| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
ANIMAL GENETICS |
* AgResearch Invermay, Puddle Alley, Private Bag 50034, Mosgiel, New Zealand
Health and Safety Compliance Office, University of Otago, Dunedin, New Zealand
 |
Abstract |
|---|
 Top Abstract INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONLITERATURE CITED |
|---|
Key Words: computed tomography • growth differentiation factor gene • lamb • meat yield • Texel • Viascan
|
INTRODUCTION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONLITERATURE CITED |
|---|
In Texel-based populations a QTL affecting lean meat yield maps to the region of the growth differentiation factor gene (GDF8 or Myostatin) on OAR2 (Broad et al., 2000
; Marcq et al., 2002; Laville et al., 2004; Johnson et al., 2005). Clop et al. (2006) identified a single G to A transition in the 3' untranslated region of GDF8 (g+6723G-A) as the probable causal polymorphism. This transition creates a target site for mir1 and mir206 microRNA, which cause translational inhibition of GDF8. The A allele is associated with lean meat yield in Australian Texels (Kijas et al., 2007) and British Charollais sheep (Hadjipavlou et al., 2008).
The GDF8 is a negative regulator of muscle growth, and any polymorphism that impairs function, results in an increase in skeletal muscle formation (Grobet et al., 1997; Kambadur et al., 1997; Marcq et al., 1998). The size of the effect of 1 copy of the QTL/SNP has been estimated to be between 0.2 and 1.5 of an SD unit, depending on the muscle and fat traits considered (Laville et al., 2004; Johnson et al., 2005; Kijas et al., 2007). The exact mode of inheritance (additive or nonadditive) has not been fully elucidated, nor has the impact of carcass weight or age on the relative size of the effect.
The aim of the present study was to use computed tomography (CT) and Viascan (a commercial 2-dimensional imaging system that estimates carcass lean content, Sastek, Brisbane, Australia; Hopkins et al., 2004) to 1) confirm the presence and effects of the A allele in New Zealand Texels; 2) determine the mode of inheritance by determining the effects of 0, 1, or 2 copies of the A allele; and 3) determine the impact of carcass weight and lamb age on the size of the effect.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONLITERATURE CITED |
|---|
Testing of the Presence of the A Allele in New Zealand Texels
The Texel sires from the study of Johnson et al. (2005) were sequenced for both the g+6723G-A and the g-2449C-G SNP described by Clop et al. (2006). Primers for the g+6723G-A SNP were derived from the bovine genome (forward: 5' TCCATATGCTAATGGT-TAGATGG3'; reverse: 5'TTGATTAACAAAATCCT-GATTTACA3'), whereas primers for the g-2449C-G SNP were those described by Clop et al. (2006). Subsequently, 1,009 animals, predominately known to be carrying at least 1 copy of the BULGE20-BM81124 microsatellite haplotype for the New Zealand Texel muscling QTL described by Johnson et al. (2005) were genotyped for the g+6723G-A SNP using a forced RFLP assay. The DNA was amplified using primers: forward, 5'GTTCGTGATGGCTGTATAACG 3'; reverse, 5'GTTAAATAAACTAATTGTTTTAGGACT 3', where the last 3 nucleotides of the forward primer create a HpyCH4IV cut site, which acts as an internal control to the same cut site surrounding the G allele at g+6723G-A SNP, using the following PCR protocol: 1 cycle (94°C for 3 min); 35 cycles (94°C for 30 s, 60°C for 30 s, 72°C degrees for 30 s) and 1 cycle (72°C for 4 min). The resulting PCR product (20 µL) was combined with 5.0 units of enzyme HpyCH4IV (NEB, Ipswich, MA; R0619) overnight at 37°C and then visualized after separation on a 3% agarose gel. Products of 240 base pairs and 189 base pairs were scored as A and G, respectively, where the restriction enzyme had successful digestions. Both the A and G alleles were detected within the New Zealand Texel population.
Animals and Traits
Two Texel-cross rams (not previously used in Johnson et al., 2005) identified as having 1 copy each of the A allele were single sire mated to 100 Texel-cross ewes, the majority of which were identified as having 1 copy of the A allele (balanced for ewe source and age). Only those ewes that conceived during the first 21-d cycle were included in the trial. These ewes and subsequently progeny were grazed together postmating until weaning. A total of 187 progeny were weaned from these matings at an average age of 75 d. The progeny were then genotyped for the A allele using the forced RFLP assay described above and were assigned into 1 of 4 slaughter groups. Each slaughter group was balanced for number of A alleles (0, 1, or 2), sex (Table 1), and sire.
|
).
All progeny were evaluated for growth, carcass dimension, and Viascan traits, and slaughter group 4 was also evaluated for CT traits. Body weights (kg) were recorded at birth, weaning, and before each of the 4 slaughter dates. Measurements recorded on the whole carcass were cold carcass weight, carcass length, leg length, circumference of the buttocks, soft tissue depth 110 mm off the mid-line in the region of the 12th rib (defined as GR by Kirton, 1989), and Viascan carcass measurements of lean yield, loin lean yield, shoulder lean yield, and total lean yield expressed as a percentage of carcass weight. Carcass length was measured from between the hind legs to the front of the neck using a set of sliding calipers with 50 mm-wide bars at each end; leg length was measured from the crotch to the end of the hind leg, which was cut through the tarsal joint. The circumference of the buttocks was measured using a flexible tape measure on the dressed carcasses hanging from their hindquarters and represented the circumference when taken in a parallel plane immediately above the anal opening. Dressing percentage was calculated by dividing the carcass weight by the preslaughter BW and multiplying by 100. The CT data were generated for carcass regions as described by Kvame et al. (2004), using a Siemens Somatom AR.C x-ray CT scanner (Erlangen, Germany). The x-ray exposures were set to 3 s at 130 kV and 70 mA. The field of view was 450 mm, and image thickness was set to 5 mm. Traits measured on the CT images of slaughter group 4 were estimates of lean, subcutaneous fat, inter-muscular fat, and bone weight for the shoulder, loin, and leg regions, plus total carcass at each time point. Measures of LM width, depth, area, and fat depth over the LM were also made on the CT images.
Statistical Analysis
Body weight and slaughter data were analyzed using the general linear model procedure (SAS Inst. Inc., Cary, NC). The models fitted included fixed effects of sire, sex (ewe or ram), slaughter group (1, 2, 3, or 4; not fitted for birth or weaning weight), and A allele status (0, 1, or 2 copies); carcass weight was fitted as a covariate for carcass length, leg length, buttock circumference, GR, and Viascan measurements. Interactions between all fixed effects and covariates were tested, but were not significant and were excluded from the final model.
The CT data were analyzed as repeated measures using the mixed model procedure of SAS. The same fixed effects were fitted as above, a covariate of carcass weight (estimated from the analysis of the CT images) was fitted for all traits excluding carcass weight, and a random effect of individual animal (1 observation per time point) was fitted. Interactions between all fixed effects and covariates were tested, but were not significant and were excluded from the final model.
Within the models fitted above, an estimate of the additive effect (a) of the A allele relative to the alternate G allele was obtained using the equation
|
RESULTS AND DISCUSSION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONLITERATURE CITED |
|---|
The 8 Texel sires described previously in Johnson et al. (2005) were genotyped for both transitions identified by Clop et al., 2006 (Table 2). Sire 150 (the sire common to all sire pedigrees used by Johnson et al., 2005) was found to be homozygous for the A allele at g+6723G-A and the C allele at g-2449C-G, as seen in the majority of Texels genotyped by Clop et al. (2006). Clop et al. (2006) reported a high level of linkage disequilibrium between these 2 SNP and had some difficulty assigning effects of g+6723G-A independently of g-2449C-G. Two exceptions to this pattern were seen when investigating the sires used by Johnson et al. (2005). Sires 1170 and 429 were segregating for muscling traits (from the work of Johnson et al. (2005), and these 2 sires were heterozygous at g+6723G-A, but were homozygous C at g-2449C-G. These results provide further supporting evidence that the A allele at g+6723G-A has an effect on muscling traits independent of alleles at g-2449C-G.
|
was tested by genotyping 1,009 animals that had previously been genotyped for the microsatellite haplotype test (Table 3). Eighty-five percent of animals genotyped were assigned the same status using the 2 tests, with 15% reclassified. Nine animals classified as carrying 2 copies by the haplotype test were shown to be single copy carriers only, and 1 animal classified as a single copy carrier by the haplotype test was shown to be a noncarrier by the SNP test. These results again confirm both the presence of the A allele within the New Zealand Texel population and that it is strongly associated with the microsatellite haplotype. The differences in assignment between the 2 tests (Table 3) indicates that, although the 2 microsatellite marker haplotype test is a good predictor of the underlying SNP, the microsatellite markers, which lie 1 cM apart and span GDF8 (and the GDF8 g+6723G-A SNP), are not in full linkage disequilibrium. In a wider context, the results emphasize care is needed with haplotype based tests, especially when they are used commercially. These results have no implications on the reported estimates from Johnson et al. (2005), as the genotypes of the sires used for the g+6723G-A SNP reflected the observed BM81124 and BULGE20 haplotypes (Table 2
). Additionally, the 9 progeny from the study of Johnson et al. (2005) with a CI haplotype from their dam were SNP tested and only 1 had the A allele at the g+6723G-A locus.
|
The least squares means for lambs having 0, 1, or 2 copies of the A allele for BW and carcass traits measured after slaughter are summarized in Table 4, and for traits assessed via CT in Table 5. No interactions initially fitted in these models were significant; the implications of this are discussed below.
|
|
This study supports the findings of Johnson et al. (2005) and Kijas et al. (2007), that the A allele does not affect any BW or growth traits. Although Walling et al. (2004) reported evidence for a QTL affecting growth on OAR2, it does not map to the same region as the GDF8 gene and is therefore likely due to a different mutation.
In general, the A allele resulted in greater values for muscle related traits, and decreased values for fat related traits as assessed on the carcass after slaughter or through measurements made on CT images. Other trials using different measures of muscling and fatness have shown similar trends (Laville et al., 2004; Johnson et al., 2005; Kijas et al., 2007). These results confirm that the A allele is associated with muscling and fat traits in New Zealand Texels and that the effect can be detected using several methods for estimating carcass muscling and fat. Of the traits directly comparable with other trials, measures on the LM from CT images showed evidence for an effect on LM width and total area, but not depth. This result is in contrast to other studies that have tended to show an effect on LM depth (versus width; Johnson et al., 2005; Kijas et al., 2007), although across all trials there was a total increase in LM area.
Estimates of the size of the effect are given in Table 6, with the additive effect ranging from 0.0 to 1.1 as a proportion of the residual SD, but consistently above 0.6 for most muscle and fat traits. These values are in broad agreement with estimates previously reported (Laville et al., 2004; Johnson et al., 2005; Kijas et al., 2007).
|
, in that most of the non-Texel genetics are derived from Romney-based breeds, whereas the non-Texel genetics of the trials of Kijas et al. (2007) and Laville et al. (2004) are Merino and Romanov, respectively. The non-Texel genetic backgrounds may have an impact on the size of the effect observed. Bunger et al. (2004) reported that the size of the effect of GDF8 polymorphisms after introgression in mice depended on the background genetics of the inbred lines. This is possibly due to other modifier loci, as Varga et al. (2005) mapped 6 loci in mice that exerted effects on muscling in the presence of the GDF8 polymorphism. Additionally, there have been reports of other QTL affecting muscling and fatness in Texel-cross sheep mapping to OAR2, approximately 15 to 20 cM downstream of GDF8 (Walling et al.,. 2004; Johnson et al., 2005). In some instances these QTL may be in linkage disequilibrium with the GDF8 SNP, and, therefore, estimates of the size of the effect may be inaccurate. Determining the Mode of Action of the g+6723G-A Transition
The mode of action is not clear when examining the least-squares means results (Tables 1 and 4). Some traits show significant differences between all 3 genotypes (0, 1, or 2 copies of the A allele), whereas for others the differences between 0 and 1 or 1 and 2 copies were not significantly different. However, in all cases where there was evidence for an effect, the apparent trend was the same in that the percentage increase/decrease from the base to 2 copies is greater than the increase/decrease from the base to 1 copy.
For the majority of muscle and fat traits, the estimates of the additive effect (a) were significant and followed the known direction of positive for muscle traits and negative for fat traits (Table 6). Although for the nonzero traits, none of the estimates of the de were significant (Table 6) and whether the results indicated a dominant or recessive mode of inheritance (same or opposite sign as the additive effect, respectively) varied within the muscle traits and the fat traits. Laville et al. (2004) similarly found that most estimates of de were not significant. Although Clop et al. (2006) and Hadjipavlou et al. (2008) concluded from their data that there is evidence for a recessive mode of action, Clop et al. (2006) did not provide SE or indications of significance to allow the estimates of de to be confirmed.
The study of Johnson et al. (2005) was based on a half-sib design, using sires heterozygous for the QTL and noncarrier dams, with evidence for the QTL coming from the comparison of the progeny carrying 1 or 0 copies of the allele. The detection of QTL effects in that study indicates that the QTL is not strongly recessive. The issue regarding the additive vs. nonadditive mode of action of the g+6723G-A transition, therefore, continues to remain unclear. However, the combined evidence available to date suggests that the benefits of the A allele are maximized in animals carrying 2 copies.
Size of the A Allele Effect Across Carcass Weights
Previous studies investigating the Texel muscling phenotype have investigated the phenotype in moderate to high BW and carcass weight animals. Carcass weights in the region of 16 to 18 kg are reported in the work of Laville et al. (2004) and Johnson et al. (2005), 22 kg in the work of Kijas et al. (2007), and BW of 51 kg in Hadjipavlou et al. (2008). There have been anecdotal reports that the effect of the A allele cannot be detected at lighter BW and carcass weights. Kijas et al. (2007) could not detect evidence for the effect of A on ultrasound measured LM depth when the animals were scanned at 7 and 9 mo of age, although the corresponding weights were not reported in the paper. However, this in itself cannot be taken as conclusive, because, as previously discussed, estimates based on ultrasound have not been consistent, and given that only LM depth was measured, they cannot conclude that it did not have an effect on the eye muscle size or weight.
This study set out to more accurately assess the impact of carcass weight on the ability to detect the effect of the A allele by slaughtering lambs at 4 different BW and serially CT scanning the fourth group at the 4 different BW. The interaction between A allele number (fixed effect) and carcass weight (covariate) for both the slaughter and CT traits was not significant, nor was the interaction between A allele number and slaughter group (fixed effect). The solutions for the carcass lean and carcass fat models from the CT scanning were used to plot the relationship between carcass weight and carcass lean/fat for animals carrying 0, 1, or 2 copies of the A allele (Figure 1). The relative differences between the A allele groups were consistent across the 4 carcass weights measured. This provides evidence that the effect of the A allele can be detected over a range of carcass weights (13 to 20 kg) and that the size of the effect across the weights is proportionately the same between the 3 genotypes.
|
|
Footnotes |
|---|
2 Corresponding author: tricia.johnson{at}agresearch.co.nz
Received for publication September 23, 2008. Accepted for publication February 24, 2009.
|
LITERATURE CITED |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONLITERATURE CITED |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), 1 (
), or 2 (
) copies of the A allele resulting from the g+6723G-A transition in the GDF8 gene.