Cloning of the complete rat immunoglobulin δ gene: evolutionary implications

Abstract

The recent discovery of a Cδ encoding gene in artiodactyls has raised questions regarding the evolution of the gene. In the present study, we have analysed the complete rat Cδ gene both at the cDNA and genomic levels, showing that the rat Cδ gene is structurally similar to the corresponding mouse gene. Analysis of the rat immunoglobulin D heavy chain cDNA tail sequences, revealed two transcripts for the secreted form with varying sizes of their 3′ untranslated region (UTR), resulting from usage of two different poly(A) addition signals. Furthermore, a membrane-bound form encoding transcript, possessing a long 3′ UTR, was also observed. Phylogenetic analysis supports that the Cδ gene appeared early in the evolution of vertebrates, and it was probably duplicated from the Cµ gene more than 400 million years ago.

Introduction

Mammals express five classes of immunoglobulin heavy chain constant region genes, including µ, δ, γ, ε, and α. The δ gene, encoding immunoglobulin D (IgD), has been thought to be a relatively novel immunoglobulin class, being present only in rodents and primates. IgD was first identified in a myeloma patient and later, it was also found in mice and rats.¹ The molecule is mainly expressed on the surface of B cells and only low concentrations of the secreted form can be found in serum.^2,3 For the past three decades, there has been no convincing evidence to support the presence of the protein or the corresponding gene in additional species of mammals. In a recent study,⁴ however, we have identified a δ gene in artiodactyl species such as cow, sheep and pig, showing that the gene in these species is structurally similar to the human δ gene, consisting of three constant region domains (CH1, CH2 and CH3) and one to two hinge encoding exons. Studies in these species have offered significant insight into IgD phylogeny, as it could be shown that the present δCH1 exon in ruminants was rather recently (approximately 20 million years ago) duplicated from the µCH1 together with its upstream sequence, leading to an approximately 96% sequence similarity between the µCH1 and δCH1 domains.⁴ The duplication of the µCH1 upstream sequence, encompassing the Sµ region, has supplied the bovine δ gene with a short functional Sδ region. Thus, in contrast to primates and rodents, the bovine δ gene can be expressed through class switch recombination, involving the bovine Sµ and Sδ. The short Sδ has so far only been demonstrated in the cow.

Based on our findings, one may speculate that the Cδ gene may be present in most, if not all mammalian species. However, in other lower vertebrates, such as birds, reptiles and amphibians, the existence of the gene remains unknown.^5,6 Surprisingly, a δ like gene with multiple constant region domains was found several years ago downstream of the µ gene in teleost fish.^7–9 The gene thus appears to have evolved differently in vertebrates.

In order to study the evolution of the vertebrate Cδ gene, it is necessary to have access to additional sequences from different species to allow a comprehensive phylogenetic analysis. As only a partial sequence of the rat δ gene cDNA was previously described¹⁰ we therefore analysed the complete rat δ gene at both the cDNA and genomic levels.

Materials and methods

RNA isolations and reverse transcriptase–polymerase chain reaction (RT–PCR) amplification of the rat IgD constant region encoding cDNA

Total cellular RNA was extracted using the RNeasy® minikit (QIAGEN, Valencia, CA) from rat (Sprague–Dawley) spleen, and subjected to cDNA synthesis with a First-Strand cDNA Synthesis Kit (Amersham Pharmacia Biotech, Uppsala, Sweden). RT–PCR was carried out under the following conditions: 94° 3 min, then 40 cycles of 94° 30 s, 55° 30 s and 72° 1 min. The primers used for amplification were: rat IgDs (5′-TAA GCT AAC CAG ACA CAG AAA G-3′) and rat IgDas (5′-CT CCT GAA AAG GAA CCT GAC T-3′).

Cloning of both the secreted and transmembrane tail of rat IgD heavy chain cDNA

3′ rapid amplification of cDNA ends (RACE) was carried out to clone both the secreted and transmembrane tail encoding sequences. About 5 µg rat spleen total RNA was reverse-transcribed into cDNA using a First-Strand cDNA Synthesis Kit (Amersham Pharmacia Biotech). The first amplification was performed using the primers rat IgD Hs (hinge exon derived, 5′-CCA ATC CAA CTC CCC AAG CAA A-3′) and cDN (designed based on the sequence of the Not I-d(T)₁₈ primer in the kit, 5′-AAC TGG AAG AAT TCG CGG CC-3′) under conditions of 94° 3 min for denaturing, 30 cycles of 94° 30 s, 58° 30 s, 72° 30 s. One µl PCR mixture from the first amplification was employed as template for the second amplification using the primers rat IgD CH3s (CH3 domain derived, 5′-ATC TGG AGT GTC CTG AGA CTG C-3′), and Fs-common (designed based on the sequence of the NotI-d(T)₁₈ primer in the kit, 5′-TGG AAG AAT TCG CGG CCG CAG GAA-3′) under conditions of 94° 3 min for denaturing, 30 cycles of 94° 30 s, 60° 30 seconds, 72° 30 s.

Long PCR amplification of genomic DNA fragments of rat immunoglubulin Cδ gene

To analyse exon–intron boundaries of the rat Cδ gene, genomic fragments spanning exon–intron–exon were amplified using a long PCR Kit (Expand^TM Long Template PCR System Kit, Roche Diagnostics Scandinavia AB, Bromma, Sweden). The primer pairs used were, Rat IgD Hs (sequence as indicated above) and Rat IgD CH3as (5′-GTG AAT GTG GTC AGG ATG TT-3′) yielding an approximately 1·4 kb fragment, Rat IgD CH3s (sequence as indicated above) and Rat IgD Secas (5′-CTC AGG GAA AGG CAG GAC CAT C-3′) yielding a 4·8-kb band, Rat IgD Secs (5′-GAT GGT CCT GCC TTT CCC TGA G-3′) and Rat IgD TMas (5′-GAC ATG CCT CCT AAT CCT ATC A-3′) yielding a 2·5-kb band. The extending temperature for all amplifications was 68°, as annealing temperature for each amplification was adjusted according to the primers used.

Cloning and sequencing of PCR products

All PCR products were run on agarose gels and purified using a QIAquick Gel Extraction Kit (QIAGEN), and subsequently cloned into the pGEM-FT vector.¹¹ Sequencing of the recombinant plasmid was performed using an ABI PRISM® BigDye™ Terminator Ready Reaction Kit (Perkin-Elmer, Foster, CA). The exon–intron boundaries were deduced by comparing the genomic DNA with the cDNA sequences.

Computational analysis of DNA sequences and construction of phylogenetic tree

All DNA sequence editing, comparisons and alignments were performed using the MegAlign program (DNASTAR, Inc, Madison, WI). Some rat Cδ genomic DNA sequences were obtained by search of the rat whole genome shotgun database (http://www.ncbi.nlm.nih.gov/blast/mmtrace.html). Construction of phylogenetic tree was made using the protpars programs from the phylip package.¹² The consensus tree were taken from 1000 bootstrapped phylogenetic trees (using IgM and IgD heavy chain constant region membrane bound tail peptide sequences). Except for the rat IgD sequence obtained in this study, all other sequences were extracted from the NCBI GenBank with the following accession numbers: cow IgM (U63637), human IgM (X14940), mouse IgM (J00443), pig IgM (U50149), sheep IgM (L04260), rabbit IgM (J00666), catfish IgM (M27230), salmon IgM (s48658), clawed frog IgM (M20484), duck IgM (U27213, AJ314754), human IgD (X57331), mouse IgD (J00447), cow IgD (AF411240), sheep IgD (AF411238), pig IgD (AF411239), catfish IgD (U67437), cod IgD (AF155203 AF155205), salmon IgD (AF141605). The membrane-bound form encoding cDNA tails of IgM heavy chain in species such as rat (BF551033), chicken (AJ395400, AJ444200, BG625537, AJ395139, AJ395400, AJ395578) and frog (BI314212), were identifieded from EST clones deposited in the NCBI GenBank.

Results

Cloning of the rat full length IgD heavy chain constant region encoding cDNA

Part of the rat IgD constant region – encoding cDNA, lacking part of the CH1 sequence, has been published previously.¹⁰ To clone the full rat IgD heavy chain constant region gene, a BLAST search against the rat EST database in the NCBI GenBank using the published δ sequence, was first carried out to look for EST clones, which may span the entire CH1 domain. One clone (AI555192) was found, containing a rat VH sequence and a putative δCH1 sequence. Two primers, rat IgDs and rat IgDas, were designed based on the sequence of the EST clone and 3′ UTR (untranslated region) of the published δ chain encoding sequence, and employed in RT–PCR amplification using total RNA isolated from rat spleen as an initial template. Sequencing of the cloned, approximately 800 bp PCR product (accession number: AY148494), revealed that the EST clone identified above indeed encoded the δ chain.

Alignment of the deduced rat IgD heavy chain constant region amino acid sequence with that of the mouse shows an overall 62·2% similarity. However, a comparison of each individual domain indicates that the CH3 domain is more conserved (76·6%) than the CH1 (44%) and hinge region (55·9%) domains. The potential disulphide bond donor, cysteine, appears at invariable positions throughout the entire constant region, if potential amino acid deletions or insertions are taken into account. In contrast, the tripeptide acceptor sequence for GlcN oligosaccharides attachment (N-X-T/S) is not positionally conserved (data not shown).

RACE amplifications of rat IgD heavy chain tail encoding sequences

Since the transmembrane tail of the rat IgD heavy chain was not known, the 3′ RACE technique was employed to clone both the transmembrane and secreted tails. The first RT–PCR amplification was performed using a hinge region derived primer, rat IgD Hs, and an anchor primer, cDN, using rat spleen total RNA as a template. The second amplification was subsequently carried out using 1 µl of the PCR mixture from the first amplification, employing the primers rat IgD CH3s (CH3 domain derived) and Fs-common. While the first amplification did not generate any visible bands, the second PCR produced two distinct bands, approximately 1·3 kb and 0·35 kb in size (Fig. 1a). Cloning and sequencing of the two bands showed that the long band represented the transcript for the membrane-anchored form, and the short band corresponded to the secreted form-encoding sequence. The transmembrane transcript has an approximately 1 kb long-3′ UTR sequence, containing (GTTT)₈ and (AC)₂₂ or (CA)₂₂ repeats representing two micro-(mini)-satellite sites (Fig. 2a), that are frequently involved in the regulation of gene expression. Furthermore, an examination of the 3′ UTR region revealed two potential poly(A) addition signal sites, indicated in Fig. 2(a).

Figure 1.

PCR amplifications of the rat Cδ gene fragments. (a) 3′ RACE PCR amplification of rat IgD heavy chain cDNA tails. 1. First PCR; 2, 3. Second PCR; 4. 1 kb DNA ladder. (b) Long PCR amplification of the rat genomic fragments. 1, 1 kb DNA ladder; 2, 3, 4, amplified genomic fragments b, c, d as indicated in Fig. 4.

Figure 2.

Nucleotide sequences of both rat IgD secreted and transmembrane encoding cDNA tails. The coding sequences, repeats and poly(A) addition signals are in bold. The repeats and poly(A) addition signal are underlined. The stop codons are indicated as stars. The sequences have been deposited into NCBI GenBank with accession numbers AY148495–148497.

Two different forms of mRNA transcripts for the secreted form were cloned, showing different lengths of the 3′ UTR and utilizations of two different poly(A) addition signal sites, located approximately 80 bp apart (Fig. 2b,c). However, the two transcripts encode the same peptide sequence of the secreted tail, and most likely, the shorter one is only rarely transcribed in rat spleen, as it is hardly visible on a gel separating the 3′ RACE products (Fig. 1a).

BLAST searches using both the rat secreted and membrane tails against the rat EST database revealed the presence of the same transcripts as we cloned (AW534436, BI291798). One EST (accession number: BI291798) clone contains an (AC)₂₀ repeat, indicating that the mini-satellite site within the 3′UTR is polymorphic in terms of the number of repeats in different animals.

Genomic organization of rat Cδ gene

Because a database for the rat whole genome shotgun sequence is available in the NCBI GenBank, it allows possibilities to map some, if not all, rat genes, by comparing the known cDNA sequence with genomic fragments deposited in the database. Employing this strategy, we found more than 10 genomic fragments containing the rat IgD CH1, Hinge and CH3 encoding sequences. One of these spans the CH1–intron–hinge region (gnl|ti|39491863 rta1lv58.y), allowing us to deduce the exon–intron boundaries and intron size. The latter was also confirmed by sequencing of cloned PCR products, whereas the hinge–CH3, CH3–Sec and Sec–TM genomic fragments were all generated by long PCR amplification using rat spleen DNA as a template and partially sequenced (Fig. 1b). We were thus able to construct a physical map of the gene (Fig. 3), where the intron sizes were determined based on the length of the PCR products. All the exon–intron boundaries were deduced based on comparisons of cDNA and genomic sequences (Table 1).

Figure 3.

Genomic organization of the rat Cδ gene. (a) A genomic fragment obtained from the rat whole genome shotgun sequence (gnl|ti|39491863 rta1lv58.y), and confirmed by sequencing of a cloned PCR product; (b, c, d) Cloned PCR products. B BamHI; E, EcoRI; H, HindIII; K, KpnI; N, NcoI; X, XbaI.

Table 1.

Exon–intron boundaries of the rat immunoglobulin δ gene

	Ex/In	Intron size (kb)	In/Ex
CH1-In-H	AGTTGCCTG/gtgagtatc	0·373 (0·371)	tgactccag/AGACAAGGA
H-In-CH3	ACATAGTAG/gtaagaagt	1·4 (1·353)	tccctgcag/GGGCTATGG
CH3-In-Sec	AAATCAGTG/gtaagtcac	4·6 (4·681)	tgattctag/GATGCTACC
CH3-In-TM1	AAATCAGTG/gtaagtcac	5·9 (6·190)	cttccccag/GCATGGTGG
TM1-In-TM2	TTCATCAAG/gtgggcctc	0·301 (0·301)	ctctggcag/GTGAAGTAG

The intron size indicated in parentheses is based on the sequence derived from the NCBI gen (accession number: AC111740).

As this paper was being revised, a long genomic rough sequence, containing the rat δ gene (accession number: AC111740), was deposited in the NCBI GenBank by the Rat Genome Sequencing Consortium. However, the sequence has not yet been annotated. Analysis of this sequence revealed that the rat Cδ gene is located 2 kb downstream of the µTM2 exon, and the exon–intron boundaries are identical to the data deduced by sequencing of the PCR products. The size of introns (as indicated in parentheses in Table 1) are also consistent with the PCR results. A restriction map of the Cδ gene was constructed based on the sequence and is consistent with the pattern of digestion of the amplified PCR products (Fig. 3).

Phylogenetic aspects of Cδ genes in vertebrates

It is known that the transmembrane tails of IgM in vertebrates are evolutionarily conserved and typically encoded by two exons. As more δ sequences are becoming available, it appears that the δTM show a high degree of sequence similarity to the µTM, consistent with the hypothesis that the Cδ gene was evolutionarily duplicated from the Cµ gene. A thorough comparison of the µTM and δTM sequences may therefore be informative for understanding the evolution of the two genes. As seen in Fig. 4(a) (including the chicken and frog µTM sequences that have not previously published, and were identified from the NCBI GenBank EST database), the known IgM and IgD membrane-bound form in vertebrates possess a nearly invariant intracellular cytoplasmic tail (KVK), with the exception of catfish δTM (KVKIA), salmon δTM (KTK) and cod δTM (KTK). In salmon and cod, the neutral amino acid, valine, is substituted by threonine^8,9 and the catfish IgD molecule contains two extra amino acids in the cytoplasmic tail.⁷

Figure 4.

Phylogeny of the δTM and µTM in vertebrates, (a) Sequence comparison of δTM and µTM from vertebrates. Dashes indicate potential deletions. The rat δTM sequence that was obtained from this study, and the rat µTM (BF551033), frog µTM (BI314212) and chicken µTM (AJ444200, BG625537, AJ395139, AJ395400 and AJ395578) were identified by us from the NCBI GenBank EST database. All other sequences were based on the following references: mouse µTM,²⁴ rabbit µTM,²⁵ cow µTM,²⁶ pig µTM,²⁷ sheep µTM,²⁸ human µTM, δTM,²⁹ duck µTM,⁶ catfish µTM,³⁰ salmon µTM,³¹ cow δTM, sheep δTM, pig δTM,⁴ mouse δTM,³² catfish δTM,⁷ salmon δTM⁸ and cod δTM.⁹ (b) An unrooted phylogenetic tree constructed using both the IgD and IgM transmembrane amino acid sequences. The tree was constructed using propars program, and taken from 1000 trees by running consenus from the phylip package.¹²

A distinct usage of stop codons in δ and µ heavy chain transcripts was also revealed when the known mammalian IgM and IgD heavy chain transcripts were compared. As shown in Table 2, the opal codon UGA is utilized in all the known secreted and membrane-bound form encoding transcripts of IgM, while the amber codon UAG appears to be used in the IgD membrane-bound encoding transcripts. None of these transcripts in mammals adopts the ochre codon UAA as a termination codon, although it is preferentially used in bird IgM transcripts. In contrast to mammals, all three teleost fish species investigated to date frequently utilize the opal codon UGA as a stop codon in IgD membrane-bound form encoding transcripts instead of the amber codon that is utilized by secreted forms of µ encoding transcripts. The differential usage of these termination codons in IgM and IgD transcripts may reflect an evolutionary difference and that the expression of IgM and IgD may be differentially regulated in vertebrates.

Table 2.

Stop codon usage of the known IgD and IgM heavy chain transcripts in vertebrates

Species	δTM	δSec	µTM	µSec
Cow	TAG		TGA	TGA
Sheep	TAG		TGA	TGA
Pig	TAG	TAG	TGA	TGA
Human	TAG	TGA	TGA	TGA
Mouse	TAG	TGA	TGA	TGA
Rat	TAG	TGA	TGA	TGA
Rabbit			TGA	TGA
Hamster				TGA
Horse				TGA
Cat				TGA
Chicken			TGA	TAA
Duck			TAA	TAA
frog			TAA	TAG
Cod	TGA		TGA	TAG
Salmon	TGA		TGA	TAG
Catfish	TGA	TAG	TGA	TAG

The rat µTM (BF551033), chicken µTM (AJ444200, BG625537, AJ395139, AJ395400 and AJ395578) and frog µTM (BI314212) were identified by us from the NCBI EST database, and the rat δTM, δSec was derived from this study. All other sequence data in this table are based on the following references: cow δTM,⁴ cow µTM, µSec,²⁶ sheep δTM,⁴ sheep µTM,²⁸ sheep µSec,³³ pig δTM, δSec,⁴ pig µTM, µSec,²⁷ human δTM, δSec, µTM, µSec,²⁹ mouse δTM, δSec,³² mouse µTM, µSec,²⁴ rat µSec,³⁴ rabbit µTM, µSec,²⁵ hamster µSec,³⁵ horse µSec,³⁶ cat µSec,³⁷ chicken µSec,³⁸ duck µTM, µSec,⁶ frog µSec,³⁹ cod δTM,⁹ cod µTM, µSec,⁴⁰ salmon δTM,⁸ salmon µTM, µSec,³¹ catfish δTM, δSec,⁷ catfish µTM, µSec.³⁰

As shown in Fig. 4(a), both the δTM and µTM sequence in vertebrates are well conserved from the lower to the higher vertebrates. We have constructed a phylogenetic tree using the peptide sequences presented in Fig. 4(b). The unrooted phylogenetic tree displayed a distinct IgD clade from the IgM, indicating that the vertebrate Cδ gene was duplicated from the Cµ gene early in evolution, probably before divergence of the teleost fish from the lineage leading to mammals, approximately 400 million years ago.^13,14

Discussion

Rodent IgD is distinguished by a short hinge region and a lack of the δCH2 domain^3,10,15,16 but the biological consequence of the short IgD heavy chain remains unclear. The absence of δCH2 domain and in rats and mice is probably caused by a genetic event that deleted a pre-existing genomic sequence containing δCH2 and one of the hinge encoding exons during evolution. However, a sequence comparison of human and mouse Cδ loci did not generate evidence to support this view¹⁷ because of the low sequence similarity between them. An examination of the Cδ gene from species such as hamster and guinea-pig, which are phylogenetically closer to rats and mice, may provide useful information for understanding the Cδ gene evolution in rodents.

As more and more Cδ gene sequences are being identified in a variety of vertebrate species, several interesting points regarding their phylogeny, expression patterns and structural diversities have emerged. In contrast to the previously held view, IgD has a wide distribution in vertebrate species, and it is probably present in most, if not all mammals, implicating that IgD is a functionally important antibody class. Our previous phylogenetic analysis indicated that teleost fish δ genes were true homologues to the mammalian genes⁴ suggesting that IgD may also be present in birds, amphibians and reptiles, which are evolutionarily located between teleost fish and mammals. However, a δ gene does not seem to be present in birds according to investigations on the bird IGHC gene loci.^5,6 In addition, a survey of an EST database representing the African clawed frog expressed genes (containing more than 220 000 clones) showed that there were no putative antibody classes other than the three known IgM, IgY, IgX (unpublished data). Furthermore, no translated sequence displayed a higher similarity to the conserved IgD transmembrane domain sequence, than the defined frog IgM transmembrane sequence. However, before a conclusion that the IgD is absent in these species is drawn, there are still some points that need to be taken into account. One major key factor is that the topology of the bird IGHC loci are distinct from the mammalian loci, as the Cα gene is situated in an inverted transcriptional orientation between the Cµ and the Cυ genes.^5,6 The unusual structure of the bird IGHC locus indicates that it may have undergone gene rearrangements during evolution, which may have deleted the Cδ gene or translocated it to downstream of the Cυ gene.

The second important point concerning IgD in vertebrates is their structural diversities. In teleost fish, up to 10 domain-encoding Cδ exons, resulting from the multiple duplications, have been observed.^7,8 Exon deletions have also played a role in the formation of the Cδ gene in Atlantic cod.⁹ Furthermore, in teleost fish, the hinge region has not yet been developed. In mammals, the lack of the δCH2 domain makes rodent IgD shorter than other known mammlian IgD molecules. Rodents also display a short, one-exon encoded, hinge region, while the hinge regions are encoded by two exons in most other mammalian species. In the pig, however, we recently found a one-exon encoded IgD hinge region, although it only shows a low degree of similarity to the rodent IgD hinge,⁴ suggesting that they arose by independent genetic events.

In all known cases, the Cδ genes are located immediately downstream of the Cµ gene. However, consistent with their structural diversities, the expression of the Cδ genes display several distinct pathways, which can basically be classified into two major categories; cotranscription with the Cµ gene and class switch recombination (CSR). The former mechanism, which is performed through alternative splicing of a long Cµ and Cδ containing transcript, is more frequently employed in the expression of mammalian Cδ genes.¹⁸ In teleost fish, however, the µCH1 is spliced onto the δ sequence to form a chimeric antibody heavy chain,^7–9 a mechanism that has never been observed in primates or rodents. CSR is a mechanism involving in expression of other antibody classes than IgM and IgD in mammals¹⁹ and only in very rare cases, direct switching to IgD has been observed in humans and mice.^20–22 The reason is the lack of an authentic switch delta region in these species.¹⁷ However, we have recently demonstrated in the cow, that there is a short Sδ region located in the Cµ and Cδ intronic region, which can direct switching.⁴

Our recent findings support the notion that IgD is an ancient antibody class, present in most mammals, indicating that there are probably other unidentified biologically important functions for IgD, in addition to its ability to replace the function of IgM in driving development of B cells as previously demonstrated in µ gene knockout mice.²³ More work is however, required to generate a complete blueprint of the role of IgD during B-cell development in order to understand the functional importance of this antibody class.

Abbreviations

CSR

class switch recombination

EST

expressed sequence tag

RACE

rapid amplification of cDNA ends

UTR

untranslated region