The evolutionary origins of mammalian immunoglobulin H chain isotypes (IgM, IgD, IgG, IgE, and IgA) are still incompletely understood as these isotypes differ considerably in structure and number from their counterparts in nonmammalian tetrapods. We report in this study that the platypus (Ornithorhynchus anatinus) Ig H chain constant region gene locus contains eight Ig encoding genes, which are arranged in an μ-δ-ο-γ2-γ1-α1-ε-α2 order, spanning a total of ∼200 kb DNA, encoding six distinct isotypes. The ο (ο for Ornithorhynchus) gene encodes a novel Ig H chain isotype that consists of four constant region domains and a hinge, and is structurally different from any of the five known mammalian Ig classes. This gene is phylogenetically related to υ (ε) and γ, and thus appears to be a structural intermediate between these two genes. The platypus δ gene encodes ten heavy chain constant region domains, lacks a hinge region and is similar to IgD in amphibians and fish, but strikingly different from that in eutherian mammals. The platypus Ig H chain isotype repertoire thus shows a unique combination of genes that share similarity both to those of nonmammalian tetrapods and eutherian animals and demonstrates how phylogenetically informative species can be used to reconstruct the evolutionary history of functionally important genes.
Immunoglobulins are the key components of the adaptive immune system in jawed vertebrates including mammals, birds, reptiles, amphibians, bony fish, and cartilaginous fish (1, 2). Eutherian (placental) mammals are known to express five classes of Igs including IgM, IgD, IgG, IgE, and IgA, among which only IgM, IgD, and IgA (or its functional equivalent, like amphibian IgX) are found in nonmammalian tetrapods such as birds, reptiles, and amphibians (1, 2, 3, 4, 5, 6, 7). Although IgM is structurally conserved in all tetrapods thus far investigated, both mammalian IgD and IgA differ from their respective counterparts in nonmammalian tetrapods by having a structurally flexible hinge and fewer numbers of constant (C) region domains.
IgD is an enigmatic Ig class, primarily because its function remains elusive. Until an ortholog was discovered in the teleost catfish (Ictalurus punctatus) a decade ago (8), IgD had been thought to be present only in primates and rodents (9). Recently, the gene encoding IgD was identified in various classes of animals including bony fish, amphibians, reptiles, and additional eutherian mammals such as pigs, cow, sheep, panda, and dogs (3, 4, 6, 10, 11, 12), thus showing a distribution across a wide phylogenetic spectrum. Based on analysis of the δ gene in Xenopus tropicalis and bony fish (3, 13), investigators uncovered a phylogenetic relationship of IgD with IgW (found in lungfish and cartilaginous fish; Ref. 14 , 15), suggesting that the IgD/W lineage is as evolutionarily ancient as IgM. Therefore, the absence of the δ gene in birds (16, 17) and rabbits (18) is probably an outcome of evolutionary loss. Unlike IgM which is structurally conserved across all jawed vertebrates, IgD exhibits extensive variation, particularly in size. The length of IgD can range from two heavy chain constant region domains (CH)5 to more than ten CH domains in different species (8, 9, 10, 13, 19, 20, 21). In addition to the length variations, in some species, such as catfish and pigs, IgD can also be expressed as a chimeric Ig H chain with an inclusion of the IgM CH1 (8, 22). Compared with their counterparts in nonmammalian vertebrates, IgD molecules from eutherian mammals contain fewer CH domains and a structurally disordered, flexible hinge region which is not seen in nonmammalian vertebrates. However, the evolutionary process that led to shortening of the IgD molecule and creation of the hinge region remains unclear.
The mammalian IgG and IgE molecules are believed to have evolved through gene duplication and subsequent evolution of IgY (23), an ancient, four-CH-domain, hinge-lacking Ig class found in birds, reptiles and amphibians. Evolution of IgY to IgE seems to be straightforward, as both Igs have four CH domains and no hinge, and essentially share sequence homology and similar functions. However, no evidence is available to explain how the hinge of mammalian IgG evolved from IgY, although two hypotheses have been proposed. The first suggests that the IgG hinge region was evolutionarily condensed from the pre-existing IgY CH2 (23), while the second argues that the hinge originated from the 3′ sequence of an intron (24).
Mammals (class Mammalia) are taxonomically classified into Prototheria (monotremes) and Theria (Metatheria: marsupials, and Eutheria: eutherians). Monotremes contain only five living species, the platypus and four species of echidna. The platypus is indigenous to Australia and displays an unusual mixture of mammalian and reptilian features, both at the genome level (25) and anatomically: females lay eggs but also provide colostral milk for their pups, and males are venomous. Monotremes last shared a common ancestor with therian mammals ∼166 million years ago (26), placing them in a prime position for reconstructing the evolutionary history of biologically important molecules.
The duck-billed platypus (Ornithorhynchus anatinus) has previously been shown to express four Ig classes: IgM, IgG1/IgG2, IgA1/IgA2, and IgE, that all structurally resemble their respective counterparts in eutherian mammals (27, 28). To explore the evolutionary origins of the mammalian Igs, we analyzed the platypus Ig heavy (H) chain constant region gene locus and identified a novel Ig isotype encoding gene (IgO, O, or ο for Ornithorhynchus) and a ten-CH domain encoding δ gene.
Materials and Methods
Platypus spleen cDNA library, cDNA, and bacterial artificial chromosome (BAC) clones
The platypus spleen cDNA library was previously constructed by Vernersson et al. (28). BAC clones (CH236-821N20, CH236-99D18, and CH236-647E09) were purchased from BACPAC Resource Center (BPRC) at Children’s Hospital Oakland Research Institute in Oakland, California. All platypus tissues were collected opportunistically from animals found dead in the wild. Tissues were frozen at –80°C before DNA/RNA extraction. RNA was extracted using TriReagent according to the manufacturer’s instructions (Molecular Research Center). RNA samples were subjected to DNase digestion using standard protocols (Promega). The Invitrogen SuperScript III First-Strand Synthesis System for RT-PCR was then used to make cDNA using the Oligo(dT)20 primer as per the manufacturer’s instructions (Invitrogen).
BLAST searches were performed against the platypus genome sequence deposited in the National Center for Biotechnology Information GenBank (http://www.ncbi.nlm.nih.gov/genome/guide/platypus/). Genomic contigs were retrieved for further analysis. To search for putative Ig domains, the genomic DNA sequences were translated into proteins, which were subsequently subject to protein to protein BLAST searches in the NCBI GenBank. The platypus BAC end sequence database (http://genome.wustl.edu/tools/blast/index.cgi?gsc_link_id = 69) is maintained by the Genome Sequencing Center at Washington University School of Medicine.
Pulsed field gel electrophoresis (PFGE)
Purified BAC DNA was digested using NotI and separated on a 1% agarose gel using a CHEF-DR III system (Bio-Rad). The DNA marker used was Midrange I PFG Marker (New England Biolabs). According to PFGE data, the insert size in CH236-821N20, CH236-99D18, and CH236-647E09 were shown to be ∼120, 135, and 150-kb, respectively.
Amplification of IgD and IgO constant region cDNA
The platypus IgD constant region cDNA (membrane-bound form) was amplified using a nested PCR amplification with the primers, JHs1 (5′ CAC TGG GGC CAA GGC ACC ATG GT 3′), JHs2 (5′ GGC ACC ATG GTC ACC ATG GT 3′), IgDTMas1 (5′ GGC TTC CTC ACT GGT GGG CAT AG 3′), and IgDTMas2 (5′ AGG GCT ACG AAG GTG GAC ACG GT 3′). The resulting 3.3-kb PCR product was directly sequenced. The IgO C region encoding cDNA was amplified using a nested PCR from a duck-billed platypus cDNA library using the primers JHs1 and IgOas1 (5′ TTT GCC GCA CTG TCT TCT GTA TG 3′), JHs2 and IgOas2 (5′ TGC AGG TGA AGG AAT CTC CGC GT 3′). In both amplifications, the DNA polymerase used was LA TaqDNA polymerase (Takara, Dalian), a proofreading enzyme.
PCR detection of the platypus Ig gene expression in different tissues
cDNA samples derived from seven tissues including gastrointestinal tract, kidney, liver, lung, reproductive tract, spleen, and testis, were used in PCR detection using normal TaqDNA polymerase (Tiangen Biotech). PCR primers were: IgM-detections: 5′ ACA AGC CTA TTC CAC GAC CTC 3′; IgM-detectionAs: 5′ GTT GAA GTG CTT GGC CTG ACA 3′; IgE-detections: 5′ GCG CTG GAA CTC TGG AGA CCT GG 3′; IgE-detectionAs: 5′ TCC TGG CAC TGT CTT GGT AGG TT 3′; IgD-detections: 5′ CTT AGC CAG TGT CCA GAT CAG 3′; IgD-detectionAs: 5′ GCA GAG CAG AGT CGG ATG ATC 3′; IgO-detections: 5′ TTC TAC AAC AGC AAA GCC CAG AC 3′; IgO-detectionAs: 5′ GGG CAA TGA AGA GGT CAG TAG GG 3′; IgG1-detections: 5′ GTA CCA AGG CGA CAA GGA ACT 3′; IgG1-detectionAs: 5′ TGT TCA TCT CAT CTC GGT GAG 3′; IgG2-detections: 5′ CGC ACA GAC GTC ATC AGA TAC 3′; IgG2-detectionAs: 5′ TGC TCA CAG ATT CCT TGC TCA 3′; IgA1-detections: 5′ GAC AGT GAC CAG GCT TCT TCC 3′; IgA1-detectionAs: 5′ GGA GAG CAC TTG CAA CAC TGT 3′; IgA2-detections: 5′ CTG ACG ACA AGA CCT ACC AAT 3′; IgA2-detectionAs: 5′ GTT AGT TCA CAG GTC AGG TTG 3′; platypus-GAPDHs: 5′ ACC GGC CAA ATA CGA TGA TAT 3′; platypus-GAPDHAs: 5′ TAC CAG GAG ATG AGC TTG ACA 3′.
DNA computations, structural prediction, and construction of phylogenetic trees
DNA and protein sequence editing, alignments and comparisons were performed using the DNAstar program (DNAstar). Phylogenetic trees were made using MrBayes 3.1.2 (29) and viewed in TREEVIEW (30). Multiple sequence alignments were performed using Clustalw. Protein disordered regions were predicted using a series of online programs (http://www. disprot.org/predictors.php).
Identification of the genomic sequences encoding IgM (μ), IgG1/G2 (γ1/γ2), IgA1/A2 (α1/α2), and IgE (ε) in the platypus
The platypus has previously been shown to express four Ig classes, encoded by six genes: IgM (μ), IgG1/G2 (γ1/γ2), IgA1/A2 (α1/α2), and IgE (ε) at the cDNA level (27, 28). Using these cDNA sequences as templates, we performed BLAST searches against the platypus genomic sequences deposited in the NCBI GenBank. This allowed us to identify a 680-kb genomic contig (NW_001794226) containing those previously known μ, γ1, ε, α2. The γ2 and α1 genes were found on two other short contigs (NW_001705893.1, NW_001777603.1, the latter contig contains only a part of the α1 genomic sequence), which are not assembled into the 680-kb contig in the current version of the platypus genome assembly. Alignments of these identified genomic sequences with their respective cDNA revealed that all genes had a similar genomic organization as their corresponding genes in eutherian mammals (Fig. 1).
Identification of the platypus IgD encoding gene
There is an ∼90-kb of DNA sequence (including sequence gaps with unknown sizes) between the μ and γ1 genes in the 680-kb contig, which is of sufficient size to accommodate additional Ig isotype encoding genes. Approximately 4.3-kb downstream of the μTM2 exon, an Ig CH domain encoding exon was identified. Further downstream, nine additional Ig CH domain-encoding exons and two typical IgM/D transmembrane exons were also observed, indicating the presence of a 10 CH domain-encoding Ig gene (Fig. 1). Blast searches using the deduced amino acid sequence revealed that it shared a relatively high degree of homology with IgD from fish, Xenopus, leopard gecko, and some mammals. Together with phylogenetic analyses (Fig. 2, supplementary Fig. 1),6 these data strongly suggest that the identified gene is the platypus δ.
We failed to amplify the δ-encoding cDNA from a platypus cDNA library (28), probably due to its low rate of expression. We thus resorted to a nested RT-PCR directly using total RNA as a template and primers derived from JH and the predicted TM1. An ∼3.3 kb PCR product was successfully amplified and directly sequenced, and was shown to contain all the ten predicted CH exons (supplementary Fig. 2). This clearly suggests that the platypus IgD can be expressed as a membrane-bound form consisting of multiple CH domains. However, we did not succeed in cloning a secreted form encoding cDNA. Like in most other mammals, the platypus δ gene is most likely expressed through alternative splicing of a long primary transcript which consists of both the δ and μ genes, as no putative switch region sequence was identified in the μ-δ intron (data not shown) (31).
Despite its apparent absence in birds (chickens and ducks), IgD has recently been identified in some reptiles (Ref. 6 , and our unpublished data) showing a roughly similar size (11 CH domains) as the platypus IgD. We subsequently performed a domain to domain sequence comparison of IgD between the platypus and two reptiles (leopard gecko and the green anole lizard), that revealed that nearly all the 10 platypus IgD CH domains have a corresponding homologous CH domain in both reptile IgD molecules (supplementary Table 1), supporting the view that these molecules are related.
A domain-to-domain sequence comparison revealed that the CH1, CH6, and CH7 domains of the platypus IgD are homologous to the CH1, CH2, and CH3 of IgD of eutherian mammals, respectively (supplementary Table 2 and Fig. 3), indicating that this molecule was shortened by a selective loss of CH domains during evolution. IgD in eutherian mammals usually contains a structurally flexible hinge in the H chain constant region (9, 12, 22). However, a detailed examination of the platypus IgD sequence reveals no hinge region (Fig. 3), indicating that the IgD hinge developed after divergence of therian mammals from monotremes.
Identification of a novel Ig gene in the platypus
In addition to the above mentioned five genes including the μ, δ, γ, ε, and α, a sixth H chain isotype encoding gene was unexpectedly found downstream of the δ. This gene consists of four CH and two TM encoding exons and is structurally similar to μ and ε. However, it has only 29.4 and 45.1% amino acid identity to the platypus IgM and IgE, respectively. We termed this gene ο (o for Ornithorhynchus anatinus, encoding IgO) (Fig. 1). RT-PCR shows that IgO is exclusively expressed in the spleen (Fig. 4). We cloned the IgO C region encoding cDNA using a nested PCR-amplification of a duck-billed platypus spleen cDNA library (Fig. 5). Surprisingly, alignment of the IgO C region with that of the platypus IgE revealed an extended CH2 domain at the N-terminal (Fig. 6). This extended sequence is abundant in prolines and very similar, in amino acid composition, to hinges of mammalian IgA and IgG (supplementary Fig. 3). Proline rich regions usually play a structural role in proteins as spacers and are normally devoid of a secondary structure, i.e., being disordered or forming a random coil. Definite Ig hinges usually display an extended (at least partially), disordered protein structure, conferring structural flexibility (32). A series of prediction programs suggested that the extended peptide in the IgO CH2 region is structurally disordered, as exemplified in supplementary Fig. 4. Therefore, it is likely to form a hinge unit together with the first cysteine residue in the CH2 domain. An Ig class with four CH domains plus a hinge has thus far not been observed in any species, suggesting that IgO represents a distinct Ig class in mammals.
IgO, a structural intermediate between IgY(E) and IgG?
To determine the phylogenetic position of IgO, we performed thorough phylogenetic analyses using its separate CH domains or the entire constant region sequence. All analyses showed that IgO is phylogenetically related to IgG, IgE, and IgY (relatively closer to IgG) (Fig. 2 and 7, supplementary Fig. 1). Despite the close phylogenetic relationship to IgG, two other lines of evidence, in addition to the obvious structural difference (4CH vs 3CH), support the notion that the IgO is not a subclass of IgG, but a distinct Ig H chain isotype. All mammalian IgG molecules contain a functionally important N-linked glycosylation site (Asn 297) in the CH2 domain. This site is conserved in homologous regions of IgM, IgE, and IgD (12, 33, 34). This site is absent in the corresponding position of the IgO constant region (Fig. 6). Moreover, the IgO hinge region lacks cysteines that are often seen in IgG hinges and used to covalently hold heavy chains. Taken together, these analyses support the notion that likely, IgO represents a structural intermediate between IgY (E) and IgG. Phylogenetic analysis indicates that the CH1, CH2, and CH3 domains of IgG correspond to the CH1, CH3, and CH4 of IgO or IgY(E), respectively (supplementary Fig. 5). The finding of a hinge region in the CH2 of IgO suggests that, in the evolutionary route from a four CH domain Ig to a three CH domain IgG, a hinge region was first developed at the N-terminal of the CH2 (of an IgY like Ig) domain and afterward, the CH2 (of an IgY like Ig) was lost. This is consistent with the hypothesis proposed by Blattner et al. (24) that the Ig hinge region may have evolved by duplication, leading to incorporation of a 3′ RNA splice site into the coding portion of a CH domain, and later, mutations creating a 5′ RNA splice site at the junction of the hinge region and the CH domain, caused detachment of the hinge exon and nonfunctionality of the CH domain involved (24). Examination of the putative IgO hinge encoding sequence did indeed reveal features of splice sites, where the 3′ end of the hinge encoding sequence is pyrimidine rich and terminated by a conserved AG di-nucleotide, exhibiting features of a 3′ splice site (supplementary Fig. 6). Immediately downstream of the hinge sequence, two cryptic 5′ splice sites (GTGAG, GTGGT) can be identified (supplementary Fig. 6). To test whether these two 5′ splice sites are functional, we designed two sense primers by fusing nucleotides upstream of the two splice sites respectively, with the first five nucleotides of the CH3 exon. These two primers were used in nested RT-PCR amplifications (data not shown), and showed that the second cryptic 5′ splice site (GTGGT) was functional (albeit less efficient) in RNA splicing based on sequencing of the PCR products. Thus, the hinge sequence can be directly spliced onto the CH3 exon, generating a transcript (VDJ-CH1-hinge-CH3-CH4) that structurally resembles an IgG H chain (supplementary Fig. 7). However, this alternatively spliced variant appears to be nonfunctional due to a shift of the open reading frame (supplementary Fig. 8). Nevertheless, it provides data to support the above-mentioned model, explaining how a genetic hinge (encoded by a separate exon) could be developed.
Localization of the γ2 and α1 genes in the platypus immunoglobulin heavy chain constant region (IGHC) gene locus
Although the genomic sequences of the platypus γ2 and α1 could be found in the genome database, they were not assembled in the main Ig gene locus (680-kb contig). This could be due to the presence of sequence gaps within the genomic contig, as in nearly all tetrapods so far investigated, all the IGHC genes are arranged in a single locus in tandem. To determine the localizations of the γ2 and α1 genes, we used several BAC clones that covered the entire IGHC gene locus. Based on BLAST searches against the platypus BAC end sequence database, BAC clone CH236-821N20 was shown to contain the μ, δ, o, and γ1 genes. Clone CH236-99D18 was physically linked to the 3′ end of CH821N20, while CH236-647E9 was located ∼12-kb downstream (Note: there is a long gap with an unknown size in between) (Fig. 1). PCR experiments showed that CH236-99D18, but not other two clones, contained the α1 gene. As there is no α1 gene in the long determined sequence downstream of the α2 in both CH236-99D18 and CH236-647E9, the most plausible location for α1 is in the gap between CH236-821N20 and CH236-647E9 (Fig. 1). This was confirmed by sequencing of PCR products amplified using primers derived from the α1 and the sequences flanking the gap. The genomic organization of the α1 was deduced by PCR and sequencing.
According to our PCR results, CH236-821N20 was shown to contain the γ2 gene. To determine the location of the γ2 gene, a couple of primers were designed covering all sequence gaps from the δ to the γ1 gene, and used together with the primers derived from γ2 in PCR amplifications using CH236-821N20 BAC DNA as a template. By sequencing of the obtained PCR products, it was deduced that the γ2 gene was located in the first gap downstream of the ο gene (Fig. 1). These data allowed us to conclude that the platypus IGHC genes are arranged in an μ-δ-ο-γ2-γ1-α1-ε-α2 order (Fig. 1).
Supporting functionality of the identified platypus IGHC locus, a JH (Ig H chain joining gene segment) locus was identified ∼10-kb upstream of the μ gene. The JH locus spans ∼4.2-kb DNA and contains 10 structurally functional JH gene segments and 1 pseudo JH (Fig. 8). The DH and VH gene loci could not be analyzed as there are too many sequence gaps upstream of the JH locus.
In the present study, we identified two Ig H chain isotypes in the duckbilled platypus, IgD and IgO, in addition to the previously known IgM, IgG1/G2, IgE, and IgA1/A2 classes, showing that the platypus expresses eight Ig H chain constant region genes, arranged in an μ-δ-ο-γ2-γ1-α1-ε-α2 order (Fig. 1). This also demonstrates that all Ig H chain isotypes expressed by eutherian mammals are present in monotremes even though these two mammalian lineages diverged ∼166 million years ago (26).
Although it has been known for some years that monotremes express typical mammalian IgM, IgG, IgA, and IgE (27, 28, 35, 36, 37, 38), the presence of an ortholog of IgD has remained elusive. We have previously tried, but failed, to clone the δ gene in the platypus using a PCR-based approach and degenerate primers covering the conserved δ gene sequence of eutherian mammals, suggesting that if it was present, it should be distinct from that of its counterparts in eutherian mammals. In this study, we show that the platypus contains a δ gene strikingly different from the δ gene in eutherian mammals. It does, however, share great similarity to the δ gene in nonmammalian vertebrates such as reptiles, amphibians, and fish. In eutherian mammals, IgD typically contains three CH domains and a hinge segment (9), which is much shorter than its counterparts in nonmammalian vertebrates, which usually have eight or more CH domains but no hinge (3, 4, 8). The platypus δ gene encodes ten CH domains, representing the first long form of IgD molecule identified in mammals. The discovery of such an IgD gene has evolutionary implications as: 1) it establishes a clear relationship between the short IgD molecules in eutherian mammals and the long IgD molecules in nonmammalian vertebrates and 2) it shows that the platypus has a unique combination of Ig genes which share similarity both to those of nonmammalian vertebrates and eutherian animals.
Sequence comparisons suggest that the platypus δ gene shares an immediate ancestor with the recently identified δ gene in a reptile, leopard gecko (6), as all its 10 CH exons are homologous to exons found in the reptile (gecko) δ gene (containing 11 CH encoding exons). We also recently identified the δ gene in another reptile, the green anole lizard (our unpublished data), displaying a similar structure to that of the gecko gene by having 11 CH encoding exons. Presence of these closely related δ genes in phylogeny in both mammals and reptiles, strongly suggests that the absence of the δ gene in birds (7, 16) is due to a recent genetic modification that occurred after their divergence from reptiles.
The platypus IgD does not contain a hinge region. Sequence comparisons suggest that the δCH1, δCH2, and δCH3 of eutherian mammals share homology with the δCH1, δCH6, and δCH7 of the platypus, suggesting a selective loss of other CH exons during the evolution. To provide insights into the timing of the loss of these domains, we searched for a δ gene in a marsupial, the gray short-tailed opossum, by using its genome sequence. Typical mammalian μ, γ, ε, and α genes are found in tandem in a single locus (data not shown). In the 200-kb of DNA between the opossum μ and γ genes (where the δ usually located), we did not find any evidence of a δ gene (or even any putative Ig encoding sequences). A δ gene was also not found anywhere in the entire opossum genome, suggesting an absence of the gene in this species. A future search for the IgD encoding gene in additional marsupials might be expected to provide more clues to this evolutionary issue.
The ο gene represents another interesting finding in platypus. It encodes four Ig CH domains and a hinge attached to the N-terminal of CH2, giving rise to an Ig H chain constant region that has thus far not been observed in vertebrates. Our data suggest that it might be an immediate evolutionary precursor of IgG. This hypothesis is based on several facts: 1) the IgO encoding gene, ο, is physically located in the position where a γ gene is usually found in the mammalian IgH locus; 2) the ο gene is phylogenetically related to the mammalian γ gene; and 3) the ο gene encodes a hinge segment. IgG is thought to have originated from IgY, an ancient, hinge-lacking, four CH domain Ig isotype. Compared with IgY, mammalian IgG usually has a hinge segment and one CH domain less (a loss of CH2 of IgY). The structure of IgO suggests that the hinge was formed before the loss of a CH domain. This hypothesis agrees with a model that explains the origin of the murine IgA hinges (24). The pivotal point in that model is that the Ig hinge region was originally derived from the noncoding 3′ end sequence of an intron that was rich in pyrimidine (especially cytidine), potentially encoding the prolines that are often seen in Ig hinge regions. This originally noncoding stretch of nucleotides was subsequently attached to the 5′ end of a CH exon and inserted in frame with the reading frame of the CH exons. A final loss of the CH exon (or a gain of a separate genetic hinge exon) involved could be achieved through mutations of hinge-CH junctional sequence generating a 5′ RNA splice site, which would result in a removal of the CH exon in mature message RNA. Nonfunctionality of the CH exon would finally cause it to be gradually lost, forming the genomic structure of the mammalian γ genes.
In summary, the discovery of IgO and IgD in the platypus highlights the important position this species holds in mammalian phylogeny and provides unique insights into the evolution of the mammalian Ig repertoire. It also provides support for the continuous evolution of immunoglobulins in vertebrates and the importance of comparative studies for reconstructing the evolutionary history of the immune system.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Erin Noonan, from the Tasmanian Department of Primary Industries and Water for providing platypus tissue samples. This work was supported by the National Science Fund for Distinguished Young Scholars (30725029), Program for New Century Excellent Talents in University of China (NCET-06-0117), the National Natural Science Foundation of China (30671497), the National Key Basic Research Program (NKBRP) (2006CB102100), the Swedish Research Council and the University of Sydney Research and Development Grant (to K.B.).
The IgD and IgO cDNA sequences reported in this study have been deposited in the National Center for Biotechnology Information GenBank under the accession numbers: EU503149-EU503150.
Abbreviations used in this paper: CH, heavy chain constant region domain; IGHC, immunoglobulin heavy chain constant region gene; PFGE, pulsed field gel electrophoresis; TM, transmembrane; BAC, bacterial artificial chromosome.
The online version of this article contains supplemental material.