Genomic sequences encoding two types of medaka hemopexin-like protein Wap65, and their gene expression profiles in embryos

Warm temperature acclimation-related 65 kDa protein, Wap65, was first identified as an abundant cytosolic protein in eurythermal fish such as carp Cyprinus carpio and goldfish Carassius auratus acclimated to 30°C (Kikuchi et al.,1993; Watabe et al.,1993). The highest accumulated levels of Wap65 mRNA have been observed in hepatopancreas and their deduced amino acid sequences resemble those of mammalian hemopexins (Kikuchi et al., 1995, 1997; Kinoshita et al., 2001). Hemopexin is a mammalian plasma glycoprotein that is mainly synthesized in liver, functioning as a scavenger of free heme with a high binding affinity to heme (Altruda et al., 1985; Nikkila et al., 1991; Morgan et al., 1993; Tolosano and Altruda, 2002). Crystal structural analysis of rabbit hemopexin revealed two essential histidine residues that are located at heme binding pockets(Paoli et al., 1999). Both carp and goldfish Wap65 lack one of the two essential histidine residues corresponding to heme axial ligands in hemopexins but the binding of Wap65 to heme had not been determined. Our recent study revealed the occurrence of two types of Wap65, mWap65-1 and mWap65-2 in medaka Oryzias latipes(Hirayama et al., 2004), and fWap65-1 and fWap65-2 in the pufferfish Takifugu rubripes (previously known as Fugu rubripes; Hirayama et al., 2003). Two types of Wap65-related protein also appear in the GenBank database for rainbow trout Oncorhynchus mykiss (Z68112 and AF281339), zebrafish Danio rerio (AI588537 and BM09558) and channel catfish Ictalurus punctatus (BM438553 and BM438613),suggesting that the duplication of the ancestral gene happened after teleosts and mammals had diverged. Although the two types of Wap65 from medaka and Fugu had primary structures similar to those of carp and goldfish,the expression patterns of Wap65-1 and Wap65-2 were not associated with variations in environmental temperatures (Hirayama et al., 2003, 2004). In terms of heme binding ability, mWap65-1 bound to heme despite lacking the two conserved histidine residues, whereas mWap65-2, with the two residues, did not(Hirayama et al., 2004). Therefore, structure-function relationships of the two types of Wap65 remain unclear.

Medaka is suitable for developmental and genetic studies as well as for transgenic experiments, because of the transparency of the embryonic chorion,high fecundity and short generation times, in addition to its small genome size (Ishikawa, 2000; Gong et al., 2001). With such advantages, in this study we screened bacterial artificial chromosome (BAC)genomic clones from medaka containing the full-length sequences of mWap65-1 and mWap65-2, determined their nucleotide sequences, including 5′- and 3′-flanking regions, and compared their genomic organizations with those of the human hemopexin gene. Then, we examined the expression profiles of mWap65s in medaka embryos using quantitative real-time PCR. Furthermore, we generated transgenic medaka expressing green fluorescent protein (GFP) driven by mWap65-1 and mWap65-2 promoters and observed their expression patterns during ontogeny.

Adults of orange-red wild-type medaka Oryzias latipes Temminck and Schlegel, called himedaka, were fed ad libitum with commercial pellets and maintained at 25°C under the artificial conditions of 14 h:10 h light:dark.

Approximately 80 embryos incubated at 25°C were collected at each of the developmental stages: stage 9 [late morula stage, 7 h post-fertilization(hpf)], 13 (early gastrula stage, 14 hpf), 15 (mid gastrula stage, 19 hpf), 16(late gastrula stage, 24 hpf), 18 (late neurula stage, 29 hpf), 20 (4 somite stage, 34 hpf), 22 (9 somite stage, 42 hpf), 24 (16 somite stage, 48 hpf), 28(30 somite stage, 70 hpf), 36 (heart developmental stage, 156 hpf) and 39(hatching stage, 240 hpf). Developmental stages were ascertained in a manner described by Iwamatsu (1994). Samples were frozen immediately in a liquid nitrogen bath and stored at-80°C until use for first strand cDNA synthesis.

Embryos at stage 32 (somite completion stage, 120 hpf) were collected and used for whole-mount in situ hybridization. For DNA microinjection,fertilized eggs were collected within 30 min after spawning and kept at 4°C to arrest development until microinjection.

A BAC library of sperm genomic DNA from medaka of HNI strain and its high-density replica (HDR) filters has been constructed at Keio University School of Medicine, Japan (Kondo et al.,2002). The BAC library was screened by colony hybridization with both mWap65s probes. mWap65-1 and mWap65-2 probes were composed of 1096-1534 nucleotides (nt) and 838-1376 nt cDNAs(Hirayama et al., 2004),respectively, and labeled with digoxigenin (DIG)-11-dUTP using DIG-High Prime DNA Labeling and Detection Starter Kit II (Roche Diagnostics, Mannheim,Germany), according to the manufacturer's instruction. 1 μg of purified BAC DNAs from positive clones were digested with HindIII and separated on 0.7% agarose gels with a TBE buffer (0.089 mol l^-1 Trisborate, 0.5 mol l^-1 EDTA). The digested DNAs were transferred to Biodyne B membranes (Pall BioSupport Division, Washington, NY, USA) and hybridized with the mWap65-1 and mWap65-2 probes by the method adopted for screening of the BAC library as described above.

Shotgun strategies were employed for sequencing selected clones, 182O24 including mWap65-1 and 107E17 including mWap65-2. About 1150 and 1550 shotgun clones from BAC clones 182O24 and 107E17, respectively, were sequenced using ABI PRISM BigDye Cycle Terminator Ready Reaction Mix diluted with 5× sequencing buffer (Applied Biosystems, Foster City, CA, USA). Excess dye-terminators were removed by gel filtration and then PCR products were automatically loaded onto an ABI PRISM 377 or 3700 DNA analyzer (Applied Biosystems). To assemble the individual shotgun sequences into contigs,computer programs, Phred, Phrap and Consed were used for base calling,assembly of sequences, and viewing and editing analysis, respectively.

The exon-intron organizations of mWap65-1 and mWap65-2were determined with genomic and cDNA nucleotide sequences (the DDBJ/EMBL/GenBank databases, accession numbers AB075198 for mWap65-1cDNA and AB075199 mWap65-2 cDNA). The sequence of a 56 kb region containing 5′- and 3′-flanking sequences of mWap65-1 with the coding sequence in the middle was compared with those of mWap65-2and the human hemopexin gene using the PipMaker(http://bio.cse.psu.edu/pipmaker/)(Schwartz et al., 2000) and RepeatMasker (Open-3.0 1996-2004; http://www.repeatmaster.org/;A. F. A. Smit, R. Hubley and P. Green, unpublished) (Smit and Green, 1999)programs. The human hemopexin gene had been mapped to chromosome 11p15.5-p15.4(Law et al., 1988) and its sequence was obtained from NCBI's LocusLink(http://www.ncbi.nlm.nih.gov/genome/guide/human/). Putative cis-elements were searched by computer program TFSEARCH version 1.3(http://www.cbrc.jp/research/db/TFSEARCH.html)that has been constructed for highly correlated sequence fragments in the TFMATRIX transcription factor binding site profile database in the TRANSFAC databases by GBF-Braunschweig (Heinemeyer et al., 1998).

Total RNA was extracted from embryos at various developmental stages using an ISOGEN system (Nippon Gene, Tokyo, Japan). To remove endogenous DNA contamination, the preparation containing total RNA was digested with DNase. An aliquot of 5 μg of total RNA was dissolved in 8.5 μl water and added with 5 U of DNase I (TaKaRa, Otsu, Japan) and 40 U of ribonuclease inhibitor(TaKaRa), then the RNA solution was incubated at 37°C for 1 h. The enzyme was inactivated at 98°C for 3 min, and the RNA solution was chilled on ice. First strand cDNA was synthesized as follows. One microlitre of 10μmol l^-1 adapter primer(5′-GGCCACGCGTCGACTAGTAC-3′), 2 μl of 10 mmol l^-1dNTP and 1 μl of water were added to the RNA solution treated with DNase. The reaction mixture was heated at 65°C for 5 min, quickly chilled on ice,added with 4 μl of 5× first-strand buffer containing 250 mmol l^-1 Tris-HCl (pH 8.3), 375 mmol l^-1 KCl and 15 mmol l^-1 MgCl₂, 0.1 mol l^-1 dithiothreitol (DTT)and 200 U SuperScript™ III reverse transcriptase (Invitrogen, Carlsbad,CA, USA), and incubated at 42°C for 50 min. The enzyme was heat-inactivated at 70°C for 15 min and first strand cDNA synthesis was completed by treatment with 1.0 μl of RNase H (Invitrogen).

Primer pairs m1-F-1250 and m1-R-1319 for mWap65-1, m2-F-111 and m2-R-173 for mWap65-2, and m-bactin-F-7 and m-bactin-R-70 for medakaβ -actin (Takagi et al.,1994) were designed using the Primer Express Software (Applied Biosystems) (Table 1). Each of 20 μl reaction mixtures contained 1× SYBR Green Master mix (Applied Biosystems), ≈25 ng of first strand cDNA and 8 pmol of each primer. Real-time PCR was performed with ABI PRISM 7300 (Applied Biosystems). Thermal cycling conditions consisted of the initial steps for 2 min at 50°C then 10 min at 95°C followed by 45 cycles of denaturation at 95°C for 15 s,annealing at 60°C for 1 min. Relative quantification was carried out by normalization the values relative to those of β-actin. Amplification specificity was examined using the melting curve following the manufacturer's instructions. Analysis and quantification using the comparative Ct method were carried out with the ABI Prism 7300 Sequence Detection Software(SDS) version 1.2 (Applied Biosystems).

Sense and antisense DIG-labeled RNA probes were synthesized from the DNA fragments of nt 1-1610 of mWap65-1 cDNA and nt 96-1510 of mWap65-2 cDNA with a DIG RNA labeling kit (Roche Diagnostics). DIG-labeling of RNA probes was carried out according to the manufacturer's instruction (Promega, Madison, WI, USA). Whole-mount in situhybridization with the DIG-labeled RNA probes was performed according to the method of Westerfield (2000)with some modifications, as follows: after fixation at 4°C for 4 h in 4%paraformaldehyde in phosphate-buffered saline (PBS) containing 0.1% Tween 20(PBSTw), the chorion was removed and embryos then refixed overnight at 4°C with the same solution. Treatment with proteinase K (TaKaRa; 5 μg ml l^-1 in PBSTw) was carried out for 10 min at 37°C and the DIG-labeled RNA probes were used for hybridization at an approximate concentration of 1 μg ml l^-1.

mWap65-1 and mWap65-2 upstream regions of ≈5 kb were amplified from BAC clones 182O24 and 107E17, respectively, by PCR using primers mWap65-1Fgfp5 and mWap65-1Rgfp5 for mWap65-1 and mWap65-2Fgfp5 and mWap65-2Rgfp5 for mWap65-2(Table 2, Fig. 1). The amplified products of mWap65-1 and mWap65-2 were double-digested with BamHI/HindIII and NotI/HindIII,respectively. The digested mWap65-1 fragment was inserted into BamHI and HindIII sites of the phrGFP vector (Stratagene, La Jolla, CA, USA), whereas that of mWap65-2 was inserted into NotI and HindIII sites(Fig. 1). Plasmids of ≈8.5 kb, containing mWap65-1 and mWap65-2 upstream regions, were designated as mWap65-1-hrGFP and mWap65-2-hrGFP,respectively.

Microinjection was carried out according to the method of Kinoshita et al.(1996). A DNA solution of≈20 ng μl^-1 in distilled water was injected into the cytoplasm of blastomeres at the single cell stage. Eggs injected with recombinant GFP plasmids were incubated at 25°C and GFP fluorescence was observed by fluorescence microscope.

An HNI strain genomic BAC library was screened to isolate mWap65-1and mWap65-2, resulting in isolation of two clones, 182O24 containing mWap65-1 and 107E17 containing mWap65-2. The shotgun strategy revealed that the genomic BAC clones 182O24 and 107E17 contained 155 kb and 196 kb, respectively, whereas the sizes of mWap65-1 and mWap65-2 were 6.1 kb and 5.4 kb, respectively. The genomic nucleotide sequences of mWap65-1 and mWap65-2 have been registered in DDBJ/EMBL/GenBank with the accession numbers AB195240 and AB195241,respectively.

Exon-intron organizations were predicted using the genomic nucleotide sequences of mWap65-1 and mWap65-2, and compared with that of the human hemopexin gene (Fig. 2). Both mWap65s consisted of 10 exons and 9 introns, as is the case for the human hemopexin gene. Although mWap65s were smaller than the human hemopexin gene (9.9 kb), their exon-intron organizations were almost identical.

The sequence of a 56 kb region containing 5′- and 3′-flanking sequences of mWap65-1 with the coding sequence in the middle was compared with those of the corresponding regions of mWap65-2 and the human hemopexin gene using the PipMaker program. As shown in Fig. 3, a high homology between mWap65-1 and mWap65-2 was found in the fifth, sixth and eighth exons, and in the seventh and ninth introns. In the 5′- and 3′-flanking regions, homology was found 15.3 and 11.6 kb upstream and 2.1 and 22.5 kb downstream of the coding region of mWap65-1. mWap65-1showed an apparent homology with the human hemopexin gene only in the fifth and sixth exons (Fig. 3). mWap65-1 also showed homology with mWap65-2 and the human hemopexin gene in a very short sequence, considered to be a repetitive element.

Previously, Hirayama et al.(2004) determined the full-length of mWap65-1 and mWap65-2 cDNAs. In this study,we found that the sequences at the 5′ ends of the cDNA clones containing mWap65-1 and mWap65-2 were all identical (data not shown),and so the putative transcription start points of mWap65s were confirmed. Thus, putative transcription start points of mWap65-1 and mWap65-2 were located at 29 and 89 bases, respectively, upstream of the translation start site (Fig. 4). Putative cis-elements were searched using prediction programs. Although a 1 kb 5′-flanking region showed no obvious sequence homology among mWap65-1, mWap65-2 and the human hemopexin gene (data not shown), various transcription factor binding sites known in vertebrates were found in mWap65s (Fig. 4). The binding sites for HNF-3β, which is a liver-enriched transcription factor, were found at -344 for mWap65-1 and -963 for mWap65-2. A Cdx1 binding site, which is an important element for developmental regulation (Subramanian et al., 1995), was abundant in the 5′-flanking regions of both mWap65-1 (-790, -741, -672, -567, -471, -370, -307, -246) and mWap65-2 (-940, -874, -812, -801, -793, -765, -707, -556, -539, -509,-494, -446, -339, -166). In addition, the 5′-flanking region of mWap65-1 contained putative binding sites for SRY (-994, -532, -463,-343, -230, -155), USF (-854), Nkx-2.5 (-838, -687), MZF1 (-654, -201), Oct-1(-516, -499), XFD-1 (-514), C/EBPα (-496), HFH-2 (-344), Evi-1 (-279),GATA-1 (-257) and AML-1a (-87), whereas that of mWap65-2 consisted of SRY (-946, -500, -463, -419, -180, -89, -48), Pbx-1 (-837), Prx-2 (-861,-819), HFH-2 (-854), Nkx-2.5 (-815, -637, +58), XFD-1 (-778), δEF1(-718), Brn-2 (-539), Oct-1 (-514), Sox-5 (-499), HNF-1 (-246) and c-Ets-1(-72).

Quantitative real-time RCR was performed to examine changes in the expression levels of mWap65s in embryos at stages 9-39, using those of β-actin as the internal standard. As shown in Fig. 5, the mWap65-1transcripts were detected in embryos at the beginning of the experiment, at stage 9, the late morula stage. The expression level at stage 13 (early gastrula) was about 30-fold higher than that at stage 9, and the highest expression level of mWap65-1 was observed at stage 16 (late gastrula). Thereafter, the amount of transcripts was reduced at stages 18(late neurula) to 24 (16 somite), but the expression level was increased again at stage 28, when 30 somites are formed in the trunk muscle, and almost constant levels of transcripts were maintained until hatching. In contrast, mWap65-2 transcripts were not detected until stage 15 (mid gastrula)(Fig. 5). The mWap65-2transcripts were first observed at stage 16. Then the expression levels were rapidly increased until hatching, although the levels were slightly decreased at stage 24 (16 somite) compared with those at stage 22 (9 somite). Thus,expression patterns of mWap65-1 and mWap65-2 were markedly different from each other.

To examine spatial expression patterns of mWap65s, whole-mount in situ hybridization was performed on embryos at stage 32, at which point somite formation is complete. mWap65-1 hybridization signals were observed along the edge of pectoral fin bud and median fin fold of the tail bud, but not in the liver (Fig. 6). However, mWap65-2 was observed only in liver, clearly demonstrating different functions of mWap65-1 and mWap65-2during ontogeny. No signal was detected with sense probes for mWap65s(data not shown).

mWap65-1-hrGFP and mWap65-2-hrGFP constructs were injected into embryos at the single cell stage. One day after injection with the mWap65-1-hrGFP construct, GFP fluorescence spots were observed throughout the blastoderm including the embryonic shield, when embryos attained stage 16 (late gastrula) (Fig. 7A). Following development from stage 16 to stage 29, GFP fluorescence was observed not only in yolk sac but also various areas of the embryonic body. Large and many small fluorescent spots were observed in the yolk sac at stages 23 (12 somite; Fig. 7C) and 26 (22 somite; Fig. 7E) as well as in the embryonic body(Fig. 7C, right part). However,the small fluorescent spots tended to disappear at stage 26 and in turn a new fluorescent spot was found in the area of the liver anlage, a primitive form of liver in embryos (Fig. 7E). The fluorescence of liver anlage was more clearly observed in the embryo at stage 29 (Fig. 7G).

Embryos injected with the mWap65-2-hrGFP construct expressed GFP at stage 16 (Fig. 8A), as was the case with mWap65-1. At this stage also fluorescent spots of mWap65-2-hrGFP were again observed in various areas of the blastoderm, including the embryonic shield. Similarly to the expression patterns of mWap65-1-hrGFP, the mWap65-2-hrGFP fluorescent spots mostly disappeared at stage 26. However, mWap65-2-hrGFP was first observed in liver anlage in the embryos at stage 30 (35 somite)(Fig. 8E) in contrast to mWap65-1-hrGFP, which was activated in this tissue as early as at stage 26 (22 somite). These differences in expression patterns between mWap65-1-hrGFP and mWap65-2-hrGFP appeared to reflect those of the accumulated mRNA levels as shown in Fig. 5. Although mWap65-1-hrGFP was examined for its activation profiles until hatching, an intense GFP fluorescence was observed only in liver(Fig. 9).

We determined the nucleotide sequences of BAC clones containing genomic sequences of two medaka Wap65s, mWap65-1 and mWap65-2. Recently we have characterized two Fugu Wap65s, fWap65-1 and fWap65-2, and demonstrated that exon-intron organizations were conserved among fWap65-1, fWap65-2 and the human hemopexin gene (Hirayama et al.,2003). Both mWap65s in this study consisted of 10 exons and 9 introns (see Fig. 2) as is the case in the human hemopexin gene and fWap65s.

Alignment of nucleotide sequences using the PipMaker program revealed that not only fWap65-1 (Hirayama et al., 2003) but also mWap65-1 showed apparent homology with the human hemopexin gene in the fifth and sixth exons (see Fig. 3). These encode a potential receptor binding site, by which rabbit hemopexin is bound to hepatoma cells (Morgan et al., 1988, 1993). It has been reported for rabbit hemopexin that eight hydrophobic residues are important in the structural stability around the heme binding site(Paoli et al., 1999). In the case of Fugu, fWap65-1 contained five, and fWap65-2 seven, out of the eight hydrophobic residues found in rabbit hemopexin and most of these residues were located in the sixth exon. These residues were also found in mWap65s (data not shown), suggesting their importance in structural integrity.

No apparent sequence homology was found in the 5′-flanking region 1 kb upstream the coding sequence among mWap65-1, mWap65-2 and human hemopexin gene. However, this region was found to contain various transcription factor binding sites (see Fig. 4). Interestingly, several transcriptional elements, which are thought to regulate development, are contained in both mWap65s. While Cdx1 is considered to be a regulator of Hox gene expression (Subramanian et al., 1995), putative binding sites for Cdx1 are rich in the 5′-flanking regions of both mWap65s. Found also in mWap65s were the binding sites for Nkx-2.5, a vertebrate homologue of tinman from Drosophila melanogaster involved in cardiac mesoderm formation (Bodmer et al.,1990; Lints et al.,1993), and Prx-2, a homeobox transcription factor possibly playing roles in development of the heart and forebrain(Leussink et al., 1995), as in the case of fWap65s (Hirayama et al., 2003). The presence of these binding sites in the 5′-flanking regions of Wap65s indicates their involvement in the transcriptional regulation during development.

Rat hemopexin transcripts were first detected in liver on day 24 of gestation and rapidly increase during the postnatal period(Nikkila et al., 1991). Chicken hemopexin is found around hatching and increases in 4-day-old chicken to more than 1000-fold over embryonic levels as revealed by electroimmunoassay of serum (Grieninger et al.,1986).

Quantitative real-time PCR in this study revealed that mWap65-1transcripts were detected even at stage 9, the beginning of the experiments(see Fig. 5). Our previous study showed that mWap65-1 transcripts were first detected at stage 24 by semi-quantitative RT-PCR (Hirayama et al., 2004). Interestingly, quantitative real-time PCR analysis showed that embryos expressed mWap65-1 transcripts even when embryonic shield was not observed, and their maximum levels were observed in embryos at stage 16, when the blastoderm covers three-quarters of the yolk sphere and the embryonic shield becomes evident(Iwamatsu, 1994). The embryonic shield is formed at stage 16, followed by the formation of somite,notochord and organs until hatching. The high expression levels of mWap65-1 in early embryonic stages indicate a possible role of this gene in early embryogenesis.

mWap65-2 transcripts were first detected at stage 16 and thereafter gradually increased during ontogeny, as embryos undergo dynamic changes of their morphology. These results, together with the presence of putative transcriptional elements involving development in the 5′-flanking region of mWap65-2, suggests that mWap65-2also plays important roles in embryonic development.

Medaka embryos injected with mWap65-1-hrGFP- and mWap65-2-hrGFP expressed GFP fluorescence at stage 16, 1 day after injection, suggesting that mWap65-1 and mWap65-2 promoters were both activated at early developmental stages. Such GFP expression in the early embryonic stages was consistent with the results obtained by real-time RT-PCR (Fig. 5).

mWap65-2 transcripts were expressed in liver as revealed by in situ hybridization (see Fig. 6). Unexpectedly, those of mWap65-1 were expressed along the edge of pectoral fin bud and median fin fold of the tail bud in embryos at stage 32, suggesting a function of mWap65-2 in the development of these two tissues. GFP fluorescence was restricted to the liver of hatched medaka injected with mWap65-1-hrGFP (see Fig. 9), shifting its localization during hatching. While the msx homeobox genes are known to be important for limb development in mouse (Muneoka and Sasoon, 1992),their homologs are expressed in zebrafish embryos along the edge of the pectoral fin bud and median fin fold of the tail bud during development(Akimenko et al., 1995).

We observed transient expression profiles of GFP driven by two mWap65s promoter in this study, but the localization of expressed GFP varied among individuals (data not shown). It is known that the F₀generation of transgenic fish show the mosaic expression patterns of transgenes (Ju et al., 1999)and the uniform expression of transgenes is normally attained in the F₁ or F₂ generation(Chou et al., 2001). We are now generating stable transgenic lines to reveal more precisely the spatiotemporal expression patterns of mWap65s.

In conclusion, we determined the genomic nucleotide sequences of two mWap65s and their flanking sequences. Furthermore, the expression profiles of mWap65s were determined using quantitative real-time PCR and in situ hybridization. We also generated transgenic medaka expressing GFP driven by two mWap65 promoters, and examined their expression patterns. The spatiotemporal patterns in embryos were different between mWap65-1 and mWap65-2, suggesting their distinct roles, at least during ontogeny.

We would like to thank T. Watanabe, The University of Tokyo, for his technical assistance. This study was partly supported by a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan.