Evaluation of gene markers for tenderness in the UK beef herd

R.I. Richardsona*, B.G. Lowmanb, N.C. Otterc, J.D., Nkrumahd, K. Suttone, K.M. Haywoodf.
a Division of Farm Animal Science, University of Bristol, Langford Bristol, BS40 5DU, UK, b Beef & Sheep Select, SAC, Greycrook, St Boswells, TD6 0EU, UK,  c Merial Animal Health, CM19 5TG, UK,  d Merial Limited, 3239 Satellite Blvd, Duluth GA 30096, USA.,  e Dovecote Park Ltd, Stapleton, Nr Pontefract. West Yorkshire. WF8 3DD, UK.,   fNational Beef Association, The Mart Centre, Tyne Green, Hexham, Northumberland, NE46 3SG, UK

*contact [email protected]

Summary

450 beef cattle were hair sampled and typed for a panel of gene markers and steaks taken from their 10d aged carcasses for measurement of colour, ultimate pH and cooked meat tenderness as measured by Warner-Bratzler shear force.  The distribution of the Igenity Tenderness 1-10 marker scores was similar to that seen in a large population of cattle in the USA.  There was a negative correlation between tenderness score and WBSF, although this was small and statistically significant at the 10% level.  Whilst the original test of the Igenity scores in the USA had shown a difference in WBSF of 1kg across scores 1-10 that measured on the UK cattle was 0.4kg.  This may have been due to superior management in the UK abattoir which included electrical stimulation, considerate chilling and meat ageing, thus producing a narrow range of meat tenderness.

Demonstration of profitable utilisation of gene markers in the UK beef herd
INTRODUCTION
The change of EU support away from direct headage payments to a Single Farm Payment has highlighted the high cost structure of UK beef producers.  Faced with cheaper imported beef, it is essential that UK producers maximise the attractiveness/quality of their product to significantly increase their returns.  The key component of this is eating quality and perhaps even more importantly consistent eating quality – the occasional bad eating experience having a disproportionately negative effect on consumer perception and hence the amount they are prepared to pay.
Genomic information technology can be employed in the management and selection of beef cattle to increase profitability and the rate of genetic progress. Recent advances in livestock genomics have resulted in the detection of numerous polymorphisms that show associations with a range of economically relevant traits in beef cattle and in particular, cooked meat tenderness.
Meat proteins are degraded post-mortem during an ageing (conditioning) period by a micromolar calcium-activated neutral protease, µ-calpain, which is regulated by an inhibitor calpastatin (Koohmaraie, 1996). These proteins are coded by genes CAPN1 and CAST respectively.  Gene mapping and discovery programmes have resulted in a number of QTL for beef traits and a number of single nucleotide polymorphisms (SNP) for tenderness have been described.  A C/G substitution at amino acid position 316 was described by Page et al. (2002) and a further C/T substitution at position 4751 (CAPN4751) by White et al. (2005). A number of SNP have also been described for CAST, one a C/T substitution (UoG-CAST, Schenkel et al., 2006) and one G/A substitution (CAST-T1, Barendese, 2002).  The two CAPN1 SNPs and CAST-T1 are used together in the GeneSTAR Tenderness panel and two CAPN1 SNPs and UoG-CAST in the Igenity TenerGENE marker panel (IGENITY TenderGENE ). As these involve 3 SNPs there are 27 potential combinations, the Igenity Tenderness score has been reduced to 10 groups based on validation experiments.  Individual associations between gene markers and cooked meat tenderness have been published by Page et al. (2004), White et al. (2005), Casas et al. (2006) and Morris et al. (2006) amongst others.  Van Enennaam et al. (2007) validated commercial DNA tests including GeneSTAR and Igenity Tenderness score on 685 and 490 animals respectively, finding a 1kg difference in WBSF between the most and least tender genotypes for both panels.
Igenity tenderness scores are presented on a 10 point scale with 1 being the least tender and 10 the most tender, summarised in the table to the left (Merial, personal communication).  The table also shows that there is a 1kg difference in Warner Bratzler shear force between a score of 1 and a score of 10.  A score of 2 does not exist because of the lack or rarity of these gene combinations. 

Similar American studies have also investigated the links between gene markers for carcass composition/yield and actual characteristic measurements.  A suite of other genetic markers (carcass merit scores) have been shown to be associated with various aspects of carcass composition and yield potential (Buchanan et al., 2002; Nkrumah et al., 2005).
Although it is likely that these relationships quantified in America, will apply to the UK, the UK gene pool may be completely different as is the system for assessing and payment of carcasses, the UK system being based on the EU carcass classification grid (MLC Yearbook, February 1983, 69 – 75). 
The objective of the project therefore was to use animals from a UK abattoir which uses best-practice for post-slaughter treatment to ensure tender meat and to:
• Quantify for UK beef cattle, the relationship between Tenderness score and shear force measurement.
• Quantify the relationship between tenderness score and the standard European classification grid.
• Check correlations between the gene markers themselves and with other important economic aspects in growing and finishing cattle, noticeably growth rates.
MATERIAL AND METHODS
Animals
In total, 450 cattle were sampled on entry to the Dovecote Park abattoir on 5 slaughter dates between May and October 2007.  Of these 367 were Angus sired (AA), 41 Limousin sired (Lim), 15 Simmental sired (Sim), 10 Charolais sired (Cha), 5 were Belgium Blue sired (BB), and 9 were South Devon sired (SD).  The remaining animals were Stabiliser.  All non-AA were slaughtered in the first batch (which also contained AA) and were reared organically.  Of the 450 cattle, 131 were heifers and 319 steers.  All cattle were suckled calves i.e. out of beef cross dams and were weaned at 6 months or more of age. From the 450 cattle, 416 steaks were measured for meat quality parameters, which is more than the 300 originally planned.  The missing 36 samples are due to either some carcasses being unavailable for sampling or certain meat quality measurements excluding the results as they are known to bias the results.  E.g. any steak with a pH greater than 5.80 (7 carcasses) were excluded as Dark Firm and Dry (DFD), high pH meat is know to have a different colour to normal meat, a reduced cookloss and to be more tender.
Date of birth and breed of sire were known for all cattle with the majority also having records on the breed cross of the dam.
Gene marker sampling and temperament scoring
Immediately prior to slaughter the cattle were held in a crate to record ear tags against passport details etc.  At this point hairs were removed from the tail for gene marker analysis.  This site was chosen because the hairs are long and they have big follicles which produce a better yield of DNA on extraction than hairs from other sites.
Holding the tail in one hand, 20 or more hairs were taken using a pair of long nosed pliers.  Hairs were grasped firmly by the based and pulled up towards the tail head.  After taking the sample, a check was made to ensure sufficient quantity of hairs and quality of follicles.  Each sample was stored using a standard hair collector card featuring a bar code and a smooth down cellophane protective cover for the follicles.  Any excess hair was trimmed sway from the card and the kill number for each animal was written on the front of the collector.  Samples from each batch were then forwarded to the laboratory for analysis.
In addition, an independent observer monitored the behaviour of each animal as it entered the crate, while it was standing in the crate and when exiting the crate using a 3 point scale. 
Entering Crush
1. Runs in fast and hits the end of the crate hard.
2. Reluctant to enter and has to be pushed in.
3. Walks into the crush itself, apparently unconcerned.
While in Crush
1. Jumps around kicking, never still and shakes head a lot.
2. Pulls back when its head is caught with occasional kicking.
3. Stands still.
When Exiting Crush
1. Runs out fast with a long flight distance before stopping.
2. Reluctant and nervous to leave the crate.
3. Walks out unconcerned.

The three individual scores were summed to provide an overall temperament score which is used in this report.
Slaughter Protocol
All the cattle arrived at Dovecote Park the previous day and were lairaged overnight.  They were subjected to the normal Dovecote Park procedure in terms of slaughter, dressing and carcass management (low voltage electrical stimulation and considerate chilling).  This included all carcasses being recorded on the full 15 point EU carcass classification grid.  After 7 days maturation the carcasses were quartered at the 10th rib and a 25mm thick steak cut from the end of the exposed m. longissimus dorsi on the hind quarter.  These samples were individually labelled, packed and transported to the University of Bristol, in cool boxes containing packs of ice, by Bristol staff.  Each pack was opened and the pH and 1h bloomed colour of each steak measured before vacuum packing and conditioning further to a total of 10 days post-slaughter at which time they were frozen to -20ºC. [Colour and pH were not measured fresh on the first 2 batches, but after thawing and being allowed to bloom before cooking].
Gene Marker Analysis
Genotyping was performed using a standard commercial assay consisting of a panel of gene markers instructive for meat tenderness and carcass grading.
The following markers were present on the assay:
• Calpain markers CAPN4751 and CAPN316
• Calpastatin marker UoGCAST1
• Leptin markers UASMS1, UASMS2 and Exon 2fb
• DGAT and BGHR markers were also present
Colour, pH and Shear Force Measurements
The steaks were over-wrapped in oxygen permeable film and placed in a cold room (2±1ºC) for 1h to bloom.  The colour of the bloomed surface was measured in triplicate using a Minolta CR300 chromameter (Minolta camera company, Milton Keynes) where L = lightness, a*= redness and b* yellowness of the sample.  Hue angle and colour saturation are calculated from a* and b*.
The pH of the meat was measured by direct probe using a Testo 230 instrument and probe.
The steaks were cooked in a George-Foreman grill with two plates such that they were heated from top and bottom simultaneously.  A single thermocouple was inserted to the geometric centre of the steak.  Steaks were placed on the pre-heated grill and cooked to an internal centre temperature of 71ºC, removed from the grill and the temperature monitored until it stopped rising.  Cook-loss was measured by weighing the steaks pre- and post-cooking.
Steaks were placed in bags and allowed to cool overnight at 2±1ºC.  On the following day, a minimum of 7 cores (12.7 mm diameter) were drilled from each steak with the fibres parallel to the long axis of the core.  The cores were sheared using a Warner-Bratzler blade (blade 1.016mm thick (cutting edge bevelled to a half-round, 60 degree V angle, corner of V rounded to a quarter round of a 2.363mm diameter circle, slot size 2.032mm, shear speed 3.82mm/s) and the force to first shear noted (WBSF).
Phenotypes and Statistical Analysis
The major traits considered in the analysis included tenderness (WBSF), meat pH, cooking loss, and colour. Carcass traits considered included carcass weight (CWT), carcass weight per day of age (CWTPDA), fatness, and conformation as assessed on the EU carcass classification grid. EU fatness and conformation classes were converted into a 15-point numerical score where E+ and P- for conformation class and 5+ and 1- for fat class were equivalent to 15 and 1, respectively. The three temperament scores (entering crush, while in crush, and exiting crush) we combined together to create a nine-point temperament score for each animal, with nine being calmer. The relationship between Igenity scores for tenderness and carcass merit were determined using a generalized linear model in SAS (SAS Institute, Inc., Cary, NC, version 9.1) that included the effects of the Igenity score, breed of sire, gender of animal, kill date, and the linear covariate of age of animal at slaughter. Mean separation among Igenity scores for the different test traits was carried out by least-squares using Type III sums of squares in SAS. The PROC CORR of SAS was used to obtain Pearson phenotypic correlations between Igenity scores and the different traits considered. In addition to the Igenity tenderness score and carcass merit scores, each marker on the panel was assessed for its association with the different traits considered in the study. In this analysis, the effect of each marker was fit individually in a SAS GLM model that included the effects of the marker genotype (coded as 2, 1, or 0 for the number of copies of one of the alleles), breed of sire, gender of animal, kill date, and the linear covariate of age of animal at slaughter effect. Significance of each statistical test was evaluated at (P < 0.05). 

RESULTS
The overall means for the carcass traits measured across all animals are shown in Table 1 and for the different breed sires in Table 2
Table 1. Summary for some of the traits measured.
Trait Animals Mean Std Dev Minimum Maximum
Carcass measurements     
Hot carcass weight (kg) 450 313 40.1 212 490
Age 450 711 116 455 947
Carcass weight per day of age† 450 0.45 0.09 0.28 0.75
Conformation class 450 6.8 1.41 4 15
Fat class 450 10.2 1.24 6 15
Flightiness -Temperament  450 8.23 1.16 3 9
Bloomed steak measurements     
Lightness 416 39.7 3.47 27.0 50.5
a* 416 23.7 4.14 11.6 32.9
b* 416 13.9 2.26 4.7 21.6
Hue 416 30.5 3.36 19.7 47.7
Saturation 416 27.5 4.47 14.0 37.4
Cooked steak measurements     
Cookloss % 413 21.8 3.79 9.2 33.1
WBSF (kg) 416 3.5 0.77 2.1 6.8
Igenity Tenderness score     
Tenderness score 447 5.6 1.86 1 10

† calculated from slaughter weight and age
There was no effect for sex except that carcasses from females were lighter and fatter.  As 82 percent of the animals had an Aberdeen Angus sire, the results from these animals will dominate.  As might be expected carcasses from BB and Lim were heavier and had a higher conformation.  AA, SD and Sim were fatter.  The one DFD BB carcass was responsible for the higher average ultimate pH for this group of only 5 carcasses.  The other DFD carcasses were 3 AA and 3 Lim.  These were excluded from the genotype analyses.
It was noted that the first group slaughtered were from organic production systems and contained the lowest proportion of Angus cross cattle.  Many of the least square means for this group were significantly different from the other groups.  The carcasses were heavier, better conformation but fatter.  Of interest were the scores for temperament.  All three measures had lower values culminating in an overall temperament score of 1 unit less than the other groups.  This resulted in a higher ultimate pH for this group.  Four of the DFD carcasses were in this group and 3 others were close to the cut-off of pH 5.8 taken as defining DFD, the other groups had 2/1, 0/0, 1/2 and 0/1 DFD and marginally DFD carcasses respectively.

Table 2.  Selected data for the sire breed groups
Cold carcass Conformation Fatness Carcase weight per day of age pH WBSF Tenderness score  Temperament
Breed weight (kg)     (kg/day)     score  
and number †LSM StdErr LSM StdErr LSM StdErr LSM StdErr LSM StdErr LSM StdErr LSM StdErr LSM StdErr
AA      367 308.6 1.96 6.49 0.06 10.41 0.06 0.44 0.00 5.62 0.01 3.44 0.04 5.6 0.17 8.44 0.05
BB         5 382.9 16.82 10.60 0.51 8.80 0.49 0.52 0.04 5.94 0.05 3.91 0.34 5.6 0.95 7.40 0.47
CHA     10 292.3 11.89 6.60 0.36 7.10 0.35 0.48 0.03 5.67 0.04 3.35 0.24 4.6 0.76 8.70 0.33
LIM       41 339.5 5.87 8.93 0.18 9.49 0.17 0.46 0.01 5.69 0.02 3.97 0.12 5.1 0.57 7.02 0.16
SD         9 292.0 12.53 6.56 0.38 10.00 0.37 0.51 0.03 5.62 0.04 3.52 0.25 6.7 0.78 8.00 0.35
SIM      15 345.9 9.71 8.27 0.29 10.20 0.28 0.49 0.02 5.66 0.03 4.13 0.19 3.8 0.67 6.60 0.27
P <0.001   <0.001   <0.001   <0.001   <0.001   <0.001   0.007   <0.001 
† Least square means
Table 3.  Correlations between Tenderness score and carcass measures and between WBSF, age and carcass measures
Traits WBSF Temperament CWT CWTPDA Conformation Fatness pH Cookloss Saturation Hue
Tenderness score -0.084 0.193 -0.059 0.025 -0.136 0.096 -0.029 0.074 0.166 -0.079
P 0.089 <0.001 0.219 0.593 0.004 0.044 0.559 0.134 0.001 0.109
          
WBSF  -0.135 0.283 0.120 0.313 -0.085 0.025 0.406 -0.023 0.236
P  0.006 <0.001 0.015 <0.001 0.086 0.611 <0.001 0.649 <0.001
          
Age 0.033 -0.103 0.166 -0.728 -0.051 0.122 -0.007 0.028 0.018 -0.062
P 0.50 0.028 0.000 <0.001 0.284 0.010 0.886 0.567 0.720 0.205

 Distribution of Tenderness Scores.
The distribution of Tenderness scores are shown in Figure1, alongside the data from 24000 US cattle used to validate the technique and Tenderness scores and associated WBSF values are plotted in Figure 2.
 
Figure 1 shows that the distribution of Tenderness scores in the UK animals sampled for this trial is almost identical to that for a much larger population that has now been tested in the USA.  As in the case of the US data there are no animals in Category 2 and very few in Categories 1, 8 and 10 (n= 5, 6 and 8 respectively).  As these numbers were obtained from a sample of 450 animals, numbers tested would have to be doubled to obtain at least 10 in these groups.  However, the original work was carried out on groups of 400-500 animals and it is expected that relationships can be found with this sort of number, but it is clear that care must be taken when assigning mean values to these groups. 
 
The WBSF has been superimposed upon Figure 2 and shows a negative relationship between Tenderness score and WBSF, i.e. moving from score 1 to score 10 then the meat is more tender.  However, this relationship is only statistically significant at the 10% level(Table 3).
Table 4.  LSMeans for WBSF for each Tenderness score for all animals and for AA alone.
TenderGene Score WBSF±SE all animals WBSF ±SE AA only
1 3.79 0.32 3.70 0.47
3 3.77 0.11 3.52 0.13
4 3.45 0.07 3.30 0.09
5 3.50 0.12 3.44 0.13
6 3.56 0.08 3.39 0.09
7 3.36 0.07 3.24 0.08
8 3.02 0.30 2.90 0.28
9 3.62 0.13 3.47 0.13
10 3.41 0.32 3.30 0.30
It is noticeable in Table 4 that the standard error for groups 1, 8 and 10 which only had low numbers were much greater than for the other groups.  Trend lines for the data were not different from one another for the AA and total animals.  Whilst bearing in mind the small numbers of animals in some of the groups and the overemphasis that any large group can have on the overall trend, it can be calculated from trend lines of the mean data values that the difference in texture (WBSF) between a score of 1 and 10 is 0.43kg.  This is about half the value found in the US experiments.  If the calculation is repeated for scores 1-8 only then the difference increases to 0.86kg.  However, in the American data e.g. Van Enennaam et al. (2007), the range of WBSF were 1.8 to 13.1 kg whilst in this data it is only 2.1 – 6.8 kg, half the range.  The American data does include Brahman phenotypes which produce tougher meat.  In this experiment the carcasses were electrically stimulated, considerately chilled and the meat aged for 10 days which would have contributed to a great reduction in the variability between samples.  But there is still the same negative association between Tenderness score and shear force value, suggesting the same genetic control of tenderness as found in the American studies.
Whilst being able to predict, or select for, tenderness was the main aim of the project, it is important that selection for increased tenderness does not produce negative effects for other traits such as carcass weights and grade.  Table 3 shows the correlations between Tenderness score and other carcass measurements between WBSF and carcass measures and between age and carcass measures.  Increasing Tenderness score from 1-10 has a positive advantage in terms of tenderness and is also associated with animals being less flighty, a better colour (saturation) but poorer conformation.  More tender meat as measured by WBSF is associated with less flighty animals, lighter, fatter carcasses, poorer conformation and a greater cook loss.  Not unexpectedly, age had a highly significant correlation with cold carcass weight and carcass weight per day of age.  Older animals were also fatter and less flighty.
There were few associations between the individual gene markers in the panel that were tested. (Table 5.)  The tenderness gene appearing most often is CAPN4751 and is associated with meat colour and temperament which are difficult to explain.  The only association with WBSF is CAPN316, another of the tenderness genes, but this is only statistically significant at the 10% level.

Table 5. Single marker associations with carcass measurements.
Marker Trait Estimate Standard Error P
BETALAC Hot carcass weight -4.122 2.266 0.070
BETALAC Conformation -0.154 0.078 0.051
BETALAC Yellowness -0.228 0.125 0.068
CAPN4751 Entry score 0.098 0.032 0.002
CAPN4751 Redness -0.612 0.189 0.001
CAPN4751 Yellowness -0.257 0.130 0.048
CAPN4751 Saturation -0.645 0.205 0.002
CAPN4751 Temperament 0.197 0.074 0.008
CAPN316 WBSF -0.113 0.067 0.094
MC1R373 Fatness -0.193 0.097 0.047
MC1R373 Conformation 0.185 0.101 0.068
GNSC319 Fatness -0.214 0.089 0.023
GNSC319 Saturation -0.467 0.234 0.047
BGHR Hot carcass weight 6.234 3.298  0.059
BGHR Temperament 0.221 0.106 0.037
UASMS2 Hue 0.462 0.234 0.049
UASMS2 Lightness 0.493 0.251 0.05

Acknowledgements
Our special thanks to members of Dovecote Park, David Gunner, Managing Director and Laurie Ibbotson, Head of Livestock Department, for their permission to access the plant facilities and their tolerance of our activities, and to Amanda Sutton and Lucy Jackson for their help in collecting the animal data.

Thanks are due to Katie Davies, Merial, who co-ordinated the onward transfer of hair samples for analysis and collecting of data, Neil Scott, SAC, who carried out all the temperament scoring and Duncan Marriott and Ivan Garcia, University of Bristol, who collected meat samples from Dovecote Park and measured all the meat quality parameters. 
Thanks are due to Borders Quality Beef Co-operative for helping to source some of the cattle for the project. 
References
Koohmaraie, M. (1996). Biochemical factors regulating the toughening and tenderisation process of meat. Meat Science, 43, S193-201.
Barendese, W.J. (2002). DNA markers for meat tenderness. International patent application PCT/AU02/00122 [International patent publication WO 99/23248]
Buchanan, F.C., Fitzsimmons, C.J., Van Kessel, A.G., Thue, T.D., Winkelman-Sim, D.C. and Schmutz SM (2002): Association of a missense mutation in the bovine leptin gene with carcass fat content and leptin mRNA levels, Genetics Selection Evolution, 34, 105-116.
Casas, E., White, S.N., Wheeler, T.L., Shackleford, S.D., Koohmaraie, M., Riley, D.G., Chase, C.C., Johnson, D.D. and Smith, T.P.L.. (2006) Effects of calpastatin and µ-calpain markers in beef cattle on tenderness traits. Journal of Animal Science, 84, 520-525.
Morris, C.A., Cullen, N.G., Hickey, S.M., Dobbie, P.M., Veenvliet, B.A., Manley, T.R., Pitchford, W.S., Kruk, Z.A., Bottema, C.D.K. and Wilson, T. (2006) Genotypic effects of calpain 1 and calpastatin on the tenderness of cooked M. longissimus dorsi steaks from Jersey x Limousion, Angus and Hereford-cross cattle, Animal Genetics, 37, 411-414.
Nkrumah, J.D., Li, C., Yu, J., Hansen, C., Keisler, D.H. and Moore SS, (2005) Polymorphisms in the bovine leptin promoter associated with serum leptin concentration, growth, feed intake, feeding behavior, and measures of carcass merit  Journal of Animal Science, 83, 20-28
Page, B.T., Casas, E., Heaton, M.P., Cullen, N.G., Hyndman, D.L., Morris, C.A., Crawford, A..M., Wheeler, T.L., Koohmaraie, M., Keele, J.W. and Smith, T.P.L. (2002) Evaluation of single-nucleotide in CAPN1 for association with meat tenderness in cattle. Journal of Animal Science, 80,3077-3085
Page, B.T., Casas, E., Quaas, R.L., Thallman, R.M., Wheeler, T.L., Shackleford, S.D., Koohmaraie, White, S.N., Bennett, G.L., Keele, J.W., Dikeman, M.E. and Smith, T.P.L. (2004) Association of markers in the bovine CAPN1 gene with meat tenderness in large crossbred populations that sample influential industry sires. Journal of Animal Science, 82, 3474-3481
Schenkel, F.S., Miller, S.P., Ziang, Z., Mandell, I.B., Ye, X., Li, H. And Wilton, J.W. (2006) Association of a single nucleotide in the calpastatin gene with carcass and meat quality traits of beef cattle. Journal of Animal Science, 84, 291-299.
Van Enennaam, A.L., Li, J., Thallman, R.M., Quaas, R.L., Dikeman, M.E., Gill, C.A., Franke, D.E. and Thomas, M.G. (2007) Validation of commercial DNA tests for quantitative beef quality traits.  Journal of Animal Science, 85, 891-900
White, S.N., Casas, E., Wheeler, T.L., Shackleford, S.D., Koohmaraie, Riley, D.G., Chase, C.C., Johnson, D.D. Keele, J.W. and Smith, T.P.L. (2005) A new single nucleotide polymorphism in CAPN1 extends the current tenderness markers test to include cattle of Bos indicus, Bos Taurus, and crossbred descent. Journal of Animal Science, 83, 2001-2008.