-to-head with 15 kb and 3 kb intervals, respectively (human in chromosome 15; chick, Buchberger et al. 2002).

In spite of these differences in genomic arrangements, the medaka mesp region is syntenic to that of mouse Mesp because genes around the mesp genes are conserved between mouse and medaka, such as Trk-C and IGF1R (data not shown). Furthermore, the expression patterns of the vertebrate Mesp gene families are nearly identical. One of the two mesp genes always expresses in both the early mesoderm and in the anterior PSM (Mesp1-type expression), whereas the other is activated in the anterior PSM (Mesp2-type) (Saga et al. 1996, 1997; Sawada et al. 2000; Fig. 2A–H). The identical expression patterns of the mesp genes among different vertebrate species suggested the presence of a conserved regulatory region, despite the differences in their gross genomic organization. However, comparisons of the upstream sequences of Mesp genes by PipMaker or VISTA software did not reveal any obvious regions of sequence conservation (data not shown). This raised the question as to whether the Mesp transcriptional regulatory mechanisms differ among species or whether the regulatory networks are in fact quite conserved, in spite of the relatively poor sequence conservation at the gene loci.

Characterization and deletion analysis of the presomitic mesoderm mesp-b enhancer

To address the above question, we localized the transcriptional regulatory region of medaka mesp-b within a 2.8 kb upstream sequence of the gene (Fig. 3E). This region almost recapitulates endogenous mesp-b expression, particularly the transcriptional initiation of the gene in the PSM and the anteriorly restricted expression in the forming somite (S0). However, the transcription by the 2.8 kb mesp-b enhancer is maintained in the segmented somites, which contrasts with the endogenous transcription of mesp-b; it is downregulated immediately after the formation of the somite. This may be attributable to the differences in the stability of the respective transcripts within somitic cells, or could be due to the lack of transcriptional silencing of the enhancer region identified in this study.

Further deletion analysis revealed that the mesp-b upstream region between −1.4 kb and −2.8 kb is likely to be critical for both the activation and patterning of mesp-b expression. This is in sharp contrast to mouse Mesp2, for which a 300 bp upstream sequence is sufficient to recapitulate endogenous expression in the PSM (Haraguchi et al. 2001). In the medaka mesp-b 2.8 kb upstream region, we found several putative binding motifs, including those for T-box transcription factors, RBPJκ, and Tcf. We also identified an RA response element (RARE). Among these mesp-b enhancer sites, the putative binding motifs for T-box transcription factors and RBPJκ were found to be crucial, whereas RARE was dispensable, for gene expression.

The Notch signaling pathway and T-box transcription factors directly regulate mesp-b expression through its 2.8 kb enhancer element

T-box transcription factors and Delta/Notch signaling pathways have been placed upstream of mesp-b transcriptional regulation in both mouse and zebrafish (reviewed in Saga & Takeda 2001; Holley & Takeda 2002; Nikaido et al. 2002). In the absence of the T-box related factor, Tbx24, the expression of zebrafish mesp-b is undetectable and essentially no somite boundary formation takes place. In the case of the medaka 2.8 kb mesp-b enhancer, two canonical T-box binding sites cooperatively function to direct the PSM-specific expression of the gene. Furthermore, the knockdown of Tbx24 abolishes reporter gene expression driven by the 2.8 kb enhancer element. These results strongly suggest that Tbx24 directly activates the transcription of mesp-b through T1 and T2 binding sites.

In Notch-defective mutants of mouse and zebrafish, the expression of the mesp genes becomes weak and adopts a broad salt-and-pepper pattern instead of a clear stripe pattern (zebrafish, Sawada et al. 2000; mouse; del Barco Barrantes et al. 1999; Takahashi et al. 2000). This indicates that Notch signaling is required for both the maintenance of high levels of the Mesp gene and for the anteriorly localized pattern of this expression. However, in transgenic medaka embryos carrying a dN1+2 mutant mesp-b enhancer, the stripe pattern of reporter gene expression is still maintained, although the overall expression levels are greatly reduced. Interestingly, the inhibition of Notch signaling with DAPT induces the salt-and-pepper pattern of expression in transgenic medaka embryos carrying the wild-type 2.8 kb as well as mutant dN1+2 mesp-b enhancers (Fig. 6G–L). These results suggest that Notch signaling regulates the levels of mesp-b expression in medaka through the RBPJκ binding sites in the gene enhancer but restricts mesp expression to the anterior compartment through yet unidentified sites also present in the 2.8 kb upstream region. Indeed, there are several less-conserved putative RBPJκ binding sites within this region and it is possible that the Notch pathway utilizes these motifs to facilitate the stripe pattern of mesp expression. Alternatively, the Notch pathway can be mediated independently of RBPJκ (Matsuno et al. 1997; Hori et al. 2004).

The effects of retinoic acid upon mesp-b expression may be indirect

The involvement of RA in the regulation of mesp genes was first reported in Xenopus but this regulatory mechanism has been somewhat controversial. RA signaling has been shown to upregulate the expression of the Xenopus mesp-related gene, Thy1 (Moreno et al. 2004), but this may not be the case for mouse Mesp2 (Morimoto et al. 2005). In our present study, we demonstrate that medaka mesp-b is positively regulated by RA signaling as the inhibition of RA synthesis reduces the levels of mesp-b expression. This suggests that the fish mesp genes are amphibian-like in terms of their RA-dependent gene regulation. Furthermore, the 2.8 kb mesp-b enhancer element is responsive to the effects of disulfiram directly or indirectly, but this is true also for the RARE mutant (dR). This implicates an indirect involvement of RA in the regulation of the mesp genes.

Other candidate factors involved in mesp-b regulation

In the current study, we did not focus on Wnt signaling because we have no genetic evidence supporting the involvement of Wnt in the regulation of fish Mesp genes. However, deletion of some of the Tcf binding sites disrupted the mesp-b 2.8 kb enhancer activity in medaka embryos. At present, we do not know any candidate Wnt ligand which is expressed in the appropriate region and may regulate this enhancer activity. The role of FGF signaling, which is another important factor that regulates the precise patterning of somites, should also be examined in this enhancer region. In the region between −2.8 and −1.4 kb upstream of medaka mesp-b, we found several motifs that can mediate Fgf signaling comprising eight E-boxes (Blackwell & Weintraub 1990), two N-boxes (Sasai et al. 1992) and 11 Ets binding sites (Hsu et al. 2004). Further analysis of the roles of these motifs in mesp-b regulation is currently underway.

Comparison of the mechanisms underlying vertebrate mesp gene regulation

As described above, Haraguchi et al. (2001) have previously identified the 300 bp fragment immediately upstream of mouse Mesp2 as a core PSM enhancer. Recent analysis further revealed that Tbx6 directly binds to T-box binding sites present in this region, which mediates Notch signaling and subsequent Mesp2 transcription (Yasuhiko et al. 2006). Indeed, a comparison of the upstream regions of medaka, fugu, zebrafish and mouse Mesp2-like genes reveals that T-box binding sites are conserved in the immediate upstream regions (Fig. 7). However, in our present study, the 1.4 kb mesp-b upstream fragment that encompasses these T-box motifs fails to recapitulate endogenous mesp-b expression. Instead, a more distant region, encompassing the sequence between −2.8 to −1.4 kb, was found to play a role in the control of medaka mesp-b. In spite of these differences, the medaka 2.8 kb mesp-b enhancer seems to adopt a similar regulatory logic to that of the mouse, as the activity of the medaka enhancer depends on the putative binding sites for both T-box factors and RBPJκ. However, those binding sites are sparsely distributed along the medaka mesp-b 2.8 kb upstream region, which prompts us to speculate that, during vertebrate evolution, such essential motifs increased in number and were widely distributed in fish, while those in mouse remained tightly clustered. Hence, the number of crucial mesp-b enhancer motifs, their arrangement and the distances that separate them, all vary considerably among vertebrate species. Similar findings have been reported for several genes including ascidian otx2 (Oda-Ishii et al. 2005). From an evolutionary point of view, it is noteworthy that the ascidian Mesp gene is directly activated by Tbx6c through its immediate upstream enhancer (Davidson et al. 2005).

In summary, our current study identifies the crucial element that drives the expression of medaka mesp-b in the PSM. Furthermore, Tbx24, Notch and RA signaling are major regulators of the activity of this enhancer and additional experiments, such as the gel-shift assays for Tbx24 and RBPJκ binding, should provide direct evidence for the roles of these signaling pathways in medaka mesp-b expression. Finally, it is still not clear why the expression of the mesp genes is confined to the anterior PSM as the activation of Notch signaling and the expression of T-box genes have been observed in the posterior to intermediate PSM (Fig. 2; Morimoto et al. 2005). Fgf signaling might have a role in this restriction (Delfini et al. 2005), but further experiments will be needed to address this question.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank Dr Kiyoshi Naruse, Dr Kyo Yamasu and Dr Kenta Sumiyama for helpful information and suggestions, Ms Kazuko Oishi and Ms Keiko Nogata for generating the shotgun library and for sequencing analysis, and Ms Yasuko Ozawa for maintenance of the fish stocks. We also thank Dr Atsuko Shimada for providing the hatching enzyme and for helpful discussions on medaka embryology. This work was supported in part by Grants-in-Aid for Scientific Research Priority Area Genome Science and Organized Research Combination System from the Ministry of Education, Culture, Sports, Science and Technology of Japan and by Research Fellowships from the Japan Society for the Promotion of Science for Young Scientists.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References