Introduction

A variety of specific immune functions have been observed in fish. Of these, allograft rejection, mixed leukocyte reaction, and delayed hypersensitivity reaction indicate that T cells are involved in these specific immune activities in a similar manner to that of mammals (Nakanishi et al. 2002). In addition, major histocompatibility complex (MHC) and T-cell receptor (TCR) genes have been identified in fish and amphibians (Hansen and Zapata 1998). It is expected that T-cell subsets exist in fish and amphibians. Nevertheless, T-cell subsets are yet to be demonstrated due to the lack of reagents for fish surface markers such as CD3, CD4 and CD8. Recently, the genes encoding polypeptides homologous to the known T-cell surface markers, CD3 and CD8, were reported in fish (Hansen and Strassburger 2000; Alabyev et al. 2000; Park et al. 2001) and amphibians (Dzialo and Cooper 1997; Gobel et al. 2000; Ropars et al. 2002). However, CD4 has not been identified in an ectothermic animal.

In avian and mammalian species, cytotoxic T cells and helper T cells are major T-cell subsets that carry CD8 and CD4, respectively. These molecules are involved in recognition of their specific antigens in association with MHC class I and class II molecules. Expression of CD8 and CD4 is critical for cell-mediated immune defense and T-cell development in the thymus. CD4 is a membrane-bound glycoprotein found on thymocytes and helper T cells (Parnes 1989), and the primary structure has been identified in several avian and mammalian species (Koskinen et al. 1999; Clark et al. 1988; Classon et al. 1986; Dumont-Drieux et al. 1992; Fomsgaard et al. 1992; Gustafsson et al. 1993; Hague et al. 1992; Maddon et al. 1985; Milde et al. 1993; Norimine et al. 1992; Romano et al. 1999). The CD4 protein is composed of four extracellular immunoglobulin (Ig)-like domains, a transmembrane region, and a cytoplasmic tail that associates with a tyrosine protein kinase, p56lck, involved in T-cell activation. CD4 genes have also been identified in human, mouse and chicken, and all are composed of ten exons (Gorman et al. 1987; Maddon et al. 1987; Koskinen et al. 2002). In these species, the promoter region of the CD4 genes does not have a TATA sequence (Koskinen et al. 2002; Salmon et al. 1993). The activation of the CD4 promoter also requires several factors, and many sequence motifs have been conserved during evolution (Koskinen et al. 2002). From an evolutionary standpoint, it is important to determine if these elements are conserved in lower vertebrates as well.

Fugu, Takifugu rubripes, has the smallest genome among vertebrates: 365 Mb, one-eighth the size of the human genome (Brenner et al. 1993). It has recently been the subject of a whole genome shotgun sequencing program (Aparicio et al. 2002) and the draft sequence data are publicly available. The availability of these data has given us a great advantage in studying the fish immune system. In this paper, we describe the cloning of the first ectothemic animal CD4 cDNA. We also report the expression and genomic organization of fugu CD4.

Materials and methods

Animals

Fugu (Takifugu rubripes) specimens were purchased from an aquaculture farm. These fish were kept in a circulating seawater tank at 20°C and were subjected to collection of various tissues and extraction of total RNA.

Screening of CD4

The fugu genomic sequence databases ver. 3.0 at the Fugu Genomics Group (http://fugu.hgmp.mrc.ac.uk) and DOE Joint Genome Institute (http://genome.jgi-psf.org/fugu6/fugu6.home.html) were screened in silico with sequences of duck and chicken CD4 (accession number AF378701 and Y12012) as a probe on a BLAST platform to obtain overlapping sequences (scaffolds) including a putative CD4.

cDNA cloning

Specific primers for RACE were designed from the putative coding region of CD4 (CD4-F1, CD4-F2, CD4-R1, and CD4-R2; Table 1). Total RNA was isolated from the thymus of fugu using RNA extraction solution (ISOGEN, Nippon gene, Toyama, Japan) following the manufacturer’s instructions. First-strand cDNA was synthesized from 1 μg of the poly(A)+ RNA with the SMART RACE cDNA amplification kit (Clontech, Palo Alto, Calif.) for 3′RACE and 5′RACE. First-round PCR amplification of 3′RACE was performed with CD4-F1 and universal primer mix-A (UPM-A, Clontech: Table 1). Nested PCR amplification of 3′RACE was performed with CD4-F2 and nested universal primer-A (NUP-A, Clontech; Table 1). Both reactions were carried out for 30 cycles of 94°C for 10 s, 60°C for 5 s, and 72°C for 1 min with a final extension of 72°C for 7 min. For 5′RACE, first-round PCR amplification was performed with CD4-R1 and UPM-A. Nested PCR amplification of 5′RACE was performed with CD4-R2 and NUP-A. Both reactions were carried out for 30 cycles of 94°C for 10 s, 60°C for 5 s, and 72°C for 30 s with a final extension of 72°C for 7 min. Products were subcloned into pCR 2.1 vector (Invitrogen, San Diego, Calif.) and sequenced using an ABI 310 genetic analyzer (ABI, Foster City, Calif.).

Table 1 Oligonucleotide primers used to amplify fugu CD4 cDNA
Full size table

Structural and phylogenetic analysis

Putative leader sequence, transmembrane and cytoplasmic regions were predicted based on SignalP (http://www.cbs.dtu.dk/services/SignalP/) and SOSUI (http://sosui.proteome.bio.tuat.ac.jp/sosuimenu0.html) programs. Immunoglobulin-like domains were predicted using InterPro (http://www.ebi.ac.uk/interpro/). Aligning the full-length amino-acid sequences of CD4 and construction of a phylogenetic tree were conducted using Clustal W (Thompson et al. 1994) and TreeView ver. 1.6.6 (Page 1996).

Genomic analysis

Genomic organization was determined by comparing the cDNA sequence and two scaffold sequences in fugu genomic database version 3.0. Transcription factor binding sites were predicted by TFSEARCH ver. 1.3 (http://www.cbrc.jp/research/db/TFSEARCHJ.html).

Reverse transcriptase PCR

Total RNA was isolated from fugu peripheral blood leukocytes (PBL), thymus, head kidney, kidney, spleen, heart, liver, gonad, intestine, skin, gill, brain, and muscle. First-strand cDNA was synthesized from 1 μg of total RNA of each organ in a 20-μl reaction. PCR amplification was performed with CD4-F2 and CD4-R3 for 30 cycles of 96°C for 10 s, 60°C for 5 s, and 72°C for 30 s (Table 1). PCR products were subjected to agarose-gel electrophoresis and stained with ethidium bromide.

Transient expression

BamHI linker-linked insert cDNA with Kozak sequence was amplified by PCR, digested, and ligated into the BamHI site of mammalian expression vector pFlag-CMV-1 (Sigma, St Louis, Mo.). COS-7 cells were transfected with 12 μg of this construct in 100-mm plates using TransFast transfection reagent (Promega, Tokyo, Japan). Transfected COS-7 cells were grown for 3 days and then harvested with PBS containing 5 mM EDTA. The cells were washed twice in PBS containing 0.5% BSA, 2 mM EDTA, and 0.01% sodium azide and adjusted to 108 cells/ml. To each 100-μl aliquot, anti-Flag antibody was added to give a final dilution of 1:1,000, and cells were incubated on ice for 30 min. The cells were subsequently washed twice and resuspended in PBS/BSA/EDTA/azide and FITC-labeled goat anti-mouse IgG (Jackson Immunoresearch Laboratories, West Grove, Pa.) added to give a final dilution of 1:100. Cells were incubated at 4°C for 30 min, and then washed as above and resuspended in PBS/BSA/EDTA/azide. After suspension, the stained cells were analyzed with a flow cytometer (Partec, Münster, Germany).

Results and discussion

Isolation of fugu CD4 cDNA

We cloned a CD4 homologue gene from a teleost species, fugu. We chose amino acid sequences of duck and chicken CD4s to query the fugu genomic sequence database by BLAST search. A sequence similar to the duck and chicken CD4 genes was found in a scaffold of fugu genome. We named the sequence as a putative fugu CD4 and designed primers for cDNA cloning. We cloned the full-length cDNA encoding a putative fugu CD4 by RACE. This cDNA sequence was subjected to a BLASTX search. The clone was most similar to the complete African green monkey CD4 (2e-5), followed by various avian and mammalian CD4 sequences.

Fugu CD4 cDNA (accession no. AB164054) consists of 2,664 nucleotide residues, encoding a single open-reading frame of 463 amino acid residues. After cleavage of the putative 30-amino-acid hydrophobic leader sequence, a predicted mature protein would be generated. The size of the putative fugu CD4 is similar to those of other CD4 molecules. The alignment of CD4 sequences and domain analysis indicates that the fugu sequence contains four Ig-like domains, a transmembrane region, and a cytoplasmic tail like avian and mammalian CD4 molecules (Fig. 1). While cloning the fugu CD4 cDNA, we found an allotypic variant (accession no. AB164055) with 6-bp difference resulting in one amino acid substitution (N299→D299).

Fig. 1
figure 1

Alignment of CD4 amino-acid sequences. Gaps, shown as dashes, have been introduced to maximize the alignment of the sequences. Residues identical with the fugu sequence are denoted by dots. Amino acids conserved in all sequences are indicated by asterisks. The conserved Cys residues are boxed. The leader sequence, the four Ig-like domains (D1–D4), and the transmembrane segment, and the cytoplasmic tail boundaries are indicated above the sequences. CD4 accession numbers are noted in the text. Other accession numbers are as follows: CD4, mouse (NM_013488); human (M12807); rabbit (M92840); white whale (AF071799); cat (AB000483)

Full size image

A multiple sequence alignment of CD4 amino acid sequences is shown in Fig. 1. Fugu CD4 has a percentage of identity with those of avian and mammalian species ranging from 15.4% to 19.6% (Table 2). Between chicken and human CD4 molecules, the percentage of identity is also very low (22.8%), with conserved overall structure (Koskinen et al. 1999). These values are much lower than those among mammalian species. These results indicate that, in spite of the low level of sequence identity in CD4 molecules, the basic structure of CD4 has been conserved during evolution.

Table 2 Amino acid identity (%) between the CD4 proteins of fugu and other vertebrates
Full size table

A phylogenetic analysis was performed using CD4 full sequences (Fig. 2). As depicted by neighbor-joining tree, the putative fugu CD4 falls into the cluster of CD4 sequences.

Fig. 2
figure 2

Phylogenetic tree of CD4, CD2, and CD22. This unrooted tree was constructed by the neighbor-joining method, based on the amino acid alignment (Clustal W) of full-length protein sequences. Node values represent bootstrap analysis of 1,000 replicants. CD4 accession numbers are noted in the text and the legend of Fig. 1. Other accession numbers are as follows: CD2, mouse (NM_013486); human (M16445); CD22, mouse (NM_009845); human (NM_001771); CD4, squirrel monkey (Q29037); rhesus monkey (D63349); African green monkey (D86589); chimpanzee (X73323); rat (NM_012705); pig (X65629); dog (X68565)

Full size image

Structural features of fugu CD4

The putative fugu CD4 structure is conserved with some differences from avian and mammalian CD4 molecules. The fugu CD4 has four Ig-like domains (D1–D4), like other known CD4s. The first and second Ig-like domains (D1 and D2) are known to be important for binding to MHC class II in mammals (Clayton et al. 1989; Bowman et al. 1990; Moebius et al. 1992, 1993; Huang et al. 1997). A disulfide bond within the D1 domain is conserved in the avian and mammalian CD4 molecules. However, in fugu D1 lacks a Cys residue capable of forming a disulfide bond within this domain, differing from the D1 domain of other known CD4s (Fig. 1). The second domain of fugu CD4 has a pair of Cys residues, like mouse and human CD4. However, in chicken, dog, cat, and whale the first Cys is replaced, such that no disulfide bond can be formed in D2 (Koskinen et al. 1999; Milde et al. 1993; Norimine et al. 1992; Romano et al. 1999). Thus, lack of a disulfide bond in the Ig-like domain is observed in some species. A disulfide bond is not likely to be necessary to form Ig-like domains and a similar type of Ig-like domain is found in the sequence of CD2 (Williams et al. 1987) and D3 of CD4 (Parnes 1989) in mammals. CD4 is expressed on helper T cells that recognize antigen with MHC class II on antigen-presenting cells (Parnes 1989). In cat, dog, and cetacean, highly constitutive expression of MHC class II molecules on T cells was observed (Deeg et al. 1982; Neefjes et al. 1986; Romano et al. 1992). In these species, different functions and interactions of T cells and antigen-presenting cells are suggested (Romano et al. 1999). Like these species, the absence of a disulfide bond in D1 of fugu CD4 may give a different conformation to this molecule and affect the binding and interaction of CD4 and MHC class II. In D3, fugu CD4 has a pair of Cys residues that might form an additional disulfide bond lacking in other animal CD4s. This disulfide bond also may give the fugu CD4 molecule a different conformation. The conserved Cys residues, which might form a disulfide bond, are found in D4 of fugu CD4. D4 is involved in the dimerization of human CD4 during antigen recognition by the TCR (Wu et al. 1997; Moldovan et al. 2002). The D4 of fugu CD4 is conserved, suggesting that dimerization of fugu CD4 may occur. The conformation of the fugu extracellular domain may be different from that of other CD4 molecules due to different disulfide bonding patterns in the first three domains. These differences may affect the function of CD4-expressing cells in fugu.

Fugu CD4 has four potential N-linked glycosylation sites, and only one is conserved and located in the D4 domain (Fig. 1). The consensus p56lck tyrosine kinase motif (lck motif, Cys-X-Cys) is conserved in the cytoplasmic domain of fugu CD4. Lck is critical for CD4 association in the cytoplasmic region and has been implicated in T-cell maturation and signaling, and the lck gene shows a tightly controlled lymphocyte-specific expression in mammals (Marth et al. 1985; Perlmutter 1989; Veillette et al. 1989). The fugu lck gene also shows lymphoid-specific expression (Brenner et al. 2002). This result is consistent with the expression pattern of the fugu CD4 gene (Fig. 3). These observations suggest that the lck may associate with the consensus lck motif in the cytoplasmic tail of fugu CD4.

Fig. 3
figure 3

Expression of fugu CD4. Tissue distribution of fugu CD4 transcripts by RT-PCR is shown in the upper panel. Primers are located in exons 8 and 9 of the fugu CD4 gene. Lower panel shows β-actin transcripts

Full size image

Expression of the fugu CD4 gene

Fugu CD4 was expressed in the thymus, head kidney, kidney, spleen, gonad, skin and gill (Fig. 3). Slight background expression of CD4 was found in the PBL, heart, liver, and brain. No expression was seen in the muscle. Fugu CD4 is expressed predominantly in the tissues of the immune system in agreement with the lymphocyte-specific expression of CD4 in mammals.

COS-7 cells were transfected with pFlagCD4 (Fig. 4). The expression of fugu CD4 was observed on the surface of transfected COS-7 cells. We found that fugu CD4 was efficiently expressed on 11% of COS-7 cells.

Fig. 4
figure 4

Cell surface expression of fugu CD4. Fugu CD4 is expressed on the surface of COS-7 cells tranfected with pFlagCD4. Mock transfected cells were negative for Flag surface expression

Full size image

Genomic analysis of fugu CD4

Fugu CD4 and putative GAPDH genes were found in the same scaffold. A syntenic relationship between the CD4 and GAPDH genes in fugu was observed, as in human and chicken (Koskinen et al. 2002). This shows conserved synteny within these regions during evolution.

Fugu CD4 gene organization was analyzed by comparing the cDNA and the scaffold sequence of fugu genomic database ver. 3.0. Fugu CD4 spans a total of 12.1 kb and is composed of 12 exons (Fig. 5A), differing from chicken and human CD4 genes, which consist of ten exons, with lengths 11.5 kb and 31.3 kb, respectively (Fig. 5A). Exons 5–12 almost identically correspond to exons 3–10 of other CD4 genes (Gorman et al. 1987; Maddon et al. 1987; Koskinen et al. 2002). The fugu CD4 gene has two additional exons within the 5′-untranslated region (5′UTR) and signal peptide coding region, respectively. In the 5′UTR, there is a long intron of 5.6 kb and 10.4 kb in chicken and human, respectively. The fugu CD4 gene also has two long introns of about 4.6 kb and 1.5 kb in the region upstream of the predicted start codon. The long introns of human, mouse, and chicken CD4 genes contain an important regulatory element (Gorman et al. 1987; Koskinen et al. 2002; Donda et al. 1996). The fugu CD4 gene may also have regulatory elements in these long introns. The fugu CD4 gene is longer than expected. Although human and mouse CD4 genes are much longer than the fugu CD4 gene, the chicken CD4 gene spans only 11.5 kb, which is almost the same size as fugu CD4. Most of the introns of the fugu CD4 gene are shorter than the corresponding human and chicken introns, but the fugu gene has additional introns, making the fugu sequence longer.

Fig. 5
figure 5

A Genomic organization of the CD4 gene. The genomic organization of the fugu CD4 gene was predicted by comparison with fugu genome draft sequence. The exons are numbered and shown in boxes separated by introns. Black and white boxes indicate coding exons and non-coding exons, respectively. B The putative promoter region of the fugu CD4 gene. The Ihr and Ihr-like sequences are underlined. The potential transcription factor binding sites are also underlined and named above the sequence. The sequence after an arrow is the first intron. An asterisk indicates transcriptional initiation site

Full size image

In the putative promoter region of the fugu CD4 gene, no TATA box was found, but we detected some potential transcription binding motifs, such as an Ets site, an Ikaros binding site, and Myb and MAZ binding sites, which have been shown to play important roles in the expression of CD4 genes (Fig. 5B, Salmon et al. 1993; Sarafova and Siu 2000). These features are also conserved in the chicken CD4 promoter (Koskinen et al. 2002).

We have cloned the first ectothermic animal CD4 gene from the fugu. The basic structure is similar to known CD4 molecules, but with a low level of overall sequence identity and some unique structural features. Our results suggest the possibility of producing specific molecules, such as antibody or nucleotide probes, to identify subtypes of lymphoid cells, which may lead to a deeper understanding of the basic structure and function of the immune system in fish.