of March 28, 2018.This information is current as

I GenesContains Five Differentially Expressed Class

)Anas platyrhynchosThe MHC of the Duck (

and Katharine E. MagorDebra A. Moon, Simona M. Veniamin, Julie A. Parks-Dely

http://www.jimmunol.org/content/175/10/6702doi: 10.4049/jimmunol.175.10.6702

2005; 175:6702-6712; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/175/10/6702.full#ref-list-1

, 0 of which you can access for free at: cites 16 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2005 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The MHC of the Duck (Anas platyrhynchos) Contains FiveDifferentially Expressed Class I Genes1

Debra A. Moon, Simona M. Veniamin, Julie A. Parks-Dely, and Katharine E. Magor2

MHC class I proteins mediate a variety of functions in antiviral defense. In humans and mice, three MHC class I loci each contributeone or two alleles and each can present a wide variety of peptide Ags. In contrast, many lower vertebrates appear to use a single MHCclass I locus. Previously we showed that a single locus was predominantly expressed in the mallard duck (Anas platyrhynchos) and thatlocus was adjacent to the polymorphic transporter for the Ag-processing (TAP2) gene. Characterization of a genomic clone from thesame duck now allows us to compare genes to account for their differential expression. The clone carried five MHC class I genes andthe TAP genes in the following gene order: TAP1, TAP2, UAA, UBA, UCA, UDA, and UEA. We designated the predominantly expressedgene UAA. Transcripts corresponding to the UDA locus were expressed at a low level. No transcripts were found for three loci, UBA,UCA, and UEA. UBA had a deletion within the promoter sequences. UCA carried a stop codon in-frame. UEA did not have a polyad-enylation signal sequence. All sequences differed primarily in peptide-binding pockets and otherwise had the hallmarks of classical MHCclass I alleles. Despite the presence of additional genes in the genome, the duck expresses predominantly one MHC class I gene. Thelimitation to one expressed MHC class I gene may have functional consequences for the ability of ducks to eliminate viral pathogens,such as influenza. The Journal of Immunology, 2005, 175: 6702–6712.

M ajor histocompatibility complex class I proteins play afundamental role in the presentation of endogenouslyderived peptides to CTLs and also serve as self-rec-

ognition elements for NK cells. As such, MHC class I genes arecritical for immune defenses against viruses. Polygeny and poly-morphism of alleles contribute to the breadth of the immune re-sponse. Comparison of vertebrate MHC genomic regions showsthat it is the most dynamic part of the genome (1). The number offunctional and defunct MHC loci can differ greatly in each species,and no orthologous genes can be identified between orders. Inhumans and mice, three MHC class I loci each contribute one ortwo alleles, presenting a broad spectrum of diverse peptides. Thiscontrasts with frogs, sharks, fish, and birds, many of which appearto predominantly express one MHC class I gene (2, 3). In manyspecies, additional genes are present in the genome; however, themechanisms accounting for the loss of expression of additionalclassical class I genes have not been examined. The limitation toone MHC class I locus is predicted to have wide-reaching func-tional consequences, influencing the positive and negative selec-tion of the TCR repertoire, the NK cell repertoire, and diversity ofpeptides that can be displayed. In light of the role of the duck asa reservoir of influenza (4, 5), we are examining MHC class I geneexpression to determine their capacity for defense against viruses.

In chickens, there are two clusters of MHC class I genes, the Blocus and the Rfp-Y locus that map to the same microchromosome (6,

7). The B locus contributes all of the hallmarks of the MHC, control-ling allograft recognition (8, 9), MLRs (10), cellular cooperation (11,12), and disease resistance (13, 14). In the Rfp-Y complex, there areadditional MHC class I genes (15), one of which is polymorphic andtranscribed (16). However, the Rfp-Y genes do not otherwise have thehallmarks of classical MHC class I genes (16).

Within the B complex, two MHC class I genes (the minor andmajor genes) flank either side of the transporter of Ag-processingTAP1 and TAP2 genes, with the dominant locus adjacent to TAP2(17). Recent nomenclature refers to the minor locus as the BF1locus, and the dominant locus is referred to as the BF2 locus (18).Both loci are considered equivalent to the classical class I MHCgenes in other species, and numerous alleles have been identified,particularly for the BF2 locus (19–23). In chickens, the TAP1 andTAP2 genes are also polymorphic (24). Kaufman (2) argues that inthis organization, the TAP and MHC class I proteins have an op-portunity to coevolve to function coordinately and evolutionaryforces select for inactivation of redundant loci.

The MHC class I regions of other birds appear to have under-gone extensive duplications, and the “minimal MHC” of thechicken is not the norm for other birds. The MHC of quail has thesame overall organization as chickens with MHC class I genesflanking TAP1 and TAP2, but is much more complex, with fourMHC class I genes, Coja-D1, -D2, -B1, and -E and three pseudo-genes -F, -G, and -H (25, 26). Of these, Coja-B1 adjacent to TAP2is the most likely ortholog of BF2, and two loci (-D1 and -D2)appear nonclassical due to weak expression (26). Sequence com-parison of the quail MHC class I genes reveals their overall sim-ilarity to each other and the authors suggest that they have dupli-cated since the divergence of these two avian species. In the redjungle fowl, the closest to the evolutionary ancestor of domesticchickens, there are 14 copies of MHC class I-like genes, 3 ofwhich are pseudogenes, in the preliminary analysis of the recentlyreleased genome sequence (27). There are reports of extensiveduplication within MHC regions of sparrows (28, 29), songbirds(30, 31), and reed warblers (32–34). These data suggest that theMHC of most birds has undergone expansion.

Department of Biological Sciences, University of Alberta, Edmonton, Canada

Received for publication February 11, 2005. Accepted for publication August16, 2005.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by a grant from the Natural Sciences and EngineeringResearch Council of Canada and an Independent Establishment Grant from the Al-berta Heritage Foundation for Medical Research (to K.E.M.). K.E.M. is an AlbertaHeritage Foundation for Medical Research Scholar.2 Address correspondence and reprint requests to Dr. Katharine E. Magor, Depart-ment of Biological Sciences, CW405 Biological Sciences, University of Alberta, Ed-monton, Alberta, Canada. E-mail address: kmagor@ualberta.ca

The Journal of Immunology

Copyright © 2005 by The American Association of Immunologists, Inc. 0022-1767/05/$02.00

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Ducks diverged earlier from the Galloanseriforme lineage (35)and comparison of their MHC class I region may shed light on theprimordial avian MHC. Previously we showed that the duck, Anasplatyrhynchos, like the chicken, also has a predominantly ex-pressed MHC class I gene, and that gene lies adjacent to the poly-morphic TAP2 gene (36). However, we had evidence of at leastfour MHC class I genes per haplotype. To resolve the discrepancybetween the number of expressed sequences and the number weestimated in the genome, we have undertaken an analysis of thegenomic organization of the MHC class I region of the same duckto allow direct identification of expressed and unexpressed genes.In this study, we show that the MHC class I region of the duckcontains five genes that show a high level of overall similarity.Surprisingly, the overall organization of MHC class I genes andTAP differed from that of chicken or quail and may resemble thelocus before a translocation event. Our analysis revealed the de-fects in some of these genes, accounting for their lower expression.Despite an expanded number of genes in the genome, the duckMHC class I region functionally resembles the minimal MHC ofthe domestic chicken.

Materials and MethodsScreening of a fosmid genomic library for MHC class I genes

A genomic library from erythrocyte DNA from a male duck, Anas platy-rhynchos (duck 26) was previously constructed in the fosmid vectorpCC1FOS (37). Approximately 124,488 independent colonies werescreened using a probe derived from a conserved region of MHC class Iexon 4 common to all cDNA sequences previously identified in the sameindividual duck (36). Overlapping primers, MHC Ia 5�-AGA TCC TGACCT TGT CCT GCC GTG-3� and MHC Ib 5�-GGG TAG AAG CCG TGAGCA CGG CAG-3� (Qiagen) were used to generate an “overgo” probe(38). Screening of the library identified only one MHC class I-hybridizingclone, clone Ap26-72A12.

Sequencing strategy

Because preliminary sequencing suggested that MHC class I genes in theclone could be very similar, a set of subclones was constructed to com-pletely sequence fosmid Ap26-72A12. EcoRI subcloning produced sevenunique clones, which completely covered the original fosmid insert: fiveMHC class I-hybridizing subclones, 5.8 kb for UAA, 5.7 kb for UBA, 4.2kb for UCA, 5.9 kb for UDA, and 4.6 kb for UEA to vector arm; a 9-kbsubclone containing TAP1 and TAP2; and a 1.5-kb subclone which con-tained only intergenic sequence. Using the EZ::TN �KAN-2� transposoninsertion kit (Epicentre Technologies) primer islands were inserted into theEcoRI subclones. The 1.5-kb EcoRI subclone was sequenced by primerwalking. Sequencing was done using Amersham DYEnamic ET terminatorchemistry (Amersham Pharmacia Biotech) on an ABI377 Automatic Se-quencer (Applied Biosystems). Contigs were created using AutoAssembler(version 2.1; Applied Biosystems) with coverage of at least 5-fold redun-dancy and included sequences on both strands. Determining the order ofthe EcoRI subclones in the final 36.8-kb contig was accomplished by sub-cloning large internal fragments, restriction mapping, and sequencing ofoverlapping fragments. The sequence of the entire 72A12 fosmid clone isdeposited in GenBank under the accession number AY885227.

Sequence analysis

The edited 72A12 final sequence was analyzed for coding regions usingGENSCAN (http://genes.mit.edu/GENSCAN.html) and through homologysearches using BLAST (www.ncbi.nlm.nih.gov/blast). MHC class I geneswithin the clone were compared with cDNA and PCR fragments previouslypublished for this individual duck (36) using GeneTool (BioTools) to as-sign haplotype and locus information to expressed genes in this animal aswell as to refine splice junctions in predicted coding regions.

RepeatMasker (http://repeatmasker.org) identified and classified mostbut not all sequence repeats. Additional WU-BLASTn and WU-BLASTx(http://proweb.org/Tools/WU-blast.html) analyses were needed to refineclassification of the embedded retroviral sequences. Chicken repeat se-quences used for comparison were from Repbase (www.girinst.org) andrecent analysis of the chicken genome (39). PipMaker (http://bio.cse.psu.edu/pipmaker) was used to produce the dot matrix plot. Gene Rec-ognition and Assembly Internet Link (GRAIL) was used to identify theCpG islands (http://compbio.ornl.gov/grailexp). Analysis of percent nucle-

otide and amino acid identity was done using GeneTool and PepTool(BioTools).

The program Synonymous/Nonsynonymous Analysis Program (http://www.hiv.lanl.gov/content/hiv-db/SNAP/) was used to calculate rates ofnonsynonymous (dN) and synonymous (dS) nucleotide substitutions be-tween loci and alleles according to the method of Nei and Gojobori (40) asimplemented (41). Peptide-binding pocket residues in the Ag recognitionsite of duck MHC class I genes were predicted by comparison to thosepredicted for trout (42) and chickens (23). Multiple alignments of MHCclass I sequences and a bootstrapped phylogenetic tree was created usingClustalW at DDBJ (http://www.ddbj.nig.ac.jp/search/clustalw-e.html).

Promoter analysis

Regions 3000 bp upstream and ending at the translational start sites of theMHC class I genes were analyzed for potential transcription factor-bindingmotifs using the following web-based programs: BioInformatics and Molec-ular Analysis Section (BIMAS) WWW Signal Scan (http://bimas.cit.nih.gov/molbio/signal/), Genomatrix MatInspector (http://www.genomatrix.de/products/MatInspector/index.html), and AliBaba2.1 (http://www.alibaba2.com/). Duck sequences were also aligned to known MHC class I promoters(human and chicken) using ClustalW and conserved sites within the promoterregions were identified manually.

Northern blot and allele-specific hybridization

Allele-specific oligonucleotide hybridization was performed to assess therelative transcript abundance. An oligonucleotide probe was designed to bespecific for the UBA MHC class I sequence (5�-CGA CAG CGA GACCAG GAA GAT-3�). The Northern blot was probed to test for expressionof transcripts matching the UBA locus using the same blot and conditionspreviously used to assess transcript abundance (36). The oligonucleotidewas end-labeled with [�-32P]dATP (PerkinElmer Canada) to �2 � 106

dpm/pmol specific activity, then purified using Sephadex G-25 Fine (Sig-ma-Aldrich) size-exclusion chromatography. The Northern blot was hy-bridized overnight at 48°C in aqueous hybridization solution (5� Den-hardt’s, 6� SSC, 0.1% SDS). Three 15 min washes were conducted at lowstringency (6� SSC, 0.1% SDS) at 50°C. The blot was exposed to KodakX-OMAT film at �80°C for 72 h.

To test the pattern of hybridization of all MHC class I in tissues, aNorthern blot was hybridized with the D26E4p probe amplified from duck26 using degenerate primers CHE4F1 (5�-CCC GAG GTS CGA GTGTSG-3�) and CHE4R1 (5�-CAC GCG GCA CYG GTA CTT GTC-3�).

RT-PCR amplification of expressed sequences

To identify expressed MHC class I alleles from other individual ducks, wesynthesized first-strand cDNA from 5 �g of duck spleen total RNA using0.5 �g of an oligo(dT) primer (5�-TCTGAATTCTCGAGTCGACATCT17-3�) and extending with the Thermoscript RNase H� Reverse Transcriptasekit according to the manufacturer’s instructions (Invitrogen Life Technol-ogies). The MHC class I internal primers (SP2, 5�-GGC TGC TGC TGGGGG TCC TG-3�) or (E2F1, 5�-GCC CCA CTC CCT GCG CTA-3�) andthe reverse primer (3� untranslated region, 5�-GCA GAT AGG AGA TGTGAG AGG TTG-3�) were used to amplify MHC class I. Products werecloned into the PCR 2.1 TOPO vector (Invitrogen Life Technologies) andat least three clones for each allele were isolated and completelysequenced. The sequences for these alleles named U*06–U*09 weredeposited in GenBank under the accession numbers AY841881–4.

To examine the tissue distribution of expression, we performed allele-specific amplification by reverse transcription PCR using tissues from aduck (D132) that was matched for the MHC alleles of the duck used for thelibrary (D26). First-strand cDNA was synthesized and used for PCR am-plification using Taq polymerase (Invitrogen Life Technologies) withTaqStart Ab (Clontech Laboratories). Templates were denatured for 3 minat 96oC followed by amplification for 30 cycles at 96oC for 30 s, 30 s atannealing temperature, followed by extension for 1 min at 72oC. Allele-specific primers had at least three positions of mismatch compared withother sequences, including the 3� end. The UAA locus (U*03 allele) wasamplified with forward (5�-ATG GGG AAG CCT TCA CA-3�) and reverse(5�-GTG AAA ACT GTT CCT G-3�) primers at an annealing temperatureof 51oC, yielding a 292-bp product. The UDA locus (U*04 allele) wasamplified using forward (5�-ACT CCC TGC GCT ACT TCT TG-3�) andreverse (5�- CTC TGC ATG TGA CGC CGC-3�) primers at 61.7oC, yield-ing a 391-bp product. All MHC class I sequences were amplified usingforward (E2F1, 5� GCC CCA CTC CCT GCG CTA-3�) and reverse (E2R1,5�-CTC TGG TTG TAG CGC TCC CG-3�) primers using the same con-ditions as UDA primers. Amplification of �-actin from cDNA using for-ward (�ACTIN1EF, 5�-ACC GCG CAA CTC CCC GAA GCC AG-3�)and reverse primers (�ACTINE2R, 5�-ATA GCT GTC TTT CTG GCC

6703The Journal of Immunology

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

CAT GC-3�) was a control for cDNA synthesis and quantity. Actin am-plification was virtually identical for spleen, liver, heart, and kidney, butslightly higher in testes. Products were separated on a 5% PAGE gel in 1�Tris-borate-EDTA and quantified using the intensity of the band divided byactin intensity using the Kodak 1D Image Analysis software, and back-ground correction was done using the minimum of perimeter intensity.

Amplification of the MHC class I and TAP2 genomic region

MHC class I sequences were amplified from genomic DNA of a secondduck (duck 64) with a primer in exon 2 D26E2F1 (5�-GCC CCA CTC CCTGCG CTA-3�) and a primer in the 3� untranslated region (5�-GCA GATAGG AGA TGT GAG AGG TTG-3�) using conditions for amplification oflonger PCR products. These products were deposited in GenBank under theaccession numbers AY854373–6. To determine the gene adjacent to TAPin duck 64, we also amplified the genomic region from within the 3� endof the MHC class I gene and the 3� end of the TAP2 gene. All PCR productswere cloned into the PCR 2.1 TOPO vector and sequenced by primerwalking.

Southern blot hybridization

High molecular weight DNA (10 �g) from erythrocytes of the White Pekinduck (duck 26) was digested to completion with restriction enzymes andseparated on a 0.8% agarose gel using the field-inversion gel electrophore-sis system (Bio-Rad). The gel was blotted onto Performa nylon transfermembrane (Genetix USA) and immobilized by UV cross-linking. Blots

were hybridized sequentially with the MHC class I and the TAP2 probes.The MHC class I probe was amplified using forward (E3F2, 5�-CCT GGAGAA CAC CTG CAT TGG G-3�) and reverse (E4R2, 5�-GTG AGC ACGGCA GGA CAA GGT-3�) primers on subclones of the genes found ongenomic clone 72A12, and PCR amplification products were pooled. Allamplifications worked except UCA. The TAP2 probe corresponded to ex-ons 6–8, amplified from cDNA clone 6.2.2 using forward primer(DT2WAF, 5�-TGA ACG GCA GTG GGA AGA GCA-3�) and reverseprimer (DT2E8R1, 5�-CAA CAC GCT GCT TCT GCC CA-3�) (36). Hy-bridizations were conducted for 16 h at 60oC in aqueous hybridizationsolution (5� Denhardt’s, 6� SSC, 5% dextran sulfate, 1% SDS, and 100�g/ml salmon sperm DNA (1� SSC: 0.15 M NaCl, 15 mM Na3C6H5O7

(pH 7.6), and 1 mM EDTA)). Washes were conducted at high stringencyin 0.1� SSC/0.1% SDS at 65oC. Probes were radiolabeled with[�-32P]dCTP by random priming (43) (PrimeIt kit; Stratagene). The blotswere exposed to Kodak X-OMAT film at �70oC for 10 days.

ResultsThe duck MHC class I region has undergone duplicationsyielding five MHC class I genes

To determine the genomic sequence of the duck MHC class I re-gion and to compare the expressed MHC class I sequences to thosein the genome, we screened a genomic library constructed from thesame duck used for construction of the cDNA library. A singlehybridizing clone was isolated, 72A12, and subjected to restrictiondigestion for mapping and subcloning (Fig. 1A). A complete contigof 36.8 kb containing TAP1 and TAP2 and five MHC class I genes(UAA, UBA, UCA, UDA, UEA) was assembled (Fig. 1B). TAP1and TAP2 are in opposite transcriptional orientation, and the fiveMHC class I genes are arranged in the same transcriptional orien-tation. The gene adjacent to TAP2 was previously identified as thepredominantly expressed locus and named UAA, and is identical insequence to the MHC-TAP2 haplotype 2 previously PCR amplifiedfrom this duck (Table I) (36). Other genes were given consecutivenames based on the proposal for naming vertebrate MHC genessuggested by Klein et al. (44). The relative abundance of tran-scripts from the MHC class I genes on this haplotype is shown(Fig. 1C), with cDNA clones U*03 and U*04 previously isolatedfrom this duck matching the UAA and UDA locus sequences ex-actly (Table I). The overall G � C nucleotide content was 58%. Ascan of the sequence using GRAIL identified abundant CpG is-lands in the 5� end of most genes extending into intron 1, with UBAbeing an exception (Fig. 1D). For comparison, we counted 37 CpGmotifs in the UAA locus, 41 in UDA, and only 17 near UBA.

We analyzed the complete sequence for the presence of repeatsthat may provide insight into the duplication events giving rise tothe five loci. RepeatMasker detected six simple repeats in inter-genic and intronic regions, nine copies of LINES resembling thechicken CR1 element, and two retroelement-derived repeats sim-ilar to the chicken Gallus gallus long terminal repeat (GGLTR)3

family (Fig. 1E). All nine copies of chicken CR1 LINE and thelong terminal repeat (LTR) elements were incomplete. The em-bedded retroviral elements similar to the chicken Gallus galluslong terminal repeat (GGLTR11) family appeared three times in

3 Abbreviations used in this paper: GGLTR, Gallus gallus long terminal repeat; LTR,long terminal repeat; ISRE, IFN-stimulated response element; GRAIL, Gene Recog-nition and Assembly Internet Link.

FIGURE 1. Gene organization of the duck MHC class I region fromTAP1 through Anpl-UEA. A, Restriction map showing the EcoRI sites usedfor subcloning and the restriction sites used to assemble the 36.8-kb contig.B, Gene map of the MHC class I region, with TAP1 and TAP2 indicated byempty boxes and the five MHC class I genes indicated by filled boxes.Arrows indicate the transcriptional orientation of the gene. C, Relativeexpression of transcripts corresponding to the five MHC class I genes.Expression was determined by allele-specific oligonucleotide hybridizationon spleen and intestinal mRNA. D, Location of CpG islands determined byGRAIL. E, Location of repetitive elements identified by RepeatMasker.These include embedded retroviral elements indicated above the line, CR1-like LINES (open boxes), and simple repeats indicated below the line. Thesequence of the entire 72A12 fosmid clone is deposited in GenBank underthe accession number AY885227.

Table I. Alleles identified in cDNA and the corresponding locus

MHC-TAP2 Haplotype UAA UBA UCA UDA UEA

Haplotype 2 U*03 NDa ND U*04 NDHaplotype 1 U*02 ND ND ND U*05

ND, Not detected.

6704 FIVE DIFFERENTIALLY EXPRESSED MHC CLASS I GENES IN DUCKS

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

the sequence, once between UDA and UEA and once between UAAand UBA, with a small fragment between UBA and UCA. The LTRappears to have inserted within CR1 LINEs before duplication ofthe locus based on the order CR1-LTR-CR1 within both intergenicregions. The location of this repeat within the UBA to UCA inter-genic region suggests that one gene of the duplicated pair givingrise to UCA was lost. The GGLTR5 sequence identified in theintergenic region of UDA to UEA is unique and has inserted sub-sequent to the duplication events. Further analysis of this elementsuggests it is closest to the chicken Hitchcock element (39) and isflanked by 7 bp direct repeats, typical of tandem site duplicationupon transposon insertion.

A dot plot analysis of the Anpl MHC class I region against itselfrevealed extended regions of duplication (Fig. 2). Unbroken diag-onal lines within the shaded regions indicated that there was con-servation of the exons and introns across all five genes. The diag-onal lines that extend into the white regions suggested that theregion of duplication included large segments of the intergenicregions as well. An extended region spanning the UEA and UDA,forms a line in comparison to the region spanning UBA to UAA,suggesting they are part of an extended region of duplication. Theembedded retroelement in this region, which appears betweenUAA and UBA and also between UDA and UEA, may have pre-dated the duplication. The only break in the line comparing UEAand UDA to UBA and UAA corresponded to the region encodingthe cytoplasmic domains, which are unique and appear disrupted inthe UEA locus.

The genes encoding the transporter for Ag processing, TAP1 andTAP2 are directly adjacent to the UAA gene of ducks. To determinewhether additional MHC class I genes lie outside the clone, weanalyzed restriction fragments hybridizing with the MHC class Iand TAP2 genes in genomic DNA from the same duck used forconstruction of the genomic library (Fig. 3). The clone 72A12 hadan EcoRI site between the dominantly expressed MHC class I geneAnpl U*03 and TAP2. A 13-kb TAP2-hybridizing EcoRI fragmentdoes not carry any class I-hybridizing genes, suggesting that thereare no class I genes immediately adjacent to TAP1. Most MHCclass I genes lie on several EcoRI fragments of 6 and one of 4.5 kb.

The hybridizing EcoRI bands of 32 or 37 kb carry the two allelicforms of the UEA gene. We were unable to determine whetheradditional MHC class I genes lie on the other side of the UEAlocus; however, all of our data taken together is consistent with theclone 72A12 representing an entire haplotype of the MHC class Iregion. Similarly, from previous analysis of the sequence for hap-lotype 1, we know there is an XhoI site between U*02 and TAP2.Therefore, the TAP2-hybridizing XhoI fragment of 24 kb belongsto haplotype 1. The fact that this fragment does not hybridize withclass I also confirms no class I genes are immediately on the otherside of TAP1. Southern blots using other restriction enzymes gavethe same conclusion (data not shown).

All five genes within the locus have the hallmarks of classicalMHC class I genes

To compare the five MHC class I genes, the deduced amino acidsequences were aligned (Fig. 4). The leaders for UAA, UCA, andUDA were 25 aa, whereas the leader of UBA was 30 aa in lengthdue to a five amino acid insertion. The leader for UEA lies outsidethe clone. The majority of polymorphic amino acid residues werewithin these two domains with 42 polymorphic residues in �1 and30 polymorphic residues within �2. The �3 domain of 91 aa washighly conserved between the five sequences, with only 11 poly-morphic residues. The transmembrane and cytoplasmic domainswere very similar, except the UBA cytoplasmic domain is oneamino acid shorter, and the UEA sequence had an unusual trun-cated cytoplasmic domain. Within the mature protein, the identityof the amino acid sequences ranges from 81.9% between UAA andUCA to 87.4% between UBA and UDA. Paired amino acid identitymatrices for each exon failed to identify loci that are more similarto each other. The loci are most different within the �1 and �2domains and all are very similar within the �3 domain.

To determine which genes could be involved in Ag presentation,we examined the amino acid sequences for hallmarks of classicalMHC class I genes. Among residues predicted to be involved in

FIGURE 2. Dot matrix analysis of the Anpl MHC class I region vsitself. Diagonals indicate the regions where contiguous sequences align.

FIGURE 3. Southern blot hybridized sequentially with MHC class I andTAP2. Genomic DNA from duck 26 (used for the genomic library) wasdigested with EcoRI and XhoI and separated on a field inversion gel.

6705The Journal of Immunology

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

peptide anchoring (Y7, Y59, Y84 (R84 in nonmammalian verte-brates), Y123, T143, K146, W147, Y159, and Y171), all wereconserved in most loci. The Y123 sequence is replaced by phe-nylalanine or leucine in all five duck genes, but is also seen in otheravian sequences and many other species (45). The UDA did nothave the conserved Y171; however, histidine at this residue(H169) was seen in other mammalian classical MHC class I se-quences. All sequences have the site for N-linked glycosylation atN85. The negatively charged residues in the two regions of the �3domain implicated in CD8 binding (46) are conserved, exceptN250 in UEA replaces aspartic acid. The glutamine directly in-volved in CD8 accessory function (47) is conserved in all se-quences except UCA, in which it is replaced by a premature stopcodon.

Evidence for diverse mechanisms of inactivation withinadditional MHC class I loci

A comparison of the genes in the locus with the sequences isolatedfrom a cDNA library constructed from the same duck allowed usto characterize the expression of these genes (Fig. 1C and Table I).The sequence of the gene adjacent to TAP2 was identical to thesequence of one of the abundantly expressed alleles, AnplU*03.The frequency with which we found transcripts corresponding toUDA was very low, consistent with it being a nonclassical or MHCclass Ib gene. The sequence of UDA matched four cDNAs of twodifferent sizes, named AnplU*04. Examination of the 3� untrans-lated region of this gene indicated the presence of two polyade-nylation sites, one at 152 bp and one at 814 bp from the stopcodon. Allele-specific oligonucleotide hybridization suggested thattranscripts of either size were expressed at a rate 10-fold less thanAnplU*03 (36) (Fig 1C). To examine the tissue distribution of thetwo expressed loci (UAA and UDA),we used reverse transcriptionfollowed by PCR amplification with allele-specific primers. Al-though this amplification is not quantitative, it was useful for com-parison between tissues. The expression of the UAA (U*03 allele)was 20–25% higher in spleen and liver than in heart, kidney, ortestis. The same pattern was seen for amplification of all alleles

using universal primers in exon 2. This correlated well with thepattern of hybridization of an exon 4 probe on a Northern blot,expected to hybridize to all sequences (Fig. 5). In contrast, theamplification of UDA was almost equal in all tissues and did notshow a significant increase in spleen and liver.

No transcripts were found corresponding to the sequence ofUEA. The U*05 cDNA sequence from this duck, differing fromUEA by one amino acid in exon 4 and one amino acid in exon 8,is most likely the allele encoded at the UEA locus on the otherhaplotype. Two transcripts of U*05 were recovered from a cDNAlibrary; however, neither had a polyadenylation signal sequence orpoly(A) tail, both arising as unusual clones (36). We noted thatUEA does not have a polyadenylation signal sequence anywherewithin the intergenic region. The end-labeled-specific oligonucle-otide used to examine the expression of the U*05 cDNA sequencewould also hybridize perfectly with the UEA sequence. The hy-bridization of the specific oligonucleotide to mRNA was undetect-able in spleen, but two weakly hybridizing bands were detectable

FIGURE 4. Alignment of predictedamino acid sequences of the five MHCclass I genes within this haplotype. Thefirst amino acid of the �1 domain isconsidered as position 1. Identity withthe UAA sequence is indicated with adot. Gaps are indicated by dashes. Con-served amino acid residues involved inbinding the peptide-terminal mainchain atoms are shown below the align-ment. The N-linked glycosylation siteis underlined. The residues predicted tointeract with CD8 are have lines abovethem.

FIGURE 5. Northern blot analysis of MHC class I expression in differ-ent tissues of duck 132. Total RNA from various tissues was separated,transferred to a blot, and hybridized with the MHC class I exon 4 probeD26E4p. The gel photograph is shown for a comparison of the amount ofRNA loaded in each lane. Position of 18S RNA is shown for size reference.Tissues correspond to heart (H), spleen (S), testes (T), kidney (K), and liver(L). The blot was washed at high stringency in 1� SSPE/0.1% SDS at65oC and exposed to Kodak X-OMAT AR film at �80°C overnight.

6706 FIVE DIFFERENTIALLY EXPRESSED MHC CLASS I GENES IN DUCKS

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

in intestine, indicating that transcripts from UEA, if present, arealso rare (36).

We were unable to detect transcripts for two loci, UCA andUBA. The sequence of UCA encoded a stop codon in-frame inexon 4 and would result in nonsense-mediated mRNA decay. UBAhad a unique exon 4 sequence, which had not been amplified fromgenomic DNA due to mismatches in the priming regions. NocDNA transcripts matching this locus were identified, although theexons of this gene are apparently intact. To examine the expressionof this gene, we hybridized the same Northern blot of spleen andintestine mRNA using an allele-specific oligonucleotide hybrid-ization (36). No hybridization to mRNA from spleen or intestinewas detectable, indicating this gene is not expressed (data notshown).

To investigate whether differences in promoter cis-elementsmight be responsible for the differential expression of the class Igenes, proximal promoter sequences were aligned for UAA, UBA,UCA, and UDA (Fig. 6). The regulation of MHC class I regiongenes is mediated by a number of conserved cis-acting regulatory

sites including enhancer A (48), an IFN-stimulated response ele-ment (ISRE) and site � (homologous to the cAMP response ele-ment) (49) and enhancer B (reviewed Ref. 86). Enhancer A isbound by NF-�B/rel family members while the ISRE site is boundby IFN-inducible transactivators (50). We identified putativetranscription factor binding sites in the promoter regions usingcomputer analysis and alignment with other promoters. An NF-�Bsite was identifiable in the UAA, UBA, and UDA promoters, whilethe sequence of UCA lacked this motif. An ISRE appeared intactin all four sequences, with one nucleotide difference in the ISREmotif for UDA.

We examined the five duck MHC class I promoters for thepresence of the conserved regulatory motif described as the S,X (both X1 and X2), and Y motif within the proximal promoter.The SXY motif is found within MHC class I, MHC class II,invariant chain, DMA and DMB, and �2-microglobulin (51).The SXY module is bound by regulatory factor X, X2 box-binding protein, and NFY bound to X1, X2, and Y respectively,while the S motif does not appear to be occupied in vivo. The

FIGURE 6. Alignment of MHC class I pro-moters. Putative transcription factor bindingsites are indicated. Enhancer elements includ-ing NF-�B and the ISRE are double underlinedand conserved motifs involved in constitutiveexpression of MHC class I sequences are linedabove in bold.

6707The Journal of Immunology

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

requirement for all three motifs and the spacing between themled to the suggestion that this motif functions as an “enhance-some” in the regulation of MHC genes (52). This motif is intactwithin the UAA, UCA, and UDA sequences, but there is a de-

letion within the X1 and X2 box of UBA. A second deletion inthe UBA promoter spans a putative CAAT box. The disruptionswithin the promoter of the UBA sequence may account for thelack of expression of this gene.

FIGURE 7. A, Phylogenetic tree showing the rela-tionship of the Anpl MHC class I sequences to the clas-sical and nonclassical MHC class I sequences ofother birds. The dendrogram was constructed withthe entire mature protein amino acid sequence of eachMHC class I sequence. Proteins used were the quailCoja-B1 (BAC82519.1), Coja-C (BAA83671.1),Coja-D1 (82515.1), Coja-D2 (BAC82516.1), andCoja-E (BAC82520.1); the chicken Gaga-BF2(CAA18972.1), Gaga-BF1 (CAA18969.1), and Gaga YF(Rfp-Y) (AAK15583.1); the Sandhill crane Grus canadien-sis Grca (AF033106); and the domestic goose Anser anserAnan (AY654899.1) were aligned with ClustalW. Boot-strap values are shown at the branch points. B, Phyloge-netic tree showing the relationship of all available AnplMHC class I sequences to the genomic loci of one hap-lotype. Sequences were obtained as cDNA clones (allelesU*02–05 with accession numbers AY24416–9 (36) oramplified by RT-PCR in this article (alleles U*06-U*09,accession numbers AY841881–4), or from othersAB115241–6 (85) and U*01 AF383511.

6708 FIVE DIFFERENTIALLY EXPRESSED MHC CLASS I GENES IN DUCKS

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Functional diversity of alleles and loci

A phylogenetic tree of the mature protein encoded by avian classI sequences shows that the sequences of the duck MHC class I locigroup closer to each other than to other avian sequences (Fig. 7A).Phylogenetic trees constructed using the �2 and �3 domain or onlythe �3 domain gave the same results, showing duck sequencesbranching separately from chicken and quail (data not shown).This suggests that the individual loci were generated after specia-tion of the duck from a common ancestor in the galloanseriformelineage. Alternatively, recombination events and/or gene conver-sion may homogenize loci, so that no locus-specific character hasbeen retained. The duplication events that gave rise to these fiveloci are not obvious from comparison of the sequences. No directlyorthologous MHC class I loci shared by duck and chicken or quailcan be definitively identified.

A phylogenetic tree of all available duck MHC class I sequencessuggests that alleles are as divergent as locus sequences (Fig. 7B).We aligned all expressed sequences, i.e., sequences isolated fromcDNA libraries and PCR amplification of reverse-transcribedcDNA, and compared them to the sequences within one haplotype.The genomic sequences form separate branches on the tree andexpressed sequence groups with each locus. Indeed, the two majoralleles of duck 26, U*02 and U*03 at the UAA locus, are on dif-ferent branches of the tree. Locus information cannot be deter-mined from similarity of sequences.

To examine whether only sequences at the major locus would beamplified by RT-PCR, we characterized the expressed andgenomic sequences present in a second duck (D64). Two alleles,U*08 and U*09, were isolated by RT-PCR from this duck. Com-parison of the genomic sequences, and PCR amplification fromMHC class I to TAP2, indicated that U*08 was at the UAA locus,whereas U*09 which groups with UBA, is an allele at a minorlocus. This was also reflected in the ratio at which clones wererecovered, with U*08 represented by 33 clones and U*09 repre-sented by 3 clones, although this bias could be due to the primersmatching one sequence better than another.

To examine the functional diversity present in one duck and thatpresent in the population, we have calculated the rate of synony-mous over nonsynonymous substitutions within the proteins (Ta-ble II). The ratio of the rate of substitution within the �3 domainis very low at 0.15, suggesting that this part of the protein is con-served. The ratio of change in the �1 and �2 domains, at 0.46, ismuch higher than that in �3, which suggests that these differencesare being selected. When we consider only the pocket residues, theratio increases to 1.00. Although we cannot argue that a value of1 is evidence of strong selection, it is strikingly different from the�3 domain and suggests that there is selection to maintain diversityin the peptide-binding pockets. The difference in rates of evolutionfor the �2 and �3 domains was noted also for chicken and mam-malian class I sequences (53). The highest ratio of nonsynonymousto synonymous substitutions between these proteins lies within thepocket residues of both the cDNA and the sequences encoded atindividual loci. Although the averages were very similar, the most

divergent sequences were the U*02 and U*06 alleles, with a dN:dS

ratio calculated at 11.1, whereas within a haplotype the maximumdifference between two locus sequences was calculated at 2.7 forthe pocket residues. Sequences within a haplotype appear no dif-ferent than alleles at a locus.

DiscussionThe dominant expression of one MHC class I gene

We screened a duck genomic library, isolated, and sequenced agenomic clone encompassing TAP1 and TAP2 and five MHC classI genes in the following gene order: UAA, UBA, UCA, UDA, andUEA. Previously, we had isolated cDNA clones from this sameduck; therefore, we were able to correlate the expressed genes withthose present in the genome. The MHC class I gene adjacent toTAP2, that we previously showed was the predominantly ex-pressed gene, we designated UAA. Three loci were not expressed(UBA, UCA, and UEA) and one locus was weakly expressed(UDA) and may be a nonclassical MHC class I gene. All MHCclass I sequences resembled alleles of classical MHC class I genesin having the conserved anchor residues for peptide terminal mainchain atoms and amino acid polymorphisms located in the �1 and�2 domains responsible for peptide binding.

We have examined the MHC class I genes for evidence of whythere is a bias for expression of the UAA locus. UCA carried a stopcodon in-frame, which would result in nonsense-mediated decay ofmRNA transcripts from this locus. In an analysis of human MHCclass I null alleles, any stop codon found (54) or introduced (55)within the first five exons of the MHC class I gene, resulted in nodetectable transcripts. The exon 4 sequence carrying the stopcodon in-frame was identified in five ducks examined. There wereno obvious defects in the MHC class I sequences encoded at otherloci, although the polyadenylation signal sequence lacking in UEAwould mean that this sequence, even if transcribed, is not trans-lated into protein.

An examination of the proximal promoters identified a defect thatmay account for the lack of expression of UBA. UBA lacked the X2box of the MHC class I region SXY regulatory element, important forconstitutive expression of MHC class I and class II genes. The UBApromoter also appeared to lack the enrichment of CpG typical offunctional promoters (56). The promoters for UAA and UDA werevirtually identical, both in sequence and CpG distribution; therefore,there was no obvious reason for the nearly 10-fold lower expressionof UDA. One difference of UDA relative to UAA, a 9-bp insertion inthe sequence between the NF-�B site and the ISRE, may place theNF-�B element out of register with the other transcriptional machin-ery on the DNA helix. Alternatively, DNA methylation or other epi-genetic mechanisms are responsible for this differential expression.DNA methylation is important in establishing and maintaining thediverse repertoires of killer cell Ig-like receptors on NK cells (57, 58)despite their similar promoters.

Table II. Rate of dN over dS nucleotide substitutions

dN/dS Values MHC Class I genes�1/�2

(180 codons)�1/�2 Pockets

(37 codons)�1/�2 Nonpockets

(143 codons)�3

(91 codons)

All sequences (18)a 0.46 1.00 0.33 0.15All cDNA (12) 0.43 0.87 0.31 0.12MHC locus sequences (5) 0.42 0.97 0.30 0.17

a In parentheses are the number of sequences analyzed for each group.

6709The Journal of Immunology

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The additional MHC class I loci in ducks are disabled throughdefects in the promoters, a stop codon in-frame, a missing poly-adenylation site, and potentially epigenetic effects such as meth-ylation. The diversity of these mechanisms suggests that these pro-cesses evolved independently. This would suggest that despiteduplication, there is strong selection for inactivation of additionalloci over time. This is counter to an intuitive reasoning that addi-tional loci could contribute greatly to an immune response by hav-ing the potential to present a greater diversity of peptides. Addi-tional MHC would also result in the positive selection of a greaterTCR repertoire, as well as NK cell receptor repertoire, since everyNK cell must have an inhibitory receptor for self-MHC class I(59). Others have argued that additional MHC would have a det-rimental effect on immunity, because it would result in the deletionof more of the TCR repertoire during negative selection (60, 61).This could be significant, since the TCR repertoire may already belimited in birds (62). There are natural examples where MHCgenes have been reduced over time. In polyploid frogs, the MHCtends toward diploidy and the duplicated copies are silenced (63).Similarly, in rainbow trout, with a tetraploid ancestry, there is ev-idence that the MHC class I region is diploid (42). The reasons forthe deletion of MHC class I genes is unknown.

The expression of a single classical MHC class I gene has beenseen in other lower vertebrates. Frogs have only a single classicalclass I gene and a large family of nonclassical MHC class Ib genesnot involved in Ag presentation (64, 65). The picture is more com-plicated in other vertebrates in which there are additional genespresent in the genome that resemble classical MHC class I loci. Insharks, most individuals expressed a single gene although othergenes were present (66, 67). Screening of cDNA libraries fromcichlid (68) and pufferfish (69) produced a maximum of four se-quences in one individual. In catfish, isolation of four classicalMHC class I sequences suggested three loci, but most transcriptscorresponded to one sequence (70). Similarly, rainbow trout haveadditional genes in the genome (71), although some appear to benonclassical (45, 72). Analysis of expressed sequences showedtrout predominantly use one MHC class I locus (42, 73, 74).

Two lower vertebrates appear to be exceptions to the expressionof one MHC class I locus, the axolotl and the cod. In the axolotl,between 10 and 17 unique sequences were amplified by RT-PCRfrom cDNA (75). Many appeared to be classical, but the analysiswas limited to the �3 domain; therefore, this remains inconclusive.In cod, 17 different cDNA sequences were identified in one indi-vidual (76). Unfortunately, most were partial clones, truncatedwithin the �2 domain, making it difficult to determine whetherthey were all capable of peptide binding. Nonetheless, the expres-sion of a large number of genes is unique in these animals.

Kaufman (2) argues that the dominant expression of a singleMHC class I locus in most lower vertebrates is attributable to theproximity of MHC class I to the TAP genes, which provides anopportunity for coevolution of these proteins involved in interre-lated functions in Ag presentation. In sharks, fish, and frogs theMHC class I genes are in close proximity to the Ag-processinggenes (LMP2 and LMP7), proteasome genes, and the Ag trans-porter genes (TAP1 and TAP2) in a “true class I region” (2). Inchickens, class I genes flank either side of the TAP genes and theLMP genes have been lost from the genome (17, 27). The majorlocus of chickens is believed to be evolving in concert with thetransporters, while the minor locus is expressed at a lower levelreported to be due to deletions within their promoters (77). Inrainbow trout, the MHC class I genes appear constrained to evolvewithin deep lineages, whereas the genetically unlinked MHC classII evolves rapidly (42). This is potentially due to linkage to theAg-processing and presentation genes. Allelic lineages have

evolved in Xenopus, the MHC class I genes presumably con-strained by linkage to TAP and the proteasome gene LMP7, sincethese genes are known to exist within extended haplotypes (78–80). Similarly, evidence of extended haplotypes of class I, TAP,and LMP genes exists in the nurse shark (67). The evidence fromthese lower vertebrates implies that the haplotypes are stable andthe proteins encoded in the MHC class I region may have an op-portunity to coevolve. Indeed, in ducks we had evidence that theTAP2 gene is significantly polymorphic (36), with alleles differingprimarily in the regions thought to confer peptide specificity. Ad-ditional MHC class I loci might then become useless because theycannot accept peptides from the mismatched TAP transporters. Al-though this provides a compelling argument for inactivation ofredundant loci, the same selective forces should favor homozygousexpression from the major locus, which is never seen.

Evidence from other birds suggests that the minimal MHC ofthe domestic chicken is not a paradigm for all birds. Indeed, thewild ancestor of the chicken, the red jungle fowl has 14 MHC classI genes (27). The quail has duplicated genes within the MHC classI region and has seven genes, including three partial pseudogenesand two weakly expressed loci (26). Southern blot evidence sug-gests that other wild birds have undergone expansion of the MHCclass I region, including sparrows (29) and reed warblers (32, 33).In the reed warbler, 14 cDNA clones representing eight differentsequences encoding MHC class I were isolated from one individ-ual, and all appeared classical (32). Relative expression has notbeen examined; however, 5 of 14 cDNA clones were identical,suggesting a bias for this locus. It has been argued that migratorybirds have the need for more MHC class I and II (both allelicdiversity and number of loci) for greater protection against thediverse pathogens that would be encountered than birds in a sed-entary lifestyle (33). The mallard duck is also a migratory bird thatthat would encounter many pathogens in its environment. Our dataon ducks would suggest that despite the evident duplication of thelocus, one gene contributes predominantly to Ag presentation.

Organization and evolution of the locus

The MHC class I genes were organized in the same transcriptionalorientation and more closely resembled each other than the MHCclass I genes of the chicken or quail. This suggested that the genesarose recently in successive rounds of duplication since the diver-gence of these species. Evidence from a dot plot suggested that oneunit of duplication included a pair of duck MHC class I genes,since the entire intergenic region spanning the UAA and UBAgenes was similar to the region spanning the UDA and UEA genes.The duplicated region includes fragments of CR1 flanking eitherside of a GGLTR11 element. The conservation of the location ofthis retroelement in the intergenic sequence in three separate lo-cations suggests that a pair of genes were involved in duplicationevents. UAA and UBA represent one pair and UDA and UEA rep-resent another pair. It is likely that UCA was originally part of athird pair; however, one gene and part of the intergenic region havebeen lost. Similarly, units of duplication were evident in the quailMHC class I region (26). Previously, no repetitive DNA elementswere identified in chicken or quail MHC class I regions. Our re-analysis of the quail MHC class I region with RepeatMasker de-tected a retroelement GGLTR-12C, which is most similar to theSoprano LTR element, immediately upstream of the pseudogeneCoja-F. This was different from the two retroelements identified inthe duck MHC class I region, including GGLTR-11 and GGLTR-5(Hitchcock). Therefore, both the MHC class I genes and the re-petitive DNA elements are unique within each species. This sug-gests that the primordial MHC class I region of birds was a simple

6710 FIVE DIFFERENTIALLY EXPRESSED MHC CLASS I GENES IN DUCKS

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

MHC, which has undergone duplication independently within eachspecies.

The duck MHC class I genes appear to be confined to the TAP2side of the transporter genes. Using Southern blot analysis, we sawno evidence of MHC class I genes on the TAP1 side of the locus,unlike the chicken and quail (26). The duck MHC gene organiza-tion may resemble the ancestral gene organization before a trans-location event that placed MHC class I genes on either side of theTAP genes. The few hybridizing bands observed on Southern blotsof duck MHC class I genes suggested that there are no other MHCgenes located in the genome. This differs from chickens in whichthere were weakly hybridizing bands that were identified as re-striction fragment pattern Y (Rfp-Y) (15, 81). Additional MHCclass I genes were found in the Rfp-y region (15, 16) nearby on thesame microchromosome (6). Thus, it is likely that there is noequivalent to the Rfp-y region in ducks.

Duck MHC class I genes reveal no locus-specific features

We suggest that until locus designation is defined for a given al-lele, all duck MHC class I sequences be given names that do notattempt to assign a locus. Our analysis of three different haplotypesshowed that the predominantly expressed gene was adjacent toTAP2 (36). These dominant alleles (U*02, U*03, and U*08) wereon separate branches of the phylogenetic tree, thus widely diver-gent alleles are expressed at a single locus. Similarity of sequencescannot be used to assign locus identity. The five MHC class I genesin one haplotype are remarkably similar to one another. Clearly anamplification of sequences from genomic DNA would result in anoverestimate of the diversity of duck MHC class I genes. Thesequences at other loci appear to resemble alleles of classical MHCclass I sequences, although the extent of polymorphism at the otherloci has not been assessed.

Why duplicate, then inactivate extra copies of the MHC class Iloci? The answer must lie in the ability to access the informationavailable in the junk nearby. MHC class I genes evolve by recom-bination through unequal crossing over, allelic exchange, and geneconversion (82), and individual alleles have a short lifetime ofusefulness (83). The duplication of genes greatly increases the po-tential for recombinatorial processes (84). In the duck, the addi-tional loci encode proteins that appear no more different than al-leles at a locus. The neighboring genes, encoding proteins with allthe hallmarks of fully capable Ag-presenting proteins, may pre-serve genetic diversity, which can then be exchanged when thespecies comes under duress from a new pathogen.

DisclosuresThe authors have no financial conflict of interest.

References1. Kelley, J., L. Walter, and J. Trowsdale. 2005. Comparative genomics of major

histocompatibility complexes. Immunogenetics. 56: 683–695.2. Kaufman, J. 1999. Co-evolving genes in MHC haplotypes: the “rule” for non-

mammalian vertebrates? Immunogenetics 50: 228–236.3. Flajnik, M. F. 2001. Comparative genomics of the MHC: glimpses into the evo-

lution of the adaptive immune system. Immunity 15: 351–362.4. Li, K. S., Y. Guan, J. Wang, G. J. Smith, K. M. Xu, L. Duan, A. P. Rahardjo,

P. Puthavathana, C. Buranathai, T. D. Nguyen, et al. 2004. Genesis of a highlypathogenic and potentially pandemic H5N1 influenza virus in eastern Asia. Na-ture 430: 209–213.

5. Chen, H., G. Deng, Z. Li, G. Tian, Y. Li, P. Jiao, L. Zhang, Z. Liu, R. G. Webster,and K. Yu. 2004. The evolution of H5N1 influenza viruses in ducks in southernChina. Proc. Natl. Acad. Sci. USA 101: 10452–10457.

6. Miller, M. M., R. M. Goto, R. L. Taylor, Jr., R. Zoorob, C. Auffray, R. W. Briles,W. E. Briles, and S. E. Bloom. 1996. Assignment of Rfp-Y to the chicken majorhistocompatibility complex/NOR microchromosome and evidence for high-fre-quency recombination associated with the nucleolar organizer region. Proc. Natl.Acad. Sci. USA 93: 3958–3962.

7. Fillon, V., R. Zoorob, M. Yerle, C. Auffray, and A. Vignal. 1996. Mapping of thegenetically independent chicken major histocompatibility complexes B and Rfp-y

to the same microchromosome by two-color fluorescent in situ hybridization.Cytogenet. Cell Genet. 75: 7–9.

8. Schierman, L. W., and A. W. Nordskog. 1961. Relationship of blood type tohistocompatibility in chickens. Science 134: 1008–1009.

9. Schierman, L. W., and A. W. Nordskog. 1963. Influence of the blood grouphistocompatibility locus in chickens on a graft-versus-host reaction. Nature 197:511–512.

10. Miggiano, V. C., I. Birgen, and J. R. Pink. 1974. The mixed leukocyte reactionin chickens: evidence for control by the major histocompatibility complex. Eur.J. Immunol. 4: 397–401.

11. Vainio, O., C. Koch, and A. Toivanen. 1984. B-L antigens (class II) of thechicken major histocompatibility complex control T-B cell interaction. Immuno-genetics 19: 131–140.

12. Vainio, O., T. Veromaa, E. Eerola, P. Toivanen, and M. J. Ratcliffe. 1988. An-tigen-presenting cell-T cell interaction in the chicken is MHC class II antigenrestricted. J. Immunol. 140: 2864–2868.

13. Bacon, L. D., R. L. Witter, L. B. Crittenden, A. Fadly, and J. Motta. 1981.B-haplotype influence on Marek’s disease, Rous sarcoma, and lymphoid leukosisvirus-induced tumors in chickens. Poultry Sci. 60: 1132–1139.

14. Lamont, S. J., C. Bolin, and N. Cheville. 1987. Genetic resistance to fowl cholerais linked to the major histocompatibility complex. Immunogenetics 25: 284–289.

15. Miller, M. M., R. Goto, A. Bernot, R. Zoorob, C. Auffray, N. Bumstead, andW. E. Briles. 1994. Two MHC class I and two MHC class II genes map to thechicken Rfp-Y system outside the B complex. Proc. Natl. Acad. Sci. USA 91:4397–4401.

16. Afanassieff, M., R. M. Goto, J. Ha, M. A. Sherman, L. Zhong, C. Auffray,F. Coudert, R. Zoorob, and M. M. Miller. 2001. At least one class I gene inrestriction fragment pattern-Y (Rfp-Y), the second MHC gene cluster in thechicken, is transcribed, polymorphic, and shows divergent specialization in an-tigen binding region. J. Immunol. 166: 3324–3333.

17. Kaufman, J., S. Milne, T. W. Gobel, B. A. Walker, J. P. Jacob, C. Auffray,R. Zoorob, and S. Beck. 1999. The chicken B locus is a minimal essential majorhistocompatibility complex. Nature 401: 923–925.

18. Miller, M. M., L. D. Bacon, K. Hala, H. D. Hunt, S. J. Ewald, J. Kaufman,R. Zoorob, and W. E. Briles. 2004. 2004 Nomenclature for the chicken majorhistocompatibility (B and Y) complex. Immunogenetics 56: 261–279.

19. Hunt, H. D., and J. E. Fulton. 1998. Analysis of polymorphisms in the majorexpressed class I locus (B-FIV) of the chicken. Immunogenetics 47: 456–467.

20. Li, L., L. W. Johnson, and S. J. Ewald. 1997. Molecular characterization of majorhistocompatibility complex (B) haplotypes in broiler chickens B-haplotype con-trol of CD4/CD8 subsets and TCR V� usage in chicken T lymphocytes. Anim.Genet. 28: 258–267.

21. Li, L., L. W. Johnson, E. J. Livant, and S. J. Ewald. 1999. The MHC of a broilerchicken line: serology, B-G genotypes, and B-F/B-L B sequences. Immunogenet-ics 49: 215–224.

22. Livant, E. J., D. Zheng, L. W. Johnson, W. Shi, and S. J. Ewald. 2001. Three newMHC haplotypes in broiler breeder chickens. Anim. Genet. 32: 123–131.

23. Livant, E. J., J. R. Brigati, and S. J. Ewald. 2004. Diversity and locus specificityof chicken MHC B class I sequences. Anim. Genet. 35: 18–27.

24. Walker, B. A., A. Hateren, S. Milne, S. Beck, and J. Kaufman. 2005. ChickenTAP genes differ from their human orthologues in locus organisation, size, se-quence features and polymorphism. Immunogenetics 57: 232–247.

25. Shiina, T., C. Shimizu, A. Oka, Y. Teraoka, T. Imanishi, T. Gojobori,K. Hanzawa, S. Watanabe, and H. Inoko. 1999. Gene organization of the quailmajor histocompatibility complex (MhcCoja) class I gene region. Immunogenet-ics 49: 384–394.

26. Shiina, T., S. Shimizu, K. Hosomichi, S. Kohara, S. Watanabe, K. Hanzawa,S. Beck, J. K. Kulski, and H. Inoko. 2004. Comparative genomic analysis of twoavian (quail and chicken) MHC regions. J. Immunol. 172: 6751–6763.

27. Consortium, I. C. G. S. 2004. Sequence and comparative analysis of the chickengenome provide unique perspectives on vertebrate evolution. Nature 432:695–716.

28. Freeman-Gallant, C. R., E. M. Johnson, F. Saponara, and M. Stanger. 2002.Variation at the major histocompatibility complex in Savannah sparrows. Mol.Ecol. 11: 1125–1130.

29. Bonneaud, C., G. Sorci, V. Morin, H. Westerdahl, R. Zoorob, and H. Wittzell.2004. Diversity of MHC class I and IIB genes in house sparrows (Passer do-mesticus). Immunogenetics 55: 855–865.

30. Edwards, S. V., E. K. Wakeland, and W. K. Potts. 1995. Contrasting histories ofavian and mammalian Mhc genes revealed by class II B sequences from song-birds. Proc. Natl. Acad. Sci. USA 92: 12200–12204.

31. Edwards, S. V., J. Gasper, D. Garrigan, D. Martindale, and B. F. Koop. 2000. A39-kb sequence around a blackbird Mhc class II gene: ghost of selection past andsongbird genome architecture. Mol. Biol. Evol. 17: 1384–1395.

32. Westerdahl, H., H. Wittzell, and T. von Schantz. 1999. Polymorphism and tran-scription of Mhc class I genes in a passerine bird, the great reed warbler. Immu-nogenetics 49: 158–170.

33. Westerdahl, H., H. Wittzell, and T. von Schantz. 2000. Mhc diversity in twopasserine birds: no evidence for a minimal essential Mhc. Immunogenetics 52:92–100.

34. Westerdahl, H., H. Wittzell, T. von Schantz, and S. Bensch. 2004. MHC class Ityping in a songbird with numerous loci and high polymorphism using motif-specific PCR and DGGE. Heredity 92: 534–542.

35. van Tuinen, M., and S. B. Hedges. 2001. Calibration of avian molecular clocks.Mol. Biol. Evol. 18: 206–213.

6711The Journal of Immunology

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

36. Mesa, C. M., K. T. Thulien, D. A. Moon, S. M. Veniamin, and K. E. Magor.2004. The dominant MHC class I gene is adjacent to the polymorphic TAP2 genein the duck, Anas platyrhynchos. Immunogenetics 56: 192–203.

37. Moon, D. A., and K. E. Magor. 2004. Construction and characterization of afosmid library for comparative analysis of the duck genome. Anim. Genet. 35:417–418.

38. Ross, M. T., S. LaBrie, J. McPherson, and V. P. Stanton. 1999. In CurrentProtocols in Human Genetics. N. C. Dracopoli, ed. Wiley, New York, p. 5.6.1.

39. Wicker, T., J. S. Robertson, S. R. Schultz, F. A. Feltus, V. Magrini, Morrison,E. R. Mardis, R. K. Wilson, D. G. Peterson, A. H. Paterson, and R. Ivarie. 2005.The repetitive landscape of the chicken genome. Genome Res. 15: 126–136.

40. Nei, M., and T. Gojobori. 1986. Simple methods for estimating the numbers ofsynonymous and nonsynonymous nucleotide substitutions. Mol. Biol. Evol. 3:418–426.

41. Korber, B. 2001. HIV signature and sequence variation analysis. In Computa-tional and Evolutionary Analysis of HIV Molecular Sequences.A. G. Rodriqo and G. H. Learn, eds. Kluwer Academic Publications, p. 55.

42. Shum, B. P., L. Guethlein, L. R. Flodin, M. A. Adkison, R. P. Hedrick,R. B. Nehring, R. J. Stet, C. Secombes, and P. Parham. 2001. Modes of salmonidMHC class I and II evolution differ from the primate paradigm. J. Immunol. 166:3297–3308.

43. Feinberg, A. P., and B. Vogelstein. 1983. A technique for radiolabeling DNArestriction endonuclease fragments to high specific activity. Anal. Biochem. 132:6–13.

44. Klein, J., R. E. Bontrop, R. L. Dawkins, H. A. Erlich, U. B. Gyllensten,E. R. Heise, P. P. Jones, P. Parham, E. K. Wakeland, and D. I. Watkins. 1990.Nomenclature for the major histocompatibility complexes of different species: aproposal. Immunogenetics 31: 217–219.

45. Shum, B. P., R. Rajalingam, K. E. Magor, K. Azumi, W. H. Carr, B. Dixon,R. J. Stet, M. A. Adkison, R. P. Hedrick, and P. Parham. 1999. A divergentnon-classical class I gene conserved in salmonids. Immunogenetics 49: 479–490.

46. Salter, R. D., A. M. Norment, B. P. Chen, C. Clayberger, A. M. Krensky,D. R. Littman, and P. Parham. 1989. Polymorphism in the �3 domain of HLA-Amolecules affects binding to CD8. Nature 338: 345–347.

47. Durairaj, M., R. Sharma, J. C. Varghese, and K. P. Kane. 2003. Requirement forQ226, but not multiple charged residues, in the class I MHC CD loop/D strandfor TCR-activated CD8 accessory function. Eur. J. Immunol. 33: 676–684.

48. Gobin, S. J., V. Keijsers, M. van Zutphen, and P. J. van den Elsen. 1998. The roleof enhancer A in the locus-specific transactivation of classical and nonclassicalHLA class I genes by nuclear factor kappa B. J. Immunol. 161: 2276–2283.

49. Gobin, S. J., A. Peijnenburg, V. Keijsers, and P. J. van den Elsen. 1997. Site �is crucial for two routes of IFN�-induced MHC class I transactivation: the ISRE-mediated route and a novel pathway involving CIITA. Immunity 6: 601–611.

50. Gobin, S. J., M. van Zutphen, A. M. Woltman, and P. J. van den Elsen. 1999.Transactivation of classical and nonclassical HLA class I genes through the IFN-stimulated response element. J. Immunol. 163: 1428–1434.

51. Gobin, S. J., A. Peijnenburg, M. van Eggermond, M. van Zutphen,R. van den Berg, and P. J. van den Elsen. 1998. The RFX complex is crucial forthe constitutive and CIITA-mediated transactivation of MHC class I and �2-microglobulin genes. Immunity 9: 531–541.

52. Gobin, S. J., M. van Zutphen, S. D. Westerheide, J. M. Boss, andP. J. van den Elsen. 2001. The MHC-specific enhanceosome and its role in MHCclass I and �2-microglobulin gene transactivation. J. Immunol. 167: 5175–5184.

53. Kaufman, J., R. Andersen, D. Avila, J. Engberg, J. Lambris, J. Salomonsen,K. Welinder, and K. Skjodt. 1992. Different features of the MHC class I het-erodimer have evolved at different rates:chicken B-F and �2-microglobulin se-quences reveal invariant surface residues. J. Immunol. 148: 1532–1546.

54. Magor, K. E., E. J. Taylor, S. Y. Shen, E. Martinez-Naves, N. M. Valiante,R. S. Wells, J. E. Gumperz, E. J. Adams, A. M. Little, F. Williams, et al. 1997.Natural inactivation of a common HLA allele (A*2402) has occurred on at leastthree separate occasions. J. Immunol. 158: 5242–5250.

55. Watanabe, Y., K. E. Magor, and P. Parham. 2001. Exon 5 encoding the trans-membrane region of HLA-A contains a transitional region for the induction ofnonsense-mediated mRNA decay. J. Immunol. 167: 6901–6911.

56. Cross, S. H., and A. P. Bird. 1995. CpG islands and genes. Curr. Opin. Genet.Dev. 5: 309–314.

57. Santourlidis, S., H. I. Trompeter, S. Weinhold, B. Eisermann, K. L. Meyer,P. Wernet, and M. Uhrberg. 2002. Crucial role of DNA methylation in determi-nation of clonally distributed killer cell Ig-like receptor expression patterns in NKcells. J. Immunol. 169: 4253–4261.

58. Chan, H. W., Z. B. Kurago, C. A. Stewart, M. J. Wilson, M. P. Martin,B. E. Mace, M. Carrington, J. Trowsdale, and C. T. Lutz. 2003. DNA methylationmaintains allele-specific KIR gene expression in human natural killer cells.J. Exp. Med. 197: 245–255.

59. Valiante, N. M., M. Uhrberg, H. G. Shilling, K. Lienert-Weidenbach,K. L. Arnett, A. D’Andrea, J. H. Phillips, L. L. Lanier, and P. Parham. 1997.Functionally and structurally distinct NK cell receptor repertoires in the periph-eral blood of two human donors. Immunity 7: 739–751.

60. Lawlor, D. A., J. Zemmour, P. D. Ennis, and P. Parham. 1990. Evolution ofclass-I MHC genes and proteins: from natural selection to thymic selection. Annu.Rev. Immunol. 8: 23–63.

61. Nowak, M. A., K. Tarczy-Hornoch, and J. M. Austyn. 1992. The optimal numberof major histocompatibility complex molecules in an individual. Proc. Natl.Acad. Sci. USA 89: 10896–10899.

62. Lahti, J. M., C. L. Chen, L. W. Tjoelker, J. M. Pickel, K. A. Schat, B. W. Calnek,C. B. Thompson, and M. D. Cooper. 1991. Two distinct �� T-cell lineages canbe distinguished by the differential usage of T-cell receptor V� gene segments.Proc. Natl. Acad. Sci. USA 88: 10956–10960.

63. Sammut, B., A. Marcuz, and L. Du Pasquier. 2002. The fate of duplicated MHCclass Ia genes in a dodecaploid amphibian, Xenopus ruwenzoriensis. Eur. J. Im-munol. 32: 2698–2709.

64. Flajnik, M. F., M. Kasahara, B. P. Shum, L. Salter-Cid, E. Taylor, andL. Du Pasquier. 1993. A novel type of class I gene organization in vertebrates: alarge family of non-MHC-linked class I genes is expressed at the RNA level inthe amphibian Xenopus. EMBO J. 12: 4385–4396.

65. Shum, B. P., D. Avila, L. Du Pasquier, M. Kasahara, and M. F. Flajnik. 1993.Isolation of a classical MHC class I cDNA from an amphibian: evidence for onlyone class I locus in the Xenopus MHC. J. Immunol. 151: 5376–5386.

66. Okamura, K., M. Ototake, T. Nakanishi, Y. Kurosawa, K. Hashimoto, andL. Lefaucheur. 1997. The most primitive vertebrates with jaws possess highlypolymorphic MHC class I genes comparable to those of humans. Immunity 7:777–790.

67. Ohta, Y., E. C. McKinney, M. F. Criscitiello, and M. F. Flajnik. 2002. Protea-some, transporter associated with antigen processing, and class I genes in thenurse shark Ginglymostoma cirratum: evidence for a stable class I region andMHC haplotype lineages. J. Immunol. 168: 771–781.

68. Sato, A., D. Klein, H. Sultmann, F. Figueroa, C. O’hUigin, and J. Klein. 1997.Class I mhc genes of cichlid fishes: identification, expression, and polymorphism.Immunogenetics 46: 63–72.

69. Timon, M., G. Elgar, S. Habu, K. Okumura, and P. C. Beverley. 1998. Molecularcloning of major histocompatibility complex class I cDNAs from the pufferfishFugu rubripes. Immunogenetics 47: 170–173.

70. Antao, A. B., V. G. Chinchar, T. J. McConnell, N. W. Miller, L. W. Clem, andM. R. Wilson. 1999. MHC class I genes of the channel catfish: sequence analysisand expression. Immunogenetics. 49: 303–311.

71. Hansen, J. D., P. Strassburger, G. H. Thorgaard, W. P. Young, andL. Du Pasquier. 1999. Expression, linkage, and polymorphism of MHC-relatedgenes in rainbow trout, Oncorhynchus mykiss. J. Immunol. 163: 774–786.

72. Shum, B. P., P. M. Mason, K. E. Magor, L. R. Flodin, R. J. Stet, and P. Parham.2002. Structures of two major histocompatibility complex class I genes of therainbow trout (Oncorhynchus mykiss). Immunogenetics 54: 193–199.

73. Aoyagi, K., J. M. Dijkstra, C. Xia, I. Denda, M. Ototake, K. Hashimoto, andT. Nakanishi. 2002. Classical MHC class I genes composed of highly divergentsequence lineages share a single locus in rainbow trout (Oncorhynchus mykiss).J. Immunol. 168: 260–273.

74. Grimholt, U., S. Larsen, R. Nordmo, P. Midtlyng, S. Kjoeglum, A. Storset,S. Saebo, and R. J. Stet. 2003. MHC polymorphism and disease resistance inAtlantic salmon (Salmo salar); facing pathogens with single expressed majorhistocompatibility class I and class II loci. Immunogenetics 55: 210–219.

75. Sammut, B., L. Du Pasquier, P. Ducoroy, V. Laurens, A. Marcuz, andA. Tournefier. 1999. Axolotl MHC architecture and polymorphism. Eur. J. Im-munol. 29: 2897–2907.

76. Persson, A. C., R. J. Stet, and L. Pilstrom. 1999. Characterization of MHC classI and �2-microglobulin sequences in Atlantic cod reveals an unusually high num-ber of expressed class I genes. Immunogenetics 50: 49–59.

77. Kaufman, J., J. Jacob, I. Shaw, B. Walker, S. Milne, S. Beck, and J. Salomonsen.1999. Gene organisation determines evolution of function in the chicken MHC.Immunol. Rev. 167: 101–117.

78. Flajnik, M. F., Y. Ohta, A. S. Greenberg, L. Salter-Cid, A. Carrizosa,L. Du Pasquier, and M. Kasahara. 1999. Two ancient allelic lineages at the singleclassical class I locus in the Xenopus MHC. J. Immunol. 163: 3826–3833.

79. Nonaka, M., C. Yamada-Namikawa, M. F. Flajnik, and L. Du Pasquier. 2000.Trans-species polymorphism of the major histocompatibility complex-encodedproteasome subunit LMP7 in an amphibian genus, Xenopus. Immunogenetics 51:186–192.

80. Ohta, Y., S. J. Powis, R. L. Lohr, M. Nonaka, L. D. Pasquier, and M. F. Flajnik.2003. Two highly divergent ancient allelic lineages of the transporter associatedwith antigen processing (TAP) gene in Xenopus: further evidence for co-evolutionamong MHC class I region genes. Eur. J. Immunol. 33: 3017–3027.

81. Briles, W. E., R. M. Goto, C. Auffray, and M. M. Miller. 1993. A polymorphicsystem related to but genetically independent of the chicken major histocompat-ibility complex. Immunogenetics 37: 408–414.

82. Parham, P., E. J. Adams, and K. L. Arnett. 1995. The origins of HLA-A,B,Cpolymorphism. Immunol. Rev. 143: 141–180.

83. Parham, P. 1994. The rise and fall of great class I genes. Semin. Immunol. 6:373–382.

84. Trowsdale, J., and P. Parham. 2004. Mini-review: defense strategies and immu-nity-related genes. Eur. J. Immunol. 34: 7–17.

85. Xia, C., C. Y. Lin, G. X. Xu, T. J. Hu, and T. Y. Yang. 2004. cDNA cloning andgenomic structure of the duck (Anas platyrhynchos) MHC class I gene. Immu-nogenetics 56: 304–309.

86. Ting, J. P.-Y., and A. S. Baldwin. 1993. Regulation of MHC gene expression.Curr. Opin. Immunol. 5: 8–16.

6712 FIVE DIFFERENTIALLY EXPRESSED MHC CLASS I GENES IN DUCKS

by guest on March 28, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from