Molecular Genetic Diversity and Evolution at the MHC DQB Locus in

Four Species of Pinnipeds

A. Rus Hoelzel,* J. Claiborne Stephens,† and Stephen J. O’Brien†
*Department of Biological Sciences, University of Durham, Durham, England; and †Laboratory of Genomic Diversity,
National Cancer Institute, Frederick, Maryland

Variation was investigated at exon 2 (including part of the putative peptide-binding region) of the class II major
histocompatibility complex (MHC) DQB locus for two congeneric phocid seal species and two congeneric otariid
seal species. Polymorphism in one phocid species, the southern elephant seal (Mirounga leonina), was comparable
to that seen in human populations, while the other phocid, the northern elephant seal (Mirounga angustirostris),
has been through a severe population bottleneck and exhibited much less variation at this locus. A phylogenetic
comparison of the four species was consistent with the trans-specific pattern of evolution described for other taxa
at this locus, and relative nonsynonymous and synonymous substitution rates suggest the maintenance of polymor-
phisms by natural selection. A comparison of sequence patterns also suggested that some variation could have been
generated through recombinational events, primarily within genera. These results suggest a pattern of evolution of
the immune response in pinnipeds similar to that in terrestrial mammal species.

Introduction

Downloaded from mbe.oxfordjournals.org by guest on September 13, 2010

Genes in the human class II major histocompati-
bility complex (MHC) at the DR, DQ, and DP loci
encode alpha and beta glycopeptide chains. These form
cell surface heterodimers on antigen-presenting cells
which bind processed antigen peptide fragments (see
reviews in Germain and Margulies 1993; Harding
1996). In many species, these loci are highly polymor-
phic, especially in the second exons of the beta genes,
and the variation is thought to be maintained by natural
selection through heterozygote advantage (Hughes and
Nei 1989; Thurz et al. 1997) or frequency-dependent
selection (Takahata and Nei 1990; Hill et al. 1992).
Natural selection for diversity is typically thought to
promote pathogen recognition. Recognition of the com-
plex between antigen-presenting cells and invading
peptides by the T-cell receptor of CD41 lymphocytes
(in humans) leads to an immune response through T-
cell activation (Guillet et al. 1987; Sette et al. 1987).
MHC diversity at these and other immune system loci
may reflect the pathogen environment of a given spe-
cies (e.g., Potts and Slev 1995), and various associa-
tions between specific infectious diseases and MHC ge-
notypes have been identified, including Marek’s dis-
ease (Longenecker et al. 1977), malaria (Hill et al.
1991), and HIV-1 (Kaslow et al. 1996). Evidence for
the maintenance of alleles over long periods of evo-
lutionary time (see Edwards and Hedrick 1998) and a
high proportion of nonsynonymous change in the pep-
tide-binding region (Hughes and Nei 1989; Yuhki and
O’Brien 1990; Miller, Withler, and Beacham 1997)
support the contention that polymorphism is main-
tained by natural selection.
Abbreviations: AFS, Antarctic fur seal; NES, northern elephant
seal; NFS, New Zealand fur seal; SES, southern elephant seal.
Key words: class II MHC, DQB, pinnipeds, molecular evolution.
Address for correspondence and reprints: A. Rus Hoelzel, De-
partment of Biological Sciences, University of Durham, South Road,
Durham, DH1 3LE, England. E-mail: a.r.hoelzel@durham.ac.uk.
Mol. Biol. Evol. 16(5):611–618. 1999
q 1999 by the Society for Molecular Biology and Evolution. ISSN: 0737-4038
It has been suggested that the pathogen environ-
ment of marine mammals may provide a diminished
selective pressure for maintaining MHC polymorphism
(Trowsdale, Groves, and Arnason 1989; Slade 1992;
Murray, Malik, and White 1995), due to a relatively
low prevalence of infectious disease in the marine en-
vironment. Slade (1992) analyzed restriction fragment
length polymorphisms (RFLPs) for four restriction en-
zymes and 24 southern elephant seals from two pop-
ulations. He used heterologous probes for MHC class
I (HLA-B7) and class II genes (DQA, DQB, and DRB)
and compared the average percentage of difference be-
tween banding patterns with those seen in various other
species. In general, there was apparently less variation
for all probes in marine mammals than in terrestrial
mammals. The highest level of variation for the south-
ern elephant seal (SES) was for the DQB probe, but
this was still much lower than comparable values seen
for other species. As Slade (1992) points out, however,
the comparison of levels of variation between taxa can
be misleading, especially for relatively low resolution
genetic analyses. Furthermore, the function of the im-
mune response is too poorly understood for most non-
human taxa to permit the identification of which loci
are critical, and the critical sequences could be short
enough to miss detection by the RFLP method.
In this study, we investigate sequence variation at
exon 2 of the DQB gene, including part of the putative
peptide-binding region, for four species of pinnipeds.
We chose this locus because it is known to be highly
polymorphic in primates (e.g., Bugawan and Erlich
1991) and a variety of other terrestrial species. We com-
pare the level of variation in the SES at this locus to
that seen in human populations and investigate the re-
lationship between alleles in different pinniped species
to evaluate whether pinnipeds have a uniformly reduced
level of diversity at the MHC relative to terrestrial mam-
mals. The study species include two congeneric species
in the superfamily Phocidea and two congeneric species
in the superfamily Otarioidea (classification as in Ried-
man 1990).

611
612 Hoelzel et al.
Table 1
DQB Allele Sequence Diversity for Four Species of Pinnipeds
SPECIES N NO. OF
Southern elephant seal . . . . . . 109
Northern elephant seal . . . . . . 69
Antarctic fur seal . . . . . . . . . . 13
New Zealand fur seal . . . . . . . 19
Materials and Methods
PCR Amplification and Single-strand Conformational
Polymorphism Analysis
Samples were collected from live animals by bi-
opsy sampling as described in Hoelzel et al. (1993), and
the DNA was extracted by standard protocol. Northern
elephant seal (NES) samples were collected at Año Nue-
vo, California; SES samples were collected at Punta
Delgada, Argentina; New Zealand fur seals (NFS; Arc-
tocephalus fosteri) were collected at Kangaroo Island,
South Australia; and Antarctic fur seals (AFS; Arcto-
cephalus gazella) were collected at Bird Island in the
South Atlantic. Primers for the exon 2 region of the
DQB gene were designed from a comparison of cow,
human, and dog sequences from the GEN-EMBL data-
base: sense—59-TCGTGTACCAGTTTAAGGGC; anti-
sense—59-ACGTCCTTCTGGCTGTTCCA. A 142-bp
segment was amplified by PCR from each individual,
and 33P alpha dATP was incorporated into the product
under the following assay conditions: 1.5 mM MgCl2;
10 mM Tris; 50 mM KCl; 100 mM dTTP, dCTP, and
dGTP; 5 mM dATP, 1 mCi alpha 33P dATP, and 50 pM
of each primer. Labeled PCR product was denatured at
958C for 5 min, chilled on ice for 1 min, and loaded
onto a nondenaturing gel (4.5% acrylamide [37.5/1 ac-
rylamide/bis], 10% glycerol) run at 208C to assay for
single-strand conformational polymorphisms (SSCPs).
The placement of the primers was limited in part by
apparent poor homology (especially for the elephant seal
species) with published sequences in much of the exon
2 region.
DNA Sequencing and Phylogenetic Analysis
PCR product was prepared for cloning using the
double gene-clean cloning kit (Bio 101). This entailed
polishing the ends using T4 DNA kinase and polymer-
ase I. Insert DNA was then cloned into bluescript sk1
vector and transformed (after Chung and Miller 1988)
into the XL1 blue strain of Escherichia coli using blue/
white selection. Five to seven clones were sequenced in
both directions for each representative of a putative ge-
notype identified through SSCP analysis. A given pu-
tative genotype was sequenced from two to four indi-
viduals with identical SSCP phenotypes. Sequencing
was by the dye-primer method for the ABI automated
sequencing system.
Mitochondrial DNA from the control region was
amplified using published primers (Hoelzel and Green
1992) and sequenced on the automated ABI system us-
ing the same primers. Four unique NFS sequences and
RANGE OF ALLELE
ALLELES Frequency % Difference Hexp Hobs
8 0.01–0.25 0.7–4.9 0.839 0.809
2 0.28–0.72 1.4 0.409 0.426
4 0.04–0.54 0.7–7.7 0.596 0.385
4 0.16–0.47 0.7–8.5 0.679 0.526
three AFS sequences were determined from six individ-
uals of each species (accession numbers AF111046-52)
and compared with published data for NES and SES
(Hoelzel et al. 1993).
Unique alleles were compared for phylogenetic re-
lationships using maximum-parsimony (Swofford and
Berlocher 1987) and neighbor-joining (Saitou and Nei
1987) analyses with the PAUP and PHYLIP computer
Downloaded from mbe.oxfordjournals.org by guest on September 13, 2010
programs. The transition/transversion ratio for the DQB
locus was approximately 1:1 (determined empirically).
For maximum parsimony, a consensus tree was derived
by bootstrap analysis with 1,000 replications. Branches
with greater than 50% support were retained. For neigh-
bor-joining analysis, genetic distances were derived us-
ing the Jukes-Cantor correction, and a bootstrap analysis
with 1,000 replications was used.
Recombination and Substitution Rate Analyses
The sequences were inspected for potential recom-
bination events, a process facilitated by the application
of published algorithms (Stephens 1985; Hudson and
Kaplan 1985) and their updated versions (see Stephens
1985 for statistical methods). Results from the statistical
tests were used to guide the subsequent inspection of
sequences and their interpretation. In addition, all vari-
able nucleotide sites were mapped onto the phylogenetic
trees suggested by PAUP and Neighbor-Joining to see
if recombination was a sensible alternative to purely
substitutional processes in the generation of these al-
leles. The inferred mosaic nature of several alleles is
best seen by juxtaposing the alternative trees derived
from these recombination analyses.
The number of synonymous substitutions per syn-
onymous site (dS) and the number of nonsynonymous
substitutions per nonsynonymous site (dN) (Nei and Go-
jobori 1986) and the standard error of those measures
(Nei and Jin 1989) were derived using the MEGA pro-
gram (Kumar, Tamura, and Nei 1993) based on an align-
ment of unique alleles.
Results
Variation
For the purpose of this study, we concentrated on
the phocid species for an assessment of levels of vari-
ation. Table 1 summarizes the diversity of genotypes in
all four species. All four species were polymorphic, and
the absolute sequence difference between alleles was up
to 8.5% within species and up to 8.5% between species.
Among 109 SES individuals, there were eight alleles,
while there were only two alleles among 69 NES indi-

FIG. 1.—DNA sequence alignment of all 16 pinniped alleles and 3 outgroups from the exon 2 region of the DQB locus. All sequences are
DQB1, except for the human DQB2 outgroup. A dot indicates the same nucleotide. Numbering across the top of the sequence is sequential for
the entire sequence, while numbering across the bottom includes only variable sites and can be cross-referenced with the numbers given in
figure 5.
viduals. All alleles were easily distinguished by the
SSCP method, except for two alleles in SES which had
similar mobilities and had to be confirmed by sequenc-
ing. One allele is shared by NES and SES, and one allele
is shared by NFS and AFS. Heterozygosity was not sig-
nificantly different than expected based on Hardy Wein-
berg for any of the four species (table 1).
Sequence Alignment and Phylogenetic Analyses
An alignment comparing the DNA sequences of all
16 DQB alleles and three outgroups is presented in fig-
ure 1, and an alignment of translated sequences is pre-
sented in figure 2. The DNA sequence alignment pro-
vided by Pileup (Devereux, Haeberli, and Smithies
1984) did not include the indels at base pairs 118 and
125; however, the inclusion of these two indels con-
served 3 bp of sequence. The amino acid alignment il-
lustrates the conservation of motifs between the pinni-
ped and human DQB1 sequences that are not shared
with the human DQB2 sequence. This and the fact that
multiple clones never revealed more than two sequences
at a locus suggest that we were successful at amplifying
just one locus from the pinniped species and that the
amplified locus is the homolog to the human DQB1 lo-
cus.
Two maximum-parsimony consensus trees are
shown in figure 3. Figure 3A shows a phylogeny based
FIG. 2.—Amino acid sequence translated from the sequence illus-
trated in figure 1. A dot indicates the same amino acid. The position
of each putative PBR site is indicated with an asterisk.
Diversity and Evolution of Pinniped DQB 613
on 142 bp of mtDNA sequence from the 59 end of the
control region (base pairs 210–351 in Hoelzel, Hancock,
Downloaded from mbe.oxfordjournals.org by guest on September 13, 2010
and Dover 1993). This tree shows a clear reciprocal
monophyletic distinction between genera and between
species, and strong bootstrap support between the Pho-
cidea and Otarioidea superfamilies. Figure 3B shows a
phylogeny derived from the 142-bp DQB sequence. For
DQB, there were 34 equally parsimonious trees. Boot-
strap analysis supported five of the nodes, shown in fig-
ure 3B. There is no clear distinction between alleles
from congeneric species, which share terminal clades for
both genera. Alleles from the two seal superfamilies are
intermingled in a single main clade. The neighbor-join-
ing tree for the DQB alleles is shown in figure 4. This
tree suggests some distinction between alleles from dif-
ferent seal superfamilies; however, the nodes separating
alleles from the two lineages are very shallow and poor-
ly supported by the bootstrap analysis, and one allele
from the fur seal species (NFS4) groups with the ele-
phant seals.
Recombination
Figure 5 illustrates the mapping of nucleotide se-
quence variants (based on variable sites; see fig. 1) to a
phylogenetic tree suggested by the neighbor-joining al-
gorithm, but in which almost all instances of homoplasy
among seals are accounted for by recombination. Figure
5A reflects the first 29 variable sites and conforms large-
ly with the neighbor-joining tree of figure 4 and the
logical taxonomic groupings. As noted in figure 4, one
fur seal sequence (NFS4) clusters with the elephant seal
sequences. Four instances of homoplasy (circled in fig.
5A) are explained by three putative recombination
events. The first of these involves AFS1 (fig. 6): the 59
end of AFS1 is identical to that of NFS1 but is otherwise
identical to AFS3 and AFS4 through nucleotide 100
(variable site 29). Similarly, NFS4 is essentially an el-
ephant seal sequence: it is nearly identical to SES7, dif-
fering only by a unique change at site 21, an inferred
parallel transition at site 29, and retention of the G at
site 1 seen in all fur seals. Hence, we invoke recombi-
nation in which a typical fur seal sequence recombined
with a sequence like SES7 to produce NFS4. Our third
proposed recombination event, in SES2/NES1, invokes
614 Hoelzel et al.
FIG. 3.—Maximum-parsimony consensus trees based on 1,000
bootstrap replications for the mtDNA control region (A) and DQB
sequences (B). Branch lengths are given above the lines, and bootstrap
support values are given below in bold type. Elephant seal genotypes
are indicated with solid letters, while fur seal genotypes are indicated
with open letters.
Downloaded from mbe.oxfordjournals.org by guest on September 13, 2010
FIG. 4.—Neighbor-joining tree based on 1,000 bootstrap replica-
tions. Branch lengths are given above the lines, and bootstrap support
values are given below in bold type. Elephant seal genotypes are in-
dicated with solid letters, while fur seal genotypes are indicated with
open letters.
replacement of the 59 end by a typical southern elephant
seal sequence, changing variable site 10 from C to G.
Each of these events invokes a recombination do-
nor that currently exists in the population. However,
consideration of additional events is also warranted. Fig-
ure 5B shows a cladogram for the 39 end of the sequence
(variable sites 30–53). Several events are suggested by
the contrast between figure 5A and B. First, note that in
figure 5B, AFS3 and AFS4 are clustered with several
northern and southern elephant seal sequences, separate
from all other fur seal sequences. There are several com-
binations of recombination events that could lead to this
shift in affinities, but it seems clear that at least two
events would need to occur. One of these would seem
to involve the common ancestor of AFS3 and AFS4 and
would lead to the acquisition of variants more typical
of elephant seals, as well as some unique variants (figs.
5B and 6). In particular, note that the 39 clustering of
variants unique to AFS3 and AFS4 is highly significant
(P , 0.006) by Stephens’ (1985) algorithm. At least one
more recombination event is needed to explain SES3
and SES4’s position with SES2 and the two NES se-
quences, distinct from the main body of SES sequences,
which now includes NFS1 in addition to NFS4.
 Diversity and Evolution of Pinniped DQB 615

Downloaded from mbe.oxfordjournals.org by guest on September 13, 2010

FIG. 5.—Variable sites are labeled (1–53) starting from the 59 end (see fig. 1). A, Phylogeny based on the first 29 variable sites. Three
instances of recombination are inferred involving the 59 ends of AFS1, NFS4, and the common sequence of SES2/NES1 and are shown encircled.
Three instances of inferred parallel change between an outgroup and a seal variant are designated with a ‘‘P’’ next to the variant. A fourth
instance of parallel change, between NFS4 and three AFS sequences, is also denoted in this manner. B, Phylogeny based on variable sites 30–
53. Three parallelisms between seals and the outgroups are inferred, as well as a reversal within seals. No recombination events within this
region are depicted, although several are inferred to explain the difference between parts A and B (see text).

Synonymous and Nonsynonymous Change
To identify a putative peptide-binding domain, the
DQB allele AFS3 was aligned with human DRB1–1105
(Apple et al. 1993). The alignment showed 80.9% ho-
mology and corresponded to amino acids 14–60 (based
on comparison with the sequence presented in Otting et
al. 1992). The crystalline structure of the human DR
alpha and beta chains has been established, and residues
in the antigen-binding groove have been identified
FIG. 6.—Mosaic structure of AFS1. AFS1 is piecemeal identical
to NFS1, AFS4, and AFS2. Note that in the central region (sites 14–
29), AFS1 is identical to AFS4 but differs from AFS2 at nine nucle-
otide sites (not shown). Diagnostic variable sites are given above
AFS4.
(Brown et al. 1993). Assuming that the structure of DQB
is similar, the same residues were considered in isolation
of the rest of the pinniped DQB exon 2 sequence (see
fig. 2).
The number of synonymous substitutions per syn-
onymous site (dS) and the number of nonsynonymous
substitutions per nonsynonymous site (dN), with the
standard error of those measures for the entire sequence,
are dN 5 0.0652 6 0.0137 and dS 5 0.0375 6 0.017.
The dN/dS ratio for the whole sequence (1.74) is similar
to that seen for the PBR region at MHC class I loci in
the domestic cat (1.81; Yuhki and O’Brien 1990). Our
sample size is small for the statistical comparison of
within the PBR versus outside the PBR. However, of the
15 amino acid residue sites among the pinniped alleles
showing DNA polymorphism (residues 1, 4, 12, 13, 15,
17, 24, 29, 34, 37, 38, 40, 42, 43, and 44; see fig. 2)
most (13 out of 15) are dominated by nonsynonymous
substitutions, and most synonymous change is found at
just one residue (number 43 in fig. 2). Site 43 is one of
the eight putative PBR sites, and all of the synonymous
616 Hoelzel et al.
changes within the PBR sites sequenced in this study
are at this residue.
Discussion
We investigated the radiation of alleles at the pin-
niped DQB1 locus and their levels of polymorphism
within species. The results show relatively high levels
of variability and a pattern of evolution that is consistent
with observations from a variety of other mammalian
species, including humans. For those species, natural se-
lection, recombination, and trans-specific polymorphism
at this locus have been invoked. Previous studies have
suggested relatively low levels of MHC variation in sev-
eral marine mammal species and large terrestrial mam-
mal species (Trowsdale, Groves, and Arnason 1989; Sla-
de 1992; Murray, Malik, and White 1995); however, for
one of the species investigated for DQB variation in this
study (SES), we found substantial levels of variation.
We suggest that variation in SES is comparable to
that in humans for the following reasons. In a study of
allelic diversification at the DQB1 locus of humans (at
a 120-bp sequence within exon 2 that overlaps with the
region sequenced in our study), Gyllensten, Lashkari,
and Erlich (1990) found 12 alleles among 30 individuals
of differing ethnic origins. The same study compared
eight primate species and found percentages of genetic
differences between alleles ranging from 0.8% to 19.2%
for this 120-bp sequence. For the same 120-bp sequence,
we found eight alleles among 109 individual SESs from
a single breeding population, and genetic differentiation
among the alleles of four species ranging from 0.8% to
11.7%. The greatest sequence divergence over this range
within a species (AFS) was 10.8%, compared with
14.2% maximum difference between human alleles at
this locus (Gyllensten, Lashkari, and Erlich 1990).
While the level of variation among pinniped alleles is
consistently lower, it appears higher at this locus than
that seen in previous studies of MHC loci in marine
mammals (e.g., Trowsdale, Groves, and Arnason 1989;
Slade 1992; Murray, Malik, and White 1995), and nearly
as high as that seen for human populations. We interpret
this to indicate a high level of variation at this locus
among the four pinniped species investigated, particu-
larly within the SES.
The phylogenetic relationship among primate DQB
alleles (Gyllensten, Lashkari, and Erlich 1990) shows a
pattern similar to that seen for the pinniped species. The
main difference is that, while the primate alleles fell into
distinct lineages (DQB1–DQB4), there is no similar
structure seen among the pinniped alleles. Instead, the
relationship appears to be an extended polytomy, with
some structure (especially in the neighbor-joining tree)
reflecting the two main groups within the pinnipeds: the
otariids and the phocids. Otariids and phocids fall into
distinct lineages in phylogenies using mtDNA (Arnason
et al. 1995; this study) and nuclear DNA markers such
as aldolase (Slade, Moritz, and Heideman 1994). The
distinction between species in these two superfamilies
is also well established based on morphological criteria
(see King 1983). The nature of the radiation of otariid
and phocid pinnipeds has been controversial (see dis-
cussion in Riedman 1990), but is likely to be at least as
early as the radiation among, for example, Old World
primate families (see Slade, Moritz, and Heideman
1994; Stewart and Disotell 1998). However, pinniped
phylogenies derived using the DQB locus do not reflect
these taxonomic differences, in that different species and
different genera are mixed together within the same al-
lelic lineages. This pattern of evolution has been referred
to as ‘‘transpecific’’ (Klein 1987). It is thought to indi-
cate the operation of natural selection through either
overdominance or frequency dependence (see Hughes
and Nei 1989; Takahata and Nei 1990).
The distribution of mutations across the DQB se-
quence is discontinuous, which can lead to a shallow
tree when relatively few sites are informative. In this
case, there are 53 variable sites, of which 21 are poly-
morphic among the pinniped species and 17 are variant
in two or more sequences. A longer nuclear DNA se-
Downloaded from mbe.oxfordjournals.org by guest on September 13, 2010
quence that includes a similar number of clustered var-
iable sites showed a phylogenetic pattern with 97%
bootstrap support separating the phocid and otariid lin-
eages (using 400 bp from the Aldolase-A gene; Slade,
Moritz, and Heideman 1994). Our mtDNA comparison
included a sequence showing a similar level and pattern
of diversity and the same overall length. MtDNA is hap-
loid and therefore has a smaller effective population
size; however, the comparison with the DQB sequence
shows a pattern that is more distinct than would be ex-
pected due to this difference alone. The DQB phylogeny
suggests that selection has maintained alleles for longer
than the average speciation event interval, although we
cannot conclude unequivocally that this is the case.
Hughes and Nei (1988, 1989) reported on the rel-
ative rates of nonsynonymous and synonymous substi-
tutions in primate species at MHC loci. They proposed
that since overdominant selection is known to increase
both the extent of polymorphisms and the rate of codon
substitution (Maruyama and Nei 1981), a greater pro-
portion of nonsynonymous substitutions within (com-
pared to outside) the codons of the peptide-binding re-
gion would indicate that the polymorphism was caused
mainly by overdominant selection. Although our study
did not focus on the region in which PBR sites are found
in the highest concentration in primates, most polymor-
phic sites (13 out of 15) showed predominantly nonsy-
nonymous change. Among these, four are putative PBR
sites. It is also possible that PBR sites other than those
determined by reference to the human data are equally
or more important, as the exact structure has not been
determined for these or for closely related taxa.
The recombinational inferences suggested in the re-
sults explain the statistical associations (Stephens 1985),
and could explain most of the homoplasy seen among
seal sequences (figs. 3–5). It is difficult to argue rigor-
ously that recombination is heavily favored over recur-
rent mutation to explain the homoplasy among seal se-
quences, and several authors have indeed inferred se-
quence convergence in the PBR (O’hUigin 1995; Ed-
wards and Hedrick 1998). Clearly, many instances of
homoplasy in the seal data set are due to one or two

variants being shared between otherwise ‘‘unrelated’’
sequences (fig. 5), so processes of parallel or reverse
mutation cannot be excluded. One could argue that
much of the homoplasy seen is due to hypervariability
at several sites. Indeed, six of the variable sites have
three or even four nucleotides present among seal se-
quences, and an additional eight sites have three or more
distinct nucleotides when outgroups are considered.
Even so, there is only a modest amount of homoplasy
between the outgroups and seals (fig. 5), which suggests
that a portion of the homoplasy seen among seal se-
quences is due to other factors, such as recombination.
Our strategy has been to determine whether recombi-
nation is a plausible mechanism given the sequences
currently seen among these species. Most of the recom-
bination events we invoke are plausible in that potential
recombination donors currently exist in the population.
Whether recombination or recurrent nucleotide
substitution is a better explanation for the pattern of al-
leleic diversity seen for this locus depends on a number
of factors, especially the relative rates of such processes.
Currently, we can only guess that the recombination rate
for this region is on the order of 1026, extrapolating from
the genomewide human estimate (see Lander and
Schork 1994). This would seem to be two or three orders
of magnitude higher than genomewide substitution rates,
although bona fide hot spots for mutation may also exist.
Regardless of which process contributes most to this
pattern of variation, both are time-dependent processes,
and it is key to note that considerable time must have
elapsed to accumulate this level of variation. The fact
that some of this variation appears to be trans-specific
further suggests that the invoked recombination events
transpired in the relevant common ancestors: that of
both fur seals, that of both elephant seals, and even in
that of otariids and phocids. The central point is that if
we assume that the recombination rate for this region is
small, the numerous instances of homoplasy among
seals suggest that many of the variants have coexisted
for a substantial period of time, allowing recombination
to shuffle them into the combinations we see today.
Such maintenance of alternative alleles over long peri-
ods of time suggests the action of natural selection, as
in terrestrial mammals (see Satta 1997).
One of the two species for which a large number
of individuals were included in this study (NES) is
known to have been through a severe population bottle-
neck approximately 110 years ago. Earlier studies have
investigated molecular genetic variation in the NES and
found greatly reduced levels of variation at both mito-
chondrial (Hoelzel et al. 1993) and nuclear markers
(Bonnell and Selander 1974; Hoelzel et al. 1993). A
simulation model approach incorporating both demo-
graphic and genetic data indicated that the population
consisted of only 10–30 seals at its nadir (Hoelzel et al.
1993). In the current study, we find only two DQB al-
leles in the northern species, compared with eight alleles
in the closely related SES. The relatively high level of
variation in the SES and the relationship between alleles
among species suggest that the low level of variation in
the NES is more a reflection of the demographic history
Diversity and Evolution of Pinniped DQB 617
of the species than a consequence of living in a marine
environment. Low levels of variation in the MHC in
endangered species have been implicated in affecting
their response to infectious disease (O’Brien and Ever-
mann 1988), although no such effect has yet been pro-
posed for the NES.
In summary, these data indicate a pattern of genetic
diversity at the class II MHC DQB locus that is com-
parable to that seen for terrestrial mammal species.
There is no indication that this locus in particular ex-
hibits less variation for these species than expected
based on this kind of a comparative analysis.
Acknowledgments
We thank Claudio Campagna for providing tissue
samples of SESs and for assistance in obtaining samples
from NESs, Tom Arnbom for samples of SESs, Burney
Le Boeuf for samples of NESs, Peter Shaughnessy for
Downloaded from mbe.oxfordjournals.org by guest on September 13, 2010
samples of NFSs, and Ian Boyd and the British Antarctic
Survey for samples of AFSs.
LITERATURE CITED
APPLE, R., C. WEELER, T. BECKER, W. KLITZ, and H. ERLICH.
1993. HLA DR-DO disease associations in hispanic cervi-
cal carcinoma. Hum. Immunol. 37:62.
ARNASON, U., K. BODIN, A. GULLBERG, C. LEDJE, and S. MOU-
CHATY. 1995. A molecular view of pinniped relationships
with particular emphasis on the true seals. J. Mol. Evol. 40:
78–85.
BONNELL, M. L., and R. K. SELANDER. 1974. Elephant seals:
genetic variation and near extinction. Science 184:908–909.
BROWN, J. H., T. S. JARDETZKY, J. C. GORDA, L. J. STERN, R.
G. URBAN, J. L. STROMINGER, and D. C. WILEY. 1993.
Three-dimensional structure of the human class II histocom-
patability antigen HLA-DR1. Nature 364:33–39.
BUGAWAN, T. L., and H. A. ERLICH. 1991. Rapid typing of
HLA-DQB1 DNA polymorphism using nonradioactive ol-
igonucleotide probes and amplified DNA. Immunogenetics
33:163–170.
CHUNG, C. T., and R. H. MILLER. 1988. A rapid and convenient
method for the preparation and storage of competent bac-
terial cells. Nucleic Acids Res. 16:3580.
DEVEREUX, J., P. HAEBERLI, and O. SMITHIES. 1984. A com-
prehensive set of sequence analysis programs for the VAX.
Nucleic Acids Res. 12:387–395.
EDWARDS, S. V., and P. W. HEDRICK. 1998. Evolution and ecol-
ogy of MHC molecules: from genomics to sexual selection.
TREE 13:305–311.
GERMAIN, R. N., and D. H. MARGULIES. 1993. The biochem-
istry and cell biology of antigen processing and presenta-
tion. Annu. Rev. Immunol. 11:403–450.
GUILLET, J. G., M. Z. LAI, T. J. BRINER, S. BUUS, A. SETTE,
H. M. GREY, J. A. SMITH, and M. L. GEFTER. 1987. Im-
munological self, nonself discrimination. Science 235:865–
870.
GYLLENSTEN, U. B., D. LASHKARI, and H. A. ERLICH. 1990.
Allelic diversification at the class II DQB locus of the mam-
malian major histocompatability complex. Proc. Natl. Acad.
Sci. USA 87:1835–1839.
HARDING, C. V. 1996. Class II antigen processing: analysis of
compartments and functions. Crit. Rev. Immunol. 16:13–
29.
HILL, A. V. S., C. E. ALLSOPP, D. KWAITOWSKI, N. M. ANSTEY,
P. TWUMASI, P. A. ROWE, S. BENNETT, D. BREWSTER, A. J.
618 Hoelzel et al.
MCMICHAEL, and B. M. GREENWOOD. 1991. Common West
African HLA antigens are associated with protection from
severe malaria. Nature 352:595–600.
HILL, A. V. S., D. KWIATKOWSKI, A. J. MCMICHAEL, B. M.
GREENWOOD, and S. BENNETT. 1992. Maintenance of MHC
polymorphism. Nature 335:167–170.
HOELZEL, A. R., and A. GREEN. 1992. Analysis of population-
level variation by sequencing PCR-amplified DNA. Pp.
159–188 in A. R. HOELZEL, ed. Molecular genetic analysis
of populations: a practical approach. Oxford University
Press, Oxford.
HOELZEL, A. R., J. HALLEY, S. J. O’BRIEN, C. CAMPAGNA, T.
ARNBOM, B. J. LE BOEUF, K. RALLS, and G. A. DOVER.
1993. Elephant seal genetic variation and the use of simu-
lation models to investigate historical population bottle-
necks. J. Hered. 84:443–449.
HOELZEL, A. R., J. M. HANCOCK, and G. A. DOVER. 1993.
Generation of VNTRs and heteroplasmy by sequence turn-
over in the mitochondrial control region of two elephant
seal species. J. Mol. Evol. 37:190–197.
HUDSON, R. R., and N. C. KAPLAN. 1985. Statistical properties
of the number of recombination events in one history of a
sample of DNA sequences. Genetics 111:147–164.
HUGHES, A. L., and M. NEI. 1988. Patterns of nucleotide sub-
stitution at major histocompatibility complex class I loci
reveals overdominant selection. Proc. Natl. Acad. Sci. USA
86:958–962.
. 1989. Nucleotide substitution at major histocompati-
bility complex class II loci: evidence for overdominant se-
lection. Proc. Natl. Acad. Sci. USA 86:958–962.
KASLOW, R. A., M. CARRINGTON, R. APPLE et al. (12 co-au-
thors). 1996. Influence of combinations of human major his-
tocompatibility complex genes on the course of HIV-1 in-
fection. Nat. Med. 2:405–411.
KING, J. E. 1983. Seals of the world. Oxford University Press,
Oxford.
KLEIN, J. 1987. Origin of major histocompatibility complex
polymorphism: the trans-species hypothesis. Hum. Immu-
nol. 19:155–162.
KUMAR, S., K. TAMURA, and M. NEI. 1993. MEGA: molecular
evolutionary genetic analysis. Version 1.01. The Pennsyl-
vania State University, University Park.
LANDER, E. S., and N. J. SCHORK. 1994. Genetic dissection of
complex traits. Science 265:2037–2048.
LONGENECKER, B. M., F. PAZDERKA, J. S. GAVORA, J. L. SPEN-
CER, E. A. STEPHENS, R. L. WITTER, and R. F. RUTH. 1977.
Role of the major histocompatability complex in resistance
to Marek’s disease: restriction of the growth of JMV-MD
tumour cells in genetically resistant birds. Adv. Exp. Biol.
88:287.
MARUYAMA, T., and M. NEI. 1981. Genetic variability main-
tained by mutation and overdominant selection in finite
populations. Genetics 98:441–459.
MILLER, K. M., R. E. WITHLER, and T. D. BEACHAM. 1997.
Molecular evolution at MHC genes in two populations of
chinook salmon Oncorhynchus tshawytscha. Mol. Ecol. 6:
937–954.
MURRAY, B. M., S. MALIK, and B. N. WHITE. 1995. Sequence
variation at the major histocompatibility complex locus
DQB in beluga whales (Delphinapterus leucas). Mol. Biol.
Evol. 12:582–593.
NEI, M., and T. GOJOBORI. 1986. Simple methods for estimat-
ing the numbers of synonymous and nonsynonymous nu-
cleotide substitutions. Mol. Biol. Evol. 3:418–426.
NEI, M., and L. JIN. 1989. Variances of the average numbers
of nucleotide substitutions within and between populations.
Mol. Biol. Evol. 6:290–300.
O’BRIEN, S. J., and J. F. EVERMANN. 1988. Interactive influence
of infectious disease and genetic diversity in natural popu-
lations. Trends Ecol. Evol. 3:254–259.
O’HUIGIN, C. 1995. Quantifying the degree of convergence in
primate MHC-DRB genes. Immuol. Rev. 143:123–140.
OTTING, N., M. KENTER, P. VAN WEEREN, M. JONKER, and R.
E. BONTROP. 1992. MHC-DQB repertoire variation in hom-
inoid and Old World primate species. J. Immunol. 149:461–
470.
POTTS, W. P., and P. R. SLEV. 1995. Pathogen-based models
favoring MHC genetic diversity. Immunol. Rev. 143:181–
197.
RIEDMAN, M. 1990. The pinnipeds; seals, sea lions and wal-
ruses. University of California Press, Berkeley.
Downloaded from mbe.oxfordjournals.org by guest on September 13, 2010
SAITOU, N., and M. NEI. 1987. The neighbor-joining method:
a new method for reconstructing phylogenetic trees. Mol.
Biol. Evol. 4:406–425.
SATTA, Y. 1997. Effects of intra-locus recombination on HLA
polymorphism. Hereditas 127:105–112.
SETTE, A., S. BUUS, S. COLON, S. A. SMITH, C. MILES, and H.
M. GREY. 1987. Structural characteristics of an antigen re-
quired for its interaction with IA and recognition by T-cells.
Nature 328:395–399.
SLADE, R. W. 1992. Limited MHC polymorphism in the south-
ern elephant seal: implications for MHC evolution and ma-
rine mammal population biology. Proc. R. Soc. Lond. B
Biol. Sci. 249:163–171.
SLADE, R. W., C. MORITZ, and A. HEIDEMAN. 1994. Multiple
nuclear-gene phylogenies: application to pinnipeds and
comparison with a mitochondrial DNA gene phylogeny.
Mol. Biol. Evol. 11:341–356.
STEPHENS, J. C. 1985. Statistical methods of DNA sequence
analysis: detection of intragenic recombination or gene con-
version. Mol. Biol. Evol. 2:539–556.
STEWART, C. B., and T. R. DISOTELL. 1998. Primate evolution
in and out of Africa. Curr. Biol. 8:R582–R588.
SWOFFORD, D. L., and S. H. BERLOCHER. 1987. Inferring evo-
lutionary trees from gene frequency data under the principle
of maximum parsimony. Syst. Zool. 36:293–325.
TAKAHATA, N., and M. NEI. 1990. Allelic genealogy under
overdominant and frequency-dependent selection and poly-
morphism of major histocompatibility complex loci. Ge-
netics 124:967–978.
THURSZ, M. R., H. C. THOMAS, B. M. GREENWOOD, and A. V.
S. HILLS. 1997. Heterozygote advantage for HLA class-II
type in hepatitis B virus infection. Nat. Genet. 17:11–12.
TROWSDALE, J., V. GROVES, and A. ARNASON. 1989. Limited
MHC polymorphism in whales. Immunogenetics 29:19–24.
YUHKI, N., and S. J. O’BRIEN. 1990. DNA recombination and
natural selection pressure sustain genetic sequence diversity
of the feline MHC class-I genes. J. Exp. Med. 172:621–
630.
AXEL MEYER, reviewing editor
Accepted January 8, 1999