Advertisement

Immunogenetics

, Volume 70, Issue 4, pp 257–269 | Cite as

Structural and functional diversity arising from intra- and inter-haplotype combinations of duplicated DQA and B loci within the ovine MHC

  • Keith T. Ballingall
  • Isabelle Lantier
  • Helen Todd
  • Frederic Lantier
  • Mara Rocchi
Original Article

Abstract

In sheep, the A and B loci encoding the α and β chains of the classical class II MHC molecules are DRA and DRB and DQA and DQB. Previous analyses described the duplication of the DQA and DQB genes. The majority of haplotypes include DQA1 and DQA2 loci, however, in a number of haplotypes, DQA1 appears absent and these haplotypes have been described as DQA1 null. In these haplotypes, the DQA2 locus is found in combination with a second locus which appeared more closely related to DQA2 than DQA1, hence the description of this locus as DQA2-like. Here we combine our previous analysis of the DQA transcripts with an analysis of the associated DQB transcripts in ten haplotypes from MHC homozygous animals. This allows the potential for surface expression of different haplotype combinations of DQA and B genes and the functional significance of DQA2-like and its predicted DQB partner to be determined. Atypical DQB transcripts (DQB2-like) were identified in haplotypes classified as DQA1-null and conserved DQB2-like orthologues were identified in other Bovidae indicating trans-species conservation of the allelic lineage. Functional combinations detected by co-transfection of DQ1, DQ2 and DQ2-like genes demonstrates the potential for a wide range of DQ molecules derived from both intra- and inter-haplotype as well as inter-locus combinations. We provide evidence that DQA2-like and B2-like genes form an evolutionary conserved pair which generates structurally distinct class II molecules that are likely to present a distinct range of peptides to CD4+ T cells.

Keywords

Ovine MHC class II DQ Haplotype Expression 

Introduction

Molecular studies in out-bred livestock species are often constrained by a lack of animals with well-defined genetic backgrounds as well as the molecular tools and reagents that are readily available in rodent models. This includes animals that control for diversity at the major histocompatibility complex (MHC), one of the most genetically diverse regions of the vertebrate genome (Parham and Ohta 1996; Trowsdale 2011). To address some of these issues an MHC-defined ovine flock was generated at the Moredun Research Institute based on four diverse MHC haplotypes. Concomitantly, five inbred lines of Prealpes sheep were developed at the INRA laboratory in Nouzilly, France with approximately 80% homozygosity across the genome (Grain et al. 1993; Viginier et al. 2012). Genetic material from MHC homozygous animals from each of these resource flocks allows accurate characterisation of MHC haplotype diversity in the domestic sheep, Ovis aries.

In sheep, the principal A and B loci encoding the α and β chains of the classical class II MHC molecules are DRA and DRB and DQA and DQB (Ballingall et al. 1992; Wright and Ballingall 1994). A single, highly transcribed and polymorphic DRB locus (DRB1) co-expresses with a single DRA locus (Ballingall et al. 1992, 2010; Ballingall and Tassi 2010). Previous analyses of the DQ repertoire in sheep described the duplication of the DQA and DQB genes (Scott et al. 1987; van Oorschot et al. 1994; Wright and Ballingall 1994; Ballingall et al. 2015). All sheep haplotypes analysed so far have two DQA genes, with the majority including a DQA1 and a DQA2 locus (Scott et al. 1991a). However, in a number of haplotypes, DQA1 appears absent and these haplotypes have been described as DQA1 null (Fabb et al. 1993; Ballingall et al. 2015). In these haplotypes, the DQA2 locus is found in combination with a second locus which, based on phylogenetic analyses of the second exon, appeared more closely related to DQA2 than DQA1, hence the description of this locus as DQA2-like.

We recently described the first full-length DQA2-like transcripts, with evidence for alleles at the DQA2-like locus having arisen through an ancient recombination between DQA1 and DQA2 loci. This recombination resulted in the 5-prime (5′) end of the gene being more closely related to DQA2 while the 3-prime (3′) end appeared closer to DQA1 (Ballingall et al. 2015). Depending on the physical location of the DQA2-like gene relative to the conserved DQA2 locus, DQA2-like alleles may represent a divergent allelic family at the DQA1 locus or a distinct DQA locus. Either way, as functional DQ genes in other mammalian species occur as closely linked A/B gene pairs, the nature of the DQB gene associated with DQA2-like remains to be determined. To further complicate the issue, a recent study of sheep class II haplotype diversity identified additional duplications and deletions at the DQA and B loci in some haplotypes (Ali et al. 2017).

An earlier functional analysis identified surface expression of DQ1 molecules following transfection of physically closely linked DQA1 and DQB1 gene pairs within a number of genomic cosmid clones (Wright and Ballingall 1994); however, the functional significance of DQA2/B2, DQA2-like and its predicted DQB gene partner remain to be determined. Combined with flow cytometry, this transfection system has been used extensively in many species, including sheep and cattle, to demonstrate the capacity of different class II A/B gene combinations to generate a heterodimer at the cell surface (Ballingall et al. 1992; Wright and Ballingall 1994). In cattle, Norimine and Brown demonstrated the ability of DQA and DQB genes within a haplotype (intra-haplotype) or from different haplotypes (inter-haplotype) to form functional class II molecules capable of presenting peptide antigens to CD4+ T cells (Norimine and Brown 2005).

Most DQB sequence data from sheep is limited to the polymorphic second exon. A single full-length DQB transcript (van Oorschot et al. 1994) and a number of partial DQB transcripts have been submitted to public databases (Herrmann-Hoesing L.M, unpublished). DQB transcript data may also be predicted from sheep genome sequence resources (Scott et al. 1991b; Wright and Ballingall et al. 1994; Hermann-Hoessing et al. 2008; Gao et al. 2010) and the latest draft of the sheep genome.

Here, we combine our previous analysis of the DQA transcripts with an analysis of the associated DQB transcripts in ten haplotypes from MHC homozygous animals. This allows the potential for surface expression of different haplotype combinations of DQA and B genes and the functional significance of DQA2-like and its predicted DQB partner to be determined.

Materials and methods

Animals and nucleic acid

All Scottish Blackface sheep were derived from a flock maintained at the Moredun Research Institute, Edinburgh, UK. MHC homozygous animals representing each of the four haplotypes (501a, 501b, 504a and 504b) carried by the two founding Scottish Blackface rams were generated by sire/daughter mating. We have previously described the class I, class II DR and DQA regions associated with these four haplotypes (Miltiadou et al. 2005; Ballingall et al. 2008a, b, 2010, 2015). Five Pré-Alpes du Sud sheep lines (hereafter termed Prealpes) were maintained at the INRA facility in Nouzilly, France. Each line was closed for up to 14 generations of inbreeding resulting in animals with approximately 80% homozygosity across the genome (Grain et al. 1993; Viginier et al. 2012 and F. Lantier, unpublished information). As previously described (Ballingall et al. 2015), genomic DNA was prepared from two DRB1/DQA homozygous animals from within each line for analysis of DQB diversity. A single Black Welsh mountain sheep was also included in the analysis as it was identified with a different DRB1/DQA2-like haplotype as part of an unrelated study. Peripheral blood mononuclear cells (PBMC) were prepared by density centrifugation according to standard methodologies. Total RNA was extracted from 2 × 106 PBMC from all Prealpe and Welsh mountain sheep or from 1 × 106 Theileria Lestoquardi immortalised mononuclear cells derived from the Moredun homozygous animals, using the Qiagen RNeasy direct kit according to the manufacturer’s instructions. First-strand cDNA was prepared using the ImProm-II RT system (Promega) in a 40 μl reaction using 200 ng of total RNA.

Defining the range of DQB genes in each haplotype

Primers for the amplification of the second exon of the DQB genes were designed from previously published genomic sequence data covering the flanking intron/exon boundaries (Scott et al. 1991a; Wright and Ballingall 1994; Herrmann-Hoesing et al. 2008; Gao et al. 2010). Primers were selected to preferentially amplify DQB1 or DQB2 sequences, although it became clear as more alleles were amplified that DQB2 was sometimes amplified with DQB1 primers and vice versa. The primer sequences are listed in Table 1. Each PCR reaction was carried out in a final volume of 50 μl containing 200 nM of each primer, 1 U OneTaq DNA polymerase (New England Biolabs) and 50 ng of DNA template. Amplification reactions were performed under the following cycling conditions; 94 °C for 4 min followed by 30 cycles of 94 °C for 30 s, 55 °C for 30 s and72°C for 30 s. A final cycle of 72oC for 5 min was added to complete the reaction.
Table 1

Primer combinations employed for the amplification of DQB alleles from genomic and cDNA

Target locus

Template

Primer sequence 5′ – 3’

Ovar-DQB1/2-like exon 2a

Genomic DNA

363F ccccgcag/aggatttcstg

364R cggcactcac/ctcgccgctgc

Ovar-DQB2 exon 2b

Genomic DNA

363F ccccgcag/aggatttcstg

406R acgctcac/ctcgccgctgcc

Ovar-DQB1/B2

cDNA

245 tgggtgttgactaccattast

Forward primers

246 attasttcttsctttgttctc

378 cagcggcctttggacrgcagytg

Ovar-DQB1/B2

cDNA

247 gacargcakctgggaattc

Reverse primers

248 acgcassyattayagaagagc

Ovar-DQB2-like F

cDNA

392 attagttgttccttttttctc

Ovar-DQB2-like R

395 aaaatatcctcaggagtcagc

Ovar-DQB2-like Rc

401caagaacacgcagctattaca

(Full-length primers)

Ovar-DQB2-like F

cDNA

393accaactggagctcctcacat

Ovar-DQB2-like R

390 ctgtgaggagctccagttggt

Internal exon 2 specific

Following convention all gene names are shown in italics

aThis primer combination also amplifies some DQB2 alleles

bThis primer combination also amplifies some DQB1 alleles

cPrimer used to amplify the 3’UTR region of DQB2-like*01:01. The solidus (/) marks the intron 1 exon 2 boundaries in the genomic forward primers and exon 2, intron 2 in the genomic reverse primers

Analysis of PCR products

The products of each PCR reaction were separated on a 1% agarose gel, stained with GelRed (Cambridge BioSciences) and visualised under a UV transilluminator. PCR products were purified from the agarose gel using the SV Gel and PCR Clean-Up system according to the manufacturer’s instructions (Promega). Eluted DNA was quantified using a Nanodrop spectrophotometer and commercially Sanger sequenced in both directions using the amplification primers. Sequence contigs were generated using the SeqManII™ programme of the DNASTAR package.

Cloning and sequence analysis

Ovar-DQB alleles were cloned into pGEM-T vector (Promega) following standard protocols and individual clones representing each allele were discriminated by colony PCR followed by Rsa I digestion of the PCR product and resolution of the allele-specific fragments on an 8% polyacrylamide gel. At least three clones representing each allele were sequenced in both directions and polymorphic positions identified using the SeqManII™ programme.

Amplification of full-length DQB1 and B2 transcripts

Full-length DQB1 and B2 transcripts were amplified from cDNA using the range of primer combinations listed in Table 1. DQB1 and B2 specific primers were designed surrounding the start codon and within the 3’UTR of a previously published ovine cDNA sequence (van Oorschot et al. 1994). No single set of DQB1 and B2 specific primers amplified all full-length transcripts, probably due to the high levels of diversity within the 5′ and 3′ UTR regions. All RT-PCR reactions were carried out in a final volume of 50 μl containing 200 nM of each primer, 1 U OneTaq DNA polymerase and 50 ng of cDNA template. The cycling profile consisted of 35 cycles of 1 min at 94 °C, 1 min at 55°C and 1 min at 72 °C. PCR fragments of the correct size were gel purified and cloned into the pGEM-T vector. Between three and five clones representing each allele were sequenced in both directions to verify integrity and the sequence was compared to the direct sequence of the purified PCR fragment for final validation.

Development of primers for the amplification of full-length DQB2-like transcripts

DQB1/B2 primers failed to amplify DQB2-like transcripts. Internal DQB2-like sequence-specific forward (393) and reverse primers (390) were designed (Table 1) using the genomic second exon sequence. These sequence-specific primers were tested in combination with the range of full-length forward or reverse primers. Fragments that amplified were sequenced and confirmed as DQB2-like through comparison with the genomic second exon sequence. These sequences were used to develop the full-length primers shown in Table 1.

Sequence analysis and validation

DQB genes were assembled from sequences obtained from both directions using the SeqManII programme in the DNASTAR package. All polymorphic sites were checked manually and each consensus sequence was fully validated by independent sequence analysis from clones derived from a second PCR reaction using different cDNA. All full-length and partial sequences have been deposited in the European Nucleotide Archive (ENA) and assigned the accession numbers LT837701-18 listed in Table 2.
Table 2

Primer combinations used for the amplification of individual DQB alleles

DQB allele

Primer combinations

Accn. number

DQB1*01:01:01

245/247

LT837701

DQB1*03:01:01

245/247

LT837702

DQB1*04:01:01

245/247

LT837703

DQB1*05:01:02

246/248

LT837704

DQB1*05:02:01

246/248

LT837705

DQB1*06:01:01

246/248

LT837706

DQB1*07:01:01

245/248

LT837707

DQB1*07:02:01

246/395

LT837708

DQB2*01:01:01

245/247

LT837709

DQB2*02:01:01

245/247

LT837710

DQB2*03:01:01

245/247

LT837711

DQB2*05:01:01

378/248

LT837712

DQB2*06:01:01

246/248

LT837714

DQB2*08:01:01

245/248

LT837715

DQB2*08:02:01

246/248

LT837716

DQB2-like*01:01:01

246/395

LT837717

392/401a

DQB2-like*02:01:01

392/395

LT837718

Following convention all gene names are shown in italics

aPrimer 401 amplified the 3’UTR from DQB2-like*0101

Class II DQB nomenclature

Discrimination between alleles at individual DQB loci was carried out by sequence and phylogenetic analysis of the full-length transcripts. Previous genomic analysis identified linked DQA1/B1 and DQA2/B2 gene pairs (Wright and Ballingall 1994; Herrmann-Hoesing et al. 2008). DQB transcripts were assigned to either B1 or B2 by sequence similarity to the relevant genomic A/B gene pairs. To maintain consistency with recent changes to HLA nomenclature (Marsh et al. 2010) (http://www.ebi.ac.uk/ipd/mhc), the following nomenclature for alleles at the Ovar-DQB loci were adopted. The first two digits following the species and locus designation (Ovar-DQB1) represents the allelic family (Ovar-DQB1*01, *02 etc). Alleles within a family differ by no more than four amino acids within the second exon and no more than four amino acid differences in the remainder of the transcript. The next two digits separated by a colon, indicate coding change within the allelic family (Ovar-DQB1*01:01) and the next two digits (Ovar-DQB1*01:01:01) may be used to indicate silent or synonymous substitutions. In this paper, we have named the divergent DQB sequences DQB2-like in order to discriminate them from DQB1 and DQB2 sequences and to reflect their functional association with DQA2-like. The allelic nomenclature shown in Table 2 is used throughout this paper. For consistency, the same nomenclature was assigned to the six full-length sheep DQB sequences obtained from public databases for comparative and phylogenetic analysis.

Evolutionary analysis

Multiple alignments of the nucleic acid and predicted amino acid sequences generated here and including other published cDNA or genomic sequences were produced using CLUSTAL Omega available on the EMBL-EBI website http://www.ebi.ac.uk/Tools/msa/clustalo/. The multiple nucleotide sequence alignment was used to estimate the maximum likelihood tree. Prior to phylogenetic tree estimation, the optimum substitution model was selected using the model selection feature (Kalyaanamoorthy et al. 2017) in IQ-TREE (Trifinopoulos et al. 2016). The optimum substitution model selected for the DQB sequences was HKY+R3 (Hasegawa et al. 1985), which reflects differences in the rates of substitutions between DQB sequences. The topology of the tree was tested with 1000 bootstrap replicates using the ultrafast booststrap method of Minh et al. (2013). The number of nucleotide and amino acid differences between sequences was calculated using MEGA6 (Tamura et al. 2013), including 1st+2nd+3rd codon positions and eliminating all positions containing gaps and missing data. There were a total of 773 positions in the final nucleotide and 255 positions in the final amino acid dataset. Analyses to estimate the evolutionary divergence between sequences were conducted using the Kimura 2-parameter model (Kimura 1980). The rate variation among sites was modelled with a gamma distribution. Standard error estimates were obtained by a bootstrap procedure involving 10,000 replicates.

Molecular modelling

Molecular models of the ovine class II DQ1, DQ2 and DQ2-like molecules were generated using Swiss-Model (https://oligo.swissmodel.expasy.org, Arnold et al. 2006). DQ1, DQ2 and DQ2-like molecules were built using amino acid sequences predicted from DQA1:01:01, DQB1*01:01, DQA2*01:01, DQB2*01:01 and DQA2-like*01:01, DQB2-like*01:01 pairs of sequences, respectively. The oligomeric structure prediction option in Swiss-Model was employed. Hetero-oligomeric assemblies were constructed using the beta-oligo-version of the SWISS-MODEL server and were generated using the human HLA-DQ8 molecule (PDB entry 5ksa.1, Petersen et al. 2016) at 2.0 Å resolution as the template, assuming conservation of the structure, as described by Biasini et al. (2014). Alignment of Ovar-DQαβ chains with the template reference sequences identified a sequence similarity of 78% for DQ2-like, 81% for DQ1 and 79% for DQ2. The global model quality estimation (GMQE) which reflects the accuracy of the model were all greater than 0.71.

Generation of mammalian expression constructs

Full-length DQA and DQB transcripts were amplified from the pGEM-T-vector clones using a proof reading Taq polymerase (Pfu, Promega) and a modified forward primer (Table 1) compatible with directional cloning into the eukaryotic expression vector pcDNA3.1/V5-His-TOPO (Invitrogen). PCR reactions were carried out in a final volume of 50 μl containing 200 nM of each primer, 1 U Taq polymerase and 50 pg of plasmid template. The cycling profile consisted of 25 cycles of 1 min at 94°C, 20 s at 60°C and 2 min at 68°C, and 1 cycle of 1 min at 94°C, 20 s at 60°C and 10 min at 68°C. Amplified products were gel purified and cloned into the expression vector according to the manufacturer’s instructions. Clones were selected and sequenced in both directions to verify orientation and integrity. Plasmid DNA for transfection was prepared from 50 ml bacterial cultures and suspended under aseptic conditions to a concentration of 500 ng/μl in sterile water. DRA*01:01 and DRB1*05:01 constructs were also generated and used as the positive control in each co-transfection experiment.

Transfection and detection of MHC gene expression

Pairs of MHC class II DQA and B genes were co-transfected into COS-7 cells using the Amaxa nucleofector system (Lonza) according to the manufactures instructions. Briefly, COS-7 cells were cultured in complete IMDM (cIMDM) plus 10% heat inactivated calf serum. Cells nearing confluence were passaged at a ratio of 1:2, 24 h prior to transfection to ensure a maximum rate of division. On the day of transfection cells were removed from the plastic using trypsin EDTA, washed and re-suspended in cIMDM to a final concentration of 5 × 105 ml−1. One millilitre of the cell suspension was pelleted by centrifugation and re-suspended in 100 μl Nucleofector solution R. Plasmid constructs containing class II MHC A and B genes (2.5 μg of each) were added and the cells shocked using programme W-001. A mock transfection control (no DNA) was also included. Following incubation at room temperature for 10 min, 0.5 ml of warm cIMDM was added and the cells transferred to one well of a six-well plate already containing 1.5 ml of warm media and incubated at 37 °C, in 5% CO2 humidified atmosphere. Products of transfected class II MHC genes were detected at the cell surface after 48 h by indirect immunofluorescence using either pan class II-specific monoclonal antibody SW73.2 (Hopkins et al. 1986), which recognises both DQβ and DRβ chains, or VPM 36, which recognises the DQα chain (Ballingall et al. 1995). Alexa fluor 488-conjugated goat anti-mouse IgG (molecular probes, 1 μg/ml final dilution) was used as a secondary antibody for VPM 36, while Alexa fluor 488-conjugated goat anti-rat IgG was used with SW73.2. Propidium iodide solution was added prior to sample acquisition (10 μl of 10 μg/ml stock) as viability discriminator. Mock transfected cells were used for both staining (primary and secondary) and unstained (no antibody) controls. Sample data were acquired using a flow cytometer equipped with a 488 nm argon-ion laser (FACSCalibur) and analysed using CellQuest (Becton Dickinson) software. In each transfection, the proportion of positive cells relative to the unstained population was determined. During the course of the study, each transfection was repeated between three and six times allowing an overall picture of the efficiency of each combination of genes to express at the cell surface.

Results

Characterisation of the range of class II MHC-DQB genes in MHC homozygous animals

The second exons of the DQB genes were amplified from genomic DNA to determine the range of diversity associated with each haplotype. Alleles at both DQB1 and DQB2 loci were identified in eight animals. In contrast, a DQB2 and a divergent DQB gene which for the purposes of this paper we term here DQB2-like, was identified in the two animals with haplotypes previously defined as DQA1 null. Knowledge of the nature and allelic composition of each haplotype allowed targeted amplification of full-length DQB transcripts from all ten animals. Consistent with the initial genotyping data, seven distinct full-length DQB1 and eight B2 transcripts were amplified, cloned and sequenced. Identical DQB1 and B2 sequences were obtained from Prealpe lines 2 and 3 and an additional divergent DQB2 transcript (DQB2*05:01) was identified in the 504a haplotype raising the possibility of a functional duplication of the DQB2 locus. Such was the diversity at the 5-prime and 3-prime ends of the second exon of the DQB2*0501 allele (Supplementary Fig. 1) amplification with the genomic primers would be unlikely. No single set of DQB1 and B2 specific primers and PCR conditions amplified all full-length transcripts, probably due to the high levels of diversity within the 5′ and 3′ UTR regions. The DQB2-like primers amplified the first full-length DQB2-like transcripts (DQB2-like*01:01 and 02:01), one from each of the DQA1 null haplotypes. The accession numbers of the DQB transcripts are listed in Table 2 along with the primer combinations used to amplify each DQB allele.

Completing the class II haplotypes

With the inclusion of the DQB diversity, we were able to complete the ten class II MHC haplotypes (Prealpe haplotypes 2 and 3 appear identical). The allelic composition of each haplotype is shown in Table 3. With the exception of haplotype 504a, each contain two pairs of DQA and DQB genes. Haplotypes 501a and BWM each include a DQA2/B2 pair and a DQA2-like/B2-like pair.
Table 3

Completed class II MHC haplotypes

Haplotype

DQA1

DQB1

DQA2

DQB2

DQA2-like

DQB2-like

DRB1

DRA

501a

  

01:02

02:01

01:01

01:01

01:01

01:01

BWM

  

08:01

02:01

02:01

02:01

13:01

01:01

501b

01:01

01:01

01:01

01:01

  

05:01

01:01

504a

02:01

04:01

03:01

05:01

  

09:01

02:01

06:01

504b

01:02

03:01

02:01

03:01

  

03:01

01:01

PA line 1

05:01

05:01:02

06:01

08:01

  

14:02

01:01

PA line 2

06:01

06:01

03:01

06:01

  

03:07

01:01

PA line 3

06:01

06:01

03:01

06:01

  

03:07

01:01

PA line 4

03:01

07:01

07:01

08:01

  

19:01

01:01

PA line 5

05:01

05:02

06:01

08:02

  

03:12

01:01

PA, Prealpe, BWM, Black Welsh mountain

Following convention all gene names are shown in italics

Sequence analysis of full-length DQB transcripts and corresponding amino acid sequences

An alignment of the nucleic acid sequences of each full-length transcript is shown in Supplementary Fig. 1, along with all other full-length sheep DQB transcripts currently available from the public databases. The corresponding amino acid sequences are shown in Fig. 1. The predicted amino acid sequences are consistent with functional DQβ proteins. DQβ1 and DQβ2 molecules consist of full-length and mature proteins of 261 and 229 amino acids, respectively. In contrast, a three-base deletion in the second exon of the DQB2-like sequences results in a mature protein with a β1 domain smaller by one amino acid. The deletion corresponds to a residue associated with the binding of peptides, according to the crystal structure of the human HLA-DR molecule (Brown et al. 1993). This is therefore likely to influence the range of peptide antigens bound and presented to CD4+ T cells. An unrelated codon deletion is also present in the Ovar-DRB2*0701 sequence, a lambda gt10 cDNA clone which was the first sheep DQB sequence submitted to the public databases in the early 1990s (L08792, van Oorschot et al. 1994). No other Bovidae DQB sequence has a similar deletion and the clone is no longer available to validate (Jill Maddox, personal communication).
Fig. 1

Pair-wise alignment of the amino acid sequences predicted from the each of the DQB transcripts described in this study along with others available in GenBank. Each sheep sequence is labelled according to the nomenclature described in the materials and methods. These include DQB1*02:01 and DQB2*04:01 which are predicted from the genomic sequence EU176819; DQB1*08:01 and DQB2-like*02:02 which are both predicted form the Mouflon genome sequencing project (XM_012173129 and XM_012159546); DQB1*05:01 which is predicted from BAC clone FJ985876 and DQB2*07:01 derived from a cDNA library clone (L08792). Representative Bos taurus cattle Bota-DQB1, (BC102959), Bota-DQB2, (D37954) and a Bos indicus DQB2-like orthologue (Boin-DQB*0601, X79349); Asian water buffalo, Bubalus bubalis, (Bubu-DQB, DQ908903) and Yak, Bos mutus (Bomu-DQB, XM_005909262) are included for comparative purposes. The boundaries of the beta-1 domain are shaded. Missing data and indels are represented by “–”

A number of features appear characteristic of DQβ1 or DQβ2 sequences. For example, all DQβ1 sequences have an SG splice motif between the final amino acid of the signal peptide and the first base of the mature protein, whereas DQβ2 and β2-like all have GR at these positions. The divergent DQβ2*0501 sequence which includes the DQβ1 SG motif is the only exception. Signal sequence prediction tools such as Phobius (http://phobius.sbc.su.se/), signal BLAST (http://sigpep.services.came.sbg.ac.at/signalblast.html), and SignalP, (http://www.cbs.dtu.dk/services/SignalP/) each suggest that SG and GR are effective splice sites (data not shown). An NQ motif at position 33–34 in the mature DQβ1 protein combined with SG is consistent in all DQβ1 sequences presented here and in all β1 sequences held within public databases. For DQβ2, GR was associated with NR at positions 33–34 in all ovine DQβ2 sequences held within public databases; APGSW at positions 86–91 also appears conserved, with the only exception being DQβ2*0501.

In addition to the deletion at position 65 described above, many unique amino acid features are associated with the DQβ2-like sequences. Overall, 12 amino acid substitutions appear unique to the DQβ2-like sequences. The number of nucleotide and amino acid substitutions and the genetic distance between DQβ1, β2 and β2-like are shown in Table 4. DQβ2-like*01:01 differs from the DQβ1*01:01 and β2*01:01 reference sequences by 38 and 40 amino acid substitutions, respectively.
Table 4

A pair-wise comparison of DQB2-Like*01:01 with DQB2*01:01 and DQB1*01:01

 

Amino acid differences

Nucleotide differences

Genetic distance (standard error)

 

DQB2-like*01:01

DQB2*01:01

40

79

0.120 (0.015)

DQB1*01:01

38

80

0.120 (0.016)

Following convention all gene names are shown in italics

While significant amino acid variation is apparent throughout the translated region, a high level of nucleotide diversity, including locus specific insertions and deletions, is evident in the immediate 3’untranslated (3’UTR) regions (Fig. 2). All DQB1 alleles have a four or five base pair deletion at 17 or 18 bases 3′ of the translation stop codon (depending on the allele). The DQB2 alleles at this location include the conserved nucleotides, ATGGA while ACGGG appears in the DQB2-like sequences. An additional single base deletion is located at position 839 in DQB2-like*01:01. In the absence of additional full-length DQB2-like sequences, we are unable to confirm the broader conservation of these motifs in sheep however, these motifs are identical in orthologous sequences from the wild Yak (Bos mutans) and Asian water buffalo (Bubalus bubalis) indicating their conservation over millions of years of evolution within the family Bovidae.
Fig. 2

Alignment showing the diversity at the immediate 3’UTR regions of sheep and other Bovidae DQB transcripts. The translation termination codon is indicated by triple number signs (###). Missing data and indels are represented by “–”

Comparative analysis

Conserved orthologues of DQB1 and DQB2 loci can readily be identified in cattle and other Bovidae through a BLAST search of the public databases. Generally, sheep DQB1 and DQB2 alleles show greater similarity to alleles at the othologous loci in cattle than to each other. Many of the locus specific features identified in sheep are conserved in cattle (Supplementary Fig. 1, Fig. 1 and Fig. 2). A similar analysis using the sheep DQB2-like sequences also identifies highly conserved orthologues in other Bovidae including Bos taurus and Bos indicus cattle, yak (Bos mutans), Asian water buffalo (Bubalus bubalis), bison, as well as mouflon (O. orientalis), a wild ancestor of domestic sheep. All nucleotide and amino acid motifs specific for sheep DQB2-like, including the codon deletion within the second exon appear to have been conserved over the 20 million years since these species shared a common ancestor.

Phylogenetic analysis

The relationship between DQB sequences was analysed in greater detail using 27 sequences derived from this study as well as orthologous sequences from the mouflon, taurine cattle, yak and buffalo. The pair-wise alignment of these DQB nucleotide sequences shown in Supplementary Fig. 1 was used to estimate the maximum likelihood tree shown in Fig. 3. The tree topology estimates that the DQB1, DQB2 and DQB2-like sequences cluster independently. It appears that the DQB2-like lineage split from the DQB2 lineage after the duplication event leading to DQB1 and DQB2, an observation which is supported by high bootstrap values. As DQB2 and DQB2-like appear together on the same haplotypes, they clearly represent independent loci. Unlike our previous observation which suggested that the DQA2-like sequences arose through an ancient recombination resulting in exons 1 and 2 appearing closer to DQA2 while the remainder of the transcript appears closer to DQA1 (Ballingall et al. 2015), this does not appear to be the case for the DQB2-like sequences.
Fig. 3

Maximum likelihood tree estimating the relationships between the 27 Bovidae DQB sequences shown aligned in Supplementary Fig. 1. The human HLA-DQB1 (M20432), murine H2-Ab1 (NM_207105) and swine SLA-DQB1 (NM_001113694) sequences were used to root the tree. Species designations associated with the MHC nomenclature are as follows; Bubu, Bubalus bubalis (Asian water buffalo); Ovar, Ovis aries (domestic sheep); Bomu, Bos mutus (Yak), Bota, Bos taurus cattle and Boin, Bos indicus cattle

Molecular modelling of the DQ molecules

Molecular models of the DQ1, DQ2 and DQ2-like molecule were generated using the oligomeric structure prediction option in Swiss-Model. The overall structure of DQ1, DQ2 and DQ2-like molecules are conserved, however, the amino acid deletion in the beta 1 domain of the DQβ2-like protein (Fig. 4a) is predicted to alter the alpha helical structure of the peptide binding groove compared with the same region of the DQ1 and DQ2 molecules. The influence of the deletion on the helical structure that makes up the side of the peptide binding groove is shown in Fig. 4b. The model predicts that the deletion exaggerates the twist in the helical structure resulting in a conformation change at this location compared with DQ1 (Fig. 4c) and DQ2 molecules (Fig. 4d).
Fig. 4

The location of the amino acid deletion in the β1 domain of the DQβ2-like protein is shown in comparison to the HLA-DQβ modelling template (4a). The position of the deletion in the model of the DQ2-like protein is shown in blue (4b). Associated conformational changes in the alpha helical structure of the peptide binding region of the DQ2-like molecule (4b) compared with DQ1 (4c) and DQ2 (4d) molecules is shown

Identification of functional DQA and DQB combinations

Eukaryotic expression constructs were generated for five DQB (two DQB1, two DQB2 and one DQB2-like) transcripts and five DQA transcripts (two DQA1, two DQA2 and one DQA2-like) previously isolated from the same haplotypes (Ballingall et al. 2015). The functional potential of each construct was assessed by co-transfection of different A/B gene combinations. The efficiency of each combination to express a class II molecule at the cell surface was determined by flow cytometry. The flow cytometry data are summarised in Table 5. Efficient surface expression of intra- and inter-haplotype combinations of DQA1 and DQB1 genes was observed. DQ1 molecules were recognised by both pan class II and DQα chain specific monoclonal antibodies. Expression of intra- and inter-haplotype combinations of DQA2 and DQB2 was also observed indicating that both DQ1 and DQ2 combinations are functional in sheep. Surprisingly, the product of the intra-haplotype pair DQA2*0101 and DQB2*0101 was not detected with the pan class II antibody SW73.2 or when tested with another pan class II reagent, SBUII 49.1 (Ballingall et al. 1995, data not shown). However, efficient expression was detected with VPM 36 indicating that in some haplotypes the combination of DQα2 and DQβ2 chains changes the quaternary structure of the protein distorting the conformational epitope recognised by SW73.2 and SBUII 49.1. Substitution of DQB2*0101 with DQB2*0201 regained SW73.2 reactivity however substitution of DQA2*0101 with DQA2*0102 did not, indicating that B2 polymorphism influences the specificity of SW73.2.
Table 5

The staining intensity and percentage of transfected cells expressing class II MHC at the cell surface

DQA construct combinations

DQB construct combinations

 

DQA1*01:01

DQA1*02:01:02

DQA2*01:01

DQA2*01:02

DQA2-like*01:01

DRA*01:01

 

DQ surface expression levels (labelling with SW73.2/VPM 36)

DQB1*01:01

++++/++++

+/++++

−/−

+/++

−/−

NT

DQB1*03:01

++++/++++

++/++++

−/−

−/+

−/−

NT

DQB2*01:01

−/+++

−/+++

−/++++

+/+

−/−

NT

DQB2*02:01

−/+++

+/+++

+++/++++

++++/++++

−/−

NT

DQB2-like *01:01

−/+++

−/+++

−/+

−/+++

−/+++

NT

DRB1*05:01

NT

NT

NT

NT

NT

++++

The intensity and percentage of positive cells from between three and six repeats of each transfection is represented as follows; − no detectable expression, + very low intensity staining in the order of 2–5% of cells, ++ low to moderate intensity staining with expression of 5–15% of cells, +++ medium to high levels of high intensity staining with expression in 16–35% of cells, ++++ high intensity staining with expression in greater than 35% of cells. NT not tested

Following convention all gene names are shown in italics

Similarly, SW73.2 failed to detect the DQ2-like molecule, however surface expression was detected with VPM 36. The DQA2-like construct only co-expressed with the DQB2-like construct suggesting that these two genes may represent an A/B gene pair. In contrast, the DQB2-like construct co-expressed efficiently with both DQA1 and DQA2 from within and between haplotypes. The class II products of these combinations were only detected with the alpha chain specific antibody VPM 36. Surface expressions of DQA2/B1 combinations were generally inefficient with only low level expression with some combinations, similarly to DQA2/B2-like. Surface expression of DQA1/B2 combinations was more efficient with intermediate levels of expression, although this was detected with VPM 36 only. Other inter-locus combinations were also observed at reduced efficiency.

Discussion

The sheep DQ sub-region is characterised by a duplication of the A and B loci resulting in haplotypes which contain both DQA1/B1 and DQA2/B2 genes (Scott et al. 1987, 1991a, b; Wright and Ballingall 1994; Ballingall et al. 2015). Unlike the duplicated DQ genes within the human MHC (HLA), the DQA/B gene pairs in sheep are both polymorphic, both are transcribed and as shown here, both appear functional.

The ovine DQ sub-region is further complicated by the identification of haplotypes that do not appear to include the DQA1 locus (Scott et al. 1991a; Fabb et al. 1993). In these haplotypes, DQA2 is accompanied by a second locus which through differential hybridisation (Scott et al. 1991a), sequence analysis of the second exon (Hickford et al. 2004) and associated phylogenetic analysis (Hickford et al. 2007), appears more closely related to DQA2 than to other DQA1 sequences. As such, these alleles were termed DQA2-like (Hickford et al. 2007) in order to differentiate them from DQA1 or DQA2. Our previous sequence and phylogenetic analysis extended these studies to include full-length DQA transcripts. This indicated that the DQA2-like allelic lineage was ancient, as conserved orthologues were identified in other members of the Bovidae family. However, the nature and function of the associated DQB loci in these haplotypes and the capacity of the various DQA and B loci to form functional class II proteins at the cell surface remained unknown.

To address these questions, we investigated the DQB loci at the genomic and transcript level in the same ten haplotypes from which the DQA diversity was described (Ballingall et al. 2015). We extended the functional analysis of the DQ loci by determining the capacity of different intra- or inter-haplotype combinations of DQA and DQB genes as well as isotype mismatched combinations (for example DQA1 and DQB2) to express class II MHC molecules at the surface of transfected cells. From the eight haplotypes previously identified with DQA1 and DQA2 loci, two transcripts corresponding to DQB1 and DQB2 were isolated. Haplotype 504a appears unusual as an additional divergent DQB2 transcript was identified (DQB2*05:01). A similar duplication of the DQB2 locus has recently been described by Ali et al. (2017). Unusual diversity at the closely linked class II DR loci has previously been identified in the 504a haplotype including the most divergent of the DRB1 alleles, DRB1*0901 (Ballingall et al. 2008b) and diversity at the usually monomorphic DRA locus (Ballingall et al. 2010).

The two DQA1 null haplotypes (501a and BWM) which included the divergent DQA2-like sequences also provided two equally divergent DQB transcripts which we termed DQB2-like. Phylogenetic analysis also clustered them independently of DQB1 and B2 alleles. Previous sequence analysis of cosmid and BAC clones indicated that the DQA1/B1 and DQA2/B2 loci are arranged as closely linked pairs (Wright and Ballingall 1994; Herrmann-Hoesing et al. 2008). While we have yet to obtain physical evidence that DQA2-like and B2-like form a closely linked A/B gene pair, as DQA2-like only co-expressed with DQB2-like in our co-transfection studies this may prove to be the case.

Comparative analysis of the DQB2-like allelic family indicated that they are highly conserved across the Bovidae family. Orthologous alleles were identified which shared the codon deletion within the second exon, identical signal sequence splice site polymorphism and conserved amino acid and 3’UTR motifs which indicates conservation of the allelic lineage over the 20 million years since these species shared a common ancestor (Matthee and Davis 2001). The shared pattern of conservation of both DQA2-like and DQB2-like supports the co-evolution of a closely linked A/B gene pair. Trans-species conservation has been a recognised feature of MHC allelic lineages for over 30 years (Klein 1987). The conservation of the DQ2-like genes over such a long period of time implies strong selective pressure which may be linked to the structural changes to the peptide binding domain evident from the modelling of the DQ2-like molecule. Forrest et al. (2010) failed to identify an association between the presence of DQA1-null haplotypes and variation in faecal egg counts suggesting that the response to gastrointestinal parasites is unlikely to provide the necessary selective pressure required to maintain the DQ2-like genes in ruminants.

We have demonstrated that both DQ1 and DQ2 pairs are able to form class II molecules at the surface of transfected cells. Similar co-transfection studies in cattle reported expression of DQ1 and DQ2 molecules and demonstrated the ability of both to present peptide antigen to CD4+ T cell lines (Norimine and Brown 2005). The pattern of surface expression described for sheep with intra- and inter-haplotype and DQ isotype mismatched A/B gene pairs appears broadly consistent with that reported in cattle (Norimine and Brown 2005), humans (Kwok et al. 1993) and mice (Braunstein and Germain 1987), where haplotype matched co-expression is efficient, while haplotype mismatched and isotype mismatched expression is dependent on the functional constraints between different pairs of DQA and B alleles.

In our sheep, DQB1 genes were restricted to surface expression with DQA1 genes within a haplotype and across haplotypes. In contrast, the DQB2 genes expressed efficiently with DQA2 genes as well as all DQA1 genes tested, albeit with a lower efficiency. The DQA2-like transcript only co-expressed with the DQB2-like transcripts while the DQB2-like transcript appears to be under less functional constraint as it also co-expressed efficiently with both DQA1 and DQA2. Isotype mismatched A/B gene combinations were generally less efficient than isotype matched. These experiments demonstrated the potential for a wide range of DQ molecules derived from intra- and inter-haplotype and inter-locus combinations of DQA and B genes. If these observations are translated in vivo, the number of MHC class II DQ molecules available for presentation of antigen to CD4+ T cells may vary considerably from a maximum of four in an MHC homozygous animal to potentially up to 16 in heterozygous animals, through both intra and inter-haplotype and isotype mismatched combinations. Such a large variation in potential class II MHC DQ molecules may significantly alter the range of antigens presented during the selection of the peripheral T cell repertoire (Klein et al. 2014).

With the increase in full-length transcript data, a number of DQB1 and DQB2 locus specific nucleotide and corresponding amino acid motifs are becoming apparent throughout the allelic families. Many such motifs are located towards the ends of the β1 domain such as the signal sequence splice site and the (HYD)APFSW motif. The 3’untranslated region also appears to be rich in locus specific sequence diversity. Insertions and deletions (indels) present in the immediate 3’UTR appear to differentiate between the allelic families represented in this study. With additional full-length sequence data becoming available from a number of sheep genome and transcriptome studies, the validity of each of these motifs for differentiating between loci will become clearer.

The physical location of the DQ2-like genes relative to the DQ2 A/B gene pair has yet to be resolved. The DQ2-like genes appear only in DQA1 null haplotypes suggesting that they may represent an ancient allelic lineage of DQA1 and B1 genes. This is not supported by our sequence and phylogenetic analysis which estimates that DQB2-like sequences have arisen from an ancient duplication of the DQB2 locus rather than DQ1. As reported for the MHC class I genes in cattle and sheep (Ellis et al. 1999; Miltiadou et al. 2005) locus specific substitutions accumulate in the 3’UTR. This also appears true for the DQB2-like sequences suggesting that they may not be unusual DQB1 alleles. These questions will only be fully resolved with the genomic sequence of the MHC region from a DQA1-null animal.

In summary, atypical DQB transcripts (DQB2-like) were identified in haplotypes classified as DQA1-null. Conserved DQB2-like orthologues were identified in other Bovidae indicating trans-species conservation of the allelic lineage for over 20 million years. Functional combinations of DQ1, DQ2 and DQ2-like genes demonstrates the potential for a wide range of DQ molecules derived from both intra- and inter-haplotype as well as inter-locus combinations. This study provides evidence that DQA2-like and B2-like genes form an evolutionary conserved pair which generate structurally distinct class II molecules that are likely to present a distinct range of peptides to CD4+ T cells.

Notes

Acknowledgements

The support of the Bioservices Division at the Moredun Research Institute is gratefully acknowledged along with funding from the Scottish Government. The research leading to these results has also received funding from the European Community’s Seventh Framework Programme (FP7, 2007-2013), Research Infrastructures action, under the grant agreement No. FP7-228394 (NADIR).

Supplementary material

251_2017_1029_MOESM1_ESM.docx (61 kb)
Supplementary Figure 1 Pair-wise alignment of DQB transcripts identified in this study along with all full-length sheep DQB transcripts available in GenBank. DQB1*02:01 and DQB2*04:01, EU176819; DQB1*08:01, XM_012173129 and DQB2-like*02:02 XM_012159546, both Mouflon, (Ovis aries musimon RefSeq Genome sequencing project); DQB1*05:01 (FJ985876). DQB2*07:01 (L08792). Representative cattle, (BoLA-DQB1, BC102959) and (BoLA-DQB, D37954), Buffalo, Bubalus bubalis, (Bubu-DQB, DQ908903) and Yak, Bos mutus (Bomu-DQB, XM_005909262). The translation initiation and termination codons are represented by ### and individual exons are labelled and sequentially shaded. Missing data and indels are represented by – . (DOCX 61 kb).

References

  1. Ali AO, Stear A, Fairlie-Clarke K, Brujeni GN, Isa NM, Salisi MS, Donskow-Łysoniewska K, Groth D, Buitkamp J, Stear MJ (2017) The genetic architecture of the MHC class II region in British Texel sheep. Immunogenetics 69:157–163CrossRefPubMedGoogle Scholar
  2. Arnold K, Bordoli L, Kopp J, Schwede T (2006) The SWISS-MODEL workspace: a web-based environment for protein structure homology modelling. Bioinformatics 22:195–201CrossRefPubMedGoogle Scholar
  3. Ballingall KT, Wright H, Redmond J, Dutia BM, Hopkins J, Lang J, Deverson EV, Howard JC, Puri N, Haig D (1992) Expression and characterization of ovine major histocompatibility complex class II (OLA-DR) genes. Anim Genet 23:347–359CrossRefPubMedGoogle Scholar
  4. Ballingall KT, Dutia BM, Hopkins J, Wright H (1995) Analysis of the fine specificities of sheep major histocompatibility complex class II-specific monoclonal antibodies using mouse L-cell transfectants. Anim Genet 26:79–84CrossRefPubMedGoogle Scholar
  5. Ballingall KT, Miltiadou D, Chai ZW, McLean K, Rocchi M, Yaga R, McKeever DJ (2008a) Genetic and proteomic analysis of the MHC class I repertoire from four ovine haplotypes. Immunogenetics 60:177–184CrossRefPubMedGoogle Scholar
  6. Ballingall KT, Fardoe K, McKeever DJ (2008b) Genomic organisation and allelic diversity within coding and non-coding regions of the Ovar-DRB1 locus. Immunogenetics 60:95–103CrossRefPubMedGoogle Scholar
  7. Ballingall KT, Tassi R (2010) Sequence-based genotyping of the sheep MHC class II DRB1 locus. Immunogenetics 62:31–39CrossRefPubMedGoogle Scholar
  8. Ballingall KT, Rocchi MS, McKeever DJ, Wright F (2010) Trans-species polymorphism and selection in the MHC class II DRA genes of domestic sheep. PLoS One 5:e11402CrossRefPubMedPubMedCentralGoogle Scholar
  9. Ballingall KT, Steele P, Lantier I, Cotelli M, Todd H, Lopez G, Martin E, Lantier F (2015) An ancient inter-locus recombination increases class II MHC DQA diversity in sheep and other bovidae. Anim Genet 46:333–336CrossRefPubMedGoogle Scholar
  10. Biasini M, Bienert S, Waterhouse A, Arnold K, Studer G, Schmidt T, Florian Kiefer F, Cassarino TG, Bertoni M, Bordoli L, Schwede T (2014) SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary information. Nuc Acid Res 42(Web Server issue):W252–W258.  https://doi.org/10.1093/nar/gku340
  11. Braunstein NS, Germain RN (1987) Allele-specific control of Ia molecule surface expression and conformation: implications for a general model of Ia structure-function relationships. Proc Natl Acad Sci U S A 84:2921–2925CrossRefPubMedPubMedCentralGoogle Scholar
  12. Brown JH, Jardetzky TS, Gorga JC, Stern LJ, Urban RG, et al. (1993) Three-dimensional structure of the human class II histocompatibility antigen HLA-DR1. Nature 364:33–39Google Scholar
  13. Ellis SA, Holmes EC, Staines KA, Smith KB, Stear MJ, McKeever DJ, MacHugh ND, Morrison WI (1999) Variation in the number of expressed MHC genes in different cattle class I haplotypes. Immunogenetics 50:319–328CrossRefPubMedGoogle Scholar
  14. Fabb SA, Maddox JF, Gogolin-Ewens KJ, Baker L, Wu MJ, Brandon MR (1993) Isolation, characterization and evolution of ovine major histocompatibility complex class II DRA and DQA genes. Anim Genet 24:249–255CrossRefPubMedGoogle Scholar
  15. Forrest RH, Zhou H, Fang Q, Smyth A, Frampton CM, Hickford JG (2010) No evidence for a universal association between variation in faecal egg count for a mixed field-challenge of gastrointestinal parasites and the presence of the Ovar-DQA1 null haplotype in sheep. Vet Immunol Immunopathol 135:303–305CrossRefPubMedGoogle Scholar
  16. Gao J, Liu K, Liu H, Blair HT, Li G, Chen C, Tan P, Ma RZ (2010) A complete DNA sequence map of the ovine major histocompatibility complex. BMC Genomics 11:466.  https://doi.org/10.1186/1471-2164-11-466 CrossRefPubMedPubMedCentralGoogle Scholar
  17. Grain F, Nain MC, Labonne MP, Lantier F, Lechopier P, Gebuhrer L, Asso J, Maddox J, Betuel H (1993) Restriction fragment length polymorphism of DQB and DRB class II genes of the ovine major histocompatibility complex. Anim Genet 24:377–384CrossRefPubMedGoogle Scholar
  18. Hasegawa M, Kishino H, Yano T (1985) Dating of human-ape splitting by a molecular clock of mitochondrial DNA. J Mol Evol 22:160–174CrossRefPubMedGoogle Scholar
  19. Herrmann-Hoesing LM, White SN, Kappmeyer LS, Herndon DR, Knowles DP (2008) Genomic analysis of Ovis aries (Ovar) MHC class IIa loci. Immunogenetics 60:167–176CrossRefPubMedGoogle Scholar
  20. Hickford JGH, Zhou H, Slow S, Fang Q (2004) Diversity of the ovine DQA2 gene. J Anim Sci 82:1553–1563CrossRefPubMedGoogle Scholar
  21. Hickford JGH, Zhou H, Fang Q (2007) Haplotype analysis of the DQA genes in sheep: evidence supporting recombination between the loci. J Anim Sci 85:577–582CrossRefPubMedGoogle Scholar
  22. Hopkins J, Dutia BM, McConnell I (1986) Monoclonal antibodies to sheep lymphocytes. I. Identification of MHC class II molecules on lymphoid tissue and changes in the level of class II expression on lymph-borne cells following antigen stimulation in vivo. Immunol 59:433–438Google Scholar
  23. Kalyaanamoorthy S, Minh BQ, Wong TFK, von Haeseler A, Jermiin LS (2017) ModelFinder: fast model selection for accurate phylogenetic estimates. Nat Methods 14:587–589CrossRefPubMedPubMedCentralGoogle Scholar
  24. Klein J (1987) Origin of major histocompatibility complex polymorphism: the trans-species hypothesis. Hum Immunol 19:155–162CrossRefPubMedGoogle Scholar
  25. Klein L, Kyewski B, Allen PM, Hogquist KA (2014) Positive and negative selection of the T cell repertoire: what thymocytes see (and don’t see). Nat Rev Immunol 14:377–391CrossRefPubMedPubMedCentralGoogle Scholar
  26. Kimura M (1980) A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J Mol Evol 16:111–120CrossRefPubMedGoogle Scholar
  27. Kwok WW, Kovats S, Thurtle P, Nepom GT (1993) HLA-DQ allelic polymorphisms constrain patterns of class II heterodimer formation. J Immunol 150:2263–2272PubMedGoogle Scholar
  28. Marsh SGE, Albert ED, Bodmer WF, Bontrop RE, Dupont B, Erlich HA, Fernandez-Vina M, Geraghty DE, Holdsworth R, Hurley CK et al (2010) Nomenclature for factors of the HLA system, 2010. Tissue Antigens 75:291–455CrossRefPubMedPubMedCentralGoogle Scholar
  29. Matthee CA, Davis SK (2001) Molecular insights into the evolution of the family Bovidae: a nuclear DNA perspective. Mol Biol Evol 18:1220–1230CrossRefPubMedGoogle Scholar
  30. Miltiadou D, Ballingall KT, Ellis SA, Russell GC, McKeever DJ (2005) Haplotype characterization of transcribed ovine major histocompatibility complex (MHC) class I genes. Immunogenetics 57:499–509CrossRefPubMedGoogle Scholar
  31. Minh BQ, Nguyen MAT, von Haeseler A (2013) Ultrafast approximation for phylogenetic bootstrap. Mol Biol Evol 30:1188–1195CrossRefPubMedPubMedCentralGoogle Scholar
  32. Norimine J, Brown WC (2005) Intrahaplotype and interhaplotype pairing of bovine leukocyte antigen DQA and DQB molecules generate functional DQ molecules important for priming CD4+ T-lymphocyte responses. Immunogenetics 57:750–762CrossRefPubMedGoogle Scholar
  33. Parham P, Ohta T (1996) Population biology of antigen presentation by MHC class I molecules. Science 272:67–74CrossRefPubMedGoogle Scholar
  34. Petersen J, Kooy-Winkelaar Y, Loh KL, Tran M, van Bergen J, Koning F, Rossjohn J, Reid HH (2016) Diverse T cell receptor gene usage in HLA-DQ8-associated celiac disease converges into a consensus binding solution. Structure 24:1643–1657CrossRefPubMedGoogle Scholar
  35. Scott PC, Gogolin-Ewens KJ, Adams TE, Brandon MR (1991a) Nucleotide sequence, polymorphism and evolution of ovine MHC class II DQA genes. Immunogenetics 34:69–79PubMedGoogle Scholar
  36. Scott PC, Maddox JF, Gogolin-Ewens KJ, Brandon MR (1991b) The nucleotide sequence and evolution of ovine MHC class II B genes: DQB and DRB. Immunogenetics 34:80–87PubMedGoogle Scholar
  37. Scott PC, Choi C-L, Brandon MR (1987) Genetic organization of the ovine MHC class II region. Immunogenetics 25:116–122CrossRefPubMedGoogle Scholar
  38. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S (2013) MEGA6: molecular evolutionary genetics analysis version 6.0. Mol Biol Evol 30:2725–2729CrossRefPubMedPubMedCentralGoogle Scholar
  39. Trifinopoulos J, Nguyen L-T, von Haeseler A, Minh BQ (2016) W-IQ-TREE: a fast online phylogenetic tool for maximum likelihood analysis. Nucleic Acids Res 44:W232–W235CrossRefPubMedPubMedCentralGoogle Scholar
  40. Trowsdale J (2011) The MHC, disease and selection. Immunol Lett 137:1–8CrossRefPubMedGoogle Scholar
  41. van Oorschot RA, Maddox JF, Adams LJ, Fabb SA (1994) Characterization and evolution of ovine MHC class II DQB sequence polymorphism. Anim Genet 25:417–424CrossRefPubMedGoogle Scholar
  42. Viginier B, Dolmazon C, Lantier I, Lantier F, Archer F, Leroux C, Terzian C (2012) Copy number variation and differential expression of a protective endogenous retrovirus in sheep. PLoS One 7:e41965CrossRefPubMedPubMedCentralGoogle Scholar
  43. Wright H, Ballingall KT (1994) Mapping and characterization of the DQ subregion of the ovine MHC. Anim Genet 25:243–249CrossRefPubMedGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany 2017

Authors and Affiliations

  • Keith T. Ballingall
    • 1
  • Isabelle Lantier
    • 2
  • Helen Todd
    • 1
  • Frederic Lantier
    • 2
  • Mara Rocchi
    • 1
  1. 1.Moredun Research InstituteMidlothianUK
  2. 2.INRA—Centre Val de Loire, UMR 1282, Infectiologie et Santé PubliqueNouzillyFrance

Personalised recommendations