Logo of geneticsGeneticsCurrent IssueInformation for AuthorsEditorial BoardSubscribeSubmit a Manuscript
Genetics. Aug 2007; 176(4): 2601–2610.
PMCID: PMC1950658

Epistasis Between Calpain 1 and Its Inhibitor Calpastatin Within Breeds of Cattle

Abstract

The calpain gene family and its inhibitors have diverse effects, many related to protein turnover, which appear to affect a range of phenotypes such as diabetes, exercise-induced muscle injury, and pathological events associated with degenerative neural diseases in humans, fertility, longevity, and postmortem effects on meat tenderness in livestock species. The calpains are inhibited by calpastatin, which binds directly to calpain. Here we report the direct measurement of epistatic interactions of causative mutations for quantitative trait loci (QTL) at calpain 1 (CAPN1), located on chromosome 29, with causative mutations for QTL variation at calpastatin (CAST), located on chromosome 7, in cattle. First we identified potential causative mutations at CAST and then genotyped these along with putative causative mutations at CAPN1 in >1500 cattle of seven breeds. The maximum allele substitution effect on the phenotype of the CAPN1:c.947G>C single nucleotide polymorphism (SNP) was 0.14 σp (P = 0.0003) and of the CAST:c.155C>T SNP was also 0.14 σp (P = 0.0011) when measured across breeds. We found significant epistasis between SNPs at CAPN1 and CAST in both taurine and zebu derived breeds. There were more additive × dominance components of epistasis than additive × additive and dominance × dominance components combined. A minority of breed comparisons did not show epistasis, suggesting that genetic variation at other genes may influence the degree of epistasis found in this system.

EPISTASIS is often assumed to be pervasive in some genetic models (Wright 1980; Gjuvsland et al. 2007) even though it is always difficult to show epistasis for genes with small phenotypic effect. Examples of epistasis, involving mutations that cause discrete changes to the phenotype, were reported soon after the principles of Mendelian inheritance were rediscovered (Bateson 1909). There are many examples (Beadle and Ephrussi 1936; Kijas et al. 1998; Olson 1999) and the subject has been reviewed recently (Carlborg and Haley 2004). Strong effects of the genetic background can also be shown on the phenotype, and these by definition contain a component of epistasis. In addition, several studies have shown that epistasis exists due to quantitative trait loci (QTL) (Wang et al. 1999; Luo et al. 2001; Carlborg et al. 2006), and epistasis has been measured for microsatellites and other DNA variants (Tiret et al. 1994; Routman and Cheverud 1997). Nevertheless, demonstrations of epistasis are not common for quantitative traits, and this may be partly due to the way epistasis is usually conceptualized (Fisher 1918; Hansen and Wagner 2001), as a residual deviation from the additive effects.

Several alternative approaches have been proposed either to estimate or to describe epistasis for quantitative variation (Cockerham 1954; Kempthorne 1954; Cheverud and Routman 1995; Hansen and Wagner 2001; Barton and Turelli 2004; Zeng et al. 2005; Mao et al. 2006) using different mathematical symbols and fundamental assumptions. The methods of Cockerham and Kempthorne were independent attempts to partition the epistasis defined by Fisher in 1918 while that of Cheverud and Routman was an attempt at defining epistasis without explicit use of allele frequencies. The Cheverud–Routman method has been criticized (Zeng et al. 2005) and the Cockerham and Kempthorne methods have both been developed further, the former by Zeng and co-workers and the latter by Mao and co-workers. The approaches of Hanson and Wagner or Barton and Turelli to describe epistasis have not been translated into formal methods, although Hanson and Wagner have reanalyzed data under their framework (Hansen and Wagner 2001). Numerical examples show that different methods will estimate different values for the components of epistasis (Zeng et al. 2005) but that different models cannot explain more variation due to epistasis.

The calpain gene family and its inhibitors are a good subject for a study of epistasis because they are known to have protein–protein interactions (Betts and Anagli 2004). This family of genes affects a range of phenotypes such as diabetes in humans (Horikawa et al. 2000), exercise-induced muscle injury (Belcastro et al. 1998), pathological events associated with degenerative neural diseases in humans (James et al. 1998), fertility and longevity (Garcia et al. 2006), and postmortem effects on meat tenderness in a variety of livestock species (Koohmaraie 1996). The calpains are inhibited by calpastatin (Goll et al. 2003), which binds directly to calpain.

To study epistasis between QTL in detail, we chose as a model system two unlinked QTL affecting the same trait (meat tenderness) and for which gene tests based on calpain 1 (CAPN1) and calpastatin (CAST) are now available. Quantitative traits are affected by an unknown number of Mendelian factors (QTL) that are in themselves small and their effects can generally only be seen when tied to chromosomal segments (Geldermann 1975). The molecular genetic bases for QTL are generally not known but are assumed to be no different in kind from mechanisms that have already been demonstrated for other traits. QTL have been reported for meat tenderness; the direct role of calpain 1 and calpastatin in postmortem meat tenderization in cattle is well known (Whipple et al. 1990; Koohmaraie 1996) and DNA variants at both CAPN1 (bovine chromosome 29) and CAST (bovine chromosome 7) affect meat tenderness (Casas et al. 2000, 2006; Barendse 2002; Page et al. 2002; White et al. 2005; Drinkwater et al. 2006; Morris et al. 2006; Schenkel et al. 2006; Van Eenennaam et al. 2007). This genetic variation could be a useful model for the study of gene–gene interaction for QTL, because these QTL have small phenotypic effects. Here we report an examination of the putative causative mutations at the calpain 1 (CAPN1) and calpastatin (CAST) genes and the direct measurement of epistatic interactions of causative mutations for these QTL, using the G2A method (Zeng et al. 2005).

MATERIAL AND METHODS

Cattle samples:

The beef CRC cattle DNA samples and the methods of measurement of phenotypes have been reported previously (Perry et al. 2001; Upton et al. 2001). The breed descriptions are as follows. The Angus, Hereford, and Shorthorn are Australian representatives of British cattle (taurine type). The Murray Grey is an Australian derivative of the Aberdeen Angus with the gray color derived from crossing of the original bulls to a light roan Shorthorn dam in 1905. The Brahman breed (zebu type) in Australia commenced in 1933 and is strongly linked to the breed in the United States, where it was founded in 1854 from a small number of breeds from India, and to Brazil. The Santa Gertrudis (old taurine × zebu composite) is a ⅜ Brahman and ⅝ Shorthorn composite begun in 1910 at the King Ranch in Texas. The Belmont Red (sanga taurine type) was founded in 1954 (Burrow 1998) and is a ½ Africander (a sanga breed from Southern Africa, seen as a taurine type), ¼ Hereford, and ¼ Shorthorn composite.

Meat tenderness was measured in kilograms using peak force measurements for the musculus longissimus lumborum (LLPF) as described previously (Perry et al. 2001). Tenderness measurements are available for Angus (N = 1409), Brahman (N = 1118), Belmont Red (N = 1475), Hereford (N = 918), Murray Grey (N = 368), Santa Gertrudis (N = 1223), and Shorthorn (N = 458) animals. The LLPF measurements were adjusted across all animals using restricted maximum likelihood in a general linear mixed model implemented through ASReml (Gilmour et al. 1995), where LLPF ~ N(μ + kill group + herd of origin + age of slaughter + sire, equation M1) and where sire was fitted as a random effect. There are several herds of origin associated with each breed so the average genetic effects due to breed and sire were included in the model as well as fixed environmental treatment effects; this avoids false associations due to inappropriate lumping of groups. A sample of 1900 animals was genotyped for each single nucleotide polymorphism (SNP), except for CAST:c.856G>A, which showed little variation in taurine animals, and CAST:c.2832A>G, which was genotyped on 4936 animals; the larger number of animals was genotyped for this SNP to validate its effects. The animals were chosen to maximize the number of sires in the study, with the aim of having as diverse a sample as possible.

CAST sequence:

Five micrograms of the total RNA from 1 g of muscle was translated to cDNA by the oligo-dT method using the Invitrogen Superscript Rnase H-Reverse transcriptase kit following the manufacturer's instructions. Ten individuals were sequenced, and individuals for sequencing were chosen on their genotype at the (NM_174003.2) CAST:c.2832A>G SNP, which is the current genetic test located in the 3′ UTR (Barendse 2002). The identity of the cDNA sequence was confirmed using BLAST (Altschul et al. 1990). These sequences were assessed, then assembled into contigs using Phred (Ewing et al. 1998) and Phrap and viewed using Consed (Gordon et al. 1998). PolyPhred (Nickerson et al. 1997) was used to identify variable bases. SNPs were described using standard nomenclature (den Dunnen and Antonarakis 2000).

To show what the likely effects of the amino acid sequence changes would be on the structure of calpastatin, we compared the amino acid changes using the translated reference sequence (NM_174003.2) for this gene. These sequences were analyzed with a nine-amino acid window using the Kyte–Doolittle hydropathy algorithm (Kyte and Doolittle 1982) and the results were plotted.

SNP analysis:

SNPs were genotyped using the Taqman MGB allele discrimination method (ABI, Foster City, CA) as before (Barendse et al. 2006) using the probes and primers listed in Table 1 by two individuals. The observed genotype frequencies were compared to expected frequencies under the Hardy–Weinberg equilibrium (HWE) using the chi-square test. Allele frequency differences were tested by comparing genotype counts between populations while LD was tested using haplotypes obtained using the EM algorithm (Weir 1996). The haplotypes were used to calculate r2 (Hill and Robertson 1968) as a measure of LD, which estimates the correlation between genotypes and has an expectation of zero for populations in linkage equilibrium (LE). Differences between mean r2 values were compared using Student's t-test with significance determined using 100,000 permutations.

TABLE 1
Primers and probes for Taqman assays for the SNP

Associations between genotypes and meat tenderness were tested using statistical methods described previously (Gilmour et al. 1995; Lynch and Walsh 1998; Barendse et al. 2007). In brief, mean residual trait values for each genotype were compared using F-tests and the most divergent genotypes were compared with a t-test, with the significance determined using 100,000 permutations. The proportion of the residual variance explained by an SNP or combination of SNP was calculated by comparing the sum of squares of the model without SNP to the model with one SNP, then the model with one SNP was compared to that with two SNPs, and so on. The order in which SNPs were added was determined by fitting all the SNPs and then removing the SNPs stepwise using the Aikake Information Criterion in S-Plus (Venables and Ripley 2000). The SNP that was removed last was fitted first and so on.

Measurement of epistasis:

Cattle populations have undergone population bottlenecks associated with domestication and subsequently with breed formation (MacHugh et al. 1997). In addition, cattle in the breeds in this study are under artificial selection in national herd recording schemes, so the populations are not large random mating groups. This may generate samples that exhibit HWD and LD between unlinked loci, so the methods that are used should either account for this or should be unaffected by population structure, or the data should be shown to be in HWE and LE. Nevertheless, methods that assume the data are in HWE and LE are only under a large disadvantage for samples of N < 200 in detecting epistasis, and spurious detection of epistasis with incorrect models will be primarily of the additive by additive component (Mao et al. 2006).

We used the full G2A method (Zeng et al. 2005) to calculate components of epistasis, i.e., additive by additive (aa), additive by dominance (ad), dominance by additive (da), and dominance by dominance (dd) components. In case of possible population structure caused by the evolutionary history of the species, the estimates of the epistatic components were tested by permutation. To determine significance, the data were permuted 10,000 times and both the phenotype and the genotype at the second locus were permuted against the genotype at the first locus with the permutation occurring within each breed; the number of times that larger components of epistasis were obtained by permutation was counted and divided by the total number of permutations. Standard errors were calculated for the epistatic components using the bootstrap, where the standard error is the standard deviation (Efron and Tibshirani 1991) of 1000 bootstrap replicates of each epistatic component. With low minor allele frequencies one of the genotypes may be absent; during permutation this absence is accounted for and does not contribute to the significance of an estimate of epistasis. The software, EPEE, performs the permutations on pairs of loci for data files that may contain a user-defined number of loci and thousands of individuals and is available at http://www.cgd.csiro.au/software.html.

RESULTS

To identify possible causative mutations in calpastatin, we sequenced animals of known genotype to find SNPs that alter the amino acid sequence. We found 12 SNPs in the 91 cDNA sequence reads from the 10 animals. Four of these change the amino acid sequence, (NM_174003.2) CAST:c.143G>A Ser48Gly, CAST:c.155C>T Pro52Leu, CAST:c.856G>A Ala286Thr, and CAST:c.1487C>T Ala496Val. CAST:c.856G>A is in the calpain inhibitory domain (CID). CAST:c.143G>A and CAST:c.155C>T are four amino acids apart in exon one. Seven of the SNPs were in the 3′UTR, and none were in the 5′UTR. The consensus sequence showed a further three synonymous DNA sequence differences compared to the reference sequence (NM_174003.2), but these were not polymorphic in the animals we sequenced.

The amino acid changes Gly48Ser and Pro52Leu make a large change to the hydropathy plot of calpastatin while the Ala286Thr and Ala496Val make small changes (Figure 1). Individually, the hydropathy associated with each amino acid change goes from −1.6 to 3.8 for Pro52Leu on the Kyte–Doolittle scale; the next best of these changes is Ala286Thr (1.8 to −0.7), then Ala496Val (1.8 to 4.2) and then Ser48Gly (−0.8 to −0.4). The combined Gly48Ser and Pro52Leu change alters the most hydrophobic region in the first 150 amino acids of the protein, increasing the average hydrophobicity and broadening the peak. The other amino acid changes do not change the overall shape of the hydropathy curves.

Figure 1.
The hydropathy plots for calpastatin for the (A) Ser48Gly and Pro52Leu, (B) Ala286Thr, and (C) Ala496Val amino acid mutations.

The genotypes of the SNP were generally in HWE within breed and the unlinked SNPs are in LE. Of 42 breed by SNP tests of HWE, one was significant (CAPN1:g.6545C>T, Belmont Red, χ2 = 8.26, P = 0.016), with two tests expected significant at the 5% threshold by chance. All allele frequencies of these SNPs were significantly different across the seven breeds, with all comparisons showing P < 0.001. The smallest range of allele frequencies was for CAST:c.155C>T with p0, the allele higher up the alphabet, ranging from 0.23 to 0.62 and the largest range of frequencies was for CAPN1:g.6545C>T with p0 ranging from 0.19 to 0.99. The mean r2 (Table 2) for SNP comparisons within CAST or CAPN1 were significantly greater than the mean r2 for SNP comparisons between CAST and CAPN1 (t = 8.98, P < 0.001, N = 105) when the within-breed estimates are compared, although the differences in mean r2 for these same comparisons calculated across breed are not significant (t = 4.02, P = 0.067, N = 15). The mean r2 for SNP comparisons within CAST or CAPN1 calculated across breed were not significantly different from those calculated within breed (t = 0.47, P = 0.37, N = 56) and the mean r2 for SNP comparisons between CAST and CAPN1 calculated across breed were not significantly different from those calculated within breed (t = 0.87, P = 0.63, N = 64). The highest value of r2 = 1.00 within breed and r2 = 0.69 across breed for a pair of SNPs. The r2 for comparisons between the unlinked pairs of SNPs between CAST and CAPN1 show values of essentially zero both within and across breed (Table 2).

TABLE 2
Linkage disequilibrium estimates between SNP

The four calpastatin protein polymorphisms, as well as the CAST:c.2832A>G SNP, showed different levels of significance although the effect of allele substitution was similar for each (Table 3). All of the polymorphisms increase the amount of residual phenotypic variance that is explained (Figure 2) even when a CAST SNP has already been added, suggesting that each SNP may affect the variation for tenderness. Combinations of only the CAST SNP explain 1.1% of the residual variance. CAST:c.155C>T (α = 0.14 σp) shows the most significant association for its sample size. CAST:c.2832A>G SNP (α = 0.13 σp) and CAST:c.143G>A (α = 0.13 σp) also show strong associations with individual breeds showing statistical significance. For CAST:c.1487C>T (α = 0.09 σp) none of the component breeds show significant associations to meat tenderness; only the combined sample does. The CAST:c.856G>A SNP (α = 0.07 σp) showed 98 GG homozygotes and only 5 AG heterozygotes in 103 European taurine animals, so those breeds were not genotyped further for this SNP. It was not associated with tenderness in the combined sample of 921 sanga taurine, zebu, and zebu-cross animals, although the Belmont Red breed by itself did show a highly significant association.

Figure 2.
The increase in residual variance explained as each additional SNP of CAPN1 and CAST is added. The increase in residual variance is compared to the residual variance that would be explained if the individual effects of each SNP were added together.
TABLE 3
Estimated sizes of effects on meat tenderness for the calpain 1 and calpastatin SNP

Both calpain SNPs, (NM_174259) CAPN1:c.947G>C Gly>Ala (α = 0.13 σp) and (AF248054.2) CAPN1:g.6545C>T (CAPN1:c.1800+169C>T) (α = 0.14 σp), explain similar-sized effects to that found for the calpastatin SNP (Table 3). The CAPN1:c.947G>C SNP shows significance in Angus and Belmont Red. The Belmont Red, the Brahman, and the Santa Gertrudis showed significant associations for CAPN1:g.6545C>T. Combinations of only the CAPN1 SNP explain 1.4% of the residual variance.

The size of effect for any of these CAPN1 and CAST SNPs varies in individual breeds (Figure 3), although some of the large differences between breeds would be due to sampling effects. Most of the very large sizes of effect are for samples of N < 500, although some of the large sizes of effect are also based on >900 animals in a breed sample (CAST:c.2832A>G). These latter effects are much larger than those found in other samples of similar size, and show that breed appears to alter the effect of the genotype. When data are combined across breeds, the effects appear to be moderated: all values reported in Figure 3 for N > 1500 are for combined samples.

Figure 3.
The relationship between sample size and the size of the gene effect for different SNPs in the different breeds and breed combinations.

There was significant epistasis between pairs of all CAST and CAPN1 SNPs (Table 4). The two-locus genotypes for CAPN1 and CAST show re-ranking of some genotypes (Figure 4), which is consistent with epistasis. The statistical analysis of epistasis in individual breeds shows that there are several breeds that have more than one significant comparison between a CAPN1 and CAST SNP. In addition, the CAST:c.1487C>TCAPN1:c.947G>C pair of SNPs showed significant epistasis in four breeds and the same component of epistasis, the additive × dominance component (ad + da), was significant in all of those breeds. Thirty percent of the comparisons were significant at the 5% level, with 3.5% at the 0.1% level. Seventy one percent of all the significant components were additive × dominance effects (ad + da) and aa were the least common. Nevertheless, some of the SNP pairs were only significant in one breed, and there were differences in the sizes of the components in the different breeds for a particular pair of SNPs.

Figure 4.
The mean tenderness values for combinations of genotypes at CAST:c.155C>T with (A) CAPN1:g.6545C>T and (B) CAPN1:c.947G>C.
TABLE 4
Components of epistasis measured between CAST and CAPN1 SNP

DISCUSSION

In this study we report that the CAST:c.155C>T SNP has the strongest association to meat tenderness compared to the four other CAST SNPs and is the most likely single causative allele. The degree of epistasis between the alleles and genotypes at CAPN1 and CAST was significant and the additive effect at one locus changed or rescaled the dominance effect at the other locus. The molecular model of the protein–protein interaction between CAPN1 and CAST is consistent with the location of CAST:c.155C>T as a causal allele and suggests how the epistatic interaction might occur.

The nonsynonymous CAST SNPs could all be causative mutations, and there is an increase in the explained residual variance as more SNPs are used. Of the changes to the amino acid sequence of calpastatin, the Pro52Leu change due to CAST:c.155C>T has the largest change in hydropathy (Kyte and Doolittle 1982) of the four amino acid changes we identified, and as it is also the SNP with the most significant association to meat tenderness tested in this study, it suggests that this is the most likely candidate for a causative mutation for meat tenderness in the calpastatin gene, if a single candidate were to be chosen. The CAST:c.143G>A SNP, which causes the smallest amino acid change, is nevertheless only four amino acids away from CAST:c.155C>T, and together they cause a significant shift in hydropathy. The a priori case for the other SNPs is not as strong. The CAST:c.856G>A, which causes the second-largest change, is only significant in one breed, the Belmont Red. While the CID, where it is located, is an important functional part of the gene, this mutation in the amino acid sequence did not show a significant association to tenderness in the combined sample, and significance may be due to linkage disequilibrium to other SNP. The CAST:c.1487C>T SNP, which causes the Ala496Val change, only shows an association to meat tenderness in the combined sample; even with many hundreds of animals in a breed, none of the within-breed associations for this marker are significant for meat tenderness. The slight increase in explained variance when several of the SNPs are added together might reflect interactions when several of the mutations are lined up, or it may reflect the effects of SNPs outside the coding sequence of the gene.

The CAPN1 SNPs have similar effects on meat tenderness, confirming the previously reported discoveries (Page et al. 2002; White et al. 2005). However, unlike the CAST:c.2832A>G SNP and CAPN1:g.6545C>T which appeared to be useful in all breed types, CAPN1:c.947G>C was thought to be more useful for taurine animals: taurine and zebu cattle are not only evolutionarily distinct, representing domestications from temperate and subtropical regions (MacHugh et al. 1997), but also have systematic differences in meat tenderness (Wheeler et al. 2001; Reverter et al. 2003). Our results confirm the breed-specific effects of the CAPN1 SNP. The c.947G>C SNP was significant only in Angus and Belmont Red, which is consistent with the previous report. The g.6545C>T SNP was significant in Brahman, Santa Gertrudis, and Belmont Red. However, none of the purebred cattle of European origin showed associations to g.6545C>T. The Belmont Red has notionally 50% sanga ancestry, derived from the Africander breed of southern Africa, and it is possible that there was a minor component of zebu gene flow into some of the sanga breeds due to the activity of herders in the past thousand years (Hanotte et al. 2002). The latter SNP is probably in linkage disequilibrium with additional variation, possibly found mainly in zebu breeds, because this SNP, 169 bp into the intron, is not likely to be functional.

The CAST and CAPN1 genotypes show significant epistasis, most of which was of the ad and da types. The evidence for epistasis is strong because it is not restricted to one breed or to one particular SNP comparison. It occurs in most of the breeds, and several of the breeds have more than one pair of SNPs showing significant epistasis. While most of the comparisons have large sample sizes, epistasis was detected in samples with <100 individuals. In addition, where the same pair of SNPs is found to be significant in different breeds, a range of genotype frequencies was found at both genes. This suggests that the detection of epistasis is not dependent upon the specific conditions of the sample and argues that the significance is not due to a chance event; the relative occurrence of significant components of epistasis is also more frequent than would be expected by chance, and only one of those was an aa component, which is the component that is most likely to be found due to chance (Mao et al. 2006). The presence of epistasis for a particular CAST–CAPN1 pair should not be interpreted as evidence of a specific functional relationship of that pair of SNPs, or that a particular amino substitution is proved to be functional because it shows a significant epistatic interaction; although the CAST–CAPN1 SNP pairs are in linkage equilibrium there will be genetic correlations within each gene due to linkage disequilibrium so several CAST–CAPN1 pairs should show significant epistasis.

Most of the epistasis was additive × dominance, which means that the dominance deviation at one locus is altered or rescaled by the additive deviation at the other locus (Hansen and Wagner 2001). This can be seen in an overall sense when the two-locus genotypes are plotted; the major shifts in phenotype occur when one of the genotypes is a heterozygote, and this occurs for both genes. The predominance of the additive × dominance components, and the consistency with which the ad and da components occur with the same pair of SNPs in different breeds, adds to the confidence in the significance of the epistasis. The numerical values for a particular epistatic component differ from breed to breed, but this would be expected since the genotype frequencies are significantly different between breeds; estimates from one population cannot be transferred to another population without losing their orthogonality (Zeng et al. 2005).

Epistasis between these genotypes suggests that genotype frequency differences between breeds should contribute to the differences in the size of the allele effect for CAPN1 and CAST SNP in different breeds. Epistasis was found in most of the breeds and for all of the SNPs at the two genes. The genotype frequencies for these SNPs differ significantly for all the breeds in this study, and the allele effects of these SNPs differ between breeds. Of course, not all of the differences in size of effect for CAST and CAPN1 alleles in the breeds will be due to epistasis—sampling error will add its share—but part of the differences in size of effect between breeds could be due to epistasis between CAPN1 and CAST SNPs as a consequence of differing frequencies. As genotype frequencies change so the frequencies of heterozygotes will alter, which will change the number of animals in which an additive × dominance epistatic effect would be observed. However, all of the differences in size of effect for these two genes in different breeds will only be properly quantified when the other genes with epistatic effects on meat tenderness are analyzed and incorporated into one calculation.

The location of the SNP in CAST is consistent with a possible role in affecting the activation of calpain 1 by calpastatin, particularly when bound to cell membranes. Calpastatin inhibits calpain 1 through the B domain, and this is the focus of most research between these molecules. However, the CAST:c.155C>T and CAST:c.143A>G SNPs are located in the L domain of calpastatin, which has until recently not been assigned a role in the interaction between these molecules. The amino terminal first sixth of the calpastatin molecule, including the L domain, was known to be responsible for binding to biological membranes at acidic phospholipids (Mellgren et al. 1989). By changing the hydropathy of this region, these mutations could influence the strength of the binding of calpastatin to cell membranes because it is thought that the mechanism by which it binds involves electrostatic interactions between the basic amino acids in the amino terminal region and the acidic phospholipids. More recently, Melloni et al. (2006) showed that the noninhibitory L domain (i.e., it is not involved in the inhibition of calpain 1 by calpastatin) binds to the catalytic DII domain of calpain and that calpain undergoes a conformational change, increasing its ability to act as a protease at physiological levels of Ca2+ ions. This is consistent with the known activation of calpain 1 when it is bound to membranes. Epistasis between CAST and CAPN1 SNP suggests that when one or the other protein has heterozygous forms it changes the overall performance of the calpain 1 calpastatin complex altering the resulting phenotype for meat tenderness. These changes may occur either where calpastatin binds to cell membranes or binds calpain 1.

Additional research could explore the physical basis of the epistatic interaction and could investigate SNPs in other genes, as they become available, to search for further examples of epistatic interactions. Further study of the protein crystal structure of the amino acid mutations, or the use of yeast two-hybrid analysis of the amino acid mutations, could yield insights into how the epistasis occurs. It may also determine whether the epistasis is specific to the particular SNP or whether epistasis between CAPN1 and CAST is likely to occur in general, and so affect a wide range of phenotypes in a range of species. In addition, other genetic variation influencing meat tenderness may also show epistatic interactions with variation at CAPN1 and CAST but this need not occur through direct protein–protein interactions. Although the example in this study is unusual in that it is for a protein–protein interaction, not because it is conceptually unlikely but because most thoroughly studied epistatic interactions are for interactions along a biochemical pathway, such as when a pigment or metabolite is sequentially altered by a series of genes, so mutations in other genes affecting the amount of muscle tissue or extracellular matrix that is deposited or the degree of breakdown of those tissues may show further epistatic interactions; with a larger range of DNA variants available showing epistatic interactions, the differences in effect of allele substitution found in different breeds might be fully calculated out. While this study cannot test the general frequency with which epistasis occurs, this model shows that epistasis can be identified for protein–protein interactions for relatively small genetic effects associated with QTL, which suggests that epistasis may be quite common given the large number of protein–protein interactions that exist.

Acknowledgments

We thank A. Reverter and T. Dixon for comments on the manuscript, and three anonymous reviewers whose comments and suggestions on methods greatly improved the manuscript through several rounds of review. We also thank G. S. Harper, V. H. Oddy, and G. Moser for discussing meat tenderness. We thank Commonwealth Scientific and Industrial Research Organization, University of New England, New South Wales Department of Primary Industries, and Queensland Department of Primary Industries and Fisheries for access to the CRC1 DNA and database, and Meat and Livestock Australia and the Cooperative Research Centre for Cattle and Beef Quality for financial support (W.B.).

References

  • Altschul, S. F., W. Gish, W. Miller, E. W. Myers and D. J. Lipman, 1990. Basic local alignment search tool. J. Mol. Biol. 215: 403–410. [PubMed]
  • Barendse, W. J., 2002. DNA markers for meat tenderness. Patent application WO02064820.
  • Barendse, W., R. J. Bunch, B. E. Harrison and M. B. Thomas, 2006. The growth hormone 1 GH1:c.457C>G mutation is associated with relative fat distribution in intra-muscular and rump fat in a large sample of Australian feedlot cattle. Anim. Genet. 37: 211–214. [PubMed]
  • Barendse, W., R. J. Bunch, J. W. Kijas and M. B. Thomas, 2007. The effect of genetic variation of the retinoic acid receptor-related orphan receptor C gene on fatness in cattle. Genetics 175: 843–853. [PMC free article] [PubMed]
  • Barton, N. H., and M. Turelli, 2004. Effects of genetic drift on variance components under a general model of epistasis. Evolution 58: 2111–2132. [PubMed]
  • Bateson, W., 1909. Mendel's Principles of Heredity. Cambridge University Press, Cambridge, UK.
  • Beadle, G. W., and B. Ephrussi, 1936. The differentiation of eye pigments in Drosophila as studied by transplantation. Genetics 21: 225–247. [PMC free article] [PubMed]
  • Belcastro, A. N., L. D. Shewchuk and D. A. Raj, 1998. Exercise-induced muscle injury: A calpain hypothesis. Mol. Cell. Biochem. 179: 135–145. [PubMed]
  • Betts, R., and J. Anagli, 2004. The beta- and gamma-CH2 of B27-WT's Leu(11) and Ile(18) side chains play a direct role in calpain inhibitions. Biochemistry 43: 2596–2604. [PubMed]
  • Burrow, H. M., 1998. The effects of inbreeding on productive and adaptive traits and temperament of tropical beef cattle. Livest. Prod. Sci. 55: 227–243.
  • Carlborg, Ö., and C. S. Haley, 2004. Epistasis: Too often neglected in complex trait studies? Nat. Rev. Genet. 5: 618–625. [PubMed]
  • Carlborg, Ö., L. Jacobsson, P. Ahgren, P. Siegel and L. Andersson, 2006. Epistasis and the release of genetic variation during long-term selection. Nat. Genet. 38: 418–420. [PubMed]
  • Casas, E., S. D. Shackelford, J. W. Keele, R. T. Stone, S. M. Kappes et al., 2000. Quantitative trait loci affecting growth and carcass composition of cattle segregating alternate forms of myostatin. J. Anim. Sci. 78: 560–569. [PubMed]
  • Casas, E., S. N. White, T. L. Wheeler, S. D. Shackelford, M. Koohmaraie et al., 2006. Effects of calpastatin and mu-calpain markers in beef cattle on tenderness traits. J. Anim. Sci. 84: 520–525. [PubMed]
  • Cheverud, J. M., and E. J. Routman, 1995. Epistasis and its contribution to genetic variance-components. Genetics 139: 1455–1461. [PMC free article] [PubMed]
  • Cockerham, C. C., 1954. An extension of the concept of partitioning hereditary variance for analysis of covariances among relatives when epistasis is present. Genetics 39: 859–882. [PMC free article] [PubMed]
  • den Dunnen, J. T., and S. E. Antonarakis, 2000. Mutation nomenclature extensions and suggestions to describe complex mutations: a discussion. Hum. Mut. 15: 7–12. [PubMed]
  • Drinkwater, R. D., Y. Li, I. Lenane, G. P. Davis, R. Shorthose et al., 2006. Detecting quantitative trait loci affecting beef tenderness on bovine chromosome 7 near calpastatin and lysyl oxidase. Aust. J. Exp. Agr. 46: 159–164.
  • Efron, B., and R. Tibshirani, 1991. Statistical data analysis in the computer age. Science 253: 390–395. [PubMed]
  • Ewing, B., L. Hillier, M. C. Wendl and P. Green, 1998. Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res. 8: 175–185. [PubMed]
  • Fisher, R. A., 1918. The correlation between relatives on the supposition of Mendelian inheritance. Trans. Roy. Soc. Edin. 52: 399–433.
  • Garcia, M. D., J. J. Michal, C. T. Gaskins, J. J. Reeves, T. L. Ott et al., 2006. Significant association of the calpastatin gene with fertility and longevity in dairy cattle. Anim. Genet. 37: 304–305. [PubMed]
  • Geldermann, H., 1975. Investigations on inheritance of quantitative characters in animals by gene markers. I. Methods. Theor. Appl. Genet. 46: 319–330. [PubMed]
  • Gilmour, A. R., R. Thompson and B. R. Cullis, 1995. Average information REML: an efficient algorithm for variance parameter estimation in linear mixed models. Biometrics 51: 1440–1450.
  • Gjuvsland, A. B., B. J. Hayes, S. W. Omholt and Ö. Carlborg, 2007. Statistical epistasis is a generic feature of gene regulatory networks. Genetics 175: 411–420. [PMC free article] [PubMed]
  • Goll, D. E., V. F. Thompson, H. Q. Li, W. Wei and J. Y. Cong, 2003. The calpain system. Physiol. Rev. 83: 731–801. [PubMed]
  • Gordon, D., C. Abajian and P. Green, 1998. Consed: a graphical tool for sequence finishing. Genome Res. 8: 195–202. [PubMed]
  • Hanotte, O., D. G. Bradley, J. W. Ochieng, Y. Verjee, E. W. Hill et al., 2002. African pastoralism: genetic imprints of origins and migrations. Science 296: 336–339. [PubMed]
  • Hansen, T. F., and G. P. Wagner, 2001. Modeling genetic architecture: A multilinear theory of gene interaction. Theor. Pop. Biol. 59: 61–86. [PubMed]
  • Hill, W. G., and A. Robertson, 1968. Linkage disequilibrium in finite populations. Theor. Appl. Genet. 38: 226–231. [PubMed]
  • Horikawa, Y., N. Oda, N. J. Cox, X. Q. Li, M. Orho-Melander et al., 2000. Genetic variation in the gene encoding calpain-10 is associated with type 2 diabetes mellitus. Nat. Genet. 26: 163–175. [PubMed]
  • James, T., D. Matzelle, R. Bartus, E. L. Hogan and N. L. Banik, 1998. New inhibitors of calpain prevent degradation of cytoskeletal and myelin proteins in spinal cord in vitro. J. Neurosci. Res. 51: 218–222. [PubMed]
  • Kempthorne, O., 1954. The correlation between relatives in a random mating population. Proc. R. Soc. Lond. Ser. B Biol. Sci. 143: 103–113. [PubMed]
  • Kijas, J. M. H., R. Wales, A. Tornsten, P. Chardon, M. Moller et al., 1998. Melanocortin receptor 1 (MC1R) mutations and coat color in pigs. Genetics 150: 1177–1185. [PMC free article] [PubMed]
  • Koohmaraie, M., 1996. Biochemical factors regulating the toughening and tenderisation processes of meat. Meat Sci. 43: 193–201. [PubMed]
  • Kyte, J., and R. F. Doolittle, 1982. A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157: 105–132. [PubMed]
  • Luo, L. J., Z.-K. Li, H. W. Mei, Q. Y. Shu, R. Tabien et al., 2001. Overdominant epistatic loci are the primary genetic basis of inbreeding depression and heterosis in rice. II. Grain yield components. Genetics 158: 1755–1771. [PMC free article] [PubMed]
  • Lynch, M., and J. B. Walsh, 1998. Genetics and Analysis of Quantitative Traits. Sinauer Associates, Sunderland, MA.
  • MacHugh, D. E., M. D. Shriver, R. T. Loftus, P. Cunningham and D. G. Bradley, 1997. Microsatellite DNA variation and the evolution, domestication and phylogeography of taurine and zebu cattle (Bos taurus and Bos indicus). Genetics 146: 1071–1086. [PMC free article] [PubMed]
  • Mao, Y. C., N. R. London, L. Ma, D. Dvorkin and Y. Da, 2006. Detection of SNP epistasis effects of quantitative traits using an extended Kempthorne model. Physiol. Genomics 28: 46–52. [PubMed]
  • Mellgren, R. L., R. D. Lane and M. T. Mericle, 1989. The binding of large calpastatin to biologic membranes is mediated in part by interaction of an amino terminal region with acidic phospholipids. Biochim. Biophys. Acta 999: 71–77. [PubMed]
  • Melloni, E., M. Averna, R. Stifanese, R. De Tullio, E. Defranchi et al., 2006. Association of calpastatin with inactive calpain: A novel mechanism to control the activation of the protease? J. Biol. Chem. 281: 24945–24954. [PubMed]
  • Morris, C. A., N. G. Cullen, S. M. Hickey, P. M. Dobbie, B. A. Veenvliet et al., 2006. Genotypic effects of calpain 1 and calpastatin on the tenderness of cooked m. longissimus dorsi steaks from Jersey x Limousin, Angus and Hereford-cross cattle. Anim. Genet. 37: 411–414. [PubMed]
  • Nickerson, D. A., V. O. Tobe and S. L. Taylor, 1997. PolyPhred: Automating the detection and genotyping of single nucleotide substitutions using fluorescence-based resequencing. Nucleic Acids Res. 25: 2745–2751. [PMC free article] [PubMed]
  • Olson, T. A., 1999. Genetics of color variation, pp. 33–53 in The Genetics of Cattle, edited by R. Fries and A. Ruvinsky. CABI, Wallingford, UK.
  • Page, B. T., E. Casas, M. P. Heaton, N. G. Cullen, D. L. Hyndman et al., 2002. Evaluation of single-nucleotide polymorphisms in CAPN1 for association with meat tenderness in cattle. J. Anim. Sci. 80: 3077–3085. [PubMed]
  • Perry, D., W. R. Shorthose, D. M. Ferguson and J. M. Thompson, 2001. Methods used in the CRC program for the determination of carcass yield and beef quality. Aust. J. Exp. Agr. 41: 953–957.
  • Reverter, A., D. J. Johnston, D. M. Ferguson, D. Perry, M. E. Goddard et al., 2003. Genetic and phenotypic characterisation of animal, carcass, and meat quality traits from temperate and tropically adapted beef breeds. 4. Correlations among animal, carcass, and meat quality traits. Aust. J. Agric. Res. 54: 149–158.
  • Routman, E. J., and J. M. Cheverud, 1997. Gene effects on a quantitative trait: Two-locus epistatic effects measured at microsatellite markers and at estimated QTL. Evolution 51: 1654–1662.
  • Schenkel, F. S., J. R. Miller, Z. Jiang, I. B. Mandell, X. Ye et al., 2006. Association of a single nucleotide polymorphism in the calpastatin gene with carcass and meat quality traits of beef cattle. J. Anim. Sci. 84: 291–299. [PubMed]
  • Tiret, L., A. Bonnardeaux, O. Poirier, S. Ricard, P. Marques-Vidal et al., 1994. Synergistic effects of angiotensin-converting enzyme and angiotensin-II type 1 receptor gene polymorphisms on risk of myocardial infarction. Lancet 344: 910–913. [PubMed]
  • Upton, W., H. M. Burrow, A. Dundon, D. L. Robinson and E. B. Farrell, 2001. CRC breeding program design, measurements and database: methods that underpin CRC research results. Aust. J. Exp. Agr. 41: 943–952.
  • Van Eenennaam, A. L., J. Li, R. M. Thallman, R. L. Quaas, M. E. Dikeman et al., 2007. Validation of commercial DNA tests for quantitative beef quality traits. J. Anim. Sci. 85: 891–900. [PubMed]
  • Venables, W. N., and B. D. Ripley, 2000. Modern Applied Statistics with S-PLUS. Springer-Verlag, New York.
  • Wang, D. L., J. Zhu, Z. K. Li and A. H. Paterson, 1999. Mapping QTLs with epistatic effects and QTLxenvironment interactions by mixed linear model approaches. Theor. Appl. Genet. 99: 1255–1264.
  • Weir, B. S., 1996. Genetic Data Analysis II. Sinauer Associates, Sunderland, MA.
  • Wheeler, T. L., L. V. Cundiff, S. D. Shackelford and M. Koohmaraie, 2001. Characterization of biological types of cattle (Cycle V): Carcass traits and longissimus palatability. J. Anim. Sci. 79: 1209–1222. [PubMed]
  • Whipple, G., M. Koohmaraie, M. E. Dikeman, J. D. Crouse, M. C. Hunt et al., 1990. Evaluation of attributes that affect longissimus muscle tenderness in Bos taurus and Bos indicus cattle. J. Anim. Sci. 68: 2716–2728. [PubMed]
  • White, S. N., E. Casas, T. L. Wheeler, S. D. Shackelford, M. Koohmaraie et al., 2005. A new single nucleotide polymorphism in CAPN1 extends the current tenderness marker test to include cattle of Bos indicus, Bos taurus, and crossbred descent. J. Anim. Sci. 83: 2001–2008. [PubMed]
  • Wright, S., 1980. Genic and organismic selection. Evolution 34: 825–843.
  • Zeng, Z.-B., T. Wang and W. Zou, 2005. Modeling quantitative trait loci and interpretation of models. Genetics 169: 1711–1725. [PMC free article] [PubMed]

Articles from Genetics are provided here courtesy of Genetics Society of America
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...