| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Departments of Structural Biology and Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA 94305;
Department of Medicine and Epidemiology, University of California School of Veterinary Medicine, Davis, CA 95616;
Department of Natural Resources, Colorado Division of Wildlife, Montrose, CO 81401;
Department of Animal Sciences, Cell Biology and Immunology Group, Wageningen University, Wageningen, The Netherlands; and
¶ Department of Zoology, University of Aberdeen, Aberdeen, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-chain gene DAB from 15 rainbow and 10 brown
trout. Both genes are highly polymorphic in both species and diploid in
expression. The MHC class I alleles comprise several highly divergent
lineages that are represented in both species and predate genera
separation. The class II alleles are less divergent, highly species
specific, and probably arose after genera separation. The striking
difference in salmonid MHC class I and class II evolution contrasts
with the situation in primates, where lineages of class II alleles have
been sustained over longer periods of time relative to class I
lineages. The difference may arise because salmonid MHC class I and II
genes are not linked, whereas in mammals they are closely linked. A
prevalent mechanism for evolving new MHC class I alleles in salmonids
is recombination in intron II that shuffles 
1 and 2 domains into
different combinations. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Structural and functional similarities of MHC class I and II molecules almost certainly reflect a common origin. For several reasons, including the restriction of class II genes to the MHC in mammals but the presence of various class I-like genes on other chromosomes, the MHC class I gene is considered to be the ancestral form and MHC class II gene to be the derivative (4). The association of class I and II genes in the MHC and the linkage disequilibrium between them suggest that there may be an advantage to this arrangement and persistent selection to maintain it.
Evidence available from the study of birds, amphibians, and cartilaginous fish indicates that these other vertebrate classes, like mammals, have an MHC that contains genes for both the polymorphic class I and II molecules (5, 6, 7). Exceptional are the bony fishes (teleosts), in which the polymorphic class I and II genes segregate as though unlinked (8, 9, 10); thus, in these species there is no single genetic region that can be called the MHC. The absence of genetic linkage between MHC class I and II genes has potentially important implications for the way that natural selection shapes the polymorphism at individual class I or II loci. To illustrate, when the two types of genes are linked, selection for a particular class II allele will inevitably change the allele frequencies at an unselected, but linked, class I locus, and vice versa. On the other hand, selection at the class II locus will have no effect on the allele frequencies at an unselected, unlinked class I locus, and vice versa. Thus, it is possible that the paradigms established from study of mammalian MHC polymorphism need not apply to their counterparts in teleosts.
To study the evolution and polymorphism of class I and II genes in bony
fish, we have compared rainbow trout (Oncorhynchus mykiss)
and brown trout (Salmo trutta), two species of salmonid. The
rainbow trout is representative of salmonids of the Pacific drainage
(genus Oncorhynchus) and the brown trout of salmonids of the
Atlantic drainage (genus Salmo) (11, 12). The
two genera are estimated to have separated for 15–20 million years
(13, 14). The salmonids are considered to be relatively
primitive teleosts, and the modern species are characterized by
pseudotetraploidy resulting from a whole genome duplication that
occurred in their common ancestor some 25–100 million years ago
(15). Consequently, four copies of a gene can be present,
as shown for a chemokine gene (16) and also for
2-microglobulin
(2m)3
genes (17) (K. E. Magor, B. P. Shum, and P.
Parham, manuscript in preparation). In addition to
2m, rainbow trout class I heavy chain and
class II genes have been described (9, 18, 19, 20, 21, 22, 23, 24, 25, 26).
Previously, we characterized two class I heavy chain genes:
UAA, which encodes a highly divergent and nonpolymorphic
gene expressed at low levels, and UBA, which is polymorphic,
highly expressed in spleen, and has structural features characteristic
of classical class I molecules that present diverse peptide Ags to T
cells (25). There are multiple UBA-related MHC
class I genes in the rainbow trout (9, 22, 23) (our
unpublished observations), with some members not linked to the main
class I region (9); however, the number of functional
genes is uncertain. Through extensive analysis of genomic DNA, Miller
and Withler classified three salmonid class I genes (A,
B, and UA) and reported limited diversity within
each gene (22). Hansen et al. (9) described
three transcribed classical class I alleles in individual fish, while
Nakanishi et al. (23) reported expression of a classical
class I gene from one haploid set of chromosomes in a diploidized
genome. For the rainbow trout MHC class II -chain, there is evidence
for two loci, although it is not clear whether both are transcribed
(18, 19, 21, 24). Recently, the rainbow trout class II
-chain, Onmy-DAA, was shown to be encoded by a single,
expressed, polymorphic locus (26).
In this study, the polymorphism of the transcribed UBA class I gene and that of the DAB class II gene in 15 rainbow trout and 10 brown trout were compared. Our results are consistent with individuals of both species expressing single, polymorphic UBA and DAB genes.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Ten rainbow trout (O. mykiss) and 10 brown trout (S. trutta) were collected from the headwaters of the Colorado River near Parshall, CO, in the spring of 1999. Five additional rainbow trout samples were from California hatcheries. All Colorado River trout were collected from the same excursion. All wild rainbow trout were spawning females; all brown trout were randomly selected. Fish were not genetically typed or selected for MHC polymorphism in any way before sequence analysis. Animals were euthanized by immersion in 0.5% 3-aminobenzoic acid ethyl ester (MS-222; Sigma, St. Louis, MO).
RT-PCR
Freshly collected spleen and liver samples were fixed immediately and stored in the reagent RNAlater (Ambion, Austin, TX). Total RNA was extracted, and first-strand cDNA were prepared by priming with oligonucleotide(dT)16 following standard protocols (27). For RT-PCR amplification, oligonucleotide primers were designed from regions conserved among full-length salmonid MHC cDNA sequences available from the GenBank database, including those from both the genera Oncorhynchus and Salmo. Primers were chosen conservatively to facilitate successful amplification from both rainbow trout and brown trout; thus, most, but not all, of the coding regions were amplified.
The MHC class II -chain DAB cDNA was amplified with the
primers DAB-LAf1 (5'-TC TCT GGA ACA GAT GGA TAT T-3') and DAB-CTr1
(5'-TA CTA CAG CAC CCC AGA AGA-3') under the following PCR conditions:
initial denaturation at 94°C for 1 min; then 22 cycles of
denaturation at 94°C for 15 s, primer annealing at 50°C for
15 s, and extension at 72°C for 1 min; and a final 10-min
incubation at 72°C included for complete DNA extension. We used a
relatively low-stringency annealing temperature for all PCRs with the
aim of increasing the number of allelic variants that could be
amplified.
For MHC class I heavy chain gene amplification, three different forward primers were paired in separate reactions with a common reverse primer, UBA-3UTr1 (5'-TGT GTT ATG TTC TTG AGA AGT TCC-3'). The PCR conditions were the same as described for DAB, except that the extension time was increased from 1 to 2 min. Together, the three PCRs produced one or two different MHC class I sequences from each fish. The combination of the forward primer UBA-5UTf1 (5'-G GAG ATA ACA TAG TTC GTC AAC AT-3') with UBA-3UTr1 generates PCR products that include the entire coding region (the first two nucleotides of the initiation codon are included in the forward primer). This primer pair worked for 9 of the 15 Onmy-UBA alleles, but for none of the Satr-UBA alleles in our trout samples. The combination of the forward primer UBA-LBf1 (5'-T ATT ATC TTG CTG GTG CTG GGA AT-3') with UBA-3UTr1 amplifies cDNA from all but two UBA alleles; however, the product lacks part of the leader peptide-encoding sequence. Therefore, all alleles amplified by the first primer set (UBA-5UTf1/UBA-3UTr1) were also amplified by the second set (UBA-LBf1/UBA-3UTr1); the reverse was not true. The combination of forward primer UBA-LCf1 (5'-T TTC ATC ATT TTG CTC CTG GGA AT-3') with UBA-3UTr1 amplified Onmy-UBA*0202 and Satr-UBA*1001 cDNA, neither of which could be amplified by the former two primer sets.
DNA sequencing
PCR-amplified cDNA was cloned into the pCR4-TOPO plasmid (Invitrogen, Carlsbad, CA), and multiple clones were selected for DNA sequence analysis. Sequences were determined on both DNA strands using the BigDye Terminator Cycle Sequencing kit (Applied Biosystems, Foster City, CA) and a 377 Automated DNA Sequencer (Applied Biosystems). Sequence assembly and initial analysis were performed with the computer program AutoAssembler (version 2.1; Applied Biosystems) and the Wisconsin Sequence Analysis software package (version 10.1; Genetics Computer Group, Madison, WI).
Nomenclature and sequences
The nomenclature used in this paper adheres to the recommended
convention for other species beginning with the prefixes
Onmy- for genes of rainbow trout (O. mykiss) and
Satr- for brown trout (S. trutta). Four digits
are used to number alleles of the UBA class I heavy chain
gene and the DAB class II -chain gene: the first two
digits are used for lineages and major alleles, and the second two
digits are used for variants or subtypes (28). A fifth
digit is added to indicate noncoding changes. The salmonid
UAA is a nonclassical, nonpolymorphic class I heavy chain
gene, the first class I gene isolated from rainbow trout (17, 25). The rainbow trout Onmy-UBA*0101 allele was
previously described by us (25). The full-length
Onmy-UBA*0201 cDNA (formerly named Onmy-UCA) was
isolated by screening a rainbow trout cDNA library in Aberdeen (R.
J. M. Stet and C. Secombes); this allele was not found in the
Colorado and California fish analyzed. New sequences reported here were
deposited into the GenBank database under the accession numbers
AF296359–AF296373 for Onmy-UBA alleles, AF296374–AF296383
for Satr-UBA alleles, AF296384–AF296397 for
Onmy-DAB alleles, and AF296398–AF296409 for
Satr-DAB alleles.
Sequence analysis
Sequences were aligned with the Pileup program of the Genetics
Computer Group package and then adjusted manually. Phylogenetic
analysis was performed using PAUP software (Phylogenetic Analysis Using
Parsimony (and Other Methods), version 4.0b4a, Sinauer, Sunderland,
MA). Uncorrected genetic distances "p" were used to generate
dendrograms with the neighbor-joining method (29).
Confidence in individual nodes was evaluated from 1000 bootstrap
replications, and only those with frequencies over 70% were retained.
For MHC class I UBA, nucleotide sequences encompassing the
entire mature protein coding region were analyzed as well as sequences
for individual domains. For MHC class II, the entire PCR-amplified
DAB sequences encoding 
97% of the mature protein
were used.
Pairwise comparisons of amino acid sequences were computed using the Distance program of the Genetics Computer Group package with no corrections. Included in the analyses were Onmy-UBA*0101, Onmy-DAB*0101, and Onmy-DAB*0201. Complete sequences for mature proteins were compared, except that the Onmy- and Satr-DAB polypeptides lacked seven residues in total from the two ends. The human HLA sequences used in this analysis were all those for which a complete sequence for the mature polypeptide is known; these include 80 HLA-A, 130 HLA-B, 27 HLA-C, and 25 HLA-DRB1 allotypes. All human sequences were obtained from the database (April 2000 update) at the Anthony Nolan Bone Marrow Trust (London, U.K.) maintained by Steven G. E. Marsh (http://www.anthonynolan.org.uk/HIG/data.html). Leopard shark class I sequences were reported by Okamura et al. (30).
The program Reticulate was used to analyze sets of allelic sequences for phylogenetic inconsistency and to identify candidate events of past recombination (31).
The rates of nonsynonymous (dN) and synonymous (dS) nucleotide substitutions were calculated according to the method of Nei and Gojobori (32) with the program Synonymous/Nonsynonymous Analysis Program (SNAP; available at http://hiv-web.lanl.gov/SNAP/WEBSNAP/SNAP.html). Peptide binding pocket residues in the Ag recognition site of salmonid MHC class I molecules were predicted by homology with mammalian structures (33, 34).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). These sequences were obtained from
clones generated by RT-PCR, thus ensuring that only transcribed genes
were studied. The class I heavy chain clones gave complete mature
protein-coding region sequences for alleles of the UBA
locus; the class II -chain clones gave mature protein-coding region
sequences for alleles of the DAB locus that lacked four
nucleotides from the 5' end and 15 nucleotides from the 3' end (Fig. 1).
|
|
). All of these alleles were novel, as
this is the first report of MHC class I and II alleles from this
species. From rainbow trout (O. mykiss), 15 class I
Onmy-UBA and 14 class II Onmy-DAB were identified
(Table II). Of these, 14 UBA and 11 Onmy-DAB
alleles were novel (Table III). None of
the fish expressed more than two class I or II alleles, and many of the
fish were heterozygotes (Table I
). These data are consistent with
diploid expression of single class I UBA and class II
DAB genes.
|
|
and II).
Several groups of investigators have previously reported MHC class I
heavy chain and class II -chain sequences from rainbow trout. These
include sequences derived from both genomic DNA (21, 22)
and cDNA (9, 17, 19, 20, 23, 24, 25). Additional heterogeneity
arises from how much of the coding region has been sequenced. Each
group of investigators has used a different nomenclature with the
result that identical loci and alleles can have different names, while
distinct loci can have the same name. Within this investigation we have
aimed for as complete coding sequences as was practically possible and
an internally consistent and parsimonious nomenclature. For rainbow
trout, the alleles identified here are compared with representative
reported sequences in Table III, where the different names are also
listed as well as their frequency in our sample of fish.
UBA class I alleles are older and more divergent than class II DAB alleles
Variation at the salmonid MHC loci was compared using phylograms
constructed from nucleotide sequences and shown in Fig. 2. The phylograms for class I (Fig. 2A) and class II (Fig. 2B) are of very distinct
topologies that differ in two important ways. First, the class II
alleles are considerably less divergent than the class I alleles, as
shown by the relatively compact size of the class II tree and the
shortness of its branches, which tend to radiate in a stellate fashion.
In contrast, the class I tree has much deeper branches and more
hierarchical structure. Second, in the class II phylogram the alleles
for the two species are completely segregated, whereas in the class I
tree there is an intermingling of Onmy- and
Satr-UBA alleles in most of the major branches. The
phylograms indicate that old lineages of MHC class I alleles, which
predate separation of the genera Salmo and
Oncorhynchus, have been retained in modern rainbow and brown
trout. In contrast, almost all of the class II DAB diversity
now seen in rainbow and brown trout is species specific and thus
originated subsequent to genera divergence. Eight positions in
the DAB amino acid sequence have species-specific patterns of
substitution (Fig. 1B). These asymmetries are the opposite
of what have been observed in the higher primates, where most of the
MHC class I variation is species specific, while older lineages of MHC
class II alleles are shared by different species (2, 3).
|
1, 2, 3, and
transmembrane/cytoplasmic regions). These comparisons showed that
highly divergent pairs of alleles, such as Onmy-UBA*0101 and
*0201 and Satr-UBA*0101 and *0501, are divergent throughout
the sequence (data not shown).
Inspection of the sequences shows that substitutions distinguishing UBA
allotypes are present throughout the coding region, but with particular
concentration in the 1 and 2 domains (Fig. 1A). In
addition, there are various insertions/deletions (indels) at the
junction between the extracellular domains and the transmembrane
anchor. For DAB allotypes, most variation (82%) is concentrated in
the 1 domain (Fig. 1B), which also contains an indel in
some brown trout alleles.
To assess further the diversity in rainbow and brown trout MHC class I
and class II allotypes, the distributions of pairwise differences in
amino acid sequence were determined for mature UBA and DAB
polypeptides. For comparison, distributions for human class I, shark
class I, and human class II allotypes were also presented (Fig. 3). Onmy-UBA and
Satr-UBA are much more diverse than HLA class I allotypes,
as assessed by both the range and the mean of pairwise differences
(Fig. 3, A–C). Whereas HLA-B, the most polymorphic HLA
class I locus, has a mean difference of 5.4%, the values are 23.8 and
23.4% for Onmy-UBA and Satr-UBA, respectively.
Even when all three highly polymorphic HLA class I genes are included
in the analysis, neither the mean nor the range approach the values
seen for salmonid class I (Fig. 1C). Because of the large
range of differences and the small number of sequences sampled, the
distribution of salmonid class I differences is flatter and more
amorphous than the well-defined shapes of the HLA class I
distributions. The distribution of pairwise differences in the class I
allotypes of the leopard shark (Triakis scyllium; Fig. 3D), the most comprehensive dataset of class I alleles from
a lower vertebrate, is more similar to that of the human class I and is
distinct from that of the trout class I. However, the intra- and
interlocus modes are not well defined in sharks due to the sharing of
an 2 lineage between all Trsc-UBA and some
Trsc-UAA allotypes (30).
|
, E–G). From this sampling
Satr-DAB is slightly more variable than HLA-DRB1, the most
polymorphic HLA class II locus, and Onmy-DAB is comparable
in diversity to HLA-DRB1. Thus, MHC class II diversity in human and the
two salmonid species appears similar, whereas the MHC class I diversity
is very different. New salmonid class I alleles have frequently been generated by recombination within intron II
We used the program Reticulate to assess the role of recombination
in the generation of MHC diversity in trout (Fig. 4). This program has been used previously
for the analysis of recombination in HLA class I and II genes
(31). The result obtained for Onmy- and
Satr-DAB resembles that reported for HLA-DRB1
(Fig. 4B) (31). Thus, as apparent from sequence
inspection, there is evidence for extensive recombination in the region
encoding the 1 domain (represented as filled squares in Fig. 4B). Indeed, Satr-DAB*0301 could be a
recombinatorial product of the parental sequences
Satr-DAB*0201 and *0601 (data not shown).
|
A). Although evidence for past recombinations can be seen
throughout the gene (presence of small filled squares), the pattern is
dominated by phylogenetic inconsistencies between sites in exon II
encoding the 1 domain and sites in the 3' exons that encode the
remainder of the mature class I heavy chain. These phylogenetic
inconsistencies provide evidence for the prevalence of recombinations
within intron II of the UBA gene and its role in the
generation of new salmonid MHC class I alleles. Additional analyses
were performed separately for rainbow and brown trout class I alleles,
yielding similar patterns (data not shown). This type of recombination,
which serves to shuffle 1 domains into novel combinations with 2
domains and the remainder of the heavy chain, has been seen in HLA
class I, but is not common (35). The species difference
could be due to variation in the size of intron II, which is 250 bp
in length in human MHC class I genes, but is estimated to be 10–20 kb
in rainbow trout.
From simple inspection of the sequences, ancestral alleles resembling
Onmy-UBA*02 appear to have been particularly active in
recombination within intron II. The 2 domain through the carboxyl
terminus of Onmy-UBA*02 is highly similar to
Onmy-UBA*0301, *0401, *0501, and *0601, but in the 1
domain they are divergent (Figs. 1A and 5). Sequences with
the reverse recombinatorial type (those resembling
Onmy-UBA*02 only in the 1 domain and in the leader
sequence) were not found in our samples, but have been reported by
Hansen et al. (e.g., Onmy-UBA*SP3 in Fig. 5) (9). Together,
Onmy-UBA*02 and allotypes related to it by recombination
(Onmy-UBA*03–06) represent one-third of the UBA allotypes
found among our rainbow trout (5 of 15). Furthermore, two related brown
trout allotypes (Satr-UBA*0901 and *1001) are also
identified, although a counterpart of Onmy-UBA*02 has yet to
be described in the brown trout (Figs. 1A and 5). To assess
whether intron II recombination in rainbow trout is specifically a
property of the Onmy-UBA*02-related alleles, reticulate
analysis was performed on UBA alleles excluding this group.
A strong phylogenetic inconsistency was still present at the boundary
between regions encoding the 1 and 2 domains (data not shown).
Thus, 1 shuffling appears to be a general phenomenon among UBA
allotypes.
|
1, but is
similar to Satr-UBA*0601 in 2 and other downstream
regions (Fig. 1A). The 1 domains are also identical in
Onmy-UBA*0101/*1101, Onmy-UBA*0601/*1501,
and Satr-UBA*0201/*0301, and the 2 through the carboxyl
termini are identical in Onmy-UBA*0202/*0401/*0501,
Onmy-UBA*0701/*0801, Onmy-UBA*1501/*1601, and
Satr-UBA*0401/*0801 (Fig. 1A). Evidence for diversifying selection in the Ag presentation site is more apparent in salmonid class II than class I
As well as their high polymorphism, the salmonid class I
and II molecules reported here have primary structures consistent with
them being Ag-presenting molecules. To search for evidence of natural
selection, we examined how the ratio between rates of
dN and dS substitution
varied within the molecule (Table IV).
For the class II DAB alleles of either rainbow trout or
brown trout, simple comparison of the sequences encoding the 1 and
2 domains revealed a role for positive selection
(dN/dS > 1) in generating
1 domain polymorphism. The 1 domain forms part of the Ag
binding site, whereas the 2 domain is not directly involved in Ag
presentation. In striking contrast, in neither rainbow trout nor brown
trout did similar comparisons of sequences encoding the class I
Ag binding site (1 and 2 domains) and those encoding the 3
domain provide any evidence for diversifying selection. This showed
that the recent and relatively small amount of sequence diversity
acquired by the class II alleles is readily seen as the result of
natural selection, whereas for the much larger amount of class I
diversity accumulated over a much longer time period, the noise appears
to have obscured any signal.
|
A). By confining
analysis to more recent, species-specific diversification within this
single lineage, evidence for natural selection was detected (Table II).
For both rainbow and brown trout this was evident from simple
comparison of the Ag binding site (1 and 2) vs 3 and was
strengthened when the pocket residues (those that directly contact
bound Ag) of the 1 and 2 domains were analyzed separately. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This study demonstrates a striking difference in the polymorphism of the MHC class I and II loci of the rainbow and brown trout that is reflective of very different evolutionary histories. The MHC class I alleles represent highly divergent and ancient lineages that predate separation of the genera Oncorhynchus and Salmo, whereas the MHC class II alleles form closely related, species-specific clusters that radiated after genera separation; Miller and Withler have made similar observations for salmonid class II alleles (21). That both rainbow and brown trout are similar in these respects indicates firstly that these properties are likely to have generality for other salmonid species (a similar situation is also seen in the Atlantic salmon (U. Grimholt and R. J. M. Stet, unpublished observations)), and secondly that as the species have been separated for 15–20 million years with no overlapping range, it is unlikely that the overall difference between the class I and II polymorphisms is a consequence of selection by particular pathogens.
MHC class I and II polymorphisms have been studied most extensively in humans and other primates, and these data have been compared here to those obtained from the salmonids. The extent of sequence diversity among the salmonid DAB class II allotypes is comparable to that of HLA-DRB1, the most polymorphic human class II locus. As for HLA-DRB1, there is clear evidence in the Onmy- and Satr-DAB sequences for polymorphism in the Ag binding site being the result of positive selection.
What distinguishes the salmonid class II DAB locus from its primate counterparts is the absence of distinct lineages common to rainbow trout and brown trout. In primates, certain human HLA-DQ and DR lineages can be recognized in prosimians, which last shared a common ancestor with humans around 85 million years ago (3).
In contrast, HLA class I lineages can only be recognized in the great
apes, which diverged from humans 6 million years ago
(2). In this context it is particularly impressive that
salmonid class I UBA has lineages that can be recognized in
rainbow trout and brown trout and thus are at least 15–20 million
years old. In addition, the extent of diversity within UBA allotypes is
much greater than that seen in human HLA-A, -B, and -C loci. Much of
this diversity was not apparent as the result of selection on the Ag
binding site, an effect that was only evident when alleles with a
single lineage were compared. These properties were consistent
with the class I lineages being ancient, and for them to have
accumulated many neutral mutations that in the analysis overwhelmed the
effects of selection. Thus current methods of comparing the rates of
synonymous and nonsynonymous substitutions might not be appropriate for
ancient lineages, a conclusion also reached by Flajnik et al.
(36).
Evidence for the divergent salmonid MHC class I lineages also comes
from the work of other investigators, using different nomenclature
(shown in Table III). From analysis of genomic DNA, Miller and Withler
(22) described three groups of salmonid MHC class I
sequence (UA, A, and B), conserved in
several salmonids species and thus predating their divergence. The
UA lineage resembles Onmy-UBA*02 and the related
recombinant alleles described here; the A lineage resembles
Onmy-UBA*01-like sequences. In contrast, the B
lineage is distinct from the sequences we found, but it is not known
whether it is expressed. In rainbow trout, Hansen et al.
(9) described three lineages of MHC class I sequence
(UAA, UBA, and UCA), each of which
resembles certain of the Onmy-UBA alleles defined here:
their UAA lineage corresponds to
Onmy-UBA*01-like alleles, their UBA
lineage corresponds to recombinant alleles with
Onmy-UBA*0201-like 1-coding sequences (e.g.,
Onmy-UBA*SP3 in Fig. 5), and their UCA lineage
corresponds to the Onmy-UBA*02–06 group of alleles that
have 2-coding sequences like that of Onmy-UBA*0201. Of
the two rainbow trout MHC class I sequences described by Nakanishi et
al. (23), one resembles Onmy-UBA*0401, and the
other resembles Onmy-UBA*1101 (Table III).
Hansen et al. (20) first noted that the 2 domain of
Onmy-UCA*C32 (identical with Onmy-UBA*0301) has
similarity with class I sequences from common carp and zebrafish. Here,
we found that such similarity between different orders of teleost is
not restricted to the 2 domain; for example, the common carp,
Cyca-UA1*01, is 74% identical in amino acid sequence with
Satr-UBA*0501 in 1 and 73% identical with
Onmy-UBA*0201 in 2 (Fig. 5) (37). This
preservation is remarkable, since the orders of salmoniforms and
cypriniforms last shared a progenitor over 120 million years
ago.
The divergence of the UBA lineages raises the question of whether they represent different loci or different alleles. Of relevance to this is the historical tetraploidy of the salmonid genome and the observation that at least some other genes are still present in four copies (16). The data we obtained are consistent with all UBA sequences being alleles of a single, diploid locus, which is also corroborated by Nakanishi et al. (23). Thus, if four copies of the UBA gene are present in an individual salmonid genome, then it appears that only two are expressed. The only piece of evidence not consistent with this view is one example of individual rainbow trout expressing three different classical MHC class I sequences (9). That could have been a situation where three copies of the UBA gene were being transcribed or where two haplotypes were expressed but one of them had duplicated UBA genes (seen for shark as Trsc-UAA/UBA). However, the method we used should have identified this phenomenon if it was present in any one of the 15 rainbow trout studied, and hence it may be infrequent. Until additional evidence supporting a more complicated model is found, the simplest interpretation of the data is that all of the UBA sequences represent a single set of alleles. Another example of divergent lineages of MHC class I alleles is to be found in the African clawed frog Xenopus laevis, with the f/f and g/g alleles sharing only 78% amino acid identity overall (36).
Characteristic of salmonid class I polymorphism is the prevalence of
alleles formed by recombination in intron II that separates the exons
encoding the two halves of the Ag binding site (the 1 and 2
domains). This type of recombinant 1-shuffled allele has been
described in humans, but is not common (35). What could
explain its increased frequency in salmonids is the presence of a much
larger intron II than in HLA class I genes. The length of this intron
in UBA is not precisely known, but there is substantial
indirect evidence suggesting that it is large. Analyses of
UBA genomic clones corresponding to transcribed genes by
both Hansen et al. (9) and ourselves (unpublished
observations) have only revealed clones containing exon III (encoding
2) and all the downstream exons. Moreover, the intron II of the
Cyca-UA1 class I gene of common carp was estimated by PCR
analysis to be about 14 kb in length (37), and the intron
II of the zebrafish class I gene Dare-UEA was estimated to
be about 10 kb and contains repetitive elements (38). In
general, the salmonid genome is rich in repetitive elements, so their
presence combined with length could explain a high rate of
recombination in the intron II of the UBA gene. It is also
intriguing that the carp Cyca-UA1*01 sequence shared
homology in its 1 domain with Satr-UBA*0501 and in its
2 domain with Onmy-UBA*0201. Thus, the mechanism for 1
shuffling could be old and operative in many teleost species.
From analysis of separately isolated genomic DNA fragments encoding the
class I 1 and 2 domains from several salmonid species, Miller and
Withler (22) concluded that salmonids have limited class I
diversity. By studying the exons separately, their study would have
missed the considerable diversity contributed by recombination in
intron II. To mimic the approach taken by Miller and Withler with our
dataset we compared the variability in the individual domains of the
peptide binding site and found it to be substantial, even when the
analysis was restricted to a single lineage of alleles. For example,
the average pairwise differences in amino acid sequence for the 1
and 2 domains of the six Onmy-UBA*01-like alleles were
10.3 and 12.8%, respectively. These differences are comparable to
those observed for HLA-B (10.4 and 7.6%, respectively). Similar
analysis of the seven rainbow trout 1 domain sequences from the A
lineage dataset reported by Miller and Withler (22) gave a
comparable value (16.0%); their dataset contained only three rainbow
trout 2 sequences from this lineage, and these had a mean pairwise
difference of 2.4%. In our dataset, we found limited polymorphism in
the Onmy-UBA*02 lineage, where the 2 domains have an
average pairwise difference of only 2.3%, a value that does not
increase when the brown trout Satr-UBA*0901 allotype is
included in the analysis (2.2%). Thus, the extent of the polymorphism
in Onmy-UBA can vary with the allelic lineage and with the
protein domain. There are also likely to be differences between trout
populations that could also contribute to the difference between our
assessment of the extent of salmonid class I polymorphism and that of
Miller and Withler (22). Our approach has been to compare
MHC class I diversity in trout with that in humans, which is generally
considered to be highly polymorphic. By that criterion, the
UBA loci of the rainbow and brown trout populations we
studied are also highly polymorphic.
Bony fish are estimated to represent about one-half of all vertebrate species (12). Whereas other species have linked class I and II genes within a single chromosomal region called the MHC, in bony fish the class I and class II genes are on different chromosomes (8, 9, 10). In teleosts this is probably a derived character because the phylogenetically more primitive cartilaginous fish have linked class I and II genes (7). Linkage, or the lack of it, has the potential to profoundly influence the evolution of polymorphism at class I and II genes. In bony fish, unlike other species, selection upon class I polymorphism need have no consequences for class II polymorphism or vice versa. Thus, the presence in rainbow trout and brown trout of old, divergent MHC class I lineages may reflect such independent modes of evolution. That this situation is seemingly the opposite of that described for humans and other higher primates illustrates the inherent problems in extrapolating the properties of MHCs from one species to another.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Benny P. Shum, Department of Structural Biology, Stanford University School of Medicine, 299 Campus Drive West, Sherman Fairchild Building, D-151, Stanford, CA 94305-5126.
3 Abbreviations used in this paper: 2m, 2-microglobulin; dN, number of nonsynonymous substitutions per site; dS, number of synonymous substitutions per site; indel, insertion/deletion event.
Received for publication August 24, 2000. Accepted for publication December 12, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2-microglobulin sequence diversity in individual rainbow trout. Proc. Natl. Acad. Sci. USA 93:2779.-chain expression in the rainbow trout. Scand. J. Immunol. 35:687.[Medline]
chain. Scand. J. Immunol. 41:365.[Medline]
2 domain within the teleostean world, MHC class I from the rainbow trout Oncorhynchus mykiss. Dev. Comp. Immunol. 20:417.[Medline]
chain of homozygous rainbow trout (Oncorhynchus mykiss). Dev. Comp. Immunol. 23:51.[Medline]
chain in salmonid fishes. Dev. Comp. Immunol. 24:751.[Medline]
resolution. J. Mol. Biol. 219:277.[Medline]
This article has been cited by other articles:
|
|
 |
|
  S. R. Garner, R. N. Bortoluzzi, D. D. Heath, and B. D. Neff Sexual conflict inhibits female mate choice for major histocompatibility complex dissimilarity in Chinook salmon Proc R Soc B, October 28, 2009; (2009) rspb.2009.1639v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U Grimholt, R Johansen, and A J Smith A review of the need and possible uses for genetically standardized Atlantic salmon (Salmo salar) in research Lab Anim, April 1, 2009; 43(2): 121 - 126. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Consuegra and C. Garcia de Leaniz MHC-mediated mate choice increases parasite resistance in salmon Proc R Soc B, June 22, 2008; 275(1641): 1397 - 1403. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. de Eyto, P. McGinnity, S. Consuegra, J. Coughlan, J. Tufto, K. Farrell, H.-J. Megens, W. Jordan, T. Cross, and R. J.M Stet Natural selection acts on Atlantic salmon major histocompatibility (MH) variability in the wild Proc R Soc B, March 22, 2007; 274(1611): 861 - 869. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Coughlan, P. McGinnity, B. O'Farrell, E. Dillane, O. Diserud, E. de Eyto, K. Farrell, K. Whelan, R. J.M. Stet, and T. F. Cross Temporal variation in an immune response gene (MHC I) in anadromous Salmo trutta in an Irish river before and during aquaculture activities ICES J. Mar. Sci., January 1, 2006; 63(7): 1248 - 1255. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. H. Bos and B. Waldman Evolution by Recombination and Transspecies Polymorphism in the MHC Class I Gene of Xenopus laevis Mol. Biol. Evol., January 1, 2006; 23(1): 137 - 143. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Moon, S. M. Veniamin, J. A. Parks-Dely, and K. E. Magor The MHC of the Duck (Anas platyrhynchos) Contains Five Differentially Expressed Class I Genes J. Immunol., November 15, 2005; 175(10): 6702 - 6712. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Consuegra, H.-J. Megens, H. Schaschl, K. Leon, R. J. M. Stet, and W. C. Jordan Rapid Evolution of the MH Class I Locus Results in Different Allelic Compositions in Recently Diverged Populations of Atlantic Salmon Mol. Biol. Evol., April 1, 2005; 22(4): 1095 - 1106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. E. Magor, B. P. Shum, and P. Parham The {beta}2-Microglobulin Locus of Rainbow Trout (Oncorhynchus mykiss) Contains Three Polymorphic Genes J. Immunol., March 15, 2004; 172(6): 3635 - 3643. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. Rise, K. R. von Schalburg, G. D. Brown, M. A. Mawer, R. H. Devlin, N. Kuipers, M. Busby, M. Beetz-Sargent, R. Alberto, A. R. Gibbs, et al. Development and Application of a Salmonid EST Database and cDNA Microarray: Data Mining and Interspecific Hybridization Characteristics Genome Res., March 1, 2004; 14(3): 478 - 490. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Boudinot, S. Boubekeur, and A. Benmansour Primary Structure and Complementarity-Determining Region (CDR) 3 Spectratyping of Rainbow Trout TCR{beta} Transcripts Identify Ten V{beta} Families with V{beta}6 Displaying Unusual CDR2 and Differently Spliced Forms J. Immunol., December 1, 2002; 169(11): 6244 - 6252. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. P. Kruiswijk, T. T. Hermsen, A. H. Westphal, H. F. J. Savelkoul, and R. J. M. Stet A Novel Functional Class I Lineage in Zebrafish (Danio rerio), Carp (Cyprinus carpio), and Large Barbus (Barbus intermedius) Showing an Unusual Conservation of the Peptide Binding Domains J. Immunol., August 15, 2002; 169(4): 1936 - 1947. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Ohta, E. C. McKinney, M. F. Criscitiello, and M. F. Flajnik Proteasome, Transporter Associated with Antigen Processing, and Class I Genes in the Nurse Shark Ginglymostoma cirratum: Evidence for a Stable Class I Region and MHC Haplotype Lineages J. Immunol., January 15, 2002; 168(2): 771 - 781. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Aoyagi, J. M. Dijkstra, C. Xia, I. Denda, M. Ototake, K. Hashimoto, and T. Nakanishi Classical MHC Class I Genes Composed of Highly Divergent Sequence Lineages Share a Single Locus in Rainbow Trout (Oncorhynchus mykiss) J. Immunol., January 1, 2002; 168(1): 260 - 273. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, Phylogenetic
compatibility between pairs of sites; 
, incompatibility. A mirror
image is defined by the diagonal line. Incompatibilities are
potentially the result of recombination. The number of informative
sites for UBA class I is 424 and that for class II
DAB is 95.