Published online by Cambridge University Press: 08 June 2005
A potential control strategy for nematode infection in sheep is the implementation of a breeding programme to select for genes associated with resistance. The Texel breed is more resistant to gastrointestinal nematode infection than the Suffolk breed, based on faecal egg count, and this difference should enable the identification of some of the genes responsible for resistance. The objective of this study was to determine if variation at the ovine MHC-DRB1 locus was associated with variation in faecal egg count in Suffolk and Texel sheep. Ovar-DRB1 alleles and faecal egg count were determined for Texel (n=105) and Suffolk (n=71) lambs. Eight Ovar-DRB1 alleles, including 1 previously unknown allele, were identified in the Texel breed by sequence-base-typing. Seven Ovar-DRB1 alleles were identified in the Suffolk breed. Two Ovar-DRB1 alleles were common to both breeds, but were among the least frequent in the Suffolk population. In the Suffolk breed 1 Ovar-DRB1 allele was associated with a decrease in faecal egg count and 2 alleles with an increase in faecal egg count. This locus accounted for 14% of the natural variation in faecal egg count in Suffolks. There was no evidence for an association between Ovar-DRB1 alleles and faecal egg count in the Texel breed and the Ovar-DRB1 locus accounted for only 3% of the phenotypic variation in faecal egg count. These results suggest that the Ovar-DRB1 gene plays an important role in resistance to nematode infection in the Suffolk breed. The difference in faecal egg counts between these breeds may be attributable in part to the different allele profile at the Ovar-DRB1 locus.
Nematode species of the genus Teladorsagia (previously Ostertagia), Trichostrongylus and Nematodirus are largely responsible for gastrointestinal parasitic infections in sheep in Ireland (Good et al. 2004). The search for control measures to supplement anthelmintics in controlling these endoparasites in sheep has involved a number of different approaches. These include the use of tanniferous forages (Niezen et al. 1998), protein supplementation (Datta et al. 1999; Coop and Kyriazakis, 2001), the development of nematode vaccines (Andrews et al. 1995; Andrews, Rolph and Munn, 1997; Knox et al. 2003), the use of nematophagus fungi to reduce larval numbers on pasture (Githigia et al. 1997; Larsen et al. 1998; Pena et al. 2002) and the replacement of anthelmintic- resistant nematodes with those of a susceptible genotype (Bird et al. 2001). However, the proven success of genetic selection for nematode resistance based on faecal egg count (FEC) has highlighted the potential of this method of nematode control (Eady, Woolaston and Barger, 2003; Kahn et al. 2003). Selective breeding for resistance diminishes the impact of infection on production, lowers the requirement for anthelmintic control and reduces larval pasture contamination (Bisset et al. 1997). Identifying loci associated with natural resistance would facilitate selection of resistant animals and aid the understanding of host-parasite interactions. Genes located within the ovine major histocompatibility complex (MHC) class I and II regions are strong candidates for resistance because of their central role in antigen presentation to T cells and their associations with nematode resistance (Hulme et al. 1991; Buitkamp et al. 1996; Outteridge et al. 1996; Stear et al. 1996). The MHC class I and II molecules are membrane-bound glycoproteins that bind foreign antigen in an antigenic cleft for display to T cells. Exon 2 of the Ovar-DRB1 gene codes for part of the MHC class II antigen binding cleft and in sheep over 80 alleles have been identified at this locus. In the Scottish Blackface breed one exon 2 allele has been associated with a significant reduction in FEC (Schwaiger et al. 1995). A number of different methods have been published for Ovar-DRB1 genotyping including direct cloning followed by sequencing (Konnai et al. 2003a), immunoprinting (Schwaiger et al. 1993), PCR-Restriction Fragment Length Polymorphism (Blattman et al. 1993; Grain et al. 1993; Konnai et al. 2003b) and Single Strand Conformational Polymorphism (Outteridge et al. 1996; Kostia et al. 1998).
Studies in Ireland have shown that the Texel breed is more resistant to nematode infection than co-grazed Suffolks, based on FEC measurements and worm burdens (Hanrahan and Crowley, 1999; Good et al. 2004). As part of a project to explain the basis for this genetic difference in nematode resistance in sheep, the allelic variation at exon 2 of the Ovar-DRB1 locus was assessed in order to determine the role of this gene in genetic resistance to nematode infection.
A total number of 155 Suffolk and 179 Texel purebred lambs were available for the study. These lambs represented 4 annual lamb crops (1999–2002). Texel lambs were derived from 11 sires and 50 different dams, Suffolk lambs were derived from 11 sires and 43 different dams. Sires were not used to produce the dams. All lambs were born in early March each year and co-grazed on pasture from birth to weaning. A single anthelmintic treatment (levamisole or benzimadazole, depending on the year) was administered at 5 weeks of age for the control of Nematodirus. Lambs were weaned at 17 weeks of age.
Faecal samples were obtained from the rectum of Suffolk (n=120) and Texel (n=148) lambs at 17 weeks of age and a 3 g aliquot of each faecal sample was processed to determine the number of worm eggs per gram of faeces (FEC) using the modified McMaster method (MAFF, 1986). Eggs of Nematodirus spp. were enumerated separately from those of other trichostrongyle species. As numbers of Nematodirus were extremely low they were not included in the analysis.
Blood samples were collected in EDTA vacutainers from lambs born from 2000 to 2002. DNA was extracted from blood using a lysis buffer and proteinase-K digestion technique. For the 1999-born lambs, a ground sample of meat was the only source of DNA but the poor quality of this DNA meant that genotype analysis was unreliable for this gene. Consequently Ovar-DRB1 genotype could only be established for 87 Suffolks and 111 Texels.
Sequence-base-typing, validated by cloning, was the method of genotyping employed in this study. Sequence-base-typing is a technique whereby each Ovar-DRB1 gene was amplified by PCR, sequenced, and alleles identified from the chromatogram. This was the technique of choice because it provided a sequence of the exon for each animal, thereby highlighting any potential single-base mutations which may be missed by other methods of analysis.
Approximately 100 ng of DNA was amplified in a 50 μl polymerase chain reaction (PCR) with 2·5 U Taq polymerase (Promega Southampton, UK). Amplification was carried out for 30 cycles in a Peltier Thermal Cycler (MJ Research, Massachusetts, USA). A cycle consisted of a 30 s denaturation at 94 °C, a 30 s annealing at 57 °C and a 1 min extension at 72 °C. A forward primer (5′-TCTCTGCAGCACATTTCCTGG-3′) (adopted from Schwaiger et al. 1995) and reverse primer (5′-CACACACACACTGCTCCACA-3′) (in-house designed primer) pair were employed to amplify the second exon of the Ovar-DRB1 gene, yielding a 308 base-pair product. These primers were located in intron 1 and 3 respectively thereby flanking the exon 2 region. PCR products were cleaned using Microcon® YM-100 centrifugal filter devices (Millipore, Massachusetts, USA) and DNA concentration determined by comparison with a low mass DNA ladder (Invitrogen, Life Technologies, California, USA). Approximately 8 ng PCR product was sequenced using the reverse primer for the Ovar-DRB1 gene with the ABI Prism® BigDye™ Terminator reaction kit version 2. Sequencing reactions were run on an ABI Prism® 377 DNA sequencer.
Sequences were viewed using Chromas software version 1.45 (Technelysium Ltd, Queensland, Australia). The identity of each Ovar-DRB1 allele was determined using custom written software (SMS Software, T. Bellenger). This software compared the query sequence to an in-house generated library of ovine DRB1 alleles and identified the alleles for which it is an exact match. This library contained 88 different allele sequences available from Genbank. The operating principle is similar to that of ‘BLAST’, except the library is self-generated, it can accommodate heterozygous sequences and does not offer a ‘best fit’ scenario but an exact match only. Allele identification from sequences resulting from cloning or from homozygous animals was clear-cut. Where the sequence obtained represented 2 different alleles it was first edited and heterozygous positions coded using the International Union of Biochemistry nucleotide translation conventions (http://www.chem.qmul.ac.uk/iupac/): M=A+C, S=C+G, R=A+G, Y=C+T, W=A+T, K=G+T. This convention is recognized by the software, which then generates all possible alleles from the heterozygous query sequence. The 2 alleles that are an exact match for the query sequence are identified. With approximately 20 potential polymorphic positions within a heterozygous sequence, the chance of 2 sets of alleles matching the query sequence is highly unlikely and did not occur. If it did, the programme would offer both sets.
In order to confirm the reliability of this procedure 3 Suffolk and 3 Texel sires, those with the largest progeny numbers, had their Ovar-DRB1 gene amplified by PCR and cloned using the TOPO TA Cloning® kit (Invitrogen Life Technologies, California, USA). To amplify the insert, PCR was carried out on DNA extracted from growing cultures using the M13 primer set. Sequencing of the insert was completed using the M13 forward and reverse primers.
The FEC values were transformed to ln (FEC+25) prior to analysis and only the FEC values based on eggs from Trichostrongyles (excluding Nematodirus species) were considered. As numbers of Nematodirus were low in both breeds, this species were not considered further. Preliminary analyses were done using Proc GLM of SAS (SAS, 1989) to fit a linear model with effects for breed, year of birth (1999–2002), sample date within year, rearing type (single or twin) and sex and breed by year interaction. A full animal model, with a pedigree file going back to the foundation of the flocks in the mid-1970s, was used to estimate the effects of the Ovar-DRB1 alleles on FEC. The allele effects were evaluated by adding regression terms to the model for each allele except the most common allele in each breed. This analysis was carried on a within breed basis, because of the difference in allele profile between the breeds, and was done using the PEST software (Groenveld, 1990). All available FEC data were included and a heritability of 0·25 was used in both breeds. In total, 71 Suffolks and 105 Texels were used in the analysis as these animals had both genotype information and FEC measurements available. The difference among the regression coefficients obtained for the alleles were evaluated using Duncan's multiple range test (Steele and Torrie, 1960).
Mean FEC for all Suffolk and Texel lambs are presented in Table 1, which show that Suffolk lambs had significantly (P<0·001) higher egg output than Texel lambs. The statistical analysis also showed that year and sampling date had highly significant effects (P<0·001 in both cases) and that the rearing type effect approached significance (P=0·07). There was no evidence of a significant year by breed interaction. The effect for sex did not approach significance and was not included in the mixed models used to estimate the allele effects.
Sequence-based-typing of lambs yielded a total of 8 alleles in Texels and 7 in Suffolks. The frequency and protein translation of each allele is shown in Figs 1 and 2 respectively. Only 2 alleles were common to both breeds and these were among the least frequent alleles in Suffolks. One previously unidentified allele was found in the Texel breed (Ovar-DRB1*GS1, Genbank Accession no. AY227049).
The sequencing was initially performed in 2 directions in all the sires in which the Ovar-DRB1 gene was cloned. In the 3 Texel sires investigated 5 alleles were identified (Ovar-DRB1*0109, Ovar-DRB1*0203, Ovar-DRB29, Ovar-DRB10, Ovar-DRB1*GS1). In the 3 Suffolk sires investigated 3 alleles were identified (Ovar-DRB33, OAMHC213, Ovar-DRB1*03411). The results from cloning confirmed the validity of the Ovar-DRB1 sequence-base-typing analysis method. Thus, sequencing was only carried out in 1 direction for all of the offspring. Without exception, 1 of the alleles carried by a sire was identified in each of its offspring. To re-confirm that the correct allele was being identified, linkage to an adjacent intronic microsatellite was also investigated, a method previously described by Schwaiger et al. (1995). Again, allele designation was consistent in all cases (results not shown).
The estimated effects of allele substitution are given in Tables 2 and 3 for Suffolks and Texels, respectively. The effect for Ovar-DRB1*0203 just failed to reach formal statistical significance (P=0·052) while those for Ovar-DRB33 and OAMHC213 were associated with P-values of about 0·12. The substitution of the most common allele in the Suffolk breed, Ovar-DRB1*03411, with Ovar-DRB1*0203 led to a large decrease (−0·81) in log-transformed FEC, relative to the most common allele. On the untransformed scale this represents a 45% reduction while the substitution of Ovar-DRB1*03411 with Ovar-DRB33 reduced FEC by approximately 30% on the untransformed scale. The substitution of Ovar-DRB1*03411 by either OAMHC213 or Ovar-DRB10 increased FEC by about 25%. The substitution effects for OAMHC205 was not analysed because only 1 copy of this allele was present in the sample. Evaluation of the differences among the allele effects showed that the effects of Ovar-DRB1*0203 and Ovar-DRB33 were both significantly different from the effect of OAMHC213 (P<0·05).
In Texels none of the allele substitution effects, relative to the most common allele Ovar-DRB1*GS1, were statistically significant.
There was no significant effect for heterozygosity in either breed. In Texels 14% of animals were homozygous and had a transformed FEC of 6·20±0·325, the corresponding figure for heterozygous Texels was 5·57±0·163. In Suffolks, 35% of animals were homozygous and had transformed FEC of 6·71±0·271, heterozygous Suffolks had a mean transformed FEC of 6·72±0·187.
Variation at the Ovar-DRB1 locus accounted for 14% of the phenotypic variation in Suffolks (after adjusting for the fixed effects of year, sample date and rearing type). The corresponding figure in Texels was 3%.
The populations of Suffolk and Texel sheep examined in this study exhibited low variation at the Ovar-DRB1 locus compared with results from other studies. Schwaiger et al. (1995) identified 19 alleles in a sample of 299 Scottish Blackface animals while Konnai et al. (2003a) found 28 alleles among 71 Suffolks, 14 alleles from 20 Cheviots and 9 alleles from 6 Corriedale sheep. While the Suffolk and Texel flocks used in the present study represent small self-contained lines the average inbreeding coefficients for the cohorts used were only 7·7% and 7·1%, respectively. Thus a small effective population size is unlikely to account for the relatively low number of alleles present. The slightly higher allelic variation in the Texel compared with the Suffolk is expected and is consistent with other findings (Walling et al. 2004).
It has been suggested that homozygosity at the Ovar-DRB1 locus may impair ability to respond effectively to diseases (Thursz et al. 1997; Sauermann et al. 2000; M. Stear, personal communication). However, in this study there was no evidence for any significant effect of heterozygosity at the Ovar-DRB1 locus on FEC (homozygotes: 6·46±0·227; heterozygotes 6·14±0·136, P=0·26).
In the Suffolk population examined here the Ovar-DRB1 gene accounted for 14% of the phenotypic variation in FEC and the evidence indicated that 3 alleles were largely responsible for this variation. This finding is consistent with disease association studies in the Scottish Blackface breed, where the Ovar-DRB1 gene accounted for 11% of the phenotypic variation in FEC and 1 Ovar-DRB1 allele was associated with a reduced faecal egg count (Schwaiger et al. 1995). It is interesting to note that the allele associated with reduced FEC in the present Suffolk population is the same allele as that associated with reduced FEC in the Scottish Blackface breed (M. Stear, personal communication).
The absence of an association between Ovar-DRB1 alleles and FEC in Texels is consistent with that of Blattman et al. (1993) who were unable to detect any evidence of an effect of the DQB or DRB regions of the MHC complex on resistance in the Merino breed. A number of explanations may be offered for the absence of evidence for an effect of the Ovar-DRB1 gene on FEC in Texels. These include linkage disequilibrium to a resistant locus that is established in Suffolks but not in Texels (linkage disequilibrium has been reported between MHC class I and II loci (Stear et al. 1996)) or favourable haplotypic combinations to combat infection which are present in Suffolks but not in Texels or age-related effects whereby the role of the Ovar-DRB1 allele is evident in Texels at ages other than 17 weeks (age-related effects previously reported by Schwaiger et al. 1995). However, it is most likely that since the Texel breed has a very different profile of alleles at the Ovar-DRB1 locus and the fact that the resistant allele in Suffolks is the second-most frequent in Texels, the majority of Ovar-DRB1 alleles in Texels are associated with resistance to nematodes. This would account for the observed breed difference in FEC and also explain why no one allele was associated with resistance in this breed.
The DRB1 exon 2 product forms part of the MHC class II antigen-binding cleft and amino acids in this region determine the size, form and charge of the cleft and, more importantly, the side-chains of foreign antigen that may be accommodated in the cleft (Stern et al. 1994; Thorsby, 1999). Thus, some Ovar-DRB1 alleles may be better suited to display antigens common to certain diseases and so generate a better immune response through a better T-cell response repertoire (Krieger et al. 1991; Coppin et al. 1993). It is argued that alleles associated with resistance, as shown in this study and by Schwaiger et al. (1995) possess the necessary amino acid combination to facilitate an effective immune challenge to combat nematode infection.
The evidence from this study together with the allele effects previously found in the Scottish Blackface breed support the hypothesis that Ovar-DRB1 gene plays an important role in the reduced resistance of Suffolk sheep to parasite infection. Further studies with larger animal numbers are desirable to confirm the present results in this breed and the utility of selection based on the genetic variation described. Specific mating plans will have to be implemented to directly establish the role of the Ovar-DRB1 gene in the large difference in FEC between Texel and Suffolk animals and, by implication, in resistance. The amino acid changes associated with the alleles shown to affect resistance support the view that these changes affect functionality of the MHC class II protein product.
We are grateful to Professor M. J. Stear for his continuous encouragement and advice. This work was supported by a grant from the Irish Department of Agriculture, Food and Rural Development (RSF16). The Wellcome Trust is acknowledged for an equipment grant (Ref. 061354).