The major histocompatibility complex (MHC) class II DRB, DQB, DPB, and DOB gene clusters are shared by different eutherian orders. Such an orthologous relationship is not seen between the β genes of birds and eutherians. A high degree of uncertainty surrounds the evolutionary relationship of marsupial class II β sequences with eutherian β gene families. In particular, it has been suggested that marsupials utilize the DRB gene cluster. A cDNA encoding an MHC class II β molecule was isolated from a brushtail possum mesenteric lymph node cDNA library. This clone is most similar to Macropus rufogriseus DBB. Our analysis suggests that all known marsupial β-chain genes, excluding DMB, fall into two separate clades, which are distinct from the eutherian DRB, DQB, DPB, or DOB gene clusters. We recommend that the DAB and DBB nomenclature be reinstated. DAB and DBB orthologs are not present in eutherians. It appears that the marsupial and eutherian lineages have retained different gene clusters following gene duplication events early in mammalian evolution.
The major histocompatibility complex (MHC) is a large multigene family that plays an important role in the regulation of immune responses in vertebrates. The basic structure and function of these molecules is now generally well understood. MHC genes can be “classical,” indicating that they are polymorphic and highly expressed, or “nonclassical,” meaning that they cluster with other MHC genes phylogenetically, but are usually not expressed at high levels, and are usually not polymorphic (reviewed in Hess and Edwards 2002). Class I genes encode receptors that are present on the surface of nearly all nucleated cells and are mainly involved in the elimination of intracellular pathogens with the help of cytotoxic T cells (Bjorkman and Parham 1990). Class II genes are expressed only in certain cells, such as B cells and macrophages, and are principally involved in the elimination of extracellular pathogens, such as most bacteria (Kappes and Strominger 1988). Helper T cells are involved in this pathway. Class III genes include complement genes as well as other genes involved in antigen presentation and peptide binding (Klein and Sato 1998).
The class II MHC molecules are composed of an α and a β chain, which are noncovalently associated (Brown et al. 1993). These are encoded by the A and B genes, respectively (Bodmer et al. 1990). Eutherians have three main classical class II gene clusters: DP, DR, and DQ, as well as the nonclassical DM and DN/DO gene clusters (Beck and Trowsdale 1999; Hughes and Nei 1990). Each of these clusters contains one or more α chain genes and one or more β chain genes. A close relationship is seen between class II gene regions in eutherians of different orders (Hughes and Nei 1990; Klein and Figueroa 1986). The DRB gene is found in humans, mice, and pigs. These DRB genes have evolved from a common ancestor. However, duplications have occurred within this gene lineage, so that some species have multiple copies of DRB; for instance, humans have nine DRB loci (Campbell and Trowsdale 1993). The resulting loci do not share a strictly orthologous relationship between species, although they do belong to the same gene cluster. Gene deletions are quite common within the mammalian MHC, as can be seen in mole rats that have lost their DO and DR gene clusters (Schopfer et al. 1987), and cats that have lost their DQ gene clusters (Beck et al. 2001).
Because of the cycles of expansion and contraction of the MHC during the evolution of vertebrates (Nei et al. 1997), the designation of orthology in these gene families is particularly difficult. Monophyletic clades containing genes from different species represent orthologs, which have been derived from a single ancestral gene. MHC loci from different eutherian gene families such as DRB, DQB, and DPB produce monophyletic clusters in phylogenetic trees (Hughes and Nei 1990). However, such an orthologous relationship is not maintained over extended periods of time; different loci are maintained independently of each other after gene duplication events. Class II β-chain genes of birds are more closely related to each other than they are to mammalian class II β chains and fall outside the mammalian MHC clade (Hughes and Nei 1990; Takahashi et al. 2000). It has therefore been suggested that the mammalian class II gene clusters arose after the separation of the synapsids and the therapsids.
To date, six marsupial MHC class II β sequences have been published: Macropus rufogriseus (red-necked wallaby) DAB and DBB (Schneider et al. 1991), Monodelphis domestica (gray short-tailed opossum) DRB (Stone et al. 1999) and DMB (O'HUigin et al. 1998), and Trichosurus vulpecula (brushtail possum) DRB1 and DRB 2 (Lam et al. 2001). The relationship of these genes to the eutherian class II gene clusters is the subject of contention. The first article to describe marsupial class II β sequences suggested that M. rufogriseus sequences are not orthologous to the eutherian β-chain gene clusters, but instead represent two new mammalian gene clusters, DAB and DBB (Schneider et al. 1991). In 1999, Stone et al. challenged this conclusion, and argued that the M. rufogriseus DAB sequences, and their own M. domestica sequence, cluster with the eutherian DRB sequences, and are quite distinct from the DQ, DO, and DP gene families. They did not comment on the phylogenetic position of the M. rufogriseus DBB, but on their tree it was basal to the DRB, DQB, and DPB clusters, with the DOB cluster basal to it. Takahashi et al. (2000) obtained a similar result. In 2001, Lam et al. described two T. vulpecula class II β clones, which were named DRB, based on results obtained using FASTA searches and the prior analysis of Stone et al. In this article we use the original nomenclature of DAB for the marsupial DAB/DRB genes. The reason for this will be justified later.
Here we describe a new marsupial MHC class II β sequence and conduct a thorough phylogenetic analysis of classical class II β-chain sequences in order to determine the gene's evolutionary origins. We examine the evolutionary relationship of the marsupial DAB and DBB genes with the eutherian DQB, DPB, DRB, and DOB gene clusters. We did not include DM genes in our analysis, as they represent a distinct evolutionary pathway of class II genes, both in terms of evolution and function (O'HUigin et al. 1998). Instead of presenting antigens, as in other class II molecules, DM molecules are involved in intracellular loading of peptides onto class II molecules (Kropshofer et al. 1999), and phylogenetically there is a high degree of divergence between the DM molecules and other class II genes (Kelly et al. 1991).
Materials and Methods
Isolation of the T. vulpecula DBB Clone
Previously we reported the amplification of a T. vulpecula DAB probe (Lam et al. 2001) using degenerate primers (MHC2F 5′-GTGCGCTTCGACAGCAACGTGGGG-3′; MHC2R 5′-CCACTCAGCATCTTACTCTGGGCAGA-3′) designed from M. domestica and M. rufogriseus DAB gene sequences. This polymerase chain reaction (PCR) product (approximately 1 kb) was used to screen a T. vulpecula mesenteric lymph node cDNA library which was constructed from tissue from a single individual (Belov et al. 1998). Filters were hybridized overnight at 55°C in Rapid Hyb buffer (Amersham). The PCR product was labeled with 32P-dATP (3000 Ci/mmol) (Amersham) using a random primed DNA labeling kit (Roche). The labeled probe was added to the Rapid Hyb buffer and incubated overnight at 55°C with constant rotation. The membranes were washed three times for 5 min in 2× SSC, twice in 2× SSC with 0.1% SDS for 30 min, and twice in 0.5× SSC with 0.1% SDS for 15 min. Secondary and tertiary screens were performed under the same conditions. Eight positive clones were isolated and sequenced in both directions by primer walking on an ABI 377 automated sequencer using Big Dye version II (Applied Biosystems) at Macquarie University. Seven of these were described as DAB (Lam et al. 2001). Here we describe the eighth clone.
Alignment of Sequences
Initially amino acid sequences were aligned using ClustalX with default parameters for gap opening and extension for pairwise and multiple comparisons and specifying substitution probabilities with the PAM 20, 60, 120, and 350 matrices (Thompson et al. 1997). The EMBOSS program Tranalign (http://www.hgmp.mrc.ac.uk/Software/EMBOSS/Apps/tranalign.html) found at ANGIS (http://www.angis.org.au) was used to align the codons of the cDNA sequences according to the amino acid alignment. We found the alignment of nucleotide sequences without this program impossible, as numerous multiresidue gaps were inserted into the alignment, resulting in the wholesale disruption of triplets coding for amino acids. Accession numbers for sequences used in the alignments are shark (GiciB_1 L20274, GiciB_2 L20275), zebrafish (BrreDAB1 L04805, BrreDAB4 U08870, BrreDEB U08874, BrreDCB U08873), Xenopus (XelaB3 D13685), chicken (Gaga2 M29763, Gaga3 M26307), wallaby (MaruDAB1 M81624, MaruDBB M81625), possum (TrvuDAB2 AF312029, TrvuDAB3 AF312030), opossum (ModoDAB1 AF010497), rat (RanoDQB X56596), mole rat (NaehDPB M16685), sheep (OvarDQB1 L08792), pig (SuscDQB_yn AY102478, SuscDRB AY191776), sea lion (ZacaDQB10 AF503406), dog (CafaDQB AF043908), horse (EqcaDQB L33910), human (HosaDQB1 M20432, HosaDRB4 MN_021983, HosaDRB3 NM_022555, HosaDOB L29472, HosaDPB1 M57466), rabbit (OrcuDPB M21468, OrcuDOB M96942), gorilla (GogoDRB1 M77154), chimpanzee (PatrDOB M24358), cat (FecaDRB U51575), and goat (CahiDRB AB008345). The MHC nomenclature used in the article is based on the recommendations of Klein et al. (1990) and involves a four-letter abbreviation of the species' scientific name: the first two letters derived from the genus, the second two from the species. The reader is advised that previously Hosa (human MHC) was known as HLA, Mumu (mouse MHC) as H-2, Rano (rat MHC) as RT1, Bota (cow MHC) as BoLa, Feca (cat MHC) as FLA, Naeh (mole rate MHC) as Smh, Patr (chimpanzee MHC) as ChLa, Gogo (gorilla MHC) as GoLA, Orcu (rabbit MHC) as RLA, Cafa (dog MHC) as DLA, Susc (pig MHC) as SLA, Eqca (horse MHC) as ELA, Ovar (sheep MHC) as OLA, and Cahi (goat MHC) as GLA. Some of these nomenclatures are still encountered in the literature, but for simplicity, only the new abbreviations are used (Klein et al. 1990).
The amino acid alignment was 284 amino acids long, while the nucleotide alignment was 852 nucleotides long and contained only coding region sequence (leader, b1, b2, transmembrane, and cytoplasmic regions). The alignment was generated with 34 sequences—7 eutherian DQB sequences, 3 eutherian DPB sequences, 6 eutherian DRB sequences, 3 eutherian DOB sequences, 4 marsupial DAB sequences, 2 marsupial DBB sequences, 2 chicken β-chain sequences, 1 Xenopus β-chain sequence, 4 zebrafish β-chain sequences, and 2 shark β-chain sequences. Base compositional differences suggested that the shark and zebrafish sequences be excluded from analysis (see below), and the Xenopus sequence was used as the outgroup.
Phylogenetic hypotheses were tested using maximum parsimony, maximum likelihood (ML), and genetic distance. Analyses were performed for nucleotide sequences and inferred amino acid sequences. Unless otherwise stated, analyses using nucleotide sequences only used the first and second positions of codons.
Genetic distance analyses of nucleotide sequences included (1) minimum evolution trees, which were calculated from Hasegawa, Kishino, and Yano (HKY) distances, with a proportion of 0.6 of sites assumed to be invariable and removing identical sites proportionally to base frequencies estimated from all sites, and with likelihood settings set to the PAUP 4.0 defaults for this model; (2) with settings as above, except that transversions only were analyzed. A total of 200 bootstrap replicates were conducted. MEGA2 analysis involved replicating the analysis of Takahashi et al. (2000) using the neighbor-joining (NJ) algorithm and the Kimura two-parameter genetic distance, with 1000 bootstrap replicates. These settings were also used in an analysis limited to transversions.
Maximum parsimony was generally performed using the default parameters in PAUP*4.0b10 (Swofford 2002), with gaps treated as unknown. For heuristic searches, 100 replicates of random taxon addition were performed. Bootstrap pseudosampling was performed with 100 replicates, with 20 replicates of random addition of taxa. When constraints were imposed, heuristic searches employed 100 replicates of random addition of taxa. Analyses taking account of alternative costings of transitions and transversions were performed assuming a relative cost of three for transversion.
For ML analyses of nucleotide sequences, MODELTEST (Posada and Crandall 1998) was used to select a model based on the Akaike information criterion (AIC) for searches. Bootstrap pseudosampling was performed with 100 replicates, with 5 replicates of random addition of taxa using the parameter settings suggested by the AIC results—model selected: TVM + G; base frequencies: A = 0.2493, C = 0.2347, G = 0.2775, and T = 0.2385; substitution model rate matrix: R[A – C] = 1.9954, R[A – G] = 2.8985, R[A – T] = 1.5323, R[C – G] = 2.1336, R[C – T] = 2.8985, and R[G – T] = 1.0000. The proportion of invariable sites was assumed to be 0.0844, and the alpha shape parameter of the gamma distribution was estimated as 1.4386. Comparison of trees using Kishino and Hasegawa (KH) (1989) likelihood tests and Shimodaira and Hasegawa (SH) (1999) likelihood tests assumed a resampling estimated log-likelihood (RELL) approximation, two-tailed for the former test and one-tailed for the latter. Imposed constraints required topologies to show a clade grouping all members of two gene clusters, generally one marsupial and one eutherian, but in one instance both marsupial clusters. Groupings within the imposed clade were not constrained.
A 2446 bp cDNA clone (GenBank accession no. AY271265) was isolated from a T. vulpecula mesenteric lymph node cDNA library. It contains the complete coding sequence of an MHC class II gene, encoding 264 amino acid residues. By comparison with other class II β-chain sequences, it is assumed that the leader peptide is encoded by the first 29 codons, the β1 domain (which contains the peptide binding region) by the next 94 codons and the β2 domain by the following 92 codons (Hughes and Nei 1990). The transmembrane and cytoplasmic regions contain 38 and 11 amino acid residues, respectively. The 3′ untranslated region contains a polyadenylation signal and a polyA tail. The NGT glycosylation site and the RFDS sequence motif in the β1 domain are conserved, suggesting that the molecule is able to bind to CD4 (Mazerolles et al. 1988).
BLAST (basic local alignment search tool) searching of GenBank using default parameters suggested that this molecule is most similar to M. rufogriseus DBB, with an E value of 0.00. The E value gives an indication of the statistical significance of a given pairwise alignment. The lower the E value the more significant the hit. Fifty of the next best matches were eutherian DQB genes (E values of 5e-42 or less), followed by DPB and DRB genes. Marsupial DAB sequences did not appear in the top 90 hits. On the other hand, sequence searching using FASTA found 88% similarity between the T. vulpecula and M. rufogriseus DBB (E = 4.1e-153), followed by 74% identity with Felis catus DRB (E = 2.8e-71), Sus scrofa DRB (E = 1.7e-69), and several other eutherian DRBs, before detecting eutherian DQB genes (E = 6.9e-69). An alignment of the T. vulpecula amino acid sequence with other marsupial MHC class II β sequences is shown in Figure 1. The M. rufogriseus DBB and T. vulpecula clone share 84% identity, while the T. vulpecula clone has only approximately 60% identity with the DAB clones. Despite this, the level of similarity between the class II molecules is very high. Cysteine residues, at positions 44 and 108 in the β1 domain, and 144 and 200 in the β2 domain are conserved. The conserved NGT and RFDS residues, necessary for interaction with CD4 (Brogdon et al. 1998), are boxed in Figure 1. It is worth noting that the T. vulpecula clone and the M. rufogriseus DBB β2 domains are two amino acids shorter than the β2 domains of the marsupial DABs. That is, they share an indel, which is not present in any of the other vertebrate sequences used in our dataset or, to our knowledge, in any vertebrate MHC class II β sequence. Based on the BLAST searches and studying the alignments, we assign this gene the nomenclature Trichosurus vulpecula DBB (Trvu-DBB).
The aligned nucleotide sequences after removal of the third base position contained 138 constant sites, 80 variable sites that were not parsimony informative, and 350 sites that were parsimony informative in a total of 568. In the amino acid alignment, there were 41 constant sites, 31 sites that were variable but not parsimony informative, and 212 parsimony informative positions in a total of 284. The fish and shark were removed from the alignment because the inclusion of sequences from either resulted in highly significant heterogeneity in base composition in the nucleotide sequences (after removal of the third base position). There was no suggestion of compositional heterogeneity in the reduced dataset (χ2 = 57.59864, df = 81, P =.97722).
Phylogenetic analysis involved the use of distance, parsimony, and ML methods, with both nucleotide and amino acid data. All trees supported the paralogy of nonmammalian and mammalian gene clusters. In fact, it appears that the class II genes of different vertebrate classes arose from distinct families of class II genes. The mammalian genes formed a monophyletic clade. This clade contained six quite separate and distinct gene clusters—eutherian DRB, DQB, DPB, and DOB, and marsupial DAB and DBB—that were always monophyletic with high bootstrap support. Only one relationship was consistently resolved within these gene clusters—eutherian DPB and DQB are more closely related to each other than they are to any other gene cluster, suggesting that they have resulted from a relatively recent gene duplication event. This finding has been noted previously (Hughes and Nei 1990; Takahashi et al. 2000). The majority of analyses were unable to resolve relationships between the other mammalian gene clusters. Topologies were produced with short internal branch lengths that were not supported by high bootstrap values. Maximum parsimony analyses supported the monophyly of the eutherian DQB, DPB, and DRB gene clusters (bootstrap 59%). This relationship is also seen in the ML results (see Figure 2), but is not supported by bootstrapping. All analyses supported a sister pairing of the DPB and DQB clusters (ML bootstrap support of 87). Otherwise bootstrapping did not support a close evolutionary relationship between any of the other gene clusters.
In the ML analysis, the marsupial DAB genes were a sister group to the eutherian DOB genes. The marsupial DAB genes did not group with the eutherian DRB clade. However, by conducting likelihood ratio tests on the optimal tree versus constrained trees, we could not statistically reject the hypothesis that the DAB and DRB genes formed a monophyletic clade, nor that DAB and DOB were monophyletic (data not shown). KH and SH tests significantly reject the hypothesis that the DAB genes are either DQBs or DPBs. The marsupial DBB genes were basal to the eutherian DPB, DQB, and DRB gene clusters. As with the DABs, clustering of the DBBs with the DPBs or the DQBs was statistically rejected, but clustering with the DOBs or the DRBs was not. The two marsupial gene clusters were no more closely related to each other than they were to the DR, DQ, DO, and DP clades.
A brushtail possum MHC class II DBB gene was isolated using a possum DAB probe at low stringency. We named this clone Trvu-DBB, based on sequence similarity to the M. rufogriseus DBB sequence using BLAST and FASTA searches. There are now two known DBBs, Trvu-DBB and Maru-DBB (M. rufogriseus). Phylogenetic analysis was conducted on a relatively large dataset in order to determine whether the DBBs formed a gene cluster on their own, as suggested by Schneider et al. (1991) or whether they grouped with one of the eutherian gene clusters. Several authors have suggested that the marsupial DAB genes belong to the eutherian DRB gene cluster (Lam et al. 2001; Stone et al. 1999; Takahashi et al. 2000). Whether DBB belongs to one of the eutherian β-chain clusters has not been addressed.
Resultant trees showed that the mammalian MHC class II sequences are monophyletic, confirming Takahashi et al.'s (2000) suggestion that all mammalian class II β-chain sequences (excluding the DM genes) evolved from a common ancestral element after the separation of the bird and mammalian lineages. As shown in the ML tree in Figure 2, within the mammalian clade, there are six clear gene clusters: DQB, DPB, DRB, DBB, DOB, and DAB. The two marsupial gene clusters do not intermingle with any of the eutherian gene clusters. The DQB and DPB gene clusters are sister groups. The relationship between the DQB/DPB, DRB, DBB, DAB, and DOB gene clusters remains unresolved. Based on the phylogenetic trees generated in this study, we feel that the nomenclature DBB is appropriate for Maru-DBB and Trvu-DBB. The DBB cluster is independent of the other clusters. The DBBs are definitely not DQBs or DPBs.
It appears that the designation of “DRB” to the marsupial DAB sequences was premature. Except for one instance, where we used the exact parameters of Takahashi et al. (NJ, Kimura two-parameter on the first and second positions, bootstrap 1000 in MEGA2), we were unable to reproduce a topology showing DAB and DRB genes as sister groups. However, even when we used these parameters, only 25% support resulted for the DRB/DAB sister grouping. Clearly this pairing is not robust against variation in analytical methodology. Stone et al. (1999) generated their tree using neighbor joining of genetic distances which were calculated from percentages of amino acid identity between the sequences. Takahashi et al. (2000) used the Kimura two-parameter model for nucleotide substitution. Differences in our resultant trees are probably due to the fact that we used more sophisticated models of nucleotide substitution. The program MODELTEST allowed us to fit a realistic model of nucleotide substitution to our dataset. This model took into account nucleotide frequencies, a rate matrix for specific substitutions, and the gamma distribution to accommodate between-site substitution rate variation. We found that the marsupial DABs are no more closely related to eutherian DRB than to any other eutherian β gene cluster. We therefore recommend that the original nomenclature for the two marsupial gene clusters, that is, DAB and DBB, be reinstated.
Studies of the marsupial MHC region have suggested that the class I genes of eutherians arose from a different ancestral element than those of marsupials (Houlden et al. 1996; Lam et al. 2000; Mayer et al. 1993). Miska et al. (2002) found that monotremes used a different gene cluster altogether. They suggested that two separate duplication events would have to have taken place, one before the separation of the monotreme lineage and one before the separation of the therians. These gene duplications would then have been followed by successive gene deletions in the three lineages, resulting in the gene clusters we see today. Our results do not support the hypothesis that the two marsupial gene clusters (DA and DB) separated after the separation of marsupials and eutherians, instead they suggest a scenario that involved multiple MHC class II clusters being present in the common ancestor of the therians. Some clusters were lost (through slow inactivation or deletion) over time, while others remained. Although the time scale for gene turnover is longer, the MHC class II genes appear to be evolving in a manner similar to the class I genes, via a birth and death model (Nei et al. 1997). Although not addressed in this article, addition of the DIB/DYB gene clusters in cattle and sheep supports the appearance of new gene clusters in specific lineages (Nei et al. 1997). At this stage, no evidence exists that DAB and DBB occur in eutherians. However, we cannot exclude the possibility that other β-chain genes are present in marsupials, particularly if they are not expressed at high levels in the spleen or lymph nodes.
Our ML tree indicates that the DRB, DQB, and DPB clusters may have evolved from a common element in the eutherian lineage, although this relationship is not supported by bootstrapping. Our results contradict the finding of Figueroa and Klein (1986), who proposed that the DP region was the first mammalian class II region to diverge, based on the orientation of the B and A genes in the DP region and in other class II regions. However, the authors later suggested that the DN/DO region diverged first, followed by DR and then the DP and DQ regions (Klein and Figueroa 1986). Our trees do support this topology.
Takahashi et al. (2000) suggested that the mammalian gene clusters originated 170–200 million years ago, and that the DOB locus was the first mammalian β chain to diverge, about 254 million years ago. They suggested that the DRB genes separated from the DPB/DQB genes about 189 million years ago, and the DQB and DPQ genes separated about 179 million years ago. Given that the marsupial and eutherian lineages are believed to have separated about 130 million years ago (Belov et al. 2002), we should therefore expect to find eutherian class II β orthologs in marsupials. However, exhaustive phylogenetic analysis, taking into account the transition/transversion ratios, and using sophisticated methods of phylogenetic analysis, did not indicate this. Further, the topology of the resultant trees was not stable, and internal branch lengths separating gene cluster clades were short, suggesting to us that the estimation of dates of divergence of these gene clusters is not prudent. We propose that the duplication and subsequent evolution of these gene clusters took place in a short interval, prior to the separation of the marsupial and eutherian lineages.
It appears that DAB and DBB are the predominant polymorphic MHC class II molecules in marsupials. We base this on the fact that these genes have been cloned several times by various groups in various marsupial species using different probes (Lam et al. 2001; Schneider et al. 1991; Stone et al. 1999). Nonetheless, marsupial DNA and DQA genes have been identified in macropods (Slade et al. 1994; Slade and Mayer 1995). If these sequences have been given the correct nomenclature, and are actually DN and DQ orthologs, then we still expect to find DOB (which pairs with DNA) and DQB genes in the macropods. To add to the confusion, the phylogenetic trees generated by mammalian α- and β-chain genes are not congruent (Takahashi et al. 2000); DNA genes appear to be more closely related to DQA genes, while DRA and DPA genes form sister groups. The B gene of the eutherian DP region may have arisen from a duplication of the DQB gene, while the A gene may have arisen from the duplication of a DRA gene (Hughes and Nei 1990).
Since the completion of this article, we have cloned some class II β genes from the monotreme lineage. We have given these genes the designation DZB, as they do not appear to be orthologous to any of the gene clusters found in marsupials or eutherians (Belov et al. 2003).
By no means is the evolutionary history of the mammalian class II region clear-cut. It is possible that detection of an orthologous relationship between the marsupial and eutherian gene clusters is not apparent using phylogenetic methods simply due to the length of time that these genes have been evolving independently. Elucidation of gene organization within the class II region may be the preferable method for determining orthology in this instance. Thorough characterization of the marsupial and monotreme class II genes and pseudogenes will provide important information about the genomic processes that led to the appearance of the eutherian MHC and will provide a clearer picture of the timing of the gene duplication events that took place in the mammalian lineages. We are now taking a systematic approach to resolving gene content and organization in the marsupial MHC. A tammar wallaby BAC library has been screened with DAB and DBB probes. Large-scale genomic sequencing of this region and elucidation of the position of the DAB and DBB genes within the MHC class II region will not only yield insights into this evolutionarily important locus in mammals, but may help resolve the nomenclature puzzle.
Corresponding Editor: Stephen J. O'Brien
This work was supported by an Australian Research Council discovery grant and postdoctoral fellowship (to K.B.).