Dev Genes Evol (1996) 206:54–63

© Springer-Verlag 1996

O R I G I NA L A RT I C L E

&roles:Isato Araki · Noriyuki Satoh

cis-Regulatory elements conserved in the proximal promoter region of an ascidian embryonic muscle myosin heavy-chain gene

&misc:Received: 14 October 1995 / Accepted in revised form: 7 December 1995

&p.1:Abstract The B-line muscle cells of the ascidian embryo are specified autonomously depending on determinants prelocalized in the myoplasm of unfertilized eggs. Expression of muscle-specific actin and myosin heavychain genes commences in the B-line presumptive muscle cells as early as the 32-cell stage. To explore the intrinsic genetic program for this differentiation, we analysed cis-regulatory elements of the Halocynthia roretzi muscle myosin heavy-chain gene (HrMHC1). Comparison of the entire amino acid sequence of HrMHC1 with those of other invertebrates and vertebrates indicated that HrMHC1 resembles myosin heavy-chain of vertebrate skeletal and cardiac muscles. A fusion gene was constructed consisting of 132 bp upstream the 5′-end of HrMHC1 gene fused to a bacterial lacZ reporter. When the fusion gene was microinjected into fertilized eggs, the reporter gene was eventually expressed only in muscle cells of tailbud embryos. It has been reported that 103 bp of sequence 5′ of the transcription start site of the ascidian embryonic muscle actin gene (HrMA4) contains information sufficient for muscle-specific expression (Hikosaka et al. 1994). Comparison of the 132 bp of sequence 5′ of the HrMHC1 gene with the 103 bp of sequence 5′ of the HrMA4 gene revealed several common motifs shared by the two genes (E-box, GATA box and Boxes A, B, T1 and T2). Point mutations inserted into these motifs suggested that the Box T1/T2 (TTTTTTCTTTCA) is critical for the promoter activity of the HrMHC1 gene.

Originally submitted to Roux’s Archives of Developmental Biology and accepted by Klaus Sander I. Araki1 · N. Satoh (✉) Department of Zoology, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-01, Japan Present address: 1 Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, 240 Longwood Avenue, Boston, MA 02115, USA&/fn-block:

&kwd:Key words cis-Elements · Myosin heavy-chain gene · Muscle-specific expression · Ascidian embryos&bdy:

Introduction In 1887, Chabry reported that destruction or removal of blastomeres from early ascidian embryos led to defective animals which were unable to compensate for the missing part. Following his report, Conklin (1905a, b) described that a certain cytoplasm of the ascidian egg, later termed the myoplasm, is segregated into blastomeres of the muscle lineage. Since those early studies, the ascidian egg has offered an experimental system to explore cellular and molecular mechanisms underlying so-called mosaic development (reviewed by Satoh 1987, 1994). In particular, muscle cell differentiation in ascidian embryos has been studied intensively (reviewed by Satoh et al. 1990; Satoh 1994). Muscle cell differentiation in ascidian embryos in an ideal experimental systems to explore the cascade of molecular mechanisms which are necessary for cellular differentiation during embryogenesis; from localization of the muscle determinants in the myoplasm to activation of muscle-specific gene expression. During embryogenesis in Halocynthia roretzi, 21 unicellular and striated muscle cells are formed on each side of the tail of the larva. Lineage analysis has revealed that each B4.1 blastomere of the bilaterally symmetrical 8-cell embryo gives rise to 14 muscle cells of the anterior and middle part of the tail, each A4.1 blastomere gives rise to 2 muscle cells of the posterior part of the tail, and each b4.2 blastomere gives rise to 5 muscle cells of the tail tip (Nishida 1987). The a4.2 cell is not engaged in the formation of larval muscle cells. Descriptive and experimental studies indicate that the B-line (primary lineage) muscle cells can develop autonomously depending on so-called muscle determinants in the myoplasm of the egg (Conklin 1905a, b; Whittaker 1973, 1982; Deno and Satoh 1984; Nishida 1992; Marikawa et al. 1994). Careful examination by in situ hybridization of whole-mount specimens revealed that the transcription of H. roretzi muscle actin

55

and myosin heavy-chain genes commence in B6.2 blastomeres of the 32-cell stage embryo (Satou et al. 1995), prior to the developmental fate restriction of muscle lineage that occurs at the 44-cell stage (Nishida 1987). One of the approaches to elucidate the molecular mechanisms of muscle cell differentiation in ascidian embryos is first to clarify cis-regulatory elements of muscle-specific genes, then to identify transcriptional factors that are responsible for muscle-specific gene expression, and finally to elucidate the transcriptional circuit. We have already analysed the 5′ flanking region of an ascidian embryonic muscle actin gene (HrMA4) and have found that a 103-bp upstream region of the gene is sufficient for muscle-specific expression of a reporter gene (Hikosaka et al. 1993, 1994). This region contain an E-box motif at −71 (Kusakabe et al. 1992). However, point mutations inserted into the E-box did not affect the muscle-specific expression of the reporter gene (Hikosaka et al. 1994; Satou et al., unpublished). In this study, we analyse the cis-regulatory elements of the 5′ upstream region of the gene for another major muscle protein, myosin heavy-chain (MHC). We obtained cDNA clones which cover nearly the entire length of the transcript. Sequence analysis indicated that the H. roretzi embryonic myosin heavy-chain gene (HrMHC1) encodes a protein that most resembles MHC of vertebrate skeletal muscle and cardiac muscle. When a fusion construct containing 132 bp of 5′ upstream sequence of HrMHC1 fused to bacterial lacZ was injected into fertilized eggs, the reporter gene was expressed specifically in muscle cells of tailbud embryos. Because the expression pattern of HrMHC1 is the same as that of an embryonic muscle actin gene HrMA4 (Satou et al. 1995), we compared sequences within the 132-bp long 5′ upstream region of the HrMHC1 gene with that of the 103-bp upstream region of the HrMA4 gene. This analysis indicated the presence of several common sequence motifs shared by these two muscle-specific genes. Finally, we made point mutations into the shared motifs of HrMHC1 to examine whether these motifs have a significant role in the specific expression of the gene.

Materials and methods Animals and embryos Halocynthia roretzi was purchased during the spawning season from fishermen in the vicinity of Otsuchi Marine Research Centre, Ocean Research Institute, University of Tokyo, Iwate and Asamushi Marine Biological Station, Tohoku University, Aomori, Japan. H. roretzi is a self-sterile hermaphrodite. Naturally spawned eggs were fertilized with a suspension of sperm from another individual. Fertilized eggs were raised at 11–13 °C. Tadpole larvae hatched at about 40 h after fertilization. Screening of cDNA and genomic libraries and nucleotide sequencing A cDNA library of H. roretzi tailbud embryos was constructed in λgt11 (Stratagene, La Jolla, Calif.; Makabe and Satoh 1989), while an H. roretzi genomic library was constructed in λFIX II

(Stratagene; Kusakabe et al. 1992). Screening of the libraries was performed according to the standard procedure (Sambrook et al. 1989). Sequences of clones were determined by the dideoxy chain termination procedure (Sanger et al. 1977) with Sequenase ver. 2.0 (United States Biochem. Co., Cleveland, Ohio). Construction of phylogenetic trees The entire amino acid sequences of myosin heavy-chain proteins from various sources shown in Fig. 3 were aligned by the Clustal V computer-program (Higgins et al. 1992). The alignment was manually corrected. The molecular phylogenetic tree was constructed by neighbour-joining (Saitou and Nei 1987). The possibility at each branch was examined by bootstrapping, the percentage of times a node was supported in 1,000 bootstrap pseudoreplications by the neighbour-joining method (Felsenstein 1985). Fusion gene constructs First, PCR (polymerase chain reaction) was performed to amplify a part of 5′ flanking region of HrMHC1 with genomic clone c403 as a template. The amplified fragments were subcloned into the multicloning site of plasmid p46.21, a version of pPD1.27, which lacks C. elegans sup-7 gene (Fire et al. 1990). The p46.21 encodes a bacterial gene for β-galactosidase (lacZ) with a nuclear localization signal in the multicloning site and was kindly provided by Dr. A. Fire (Carnegie Institution of Washington). The PCR conditions were as follows. Fifty picomole of each primer described below was added to the reaction mixture for PCR [50 µl final volume; sample overlaid with one drop of mineral oil (Aldrich Chemical Co., Milwaukee, Wis.)] which contained 0.5 µg of c403 cloned in pBluescript II SK(+) (Stratagene), 1× PCR buffer [10× PCR buffer (Toyobo Biochem., Osaka, Japan): 500 mM KCl, 100 mΜ TRISHCl, pH 9.0 at 25 °C, 1% Triton X-100], 1.5 mM MgCl2, 0.2 mM each of dNTP (deoxyribonucleoside triphosphate) and 1 unit of Taq polymerase (Toyobo). Twenty-three cycles of the PCR were performed in an automated DNA thermal cycler (Perkin-Elmer Co., Norwalk, Colo.) under the following conditions: denaturation at 94 °C for 1 min, primer annealing at 50 °C for 2 min, and primer extention at 72 °C for 3 min. All amplified PCR fragments mentioned here were treated with T4 nucleotide kinase [New England Biolabs (NEB), Beverly, Mass.] and T4 DNA polymerase (NEB) before restriction enzyme-treatment. Oligonucleotides used in the procedure were as follows: primer L, 5′-GTGGATCCTGCACCTTGGAGTGGTA-3′; primer J, 5′-ATCGAATCAGCAGTTGC-3′; primer N, 5′-CACACAGTGATAGGTGA-3′; primer X, 5′-GTCCCGGGACAAATCGAGGGTTGCC-3′; primer cA, 5′-CTTCTCTGTTTTTCATT-3′; primer mA, 5′-GTACGTGCGAATCAGCAGTTGCGC-3′; primer cE, 5′-CTGATTCGATTCGTACT-3′; primer mE, 5′-AGGTTGCGCTTTTTTCTTTC-3′; primer cB, 5′-CTGCTGATTCGATTCGT-3′; primer mB, 5′-AGTACATTTTTTCTTTCACACAG-3′; primer cT, 5′-GCGCAACTGCTGATTCG-3′; primer mT, 5′-AGCATAGGTACACACAGTGATAGGTGA-3′; primer cG, 5′-CACTGTGTGAAAGAAAA-3′; and primer mG, 5′-CGTCGTGATGAACGTTTAATA-3′. To make pHrMHC1(−132)-Z, primers L and X were used for the PCR amplification. The amplified fragments were digested with BamHI and XmaI, and ligated to BamHI/XmaI digested p46.21. pHrMHC1(−90)-Z and pHrMHC1(−61)-Z were made with primers J and N, and primers J and X, respectively. The fragments were digested with XmaI, then ligated to BamHI(blunt-ended)/XmaI digested p46.21. At the 5′ end of the amplified fragments for pHrMHC1(−90)-Z and pHrMHC1(−61)-Z, there was a G residue probably because of an artifact during the subcloning. To construct pHrMHC1(−132)-µA, two independent PCR amplifications were performed with primers L and cA, and primers mA and X, respectively. The fragments amplified with the former

56 Table 1 Mutated sequences in the conserved motifs (bold letters mutated nucleotides, hyphen in µG represents the deletion of the nucleotide in that position)&/tbl.c:&

Motifs

Wild-type sequences

Mutated sequences

Box A

TACGAAT

µA GTACGTG

E-box/Box B E-box (CANNTG) Box B

CAGTTGCGC CAGTTG TTGCGC

µE AGGTTG µB AGTACA

GATA site (WGATAR)

TGATAG

µG T-CGTC

Box T1/T2 (Box T) Box T1 Box T2

TTTTTTCTTTCA TTTTTTC TTTCTTTCA

µT AGCATAGGTACA

&/tbl.: primers were digested with BamHI, and those amplified with the latter primers were digested with XmaI. They were ligated to BamHI/XmaI digested p46.21. The same procedure was used for making other constructs bearing point mutations. Primers cE and mE, cB and mB, cT and mT, cG and mG were used to make pHrMHC1(−132)-µE, -µB, -µT, -µG, respectively (Table 1). The PCR fragments derived from primers L and cG lack the 3′ most G probably because of an artifact during the subcloning. The promoter regions were sequenced to confirm the entire nucleotide sequences. Microinjection of fusion gene constructs and histochemical detection of β-galactosidase (β-gal) activity Microinjection of fusion constructs and histochemical detection of β-gal activity were performed as described previously (Hikosaka et al. 1994; Kusakabe et al. 1995). Ascidian eggs are enclosed by a vitelline coat, or chorion. Fertilized eggs are treated with a solution [1% sodium thioglycorate, 0.05% actinase E (Kaken Pharmaceutical Co., Ltd., Tokyo), adjusted pH 10 with NaOH] for 2 min with pipetting to remove follicle cells. After this treatment, the eggs are easily fixed on coverglasses, which make microinjection of recombinant DNAs into the eggs easy. The microinjection was performed through the vitelline coat. Plasmid DNAs were linearized by digestion with PstI and dissolved in 1 mM TRIS-HCl, 0.1 mM EDTA, pH 8.0. Microinjection was carried out with injection pipettes held by micromanipulators (model MN-151; Narishige Scientific Instruments Lab., Tokyo) on a coverglass under a stereomicroscope. Micropipettes were made on a horizontal puller (model PG-1; Narishige) from 1.2-mm fibre-filled glass capillary tubing (Microcaps; Drummond Sci. Co., Broomall, Pa.) and were sterilized. The DNA solution (2.5 µl/ml) was injected into the cytoplasm of fertilized eggs under pressure. Injected eggs were incubated on plastic dishes coated with 1.5% agar in Millipore-filtered seawater containing 50 mg/l streptomycin sulfate (Meiji Seika Co. Ltd., Tokyo) until late tailbud stage. Injected embryos were fixed for 30 min at room temperature in 0.5 M NaCl, 27 mM KCl, 2 mM EDTA (pH 8.0) containing 1% glutaraldehyde. Fixed embryos were rinsed in phosphate-buffered saline (PBS) and incubated in PBS that contained 250 µM 5-bromo4-chloro-3-indolyl-β-D-galactopyranoside (X-gal), 0.1% Triton X100, 1 mM MgCl2, 3 mM K4[Fe(CN)6] and 3 mM K3[Fe(CN)6] at 37 °C for 30 min. The stained embryos were washed in PBS to stop the staining reaction and observed under an Olympus stereomicroscope.

Results Isolation and characterization of cDNA clones for an ascidian myosin heavy-chain gene HrMHC1 In a previous study, Makabe and Satoh (1989) isolated a cDNA clone, cM11, from an H. roretzi tailbud-embryo cDNA library. The cM11 corresponded to a central portion of the transcript for an ascidian myosin heavy-chain

Fig. 1 Restriction enzyme map of the cDNA fragment for H. roretzi myosin heavy-chain (MHC) gene. The four cDNA clones cover the entire length of the cDNA. cM11 was originally isolated by Makabe and Satoh (1989) with a specific antibody Mu-2, while the other three clones were isolated in this study. B, E, K, N, P, S and X correspond to restriction sites for BamHI, EcoRI, KpnI, NotI, PstI, SacI and XhoI. The head and tail junction of the predicted MHC protein is shown above the cDNA clone&ig.c:/f

gene (HrMHC1; Fig. 1). To obtain clones that contain sequences more 5′ (and also more 3′) of the cloned cDNA, we again screened the library with cM11 as probe. Among the positive clones, we selected two clones, c2601 and c902, to characterize their primary structures. However, as shown in Fig. 1, they did not contain the 5′ most region of the transcript. By further screening of the library with the 5′ most part of c2601 as the probe, we obtained c4111, which contained the 5′ most part of the cDNA (Fig. 1). In total, these cDNA clones encompassed about 6.0 kb, which covered almost the entire length of the cDNA for the HrMHC1 gene, judging from Northern analysis (Makabe and Satoh 1989). The complete amino acid sequences of HrMHC1 were deduced from the composite entire nucleotide sequences of the clones. As shown in Fig. 2, the clones encoded a polypeptide of 1,927 amino acids. The calculated relative molecular mass of the predicted protein was about 222 kDa. The full extent of amino acid identity was 69.7% between HrMHC1 and the mouse cardiac αMHC. The sequence analysis showed that the HrMHC1 is a member of the heavy chain of conventional myosin or myosin II (reviewed by Mooseker 1993). To examine the relationship of HrMHC1 with MHC proteins of other animal groups, we constructed a phylogenetic tree by use of the full amino acid sequences of various MHCs. As shown in Fig. 3, HrMHC1 formed a discrete group with vertebrate skeletal and cardiac

57

Fig. 2 Amino acid sequence (upper) of HrMHC1 deduced from nucleotide sequence of the cDNA clones and its comparison with that of mouse cardiac MHC (lower). Sequence identities are indicated by asterisks. For maximal similarity gaps were introduced. Protein ATP binding site, 25 kDa/50 kDa junction, 50 kDa/25 kDa junction, actin binding sites and MLC1 binding site are underlined. The head and tail junction and S-2 hinge regions are shown above the sequence. (The sequence of HrMHC1 will appear in the DDBJ, EMBL and GenBank Nucleotide Sequence Databases with the accession number D45163)&ig.c:/f

MHCs, suggesting that HrMHC1 is most similar to vertebrate muscle MHCs. The 132-bp 5′ flanking region of HrMHC1 contains the proximal cis-elements necessary for muscle-specific expression To examine the proximal cis-regulatory elements in the 5′ flanking region of the HrMHC1 gene, we screened 2.4×105 pfu of an H. roretzi genomic library with the 5′

58 Fig. 3 A molecular phylogenetic tree calculated by the neighbour-joining method (Saitou and Nei 1987) using the entire amino acid sequence of MHC proteins of various sources. Numbers at each branch indicate the percentage of times a node was supported in 1,000 bootstrap pseudoreplications by the neighbour-joining method. The tree suggests an affinity of the HrMHC1 with vertebrate skeletal and cardiac muscle MHCs&ig.c:/f

Table 2 Expression of the reporter gene in H. roretzi tailbud embryos that developed from eggs injected with various fusion gene constructs&/tbl.c:& Fusion gene construct

No. of expts.

1 Embryos scoreda

Region of expression 2 Muscleb

3 Muscle & ectopicc

4 Ectopicd

(2+3)/1

Ectopic region

17.4%

Epidermis pigment cell Epidermis

−132

3

144

23 (16.0%)

2 (1.4%)

2 (1.4%)

−90 −61

3 2

110 121

4 (3.6%) 1 (0.8%)

1 (0.9%) 0

0 0

4.5% 0.8%

µA µE

3 4

184 117

13 (7.1%) 11 (9.4%)

0 2 (1.7%)

0 0

7.1% 11.1%

µB µT µG

4 6 1

197 234 136

27 (13.7%) 1 (0.4%) 18 (13.2%)

1 (0.5%) 0 4 (2.9%)

0 0 5 (3.7%)

14.2% 0.4% 16.2%

a b c d

Epidermis endoderm Epidermis Epidermis pigment cell

We scored only normally developed embryos The number of embryos which expressed the transgene only in muscle The number of embryos which expressed the transgene in muscle and non-muscle tissues The number of embryos which expressed the transgene only in non-muscle tissues&/tbl.:

most KpnI fragment of c4111 as probe. Several positive clones were isolated. After restriction enzyme mapping and sequencing, it was found that one of the clones, c403 contained the 5′ flanking region of the gene. The spatio-temporal expression of HrMHC1 was almost identical to that of an ascidian embryonic muscle actin gene HrMA4 (Satou et al. 1995). In addition, it was shown that the 103-bp upstream region of HrMA4 is sufficient for appropriate expression of the reporter gene (Hikosaka et al. 1994). These results suggested that muscle-specific expression of HrMHC1 might also be controlled by a rather short upstream region of the gene. We, therefore, constructed pHrMHC1(−132)-Z, in which only the 242-bp upstream region of HrMHC1 (from −132 to +110 including first 19 bp of coding region) was fused with lacZ. We injected about 0.2 nl of DNA into an egg

about 30–90 min after insemination. Injection of DNA at a concentration of 2.5 µg/ml delivers about 105 copies of pHrMHC1-Z per egg. On average, about half of the eggs injected with fusion constructs cleaved normally and developed to tailbud embryos with normal morphology, although the success rate varied among batches. We scored the expression of the reporter gene only in manipulated tailbud embryos that exhibited normal morphology. No endogenous β-gal activity was detected in control noninjected embryos (data not shown). In about 15% of the tailbud embryos that developed from eggs injected with pHrMHC1-Z, β-gal activity was detected in the nucleus and cytoplasm of the tail muscle cells (Table 2; Figs. 4A and 5). No other structures, including endoderm, mesenchyme, and notochord, were stained, although a few cases of ectopic expression were

59

Fig. 4A–D Expression of lacZ in H. roretzi embryos that developed from eggs and embryos injected with pHrMHC1-lacZ fusion gene constructs. About 105 copies of pHrMHC1-lacZ were injected into: A a fertilized egg, B B4.1, C A4.1 and D b4.2 blastomeres of the 8-cell embryo. Injected embryos were fixed at the mid-tailbud stage and stained for β-galactosidase activity with X-gal. The β-gal activity was only observed in muscle cells (Mu) of the tail in the embryos (En endoderm, Epi epidermis, Nt notochord, bars 100 µm)&ig.c:/f

seen in epidermis and pigment cells. Cells positive for βgal usually formed clusters, suggesting that the injected plasmid DNAs were segregated into specific lineages as a mosaic. Sometimes, β-gal activity was found in many of the muscle cells of the embryo (Fig. 4A), while in other embryos reporter gene expression was restricted to a few muscle cells in certain regions. Next, we examined the basic promoter activity of pHrMHC1(−90)-Z and pHrMHC1(−61)-Z, in which 200 bp (from −90 to +110) and 171 bp (from −61 to +110) of the HrMHC1 upstream region was fused to lacZ, respective-

ly. Injection of pHrMHC1(−90)-Z resulted in muscle-specific expression of the reporter gene, although the frequency of positive embryos decreased to about 5% of the injected embryos (Table 2; Fig. 5). Injection of pHrMHC1(−61)-Z into fertilized eggs did not promote reporter gene expression (Table 2; Fig. 5), suggesting that the region between −132 and −61 upstream of the HrMHC1 initiation site is important for promoter activity. Among the 42 muscle cells of H. roretzi tailbud embryo, 28 are derived from the B4.1-pair of the 8-cell embryo, 4 from the A4.1-, and 10 from b4.2-pair. The a4.2pair is not involved in the formation of muscle cells. As described before, B-line presumptive muscle cells differentiate autonomously, while the A- and b-line presumptive muscle cells show less potential for autonomous differentiation (Nishida 1991). We examined reporter gene expression by injecting pHrMHC1(−132)-Z into each of the four blastomere-pairs of an 8-cell embryo. As shown in Table 3 and Fig. 4, reporter gene expression in muscle was observed when the construct was injected into either

60

Fig. 5 Diagram illustrating relative activities of the various promoter constructs in embryonic muscle cells. The promoter activity of pHrMHC (−132)-Z was regarded as 100%. µA, µE, µB, µT and µG are mutations inserted into shared motifs shown in Table 1 and Fig. 6. It is evident that mutations in Box T (µT) diminished the promoter activity drastically&ig.c:/f

B4.1 (Fig. 4B) or A4.1 (Fig. 4C) or b4.2 (Fig. 4D). No β-gal expression in muscle was apparent when pHrMHC1(−132)-Z was injected into a4.2 (Table 3). In addition, the percentage of β-gal positive embryos was 63.8, 17.9 and 25.0% when pHrMHC(−132)-Z was injected into B4.1, A4.1 and b4.2, respectively. Interestingly, percentage of β-gal positive embryos correlates with the percentage of muscle cells in the each sublineage (66.7, 9.5, 23.8%). This result suggests that there is no significant difference in the promoter activity of the 5′ upstream region of HrMHC1 between the various muscle sublineages.

The 5′ flanking regions of HrMHC1 and HrMA4 share several common sequence motifs Transcripts of both HrMHC1 and HrMA4 genes became detectable as early as the 32-cell stage in B6.2 cells (a progenitor of B-line muscle cells). About 130 bp of the 5′ flanking region of HrMHC1 (this study) and about 100 bp of the 5′ flanking region of HrMA4 (Hikosaka et Table 3 Expression of the reporter gene in H. roretzi tailbud embryos that developed from the 8-cell embryos injected with pHrMHC1(−132)-Z&/tbl.c:&

Blastomere injected

&/tbl.:

B4.1 A4.1 b4.2 a4.2

Embryos scored 58 39 28 26

al. 1994) is essential for their transcriptional activity, respectively. These results suggested to us that the proximal 5′ flanking regions of both muscle-specific structural genes share common sequence motifs that drive musclespecific expression. We therefore compared sequences of 5′ flanking regions of the HrMHC1 and HrMA4 genes. As shown in Fig. 6A, there were several conserved motifs in the 5′ proximal regions of these genes. An Ebox was found at −80 in HrMHC1 and at −71 in HrMA4. The E-box in the HrMHC1 gene is CAGTTG, which has been shown to be a preferential binding site for the MyoD-E2A complex (Blackwell and Weintraub 1990); in contrast, the E-box in the HrMA4 gene is different (CAACTG). A GATA binding site (WGATAR) is present at −54 in HrMHC1 and at −77 in HrMA4. This motif is recognized by GATA transcription factors and has been shown to be necessary for expression of some cardiac muscle-specific genes (Grépin et al. 1994; Ip et al. 1994; Molkentin et al. 1994). In addition to these known sequence motifs, the 5′ upstream regions of HrMHC1 and HrMA4 share at least four other sequence motifs; TACGAAT, TTGCGC, TTTTTC and TTTCTTTCA, designated Box A, B, T1 and T2, respectively (Fig. 6A). The motif alignment of Box A, the E-box and Box T1/T2 is conserved between both promoters (Fig. 6A). These motifs were also seen in the ascidian actin genes HrMA2, 5 and 6, although Box T2 in HrMA2 and Boxes B and T2 in HrMA6 are slightly modified and there is no E-box in HrMA6 (Fig. 6B). Box T1/T2 is essential for the transcriptional activity of HrMHC1 promoter We examined whether the above described motifs play significant roles in the transcriptional activity of the HrMHC1 promoter in the embryo. Point mutations were introduced into each of the motifs in pHrMHC1(−132)-Z (Table 1; Fig. 6A). Mutation of the GATA-binding motif (µG) did not affect the transcriptional activity (Table 2; Fig. 5). Expression of the reporter gene was scored at a level similar to that of pHrMHC1(−132)-Z. An introduction of point mutation in the E-box (µE) did not diminish the activity, nor affected the tissue specificity of the HrMHC1 promoter, although the activity decreased to some extent (Table 2; Fig. 5). A sequence alternation inserted into Box B (µB) resulted in a slight decrease in the activity. Since destruction of the Box B also affected the E-box sequence, this

Regions of expression

Ectopic region

Muscle

Muscle & ectopic

Ectopic

32 (55.2%) 7 (17.9%) 6 (21.4%) 0

5 (8.6%) 0 1 (3.6%) 0

0 2 (5.1%) 0 2 (7.7%)

Endoderm Trunk lateral cells Epidermis Epidermis pigment cell

61

Fig. 6 A Nucleotide sequences of the 5′ flanking regions of HrMHC1 and HrMA4 genes. Nucleotide positions are indicated relative to the start site of transcription (+1). TATA box, GATA-1 motif and E-box are shown by boxes. In addition, sequence motifs (Boxes A, B, T1 and T2) shared by the two genes (see also B) are shown by boxes. B Schematic representation of common sequence motifs shared by the 5′ flanking regions of HrMHC1 and five muscle actin genes, HrMA4a, 4b, 5, 2 and 6. Solid boxes indicate the TATA box. A, B, E, G, T1 and T2 indicate boxes A, B, E, G, T1 and T2, respectively. The T2 box motifs are modified in HrMA2 and HrMA6, while in HrMA6 the E box diminishes and Box B and T2 are slightly modified&ig.c:/f

decrease could be attributed to an alternation of the Ebox itself. A greater decrease in the transcriptional activity of the transgene was detected when a mutation was introduced into Box A (µA). Mutation of Box T1/T2 (µT) nearly completely inhibited transgene expression (Table 2; Fig. 5).

Discussion In this study, we determined the complete nucleotide sequence of cDNA clones encoding an ascidian myosin heavy-chain gene HrMHC1. Comparison of the complete amino acid sequence of HrMHC1 with that of vertebrate and invertebrate MHCs demonstrated that HrMHC1 most closely resembles MHC of vertebrate skeletal and cardiac muscles. The ascidian larval muscle is striated despite being unicellular (reviewed by Satoh 1994). Although alternative splicing has been reported in some muscle myosin heavy-chain genes (Nagai et al. 1989; George et al. 1989; Sindhwani et al. 1994), our analysis suggests that the H. roretzi myosin heavy-chain is not al-

ternatively spliced. Although we can not rule out the presence of other muscle myosin heavy-chain in ascidian embryonic muscle, HrMHC1 is the most prominent muscle myosin heavy-chain in the muscle. During H. roretzi embryogenesis, transcripts of both HrMA4 and HrMHC1 become detectable in the B6.2 blastomeres of the 32-cell stage embryo (Satou et al. 1995). This coordinated temporal and spatial expression of HrMA4 and HrMHC1 suggests that their expression is regulated by common cis-regulatory elements shared by both genes. Because HrMA4 contains the basic regulatory elements immediately upstream of the transcription initiation site (Hikosaka et al. 1994), it seemed probable to us that HrMHC1 might also contain muscle-specific regulatory sequences near the transcription initiation site. As shown in Figs. 4 and 5, a 132-bp 5′ flanking region of HrMHC1 is capable of promoting muscle-specific expression of lacZ fusion constructs. However, the frequency of embryos exhibiting the reporter gene expression was not always high (nearly 17% of injected embryos) compared to that of HrMA4 (nearly 61%; Hikosaka et al. 1994). Although this lower percentage of expression of the MHC transgene could be due to injection of smaller amount of DNA in this study and the difference of the vector, it also raises the possibility that there may be additional enhancers in the MHC gene. An analysis of cisregulatory elements of four C. elegans muscle myosin heavy-chain genes suggested that the promoter and the enhancer independently regulate the muscle-specific expression of the genes (Okkema et al. 1993). Therefore, it is important to examine whether HrMHC1 contains enhancer motifs that act independently of the promoter shown in this study.

62

A comparison of nucleotide sequence of the 132-bp 5′ flanking region of HrMHC1 with that of the 103-bp 5′ flanking region of HrMA4 revealed that the two genes share several common motifs in the regions. The E-box is binding-site for myogenic bHLH proteins and is present in the enhancer/promoter region of several skeletal muscle-specific genes (reviewed by Lassar and Münsterberg 1994). The sequence of the HrMHC1 E-box is CAGTTG, which is identical to a preferentially binding sequence for MyoD-E2A heterodimer (Blackwell and Weintraub 1990). However, disruption of the E-box of HrMHC1 did not significantly affect the transcriptional activity, nor the tissue specificity of the promoter. Similar dispensability of the proximal E-box has also been found for HrMA4 (Hikosaka et al. 1994). Therefore, the E-box motif, and perhaps AMD1, a member of myogenic bHLH protein in ascidian (Araki et al. 1994), might not play a direct role in tissue-specific transcriptional activity of HrMHC1 and HrMA4 in embryonic muscle. However, we cannot exclude the possibility that they act indirectly on the promoters. Furthermore, other musclespecific transcription factors and cis-elements might compensate the function of the myogenic bHLH factors which might act directly on the promoters at the early stage of activation in normal embryos, because so far we do not know when these lacZ reporter transcripts appear. In addition, there is a possibility that the point mutations do not efficiently inhibit the binding of AMD1 protein to the E-box. The binding of GATA factors to a cognate binding site is apparently important for cardiac-specific expression of vertebrate cardiac muscle genes (Grépin et al. 1994; Ip et al. 1994; Molkentin et al. 1994). Both HrMHC1 and HrMA4 contain a GATA binding motif in the 5′ flanking region close to the transcription initiation site. Mutation of this site in HrMHC1 did not affect the transcriptional activity of the gene. However, it should be determined in future studies whether the mutation seriously affects the binding of the GATA factors to this site. Among four other conserved motifs (Boxes A, B, T1 and T2) shared by HrMHC1 and HrMA4, mutation of the Box T1/T2 nearly extinguished the transcriptional activity of this gene (Table 2; Fig. 5). Therefore, Box T1/T2 plays a crucial role in the muscle-specific transcriptional activity. Although the Box T1/T2 is T-rich (TTTTTTCTTTCA), the sequence is different from MEF2 site (YTWWAAATAR; Gosset et al. 1989; Black et al. 1995). Box T1/T2 motif is also present in the shared 5′ flanking region of HrMA1a/b muscle actin gene pair (Kusakabe et al. 1995), suggesting that this sequence motif is a common and important element for muscle-specific expression of ascidian muscle-specific genes. There is no report so far that the Box T motif has a pivotal role in the control of transcription of musclespecific genes in embryos of other animal groups. In contrast to vertebrate skeletal muscle, most of the ascidian embryonic muscle cells develop autonomously (Conklin 1905a, b; Whittaker 1973, 1982; Deno and Satoh 1984; Nishida 1992; Marikawa et al. 1994). The

question whether Box T and its binding factor(s) are specific to ascidians should be addressed in future. A great deal of evidence has established that there are cytoplasmic determinants in ascidian eggs which dictate embryonic muscle development (Conklin 1905a, b; Whittaker 1973; 1982; Deno and Satoh 1984; Nishida 1992; Marikawa et al. 1994). HrMHC1 and HrMA4 are activated at the end of the 32-cell stage (Satou et al. 1995). The analysis of trans-acting factors which bind to the cis-elements in promoters of these genes will shed light on whether the cytoplasmic muscle determinants directly activate these muscle genes, or activate them via an intervening cascade of regulators. &p.2:Acknowledgements We thank Dr. Andrew Fire of the Carnegie Institution of Washington for his generous gift of the plasmid 46.2, Dr. Andrew Lassar and Dr. Tom Schultheiss for critical reading of the manuscript, Dr. Takaharu Numakunai and all other staff members of Asamushi Marine Biological Station for their hospitality, Dr. Tatsuya Ueki for his help in the DNA motif analysis, and Hitoyoshi Yasuo for his technical advice. This study was supported by a Grant-in-Aid for Specially Promoted Research (No. 07102012) from the Ministry of Education, Science, Sports and Culture, Japan to N.S. I.A. was supported by a Predoctoral Fellowship from the JSPS (Japan Society for the Promotion of Science) for Japanese Junior Scientists with a research grant (no. 2327).

References Araki I, Saiga H, Makabe KW, Satoh N (1994) Expression of AMD1, a gene for a MyoD1-related factor in the ascidian Halocynthia roretzi. Roux’s Arch Dev Biol 203:320–327 Black BL, Martin JF, Olson EN (1995) The mouse MRF4 promoter trans-activated by muscle-specific transcription factors. J Biol Chem 270:2889–2892 Blackwell TK, Weintraub H (1990) Differences and similarities in DNA-binding preferences of MyoD and E2A protein complexes revealed by binding site selection. Science 250:1104–1110 Chabry L (1887) Contribution a l’embryologie normale et teratologique des Ascidies simples. J Anat Physiol (Paris) 23:167– 319 Conklin EG (1905a) The organization and cell lineage of the ascidian egg. J Acad Natl Sci Philadelphia 13:1–119 Conklin EG (1905b) Organ forming substances in the eggs of ascidians. Biol Bull 8:205–230 Deno T, Satoh N (1984) Studies on the cytoplasmic determinant for muscle cell differentiation in ascidian embryos: an attempt at transplantation of the myoplasm. Dev Growth Differ 26:43–48 Felsenstein J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–791 Fire A, Harrison SW, Dixon D (1990) A modular set of lacZ fusion vectors for studying gene expression in Caenorhabditis elegans. Gene 93:189–198 George EL, Ober MB, Emerson CP Jr (1989) Functional domains of the Drosophila muscle myosin heavy chain gene are encoded by alternatively spliced exons. Mol Cell Biol 9:2957–2974 Gossett LA, Kelvin DJ, Sternberg EA, Olson EN (1989) A new myocyte-specific enhancer-binding factor that recognizes a conserved element associated with multiple muscle-specific genes. Mol Cell Biol 9:5022–5033 Grépin C, Dagnino L, Robitaille L, Haberstroh L, Antakly T, Nemer M (1994) A hormone-enconding gene identifies a pathway for cardiac but not skeletal muscle gene transcription. Mol Cell Biol 14:3115–3129 Higgins DJ, Bleasby AJ, Fuchs R (1992) Clustal V: improbed software for multiple sequence alignment. Comput Appl Biosci 8:189–191

63 Hikosaka A, Satoh N, Makabe KW (1993) Regulated spatial expression of fusion gene constructs with the 5′ upstream region of Halocynthia roretzi muscle actin gene in Ciona savignyi embryos. Roux’s Arch Dev Biol 203:104–112 Hikosaka A, Kusakabe T, Satoh N (1994) Short upstream sequences associated with the muscle-specific expression of an actin gene in ascidian embryos. Dev Biol 166:763–769 Ip HS, Wilson DB, Heikinheimo M, Tang Z, Ting C-N, Simon MC, Leiden JM, Parmacek MS (1994) The GATA-4 transcription factor transactivates the cardiac muscle-specific troponin C promoter-enhancer in nonmuscle cells. Mol Cell Biol 14:7517–7526 Kusakabe T, Makabe KW, Satoh N (1992) Tunicate muscle actin genes. Structure and organization as a gene cluster. J Mol Biol 227:955–960 Kusakabe T, Hikosaka A, Satoh N (1995) Coexpression and promoter function in two muscle actin gene complexes of different structural organization in the ascidian Halocynthia roretzi. Dev Biol 169:461–472 Lassar AB, Münsterberg A (1994) Wiring diagrams: regulatory circuits and the control of skeletal myogenesis. Curr Opin Cell Biol 6:432–442 Makabe KW, Satoh N (1989) Temporal expression of myosin heavy chain gene during ascidian embryogenesis. Dev Growth Differ 31:71–77 Marikawa Y, Yoshida S, Satoh N (1994) Development of egg fragments of the ascidian Ciona savignyi: the cytoplasmic factors responsible for muscle differentiation are separated into a specific fragment. Dev Biol 162:134–142 Molkentin JD, Kalvakolanu DV, Markham BE (1994) Transcription factor GATA-4 regulates cardiac muscle-specific expression of the α-myosin heavy-chain gene. Mol Cell Biol 14:4947–4957 Mooseker M (1993) A multitude of myosins. Curr Biol 3:245–248 Nagai R, Kuro-o M, Babij P, Periasamy M (1989) Identification of two types of smooth muscle myosin heavy chain isoforms by cDNA cloning and immunoblot analysis. J Biol Chem 264: 9734–9737 Nishida H (1987) Cell lineage analysis in ascidian embryos by intracellular injection of a tracer enzyme. III. Up to the tissue restricted stage. Dev Biol 121:526–541

Nishida H (1991) Determinative mechanisms in secondary muscle lineages of ascidian embryos: development of muscle-specific features in isolated muscle progenitor cells. Development 108: 559–568 Nishida H (1992) Regionality of egg cytoplasm that promotes muscle differentiation in embryo of the ascidian: Halocynthia roretzi. Development 116:521–529 Okkema PG, Harrison SW, Plunger V, Aryana A, Fire A (1993) Sequence requirements for myosin gene expression and regulation in Caenorhabditis elegans. Genetics 135:385–404 Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4:406–425 Sambrook J, Fritsh EF, Maniatis T (1989) Molecular cloning: a laboratory manual, 2nd edn. Cold Spring Harbor Laboratory, New York Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74: 5463–5467 Satoh N (1987) Towards a molecular understanding of differentiation mechanisms in ascidian embryos. BioEssays 7:51–56 Satoh N (1994) Developmental biology of ascidians. Cambridge University Press, Cambridge Satoh N, Deno T, Nishida H, Nishikata T, Makabe KW (1990) Cellular and molecular mechanisms of muscle cell differentiation in ascidian embryos. Int Rev Cytol 122:221–258 Satou Y, Kusakabe T, Araki I, Satoh N (1995) Timing of initiation of muscle-specific gene expression in the ascidian embryo precedes that of developmental fate restriction in lineage cells. Dev Growth Differ 37:319–327 Sindhwani R, Ismail-Beigi F, Leinwand LA (1994) Post-transcriptional regulation of rat α cardiac myosin heavy chain gene expression. J Biol Chem 269:3272–3276 Whittaker JR (1973) Segregation during ascidian embryogenesis of egg cytoplasmic information for tissue-specific enzyme development. Proc Natl Acad Sci USA 70:2096–2100 Whittaker JR (1982) Muscle lineage cytoplasm can change the developmental expression in epidermal lineage cells of ascidian embryos. Dev Biol 93:463–470