OUP user menu

Divergent Patterns of Variation in Major Histocompatibility Complex Class II Alleles Among Antarctic Phocid Pinnipeds

Niles Lehman, Debra J. Decker, Brent S. Stewart
DOI: http://dx.doi.org/10.1644/BDW-010.1 1215-1224 First published online: 21 December 2004


The 4 species of phocid pinnipeds that live in pack-ice habitats of the Antarctic have been relatively isolated from infectious diseases that are prevalent in mammals elsewhere. Consequently, patterns of genetic variability at key immune system loci in these seals might be primarily related to other selective factors correlative with interspecific differences in life history, demography, and ecological niches. To test that hypothesis, we investigated nucleotide-sequence variation in a 162-base pair region of DQα exon 2 in major histocompatibility complex (MHC) class II genes of Weddell seals (Leptonychotes weddellii), crabeater seals (Lobodon carcinophaga), Ross seals (Ommatophoca rossii), and leopard seals (Hydrurga leptonyx) from the Ross Sea of Antarctica. We found substantial differences in patterns of population genetic diversity among these seals. Crabeater seals were the most diverse, with 2 DQα loci, 39 distinct alleles among 30 seals, and at least 93% observed heterozygosity. Leopard seals were the least diverse, with only 1 allele detected in 13 seals. Weddell and Ross seals had intermediate diversity, with 11 alleles in 35 Weddell seals (89% observed heterozygosity) and 2 alleles in 42 Ross seals (21% observed heterozygosity). These patterns are congruent with extant knowledge of interspecific variation in life histories, population biology, and ecological niches of these species and consistent with previously reported patterns of microsatellite genetic variability. Moreover, our findings suggest that responses to past evolutionary pressures have differed in each species and that interspecific variation in future responses to introduced infectious diseases also may vary substantially.

Key words
  • Antarctica
  • crabeater seal
  • DQα
  • genetic diversity
  • leopard seal
  • major histocompatibility complex
  • natural selection
  • pinnipeds
  • Ross seal
  • Weddell seal

The major histocompatibility complex (MHC) of mammals is a suite of loci that codes for proteins that have key roles in the responses of the immune system to infectious disease. Proteins that are expressed by MHC loci of both class I (directed toward intracellular pathogens such as viruses) and class II (directed toward extracellular pathogens such as eukaryotic parasites and most bacteria) bind fragments of foreign antigens and present them to T-lymphocytes to stimulate a strong phagocytotic immune response. The chemical interactions between expressed MHC proteins and antigens are highly specific. Each potential antigen is presumed to be recognized and most tightly bound by the protein products of a particular MHC allele. Such a direct correlation between specific MHC alleles and pathogen-specific immunity has been observed in a few mammals. For example, the human HLA-B53 allele at a class I locus has been shown to provide enhanced protection against the malarial parasite Plasmodium (Hill et al. 1991, 1992) and is present in higher frequencies in populations more prone to malarial infestation. Consequently, high allelic MHC diversity in a population is selectively favored in a multifaceted pathogenic environment as a function of herd immunity, where allelic diversity is expected to be maintained by heterozygote advantage (e.g., Parham and Ohta 1996; Potts and Slev 1995). This suggests that MHC allelic diversity might be higher in outbred populations exposed to a broad range of pathogens than in either inbred populations or in populations that are relatively naïve immunologically (e.g., Ditchkoff et al. 2001; Hedrick 1994; Hughes 1991; Van Den Bussche et al. 2002). Recent studies have documented the ability of new pathogens in a mammalian niche to greatly alter ecological dynamics (e.g., Tompkins et al. 2003), heightening the evolutionary significance of immunological loci. In any event, MHC loci have the potential to expose differences in the ecological histories of various mammals, even those that share a recent common ancestor and a common biogeographical experience (Hedrick and Kim 1999; Yeager and Hughes 1999).

The 4 species of phocid pinnipeds that live in Antarctic pack-ice habitats are ideally suited to test the ability of MHC variation to reflect species-specific ecological pressures. Weddell seals (Leptonychotes weddellii), Ross seals (Ommatophoca rossii), crabeater seals (Lobodon carcinophaga), and leopard seals (Hydrurga leptonyx) are closely related and form a monophyletic clade (Fig. 1; see also Heyning and Lento 2002), yet life-history pattems and biogeography appear to differ strikingly among them (e.g., Davis et al. 2000; Reeves et al. 2002; Reeves and Stewart 2003). Because of their circumpolar Antarctic distribution and relative isolation in extreme ice-dominated habitats, these Antarctic pack-ice seals may not have been exposed to a relatively high diversity of pathogens compared to temperate and tropical pinnipeds. Moreover, their populations have, as yet, been little affected by humans, in contrast to long-term and substantial reductions of populations of most other pinnipeds (Reeves et al. 1992, 2002).

Fig. 1

Gross phylogeny of tribe Phocinae showing Weddell-Ross seal and crabeater-leopard seal clades based on ecomorpho-logical studies (Bininda-Emonds and Russell 1996; de Muizon 1982) and mitoochondrial DNA sequences (Ámason et al. 1995).

The recent shared ancestry, but divergent ecologies, of the 4 Antarctic phocids may permit a direct test of the relationship between gene and/or allelic diversity and ecological differentiation. The few genetic studies of these species do suggest substantial species-dependent trends. Allozyme analyses of Weddell seals indicated a relatively low level of within-population diversity and some degree of population substruc-turing (Shaughnessy 1969; Testa 1986) that might indicate breeding site fidelity. Indeed, clear interspecific differences in levels of genetic diversity were reported at 18 microsatellite loci in nearly 200 Weddell, crabeater, and leopard seals. Crabeater seals were the most variable (13 alleles per locus and 87% observed heterozygosity), followed by Weddell seals (10 and 73%) and leopard seals (7.7 and 72%). Yet nucleotide-sequence variation in these mammals remains largely unknown, especially at immunologically relevant loci.

Here we report our findings of genetic diversity of Ross, Weddell, crabeater, and leopard seals from the Ross Sea, Antarctica, to allow a comparison with microsatellite data and to provide a framework for establishing molecular-ecological hypotheses. We focused on the class II loci to examine bacterial and eukaryotic parasite pathogenic exposure. We investigated polymorphism in 162 base pairs (bp) of the 2nd exon of the DQα locus, which contains the peptide-binding region of the antigen-presenting subunit (α1) of the MHC and is known to be polymorphic in populations of many other mammals (Hedrick and Kim 1999). Although variation at the DQβ locus is generally greater compared with variation at the DQα locus in most mammals studied so far, we focused on the latter because it may be more likely to reveal distinct alleles that may be maintained by either diversifying or positive selection.

Materials and Methods

Sample collection.—We collected samples of skin (surface biopsies, 0.5-1.0 cm3) from free-ranging Ross (n = 42), Weddell (n = 202), crabeater (n — 186), and leopard (n — 13) seals during a multidisci-plinary research cruise in the Ross Sea, Antarctica, between December 1999 and February 2000 (Fig. 2; see also Ackley et al. 2003). All samples were stored frozen until they were later processed for DNA extraction in our laboratory in the United States. We chose, at random, samples of 35 Weddell seals and 30 crabeater seals for DNA extraction and analyses of MHC allelic diversity, with the goal of surveying 15–20% of the available samples for these species. We analyzed all samples of Ross and leopard seals. We also included 1 sample from a southem elephant seal (Mirounga leonina) as the most likely close outgroup species to the strictly Antarctic phocid pinnipeds (Fig. 1).

Fig. 2

Map of Antarctica showing sampling area and trace of sampling cruise. Symbols denote sampling localities for crabeater (C), leopard (L), Ross (R), and Weddell (W) seals.

DNA extraction and amplification.—We used 0.125-0.5 cm3 of each tissue sample for DNA extraction. The epidermis, dermis, and approximately 1 mm of the hypodermis were removed with a sterile razor blade, and the remaining tissue was chopped into 4–6 small cubes. DNA was extracted from these cubes by the use of DNeasy Tissue Kits (Qiagen, Valencia, Califomia), eluted in 50 ρl of Tris-ethylenediaminetetraacetic acid buffer, and stored at — 20°C prior to genetic analysis. Prior to amplification via polymerase chain reaction (PCR), the genomic DNA was typically diluted 1:10 in sterile water. We amplified DQα exons 2 and 3 and the intervening intron (Fig. 3) by using mammalian universal DQα primers MDQAl (5′-CCGGATCCCAGTACACCCATGAATTTGATGG-3′) and MDQA2 (5′-CCGGATCCCCAGTGCTCCACCTTGCAGTC-3′—Slade et al. 1993), which produce a product approximately 900 bp long. We were able to amplify the intemal portion of DQα exon 2 alone, including the peptide-binding regions, in most cases by using MDQA1 and human primer GH27 (5′-CACGGATCCGGTAGCAGCGGTAGAGTTG-3′—Scharf et al. 1986). All amplifications were carried out in 50-ρl volumes in a PTC-100 PCR machine (MJ Research, Incline Village, Nevada) by using 1.25 units of AmpliTaq DNA polymerase (Perkin-Elmer, Foster City, Califomia) and the following cycle parameters: 92°C for 10 min, (92°C for 1 min, 50°C for 1 min, 72°C for 1 min) × 40, and 72°C for 10 min. We reanalyzed samples from a few seals by using PCR products obtained from a higher annealing temperature. We were particularly cautious in all of our extractions and amplifications to prevent cross-contamination and contamination of PCR reaction mixtures. Tissue extraction, PCR setup, and all post-PCR analyses were carried out in 3 physically separate rooms; barrier pipette tips were used at all times; and PCR setup was performed in a dedicated ultraviolet-irradiation hood with dedicated instruments and reagents.

Fig. 3

Schematic of major histocompatibility complex (MHC) class II DQα locus and primer-binding sites. Abbreviations: L, exon 1 (leader peptide); i, intron (about 450 base pairs [bp] in the Antarctic phocids); TM, transmembrane domain; C, constant domain; PBR, peptide-binding region (alpha-helical portion). The primer pair MDQA1+MDQA2 amplifies a 900-bp product in Antarctic phocids, spanning exon 2, the intron, and exon 3. Pair MDQA1+GH27 amplifies a 220-bp product, spanning the peptide-binding region of exon 2. The sequences reported in this paper begin 8 nucleotides downstream of the 3′ end of MDQA1, or in the 91st nucleotide (31st amino acid) of exon 2, assuming exon 1, the leader peptide, is 26 amino acids long as in many other mammals, although the exact manner in which MHC peptides are spliced is not yet known for Antarctic phocids.

Allele cloning and sequence determination.—We ligated the PCR products into the vector pCR2.1 by using the TA Cloning Kit (Invitrogen, Carlsbad, Califomia) and then transformed the vector into competent Escherichia coli cells by standard procedures. Because these MHC loci were likely to be substantially heterozygous, we identified 10 colonies per seal that contained MHC inserts. Plasmid DNA was isolated from each colony by using the boiling-lysis method of Sambrook et al. (1989), and between 2 and 10 of those plasmids were subjected to nucleotide sequence analysis.

All samples were analyzed for sequence variation at the DQα locus. Nucleotide sequence variation in exon 2 was determined by the use of the MDQA1 primer in manual sequencing reactions by using the Sanger dideoxy method via Sequenase 2.0 DNA polymerase (United States Biochemical, Cleveland, Ohio) and the accompanying reagents supplied by the manufacturer. Sequence determination was made by using [α35S]deoxyadenosine triphosphate, 6% denaturing Polyacrylamide gel electrophoresis, and autoradiography. Sequences were confirmed for approximately 10% of the seals by using the GH27 primer in a 2nd reaction. We characterized the nucleotide sequences of a subset of DQα clones and uncloned PCR products, particularly those from seals with highly divergent sequences, at both exon 2 (with the MDQA1 primer) and exon 3 (with the MDQA2 primer) by using the BigDye Terminator cycle sequencing protocol and analysis on a Prism 310 capillary automated sequence apparatus (Applied Biosystems, Foster City, Califomia).

Restriction endonuclease analysis.—In the case of variation at the DQα locus in Ross seals, not all seals were subjected to complete nucleotide sequence analysis because of the failure of 1 putative allele to amplify with the MDQA primers (see “Results”). We assayed all 42 Ross seals for the presence or absence of the null allele by using restriction fragment length polymorphism analyses that could reveal the diploid MHC genotypes of each individual. Between 8 and 10 ρl of the MDQA1+MDQA2 or MDQA1+GH27 PGR products were subjected to restriction analysis with the endonucleases AluI, AvaII, BspHI, MseI, NdeI, PvuII, RsaI, Sau96I, and ScrFI (New England Biolabs, Beverly, Massachusetts) in 20-ρl reaction volumes for 2 h at 37°C under the conditions specified by the manufacturer. Restriction fragments were separated on 1–2% agarose gels stained with ethidium bromide and visualized by ultraviolet transillumination. We determined the nucleotide sequence for the 1 allele that could be amplified with the MDQA1 primer by performing an amplification of genomic DNA with the MDQA1+MDQA2 primer pair, restricting the products with the XbaI endonuclease, diluting the digested DNA 106-fold, reamplifying with MDQA1+MDQA2, and performing a BigDye Terminator sequence reaction with the MDQA1 primer. We also amplified all 13 leopard seal samples with the MDQA1+GH-27 primer pair and screened the products with 4 restriction enzymes (AluI, MseI, PvuII, and ScrFI) for allelic variation.

Sequence analysis.—All sequences were aligned by eye and compared to previously determined sequences from elephant seals (Slade et al. 1993, 1998; Weber et al. 2004) and humans. Percent heterozygosities, average pairwise nucleotide substitutions per site among alleles (δxy, corrected from observed values by using the method of Jukes and Cantor [1969]), and nucleotide diversities (π) all were calculated with DnaSP software, v. 3.51 (Rozas and Rozas 1999). MHC DQα allele sequences were compared to published sequences (southem elephant seal, GenBank accession no. U91907; domestic dog, M74907; human, M26041; cotton rat, AF155914; and mouse, M21931) by using BLAST software searches (http://www.ncbi.nlm.nih.gov/BLAST). Observed and expected values of heterozygosity were calculated by hand and tested for significance with X2 tests with corrections for small sample sizes (Hedrick 2000). Sequences from representative alleles from each species were deposited in GenBank.

The research was authorized by research permits 976 under the United States Marine Mammal Protection Act and 2000–01 under the United States Antarctic Conservation Act. The research was approved by the Institutional Animal Care and Use Committee of Hubbs-SeaWorld Research Institute, which is registered as a Research Facility with the United States Department of Agriculture-Animal and Plant Inspection Service, and complied with American Society of Mammalogists animal care and use guidelines (Animal Care and Use Committee 1998).


We detected 52 alleles among 120 seals at the DQα locus in approximately 800 clones of the 162-bp region that surround the peptide-binding region of exon 2 (Fig. 3). Alignment of the most common alleles from each species with an allele from the southem elephant seal and with published data for Califomia seal lions (Bowen et al. 2002), domestic dogs (Sarmiento et al. 1992), and humans confirms the assignment of the alleles from Antarctic pack-ice seals to DQα (Figs. 4 and 5). This is the 1st systematic study of nucleotide level immunogenetic variation in Antarctic pack-ice seals and, in short, diversity varied greatly among Ross, Weddell, crabeater, and leopard seals (Table 1).

Fig. 4

Alignment of nucleotide sequences of all Antarctic pack-ice seal DQα alleles. Alleles W1-W11 are from Weddell seals, L1 is from leopard seals, R1 is from Ross seals, and C1-C39 are from crabeater seals. Numbering scheme is based on nucleotide positions in the complete exon 2. Dots indicate residues identical to the top sequence, allele W6. Asterisks indicate missing data. Dashes indicate deleted nucleotides relative to W6. (Nucleotide sequence data are available from N. Lehman, niles@pdx.edu.)

View this table:
Table 1

Diversity statistics for Antarctic phocid DQα loci (based on 162 base pairs of exon 2).

Seal speciesNo. individuals genotypedNo. apparent DQα lociNo. alleles in totalHeterozygous individuals (%)Mean allele identity (%)Nucleotide diversity
  • a 49% (17/35) of individuals confirmed heterozygotes by sequence analysis; 51 % (18/35) of individuals typed by a combination of sequencing of clones and batch polymerase chain reaction products.

  • b Assuming null allele (R2) is distinct from R1.

  • c Not determined.

Fig. 5

Alignment of amino-acid sequences from selected Antarctic pack-ice seal DQα alleles and comparison sequences from other species. Residue numbering scheme is based on cotton rat (Sigmodon hispidus—Pfau et al. 1999). Dots indicate residues identical to the top sequence, allele W6 from Weddell seals. Asterisks indicate missing data. Boldface sequences are from Antarctic pack-ice seals: Lewe = Weddell seal, Omro = Ross seal, Loca = crabeater seal, Hyle = leopard seal. Other sequences are from published sources: Mile = southem elephant seal (Mirounga leonina—Slade et al. 1998; except that the current study provided the last 4 residues, plus a confirmation of the 1st 50), Cafa = domestic dog (Canis lupus familiaris), Hosa = human (Homo sapiens sapiens), Sihi = cotton rat, and Mumu = house mouse (Mus musculus). Underline indicates approximate span of the alpha helix making peptide contact in class II molecules (Brown et al. 1993).

Weddell seals.—We analyzed 35 Weddell seals, including roughly half (17) by direct nucleotide sequence analysis of clones from PCR products. Between 3 and 7 (average = 4.4) DQα clones were assayed for each seal. No more than 2 alleles were detected in any seal, indicating that there is only 1 DQα locus in Weddell seals. Of those 17 seals, all were unambiguously heterozygous because 2 distinct alleles were present among the clones analyzed for these individuals. We detected a distinct MHC locus (DOα) on occasion, but these clones could easily be removed from the data set because of their low (<50%) sequence similarity to the DQα alleles (Decker et al. 2002). We genotyped the remaining 18 seals by direct cycle sequencing of MDQA1+MDQA2 PCR products obtained with a 55°C annealing temperature (instead of 50°C) to exclude DOα alleles. All of those seals, including the one for which no genotype data were available previously, could be clearly assigned a diploid genotype by this method, often by subtracting the known allele from the sequence chromatogram. Consequently, at least 31 of the seals were heterozygous, yielding a minimum estimate of observed heterozygosity of 86%.

We detected 11 distinct alleles among sequences spanning 162 nucleotides of exon 2 from the 35 Weddell seals (Table 2). All showed high sequence similarity, varying at only 10 sites (6.2% polymorphic sites). A parsimony network of all alleles (Fig. 6) indicated that these 11 alleles are only 1 or 2 substitutions distinct from another allele. The average percent sequence identity among these alleles was consequently high (δ = 98.1%), and the nucleotide diversity was accordingly low (δ = 0.0189). Allele W6 (Genbank accession no. AY283565), was the most common, occurring in about one-third of the seals, and it occupied a central position in the parsimony network (Fig. 6). Despite having similar nucleotide sequences, the other alleles may be functionally distinct from W6. Six of the 10 variable sites had polymorphisms that result in amino-acid substitutions when compared to W6. Consequently, only 1 allele (W2) would not exhibit at least 1 amino-acid substitution when compared to W6. Moreover, the next 2 most common alleles, W8 and W9, which together equal W6 in frequency, both had 2 significant amino-acid substitutions in juxtaposition to W6: a Leu ➛ Ser at position 197 in the nucleotide sequence and a Met Leu at position 217. Further, those 2 mutations always occurred together, resulting in a dichotomy of alleles that had both mutations (W1, W3, W4, W5, W7, W8, W9, and W10) and those that had neither (W2, W6, and W11), 2 groups found in almost identical frequencies (51% and 49%, respectively).

View this table:
Table 2

DQα allele frequencies in 4 species of Antarctic phocids.

AlleleFrequencyNo. seals with alleleNo. seals homozygous for allele
C1-C19, C11,
C13, C15, C16,
C20, C22-C390.01b11” (C15)
  • a Assuming R1-containing seals are heterozygous R1R2 (see text).

  • b Calculated by dividing number of seals with allele by total number of scored alleles (88).

  • c Maximum values; crabeater seals with only 1 allele found are tentatively homozygous.

Fig. 6

Parsimony network of Antarctic pack-ice seal DQa alleles. Each crosshair along the arrows represents 1 nucleotide substitution. Empty circles represent alleles not yet discovered among Antarctic phocids. All crabeater seal alleles cluster 1 or 2 mutations away from C10 or C12 and thus are not explicitly shown.

Ross seals.—We analyzed 35 Ross seals, 1st by cloning and sequence analysis of MDQA1+MDQA2 PCR products, and then by restriction fragment length polymorphism analysis. In a preliminary sequence-based survey of 13 Ross seals, only 1 DQα allele (R1, accession no. AY283567) and 1 DOα allele (Decker et al. 2002) were found. The same was found when the MDQA1+GH27 primer pair was used. Clearly, the DOα locus was outcompeting 1 or both of the MDQA primers in some seals. We suspected, therefore, that at least 1 alternative DQα allele to R1 exists in these seals, but that amplification was precluded because of substantial nucleotide substitutions at 1 or more PCR primer-binding sites. We were unable to amplify this allele with modified primers, suggesting that the alternative allele(s) differ from R1 at more than 1 nucleotide site. Consequently, we assayed overall variation in Ross seals by restriction fragment length polymorphism analysis. We amplified all 42 Ross seal samples with MDQA1 and MDQA2 and screened the products with the AvaII restriction enzyme. From this screen it was apparent that the R1 allele occurred in only 9 of the seals. If those 9 seals also had an R2 allele that would not amplify, this would translate into an observed heterozygosity of 21%, although we cannot exclude the possibility that some of those 9 seals were homozygous for R1. Scoring 9 seals as R1R2 heterozygotes, and the other 33 seals as homozygous for an R2 allele yields allele frequencies of 11% for R1 and 89% for R2. The Hardy-Weinberg frequency expectations cannot be rejected for these proportions (G-test, P > 0.1).

The DOα allele contains an XbaI site not found in the DQα PCR products. Consequently, to ensure that there were no amplifiable DQa alleles other than R1, we amplified genomic DNA from the 9 R1-containing seals with MDQA1+MDQA2, digested the products with XbaI, diluted, and reamplified with MDQA1+MDQA2. We then screened the resulting PCR products with a battery of 9 restriction enzymes. No restriction fragment length polymorphism variation was seen, and we concluded that the most likely interpretation of these data is that R1 is segregating with another, undetermined, allele in Hardy-Weinberg equilibrium in Ross seals. The R1 allele differs from the W6 allele by only 2 nucleotides (Fig. 4), neither of which results in an amino-acid substitution (Fig. 5).

Crabeater seals.—We analyzed 30 crabeater seals for DQα variation, all by direct sequencing of cloned MHC PCR products. The 1st few crabeater seals examined revealed more than 2 alleles per seal. Consequently, we speculated that there may be more than 1 locus for DQα in this species, similar to findings in some other mammals (e.g., Califomia sea lion [Zalophus californianus]—Bowen et al. 2002). Therefore, more clones per individual than in the Weddell seal were needed to assess the DQα genotype and we chose between 5 and 8 clones for analysis. In 30 seals, chosen at random, we detected between 2 and 4 alleles per seal, indicating that the DQα locus has been duplicated in crabeater seals. All but 2 individuals were heterozygous because they had more than 1 MHC allele, despite some uncertainty about the distribution of alleles among uniparental haplotypes (Table 2).

The alleles in crabeater seals shared a high amount of sequence similarity, as they did among the W6-like alleles in Weddell seals. In the 30 seals surveyed, we found 39 alleles (Table 1), all of which were similar in sequence to W6, although found as a disjunct group falling on average 6 nucleotide substitutions away from W6 (Fig. 6). The average percent similarity (98.6%) and nucleotide diversity values (0.0192) among these alleles in crabeater seals were similar to those for Weddell seals, although the number of polymorphic sites (17) in crabeater seals was somewhat higher. Alignment with alleles from Weddell, Ross, and leopard seals (Fig. 4) indicates that there are more DQα positions (7) that have substitutions specific to crabeater seals than any of the other 3 species. This suggests that crabeater seals may be an outlier for MHC of the clade, although this locus does not lend itself well to formal phylogenetic analysis for species trees.

Among the 39 alleles we detected in the crabeater seal, all but 7 (C1O, C12, C14, C17, C18, C19, and C21) were unique to single seals. Only 2 alleles (C10, accession no. AY283568, found in 17 of the 30 seals, and C12, which differs from C10 by only 3 nucleotides, found in 15 seals; Fig. 4) could be considered frequent. As in Weddell seals, there were 2 general clusters of alleles based on amino-acid sequences. The C10-like alleles account for about two-thirds of the sequences, whereas the C12-like alleles account for about one-third of the sequences (Fig. 4). These 2 alleles differed at 2 amino-acid sites, positions 66 and 75 in the DQα protein, by murine numbering (Fig. 5). A 3rd amino-acid variant, C19, and 6 related alleles had a 3rd amino acid at position 66. Positions 66 and 75, and in fact all polymorphic amino-acid sites in the 4 Antarctic phocids except for terminal position 84, lie within the DQα alpha helix that makes contact with bound peptide, based on homology with known class II structures (Batalia and Collins 1997; Brown et al. 1993). Thus, these alleles may be the result of diversifying selection at the MHC.

Leopard seals.—We analyzed all 13 leopard seals for DQα variation. Sequence analysis . of 5–8 clones from MDQA1+MDQA2 PCR products revealed a single allele, L1 (accession no. AY283566), that falls squarely within the W6 constellation, being a putative mutational intermediate between W6 and W11 (Fig. 6). This allele has 1 amino-acid substitution (Gly ➛ Arg at residue 52) compared to W6 and R1. To further insure that no alternative alleles had been missed in the cloning, we then screened all 13 leopard seals for variation by using a 4-enzyme restriction fragment length polymorphism screen on MDQA1+GH27 PCR products, which span only the 162 bp of exon 2 examined in this study. No restriction fragment length polymorphism was found in this screen, and thus we concluded that all 13 leopard seals were homozygous for the L1 allele at the DQα locus.


We found that pattems of MHC genetic diversity differed substantially among the 4 species of Antarctic phocid pinnipeds. This finding is qualitatively consistent with results from previous studies on other loci, and we think that it reflects differences among them in life-history pattems, biogeography, demography, and ecological niches, and perhaps secondarily to correlative differences in susceptibility to exposure to infectious disease.

Leopard seals were the least polymorphic, with only 1 DQα alelle. Although our sample size is small, the MHC data mirror microsatellite data demonstrating fewer alleles and less heterozygosity in leopard seals compared to Weddell and crabeater seals (Davis et al. 2000). The leopard seal, an aggressive apex predator in Antarctic marine ecosystems that eats other seals in addition to fish, squid, and krill that are prey for the other species, is not well studied but it appears to be far less abundant than the other species. Thus, the relative lack of variation in this species could be attributed to random genetic drift. However, the intriguing possibility exists that the camivorous (rather than strictly piscivorous [i.e., fish eating], teutophagous [i.e., squid eating], or krill eating) ecology of this species has exposed it to a superantigen that drove the fixation of 1 particular MHC allele at DQα.

Crabeater seals were the most polymorphic, displaying an abundance of alleles (39 in 30 seals), heterozygosity near 100%, and 2 apparently active DQα loci. Crabeater seals are very abundant in the Southem Ocean and are evidently not philopatric. In fact, some estimates put the total crabeater seal population size at 10–15 million, making it the world's most abundant pinniped (Reeves et al. 2002). Davis et al. (2000) detected more alleles and more heterozygosity at microsatellite loci in this species than in leopard and Weddell seals, and they suggested that the abundance and mobility of this species has engendered and preserved a large pool of genetic diversity. Our DQα data are consistent with that hypothesis. There are 2 common alleles, C10 and C12, but they apparently are not geographically partitioned. There also is no indication that they are associated with each other, except that they may reside on physically different loci (i.e., not be allelic) because when we found only 2 alleles in a seal, they were more likely to be C10 or C12 than any other pair. However, these 2 alleles differed at 2 or more amino-acid positions in the peptide-binding groove, leaving open the possibility of heterozygote advantage extending to 2 homologous loci. No other allele was found in more than 5 crabeater seals.

Weddell seals had intermediate levels of diversity, with a modest number of alleles (11 in 35 seals) and 89% observed heterozygosity. Weddell seals are the most extensively studied of the 4 Antarctic pinnipeds, with examination of allozyme (Testa 1986), multilocus DNA fingerprinting (Testa and Scotton 1999), and microsatellite (Davis et al. 2000; Gelatt 2001; Gelatt et al. 2001) data suggesting intermediate levels of diversity forged through a balance between strong philopatry and large population size and circumpolar distribution. Our data present no strong evidence of Weddell seals being significantly out of Hardy-Weinberg equilibrium at the DQα locus. Homozygotes of the most frequent allele (W6 at 35.7%) had a frequency of 4 of 30, matching the Hardy-Weinberg equilibrium expectation of 3.8. However, there was a slight excess of 3 types of heterozygotes: W6W9 (5, compared to an expected number of 2.7), W6W8 (6, compared to an expected 2.0), and W6W11 (5, compared to an expected 1.4). No other genotype would be expected more than once and none was found more than twice. Overall, the small sample size precluded an exact test of Hardy-Weinberg equilibrium. Nevertheless, we did detect 2 allelic constellations in terms of amino-acid sequences in and about the MHC peptide-binding region (Fig. 5). In conjunction with the excess of certain heterozygotes, these distinct constellations could reflect past or present heterosis resulting from diversifying selection imposed by nonviral pathogens.

Ross seals showed a 4th distinct pattem of MHC diversity. At the DQα locus, we detected only 2 alleles among 42 seals. These 2 alleles combine to create 2 genotypes that are in Hardy-Weinberg proportions regardless of their degree of divergence, which is yet to be determined. Little is yet known about the ecology and genetics of Ross seals; in fact, the data here are the 1st population-level DNA sequences to be reported. Recent observations suggest Ross seals may have patchy distributions, be polygynous and faithful to limited breeding sites in heavy pack-ice habitats, and may spend most of the nonbreeding season foraging in ice-free pelagic Southem Ocean systems. Such life-history characteristics could account for the pattems of low genotypic diversity that we observed at the DQα locus.

It is not unusual to detect divergent genetic signatures among mammalian species that vary greatly in morphological, behavioral, and demographic characteristics, which can strongly influence evolutionary trajectories. It is remarkable though, that the 4 Anatarctic phocids have such disparate pattems at the MHC DQα locus, given their monophyly, common macro-biogeography, and historical isolation from other mammals. We think that these data are indicative of the power of ecological factors, particularly disease, in modulating allelic diversity, population-genetic subdivision, and even the number of homologous loci in short evolutionary time frames. Ámason et al. (1995) postulated a similar explanation for the high similarity in mitochondrial DNA sequences between Weddell and leopard seals despite their obvious morphological divergence.

In Antarctic seals, the expectations of high polymorphism and heterozygosity are apparently realized in at most 2 of the 4 species. Yet when all 4 species are compared, it appears that they may have been challenged by substantial environmental changes. Although recent human intmsion into Antarctic ecosystems may have presented some additional challenges, we think that the present suite of MHC alleles and genotypes more likely reflects adaptive episodic responses to historic selective factors that may no longer be operating (cf. Hedrick and Kim 1999). Little is yet known about the incidence and prevalence of infectious disease in these Antarctic pinnipeds, although some evidence suggests recent exposure to certain pathogens (e.g., Bengtson et al. 1991; Harder et al. 1991; Laws and Taylor 1957; Linn et al. 2001; Stenvers et al. 1992). Those pathogens could have been introduced by humans directly or through the use of dogs by explorers and scientists in the 19th and 20th centuries, or naturally by other wildlife (e.g., Austin and Webster 1993; Gardner et al. 1997; Gauthier-Clerc et al. 2002; Moore and Cameron 1969; Morgan and Westbury 1981; Morgan et al. 1981; Oelke and Steiniger 1973; Parmelee et al. 1978). Some of these diseases can spread rapidly and cause mass mortality and, arguably, intense short-term selection and population-genetic response (e.g., Califomia sea lions— Acevedo-Whitehouse et al. 2003). Such selection pressure may have been ultimately anthropogenic, but the detailed history of the immune loci in these seals will require further genetic, pathological, and epidemiological study.


We thank the staff of the United States National Science Foundation office in Christchurch, New Zealand, the Crary Laboratory at McMurdo Sound, Antarctic Support Associates, Raytheon Polar Services, the pilots and staff of Petroleum Helicopters Inc., and the officers and crew of the RVIB Nathaniel B. Palmer for logistic support in McMurdo Sound and during surveys in the westem Ross Sea. We also thank P. Yochem, M. Koski, T. Gelatt, C. Davis, D. Siniff, and I. Stirling for field assistance and A. Burton, M. Swinton, A. Krummel, D. Weber, P. Yochem, and W. Yu for laboratory assistance. The research was supported by grants from the United States National Science Foundation (grants OPP-9816011 to B. S. Stewart and N. Lehman and OPP-9816035 to P. K. Yochem and B. S. Stewart), from the Hubbs Society of the Hubbs-SeaWorld Research Institute, and from SeaWorld Inc. (Anheuser-Busch Corp. parent company).


  • Associate Editor was David Weller.

Literature Cited

View Abstract