Materials

Collection of gametes of A. crassispina and culture of embryos were performed as described before (Yamasu et al., 1995).

Hybridization analyses

Total RNA was isolated from adult tissues or embryos at different developmental stages by the acid guanidinium phenolchloroform method (Chomczynski and Sacchi, 1987). Southern and northern analyses were done according to the standard protocol (Sambrook et al., 1989) using as probe the 5′-terminal cDNA fragment (+1 to +270/BamHI site) labeled with [α-³²P]dCTP. Final wash was done with 0.2×SSC/0.1% SDS for 15 min at 65°C.

Whole mount in situ hybridization was performed as described previously (Onodera et al., 1999) using as probe the entire AcEFP cDNA.

Sequence analysis

Nucleotide sequences of cDNAs and deduced amino acid sequences were analyzed by Genetyx-Mac (Ver. 9.0; Software Development Co., Ltd.). The following amino acid sequences of EF-1α registered in GenBank were used for comparative analyses: Homo sapiens (EF-1α-1, X03558; EF-1α-2, X70940), Xenopus laevis (EF-1α-S, M25697; EF-1α-O, M75873; EF-1α-O1, Z19545; 42Sp50, X56699), Danio rerio (X77689), Drosophila melanogaster (EF-1α F1, X06869; EF-1α F2, X06870), Onchocerca volvulus (M64333), Arabidopsis thaliana (X16430), Tetrahymena pyriformis (D11083), Dictyostelium discoideum (X55972), Neurospora crassa (D45837), Saccharomyces cerevisiae (M10992). In addition, the sequence for a EF-1α-related protein (SU5 antigen) of another sea urchin Strongylocentrotus purpuratus deposited in the PIR data base (A37159) was also employed as well for comparison.

Screening of cDNA clones for the sea urchin EF-1α-related protein

A PCR fragment with high similarity to EF-1α cDNAs of other organisms was fortuitously obtained in our previous study searching for protein tyrosine kinase genes (Sakuma et al., 1997). A λZAPII cDNA library from sea urchin prism embryos (Yamasu et al., 1995) was screened using this fragment as probe. The longest cDNA clone obtained (pEF29) contains a fragment of 1754 bp and includes an ORF of 462 amino acids (aa) with an estimated size of 50.8 kDa (Fig. 1). Ten bp upstream to the initiation codon is seen an in-frame termination codon (data not shown), showing that the ORF codes for a full structure of the protein.

Fig. 1

Comparison of sea urchin AcEFP with similar proteins from other eukaryotes. The sequence of AcEFP of A. crassispina is shown at the top. Residues identical to those of sea urchin AcEFP are indicated by dots, and gaps introduced to optimize alignments are shown by dashes. Residues conserved in all the five sequences are shown with asterisks. The putative effector regions interacting with the ribosome (thin line) and actin-binding region (thick line) are indicated above the sequences. Residues proposed to interact with tRNA are marked with open circles. The 12-aa segments (12 aa) unique to animal and fungal EF-1α and the three putative protein kinase C phosphorylation sites are indicated with solid boxes. Asp (D) which has been shown to be covalently linked to phosphatidylinositol is marked with a solid arrowhead, and the three putative GTP binding domains are indicated with open rectangles above the sequences. The 2-aa insertion observed in animal EF-1α is marked with a horizontal line plus an open arrowhead. The two highly conserved sequences utilized for RT-PCR are shown with arrows. SU5 antigen, partial sequence of the EF-1α-related protein from Strongylocentrotus purpuratus; Homo, human EF-1α-1; Drosophila, Drosophila melanogaster EF-1α F1; Saccharomyces, Saccharomyces cerevisiae EF-1α.

Structural analysis of the sea urchin EF-1α cDNA

The deduced protein (AcEFP for EF-1α-related protein of A. crassispina) shows significant homology to EF-1α that has been reported so far (Table 1). Curiously, however, the similarities of the AcEFP protein to EF-1α of other animals (78–81%) are not clearly distinguishable from those to EF-1α of fungi (76–79%) in contrast to the higher similarities among typical EF-1α of other animals such as vertebrates, arthropods, and nematodes (Table 1; additional data not shown; see also Fig. 2). Actually, though vertebrates and echinoderms including sea urchins constitute one phylogenetic group (deuterostome), EF-1α of another group (protostome) show higher identities of more than 83% to vertebrate EF-1α. This raises the possibility that AcEFP might be a specialized protein related to, but different from, the authentic EF-1α.

Table 1

Comparison of eukaryotic EF-1α proteins^a

Fig. 2

Phylogenetic tree of EF-1α and related proteins. The tree was built by the neighbor-joining method (Saitou and Nei, 1987).

In its entirety, however, the AcEFP protein shows all the features of typical animal EF-1α (Fig. 1): 1) The 12-aa segment found in AcEFP from +214 to +225 with respect to the start codon is unique to animal and fungal EF-1α sequences (Nordnes et al., 1994). 2) The full complement of the three PKC phosphorylation sites, which is present only in those of animals and fungi (Nordnes et al., 1994), is also seen in the sea urchin protein (+242, +316, +383). 3) The 2-aa insertion in the sea urchin protein (+158 to +159) has been seen in animal EF-1α but not in fungal EF-1α. 4) The actin binding sequence in human EF-1α can also be seen in the sea urchin protein (Yang et al., 1990). 5) Basic amino acids considered necessary for binding to tRNA can also be found in the sea urchin protein (Berchtold et al., 1993). 6) The amino acid required for binding to the endoplasmic reticulum (Hayashi et al., 1989) also occurs in the sea urchin protein (D at +306).

Kuriyama et al. (1990) obtained monoclonal antibody (SU5) that recognizes a centrosphere protein of 50 kDa and cloned a partial cDNA for this antigen. The SU5 antigen also show significant similarity to EF-1α, though the antigen has several distinct features such as a one amino-acid insertion at a site that corresponds to between +103 and +104 of AcEFP and lack of the first PKC target site among the three (Fig. 1). Despite these structural uniqueness, the SU5 antigen is most close to AcEFP (Figs. 1, 2, Table 1), suggesting that the SU5 antigen and AcEFP diverged during the evolution of the sea urchin after the duplication of a common ancestor gene.

To test if other genes related to EF-1α are expressed in sea urchin embryos, we performed RT-PCR on cDNA from gastrulae employing degenerate primers designed for the two regions (GEFEAG, QDVYKI; Fig. 1) which are highly conserved among EF-1α of diverse arrays of organisms, and the PCR fragment of the expected size (ca.430 bp) was cloned into plasmid. Ten randomly selected clones showed essentially the same sequences as the AcEFP cDNA clone, demonstrating that AcEFP represents the main EF-1α family gene at least in gastrulae of A. crassispina.

In the genomic Southern analysis (Fig. 3A), only one or a few intense bands were seen for different restriction enzymes, showing that there are at most two genes for AcEFP in the genome.

Fig. 3

Southern and northern analyses. (A) Genomic Southern analysis of AcEFP. Genomic DNA from sea urchin sperm was digested with BamHI (B), DraII (D), and EcoRI (RI). Two μg digested DNA per lane were separated on a 1% agarose gel. (B) Developmental northern analysis. Five μg of total RNA per lane were run on a formaldehyde agarose gel. E, unfertilized eggs; 8c, 8-cell stage; 16c, 16-cell stage; HB, hatching blastula stage; MB, mesenchyme blastula stage; EG, early gastrula stage; MG, mid-gastrula stage; LG, late gastrula stage; EP, early pluteus stage, Pl, pluteus stage. (C) Expression of the AcEFP gene in adult tissues. One μg of total RNA per lane were run on formaldehyde gels. Pr, prism embryos; Om, mature ovary; Oim, immature ovary; Tim, immature testis; G, gut; Co, coelomocytes. Patterns of 18S rRNA stained with ethidium bromide are shown to the bottom. For coelomocytes, the expression was significantly low and detected only after an extended exposure. (D) Effects of ECM destabilizer on the expression of AcEFP. BAPN or cis-hydroxyproline of the indicated concentration was added to the culture at 1 hr postfertilization. After culture till 24 hr postfertilization, RNA was obtained from embryos and subjected to the northern analysis. Membranes were sequentially hybridized with AcEFP and AcSrc1 probe.

Expression of the AcEFP gene in embryos and adults

Zygotic expression of the AcEFP gene was first detected as a 2.2 kb band as early as in hatching blastulae and increased in amount until the early gastrula stage, when the expression leveled off and remained constant through the pluteus stage (Fig. 3B). Longer exposure revealed faint bands of the same size from the unfertilized egg through the cleavage stage (data not shown). The spatial regulation of the gene in the embryo was addressed by the whole-mount in situ hybridization (Fig. 4). From mesenchyme blastulae (data not shown) to early gastrulae (Fig. 4B), the expression was observed rather ubiquitously. In later stages such as the prism stage, the transcript was still widely observed, though relatively intense expression was detected in the gut and oral ectoderm. In keeping with this, we observed a slight increase in the amount of the transcript after treatment with lithium ions, which are known to respecify the prospective ectoderm to endoderm in sea urchin embryos (data not shown). In adult sea urchins, the transcript of AcEFP exist in all the tissues examined (Fig. 3C), with relatively high expression in immature testes and low expression in coelomocytes. Generally speaking, the expression levels in adult tissues are lower than that in prism embryos.

Fig. 4

Whole mount in situ hybridization analyses of AcEFP expression in sea urchin embryos. (A, C) An early gastrula (A, lateral view) or a prism embryo (C, frontal view) was stained using sense probes as negative controls. (B, D, E) An early gastrula (B, lateral view) or prism embryos (D, anal view; E, lateral view) were stained using antisense probes. A, animal pole; AE, aboral ectoderm; G, gut; OE, oral ectoderm; PMC, primary mesenchyme cells; V, vegetal pole.

ECM-independent expression of AcEFP gene during embryogenesis

To test if the AcEFP gene is regulated by external signals, embryos were treated with β-aminopropionitrile (BAPN) or cis-hydroxyproline, which destabilizes the collagen polymer, leading to disruption of the embryonic ECM (Wessel and McClay, 1987). Both treatments significantly reduced the expression of a gut-specific tyrosine kinase gene, AcSrc1 (Onodera et al., 1999), in a dose-dependent manner (Fig. 3D). In contrast, repression of AcEFP expression by hydroxyproline and BAPN was only limited. Thus, the expression of AcEFP does not depend strongly on the integrity of ECM in the embryo.

EF-1α-related protein of the sea urchin

In the present study we have cloned a cDNA for a sea urchin protein that is highly related to EF-1α. Comparison of this protein, AcEFP, with EF-1α of other organisms renders it doubtful that AcEFP is the sea urchin EF-1α. On the other hand, since the rate of base substitution is known to be rather high in the sea urchin genome (Hall et al., 1980), the possibility cannot be excluded that AcEFP is the sea urchin EF-1α. Actually AcEFP possesses all the features of EF-1α known till now, and RT-PCR identified only the mRNA for AcEFP in gastrulae, favoring the view that AcEFP is the genuine EF-1α. Though this issue awaits further study, it is highly likely that AcEFP is in some ways involved in protein synthesis.

Comparison with other EF-1α-related proteins from sea urchins

From eggs of different sea urchin species, Hemicentrotus pulcherrimus and Pseudocentrotus depressus, GTP-binding proteins of 51 kDa with EF-1α-like activities and antigenicity were isolated as mitotic apparatus-associated proteins (Ohta et al., 1990). This is not surprizing since EF-1α is now known to associate with the cytoskeleton (Shiina et al., 1994; Yang et al., 1990). It is possible that the 51-kDa protein corresponds to AcEFP of A. crassispina, though it should be addressed in the future. Meanwhile, Kuriyama et al. obtained monoclonal antibody (SU5) which recognizes a centrosphere protein of 50 kDa and cloned a partial cDNA for this antigen (Kuriyama et al., 1990). Despite its unique structural features, the SU5 antigen is most close to AcEFP (Figs. 1, 2, Table 1), suggesting that the duplication of the ancestor gene of AcEFP and SU5 antigen took place after the branching of the sea urchin or echinoderm from other phylogenetic groups in the deuterostome. Since SU5 antigen seems to have accumulated more amino acid substitutions than AcEFP, SU5 antigen might have acquired novel functions after gene duplication such as for the regulation of mitotic spindle formation in sea urchin cells. We failed to obtain cDNAs for any SU5-type protein from A. crassispina by the RT-PCR probably because it is not expressed abundantly at the gastrula stage or the primers used here are not suitable for amplification of cDNA for such highly deviated proteins. Though cloning of another partial cDNA clone for sea urchin EF-1α-related gene was reported previously, no sequence data was provided, making it impossible to be compared with AcEFP (Peeler et al., 1990).

In fact, it is well-known that a given single organism contains multiple EF-1α genes and related genes. In Xenopus, one of the four EF-1α-related proteins (42Sp50) is considered involved in long-term storage of 5S RNA and aminoacyltRNA (Deschamps et al., 1991; Viel et al., 1991). It is possible, therefore, that the sea urchin also possesses multiple EF-1α family proteins with different roles. Possiblity of the presence of multiple genes closely related to AcEFP was, however, excluded by the result of the genomic Southern analysis (Fig. 3A). In adult sea urchins, the transcript of AcEFP exists in all the tissues examined (Fig. 3C), with relatively high expression in immature testes and low expression in coelomocytes. In embryos, AcEFP is expressed also ubiquitously, with some bias to the gut and oral ectoderm. Its universal expression is in line with a view that AcEFP is involved in various aspects of protein synthesis per se or its regulation. The bias of the expression to the gut and oral ectoderm is known for SpS24, a small ribosomal subunit protein in the sea urchin S. purpuratus, which is also involved in protein synthesis (Angerer et al., 1992). Both tissues are highly proliferative in sea urchin embryos (Angerer et al., 1992 and references therein), and may require active translation for their growth. Among sea urchin genes expressed during embryogenesis, some are known to depend on ECM for its full expression (Spec1, LpS1) while others (SM50, LyUSF) do not (Benson et al., 1991; George et al., 1996). Our present result suggested that the expression of AcEFP is partially independent of the external signal. In fact, we observed that AcEFP expression still persisted, though at a reduced level, when embryos were dissoicated into cells and cultured succeedingly (unpublished data).

Translational control in sea urchin embryos

Translational control plays important roles during sea urchin embryogenesis, especially in the enhancement of protein synthesis after fertilization. Total rate of protein synthesis increases more than 100-fold (Regier and Kafatos, 1977), and enhancement of the rate of translational elongation was estimated to be 2.5-fold (Hille and Albers, 1979). Since AcEFP belongs to the EF-1α family, it will be involved in the protein synthesis itself or its regulation. Thus, cloning of the AcEFP cDNA will certainly contribute to the elucidation of this longstanding issue of the sea urchin development.