Effects of the Expressions and Variants of the CAST Gene on the Fatty Acid Composition of the Longissimus Thoracis Muscle of Grazing Sonid Sheep

Simple Summary This study aimed to evaluate the relationship between the expression levels of the CAST gene and the fatty acid (FA) composition in the longissimus thoracis (LL) muscle and to identify novel variants of CAST and perform association analysis with the FA composition in grazing Sonid sheep. The correlation results showed that high expression levels of CAST are correlated with better FA compositions and classes in LL. Four c.646G>C (G216R), c.1210C>T (R404C), c.1437G>A (479T), and c.2097C>T (699G) mutations were identified in the CAST gene of Sonid sheep. The association studies showed that c.1210C>T is associated with C14:0, C18:0, C18:1n9c, C18:3n3, n3, C12:0, n6, and n6/n3; c.646G>C and c.1437G>A in linkage disequilibrium-Mongolia (r2 = 0.964) were associated with C14:0, C18:0, SFA, C18:1n9c, n3, C10:0, C18:1n9t, and n6/n3; and c.2097C>T was associated with C18:3n3, n3, C10:0, and n6/n3 in the LL of Sonid sheep. Thus, the correlation results and associated mutations were expected to be genetic selection markers for the FA composition and meat quality of Sonid sheep muscle and provide new insight into sheep meat quality traits influenced by the ovine CAST gene. Abstract Fatty acid (FA) composition has an important impact on the nutrition and flavor of meat, and on consumer health, and is receiving more attention in the sheep industry. This study aimed to evaluate the relationship between the expression levels of the CAST gene and the FA composition in the longissimus thoracis (LL) muscle, to identify novel variants of CAST, and to perform association analysis with the FA composition in grazing Sonid lambs. The correlation results showed that high expression levels of CAST are correlated with better FA compositions and classes in LL. For association studies, the results showed that c.1210C>T and c.1437G>A in LD-M, and c.2097C>T mutations are associated with some compositions and classes of FA in the LL of grazing Sonid sheep. Two missense c.646G>C (G216R) and c.1210C>T (R404C) mutations were predicted to influence the Calpain_inhib domains of CAST. Thus, the correlation results and associated mutations are expected to be genetic selection markers for the FA composition and meat quality of grazing Sonid lamb muscle and provide new insights into sheep meat quality traits influenced by the ovine CAST gene.


Introduction
In China, the annual lamb consumption per person increased from 2.99 kg in 2011 to 3.93 kg in 2021. China has the largest number of sheep in the world, with nearly 200 million before 2022. Among them, the Mongolia sheep population is the most widely distributed and the most numerous. Mongolia sheep, an old and primitive sheep breed, is mainly distributed in the grassland, desert, and agricultural areas in northern China, Mongolia, and Central Asia and is the common ancestor of Chinese short fat-tailed sheep breeds, such as the Sonid, Ujimqin, Hulunbuir, Tan, Bayanbulak, Small-tailed Han, Duolang, and Hu sheep (China National Commission of Animal Genetic Resources) [1]. Recently, the meat of Mongolia sheep populations was recognized as a natural green food and has become increasingly popular in China.
The fatty acid (FA) composition of lamb meat is an important trait for the determination of flavor and nutrition, and this meat can provide essential FA for human health [2]. The types and levels of FA are influenced by polygenic and environmental factors [3,4]. A better knowledge of the molecular architecture of meat FA composition is important as it may generate new opportunities for more effective marker-assisted breeding, leading to economic benefits for the sheep industry [5]. Nevertheless, compared to beef and pork, the studies of genetic effects, including the mutation and expression of candidate genes, on the FA composition of lamb are few.
Calpastatin (CAST) is involved in the calpain-calpastatin system, which influences many important processes, including muscle development and growth [6], and is also known to regulate the degradation of myofibrillar proteins both in living and in postmortem muscle tissue [7,8]. To date, many polymorphisms of the CAST gene have been identified and associated with meat quality traits, such as tenderness, color, and intramuscular fat (IMF) content, as well as carcass traits in cattle and pigs [9][10][11][12][13][14]. However, research focusing on the relationship between the CAST gene and meat quality, especially FA composition, in sheep is limited. A report showed that a polymorphic variant by PCR-SSCP in intron 12 of the ovine CAST gene is associated with IMF content in two Polish synthetic lines of sheep [15]. Therefore, the CAST gene could be considered a potential candidate gene for meat quality traits in sheep.
Thus, the aims of this study reported herein were to (1) evaluate the effects of the expression levels of CAST on the FA composition in the longissimus thoracis muscle of sheep, (2) identify novel variants of CAST and perform association analysis with the FA composition in Sonid sheep, and (3) predict the effect of the novel variants on the feature and structure of the mRNA and protein of the ovine CAST gene. Our study may facilitate effective marker-assisted selection to promote the meat quality in Mongolia sheep populations and provide new insights into the effect of the ovine CAST gene on sheep meat quality traits.

Animals and Samples
The study was carried out on 378 castrated ram lambs (approximately 6 months old) of Sonid sheep, born in 2020. The lambs were raised until weaning in an outdoor extensive grazing system, feeding on natural pastures as usual in the Sonid grassland of Inner Mongolia. After weaning, the castrated male lambs were grazed under the same conditions as usual in the Sonid grassland until 6 months old. All the experimental animals were healthy. The 378 lambs were the progeny of different sires and dams, and any relationship between the parents was not recorded. These lambs were the progeny of more than 20 unrelated sires.
Lambs were slaughtered in accordance with the Chinese industry standard (NY/T 1564-2021). After being slaughtered, the longissimus thoracis (LL) muscle tissues (at least 500 g between the 12 and 13th ribs) of 378 animals (left half of the carcass) were removed on ice, taken to the laboratory, frozen, and stored at −35 • C until analysis for FA composition. Of the 378 lambs, the LL samples of 90 animals were randomly selected and frozen in Animals 2023, 13,195 3 of 17 liquid nitrogen immediately after slaughter and stored at −80 • C until used for RNA extraction. Three lambs were randomly selected to sample the LL, semitendinosus, heart, subcutaneous fat, liver, lung, spleen, kidney, large intestine, small intestine, and stomach, which were taken and stored in liquid nitrogen for tissue expression profile analysis of the CAST gene.

Determination of Fatty Acid Composition
The fatty acid profile was determined after extracting the total lipids and performing esterification and methylation processes using the method in accordance with Folch et al. [16]. After the extracted lipid was converted to fatty acid methyl esters (FAMEs), the FA profiles were analyzed using a gas chromatography-flame ionization detector (Varian 450-GC, Bruker Daltonics Inc., Fremont, CA, USA). Injector and detector temperatures were held at 260 • C. The oven temperature was initially 120 • C for 5 min and was increased to 230 • C at 3 • C/min and held at that temperature for 3 min, and then increased to 240 • C at a rate of 1.5 • C/min and held for 5 min. The samples containing FAMEs in hexane (1.0 µL) were injected through the split injection port (10:1) onto an RT-2560 capillary column (100 m length, 0.25 mm internal diameter, 0.20 µm film thickness; RESTEK, Bellefonte, PA, USA). Individual FAMEs were identified by comparing their retention times with those of authenticated standards, and the results were expressed as a percentage of the total FAMEs. In addition, individual FAs were used to calculate the sums of saturated fatty acids (SFA), monounsaturated fatty acids (MUFA), and polyunsaturated fatty acids (PUFA).

RNA Extraction and Real-Time Quantitative Amplification
Total RNA from the LL of 90 Sonid sheep was extracted using RNAiso Plus (Takara Bio Inc., Dalian, China), and cDNA synthesis was performed using a PrimeScript TM RT Reagent Kit with gDNA Eraser (Takara Bio Inc., Dalian, China) according to the manufacturer's protocol. Real-time polymerase chain reaction (PCR) analyses were performed on a BIO-RAD Real-time PCR system in a 20 µL reaction volume, including 10 µL of 2 × SYBR ® Premix Ex Taq TM II (Takara Bio Inc., Dalian, China), 0.8 µL of forward and reverse primers (10 µL), 0.4 µL of ROX Reference Dye, 2.0 µL of template cDNA, and 6.0 µL of nuclease-free water. The real-time PCR conditions were as follows: 95 • C for 30 s, 40 cycles of 95 • C for 5 s, and 60 • C for 34 s. We tested the gene suitability of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene as the internal housekeeping gene in this experiment. The gene expression stability value was less than 0.05, which met the stability required to be a housekeeping gene [17]. The expression of each gene was normalized against GAPDH (Table 1), and relative expression levels were determined using the 2 −∆∆Ct method [18].

Resequencing and Variant Detection in CAST
Genomic DNA was isolated from the LL using the Wizard ® Genomic DNA Purification Kit (Promega Corp., Madison, WI, USA). The quality and quantity of the extracted DNA were tested with a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). The quality of the genomic DNA preparations was also verified using agarose gel electrophoresis. Genomic DNA with absorbance at OD260 nm/OD280 nm of ≥1.80 and no RNA or protein in the agarose gel was used as a template. The DNA of 10 Sonid sheep was randomly selected and sent to Beijing Novogene Company for 10 × genome resequencing.

Polymorphism Genotyping Using iPLEX MassARRAY
Four target variants were genotyped with the MassARRAY ® SNP genotyping system (Agena Bioscience, San Diego, CA, USA) in the 378 Sonid sheep. PCR and extension primers were designed from sequences containing each target mutation and~100 upstream and downstream bases with Assay Design Suite (http://agenabio.com/assay-design-suite-20 -software) using the default settings ( Table 2). The genotype of each SNP was analyzed using the Sequenom MassARRAY iPLEX platform (Sequenom, San Diego, CA, USA) [19]. The resulting data were analyzed using MassARRAY Typer 4.0 Analyzer software (Agena Bioscience, San Diego, CA, USA).

Statistical Analysis
Genotypic and allelic frequencies and Hardy-Weinberg equilibrium were calculated for the Sonid sheep in this study. Population genetic indices, observed heterozygosity (H o ), expected heterozygosity (H e ), effective allele numbers (n e ), and the polymorphism information content (PIC) were calculated using Nei's methods [22]. The allelic frequency of each variant was compared using a χ 2 test. The LD, including D' and r 2 , was assessed using HaploView 4.2 software [23]. Haplotypes were obtained using SHEsis [24]. The associations between FA composition/class and different genotypes of the four novel variants and three haplotypes in the Sonid sheep were analyzed using SPSS 24.0 (SPSS, Inc., Chicago, IL, USA). The statistical linear model was as follows: Y i = µ + G i + e i , where Y i is the observed value of fatty acid composition traits, µ is the mean of each measurement, G i means the fixed effect of the genotypes, and e i means standard error. When the number of sheep with a given genotype was less than 10, their associations and effects could not be reliably estimated. Therefore, animals with this genotype were excluded from the analysis. The Bonferroni correction was used to adjust p values [25]. Correlation analyses between any two traits of FA compositions and FA classes, and between expression levels of CAST and each of the FA compositions and FA classes, were calculated using SPSS 24.0, and results are presented as the Pearson correlation coefficient (SPSS, Inc., Chicago, IL, USA).

CAST Gene Expression Profiles in Sheep
CAST gene expression profiles in various tissues of Sonid sheep are shown in Figure  2a. CAST gene expressions were significant higher in heart tissue, subcutaneous fat, the LL, and semitendinosus muscle than in other tissues ( Figure 2a; statistical difference of the CAST gene expression between different tissues are shown in Table S3).

CAST Gene Expression Profiles in Sheep
CAST gene expression profiles in various tissues of Sonid sheep are shown in Figure 2a.
CAST gene expressions were significant higher in heart tissue, subcutaneous fat, the LL, and semitendinosus muscle than in other tissues (Figure 2a; statistical difference of the CAST gene expression between different tissues are shown in Table S3). CAST gene expressions were significantly positively correlated with C18:0 and C18:3n3 (p < 0.01; Figure 2b,c).      (Table S4).

Linkage Disequilibrium Analysis of Novel Variants of CAST
To identify the linkage relationships among the four variants, D' and r 2 were estimated for the experimental Sonid sheep. The c.646G>C and c.1437C>T variants were in nearly complete LD in the experimental Sonid sheep by r 2 = 0.964 ( Figure 5). Thus, these LD groups were analyzed together and marked as a single locus, designated LD-M. D' and r 2 for the experimental Sonid sheep are shown in Table S5.  (Table S4).

Linkage Disequilibrium Analysis of Novel Variants of CAST
To identify the linkage relationships among the four variants, D' and r 2 were estimated for the experimental Sonid sheep. The c.646G>C and c.1437C>T variants were in nearly complete LD in the experimental Sonid sheep by r 2 = 0.964 ( Figure 5). Thus, these LD groups were analyzed together and marked as a single locus, designated LD-M. D' and r 2 for the experimental Sonid sheep are shown in Table S5.

Associations between Variants in CAST and Fatty Acid
The significant effects of the four novel variants of the CAST gene on FA compositions and FA classes in the LL of Sonid sheep are shown in Table 3, and other results are presented in Table S6. For c.1210C>T, the contents of C14:0 (p < 0.01), C18:0 (p < 0.05), C18:1n9c (p < 0.05), and C18:3n3 (p < 0.05) of the CC genotype individuals were signifi-

Associations between Variants in CAST and Fatty Acid
The significant effects of the four novel variants of the CAST gene on FA compositions and FA classes in the LL of Sonid sheep are shown in Table 3, and other results are presented in Table S6. For c.1210C>T, the contents of C14:0 (p < 0.01), C18:0 (p < 0.05), C18:1n9c (p < 0.05), and C18:3n3 (p < 0.05) of the CC genotype individuals were significantly higher than those of the CT genotype individuals; the contents of n3-PUFA of the CC genotype individuals were significantly higher than those of the CT and TT genotype individuals (p < 0.01); the content of C12:0 of the CC genotype individuals was significantly lower than that of CT genotype individuals (p < 0.05); the content of n6-PUFA of the CC genotype individuals was significantly lower than that of TT genotype individuals (p < 0.01); and the ratio of n6/n3-PUFA of the CC genotype individuals was significantly lower than that of the CT and TT genotype individuals (p < 0.01; Table 3). For c.1437G>A in LD-M, compared to the GA genotype, the GG genotype individuals had significantly higher contents of C14:0 (p < 0.01), C18:0 (p < 0.01), SFA (p < 0.05), C18:1n9c (p < 0.05), and n3-PUFA (p < 0.01), and the GG genotype individuals had significantly lower C10:0 (p < 0.05) and C18:1n9t (p < 0.05) contents and the ratio of n6/n3-PUFA (p < 0.01; Table 3). For c.2097C>T, sheep with the CC genotype had significantly higher C18:3n3 and n3-PUFA levels when compared to sheep with the CT genotype (p < 0.05), and the content of C10:0 and the ratio of n6/n3-PUFA of the CC genotype individuals were significantly lower than those of the CT genotype individuals (p < 0.05 and p < 0.01, respectively) in the LL of Sonid sheep (Table 3).

Associations between Haplotypes of CAST and Fatty Acid
Using the online tool SHEsis, different haplotypes were constructed in the experimental population of Sonid sheep in order to analyze the associations between FA composition and haplotypes. A haplotype with a frequency of >3% was considered a distinguishable haplotype, while haplotypes with a relative frequency of <3% were pooled into a single group. Thus, haplotype 1 (GCGC, H1) had the highest frequency (0.798), followed by haplotype 2 (CTAT, H2, 0.101) and haplotype 3 (GTGC, H3, 0.053); see Table 4.  Table 4 shows the significant effects of the haplotype of the CAST gene on FA composition and FA classes in the LL of Sonid sheep, and other results are presented in Table S7. By haplotype-based association analyses, we found that the content of C12:0 of the H1H1 haplotype individuals was significantly lower than that of the H1H3 haplotype individuals (p < 0.05), and the content of heneicosylic acid (C21:0) of the H1H3 haplotype individuals were significantly higher than that of H1H1 (p < 0.01) and H1H2 (p < 0.05) haplotype individuals (Table 5). Compared to the H1H2 haplotype, the H1H1 haplotype individuals had significantly higher contents of C18:3n3 (p < 0.05) and n3-PUFA (p < 0.01; Table 5). The ratio of n6/n3-PUFA of the H1H1 haplotype individuals was significantly lower than that of H1H2 (p < 0.05) and H1H3 (p < 0.05) haplotype individuals (Table 5).

Amino Acid Sequence Analysis of Ovine CAST
ProtParam was used to predict the physicochemical properties of amino acid sequences. The molecular weight and isoelectric point of ovine CAST were 78,998.90 Da and 6.08, respectively. The amino acid composition of the CAST protein showed that the highest proportion of the CAST protein was 13.4% for lysine and the lowest was 0.1% for tryptophan.
TMHMM was used to predict subcellular localization based on the protein function and the physical and chemical environment in vivo. There was no transmembrane helix position on the CAST protein (Figure 7a). We predicted 1 N-glycosylation site (Asn) and 95 phosphorylation sites (composed of 59 Ser, 34 Thr, and 2 Tyr) in CAST (Figure 7b,c).

Amino Acid Sequence Analysis of Ovine CAST
ProtParam was used to predict the physicochemical properties of amino acid sequences. The molecular weight and isoelectric point of ovine CAST were 78,998.90 Da and 6.08, respectively. The amino acid composition of the CAST protein showed that the highest proportion of the CAST protein was 13.4% for lysine and the lowest was 0.1% for tryptophan.
TMHMM was used to predict subcellular localization based on the protein function and the physical and chemical environment in vivo. There was no transmembrane helix position on the CAST protein (Figure 7a). We predicted 1 N-glycosylation site (Asn) and 95 phosphorylation sites (composed of 59 Ser, 34 Thr, and 2 Tyr) in CAST (Figure 7b,c).

Multiple Sequence Alignment and Phylogenetic
To construct phylogenetic trees for the gene identified as homologous to ovine CAST, the amino acid sequence of CAST in some models and domesticated animals was used. After that, a molecular phylogenetic tree was constructed using the maximum-parsimony method, and 1000 Bootstrap replications were carried out using MEGA software (Version X).

Effect of Variants on the mRNA Secondary Structure of the Ovine CAST Gene
Using minimum free energy (MFE)-based RNAfold platform analysis, the results can show a difference in the secondary structure of a point mutation [26]. The MFE of exonic mRNA sequences with wild types at the c.1437G>A in exon 19 and c.2097C>T in exon 26 were −16.20 kcal/mol and −13.40 kcal/mol, respectively. The MFE of the c.1437G>A and c.2097C>T mutations reduced to −11.10 kcal/mol and −11.50 kcal/mol, respectively. As a result of these mutations, the mRNA secondary structure of CAST also changed (Figure 8).
show a difference in the secondary structure of a point mutation [26]. The MFE of exonic mRNA sequences with wild types at the c.1437G>A in exon 19 and c.2097C>T in exon 26 were −16.20 kcal/mol and −13.40 kcal/mol, respectively. The MFE of the c.1437G>A and c.2097C>T mutations reduced to −11.10 kcal/mol and −11.50 kcal/mol, respectively. As a result of these mutations, the mRNA secondary structure of CAST also changed ( Figure  8). The secondary structure for the wild-type mRNA sequence of exon 26 with the C allele at the c.2097C>T site and the secondary structure for the mRNA sequence of exon 26 with the T allele at the c.2097C>T site. The structure above is colored according to base-pairing probabilities.

Effect of Variants on the Secondary and Tertiary Structure of the Ovine CAST Protein
SOPMA software was used to analyze the secondary structures of the coding proteins of ovine CAST, and the results showed that the percentages of the alpha helix, beta turn, random coil, and extended strand of CAST (CAST Protein ID: NP_001009788.1) were 31.12%, 2.90%, 63.49%, and 2.49%, respectively (Figure 9a). The black arrows indicate four changes in the secondary structure of the protein when the sequence of amino acids changed from glycine to arginine at the c.646G>C (G216R) locus (Figure 9b), and the percentages of the alpha helix, beta turn, random coil, and extended strand of CAST changed to 31.26%, 2.49%, 63.62%, and 2.63%, respectively. The black arrows indicate three changes in the secondary structure of the protein when the amino acid sequence changed from arginine to cysteine of at the c.1210C>T (R404C) locus. (Figure 9c), and the percentages of the alpha helix, beta turn, random coil, and extended strand of CAST changed to 31.40%, 3.04%, 62.79%, and 2.77%, respectively.

Effect of Variants on the Secondary and Tertiary Structure of the Ovine CAST Protein
SOPMA software was used to analyze the secondary structures of the coding proteins of ovine CAST, and the results showed that the percentages of the alpha helix, beta turn, random coil, and extended strand of CAST (CAST Protein ID: NP_001009788.1) were 31.12%, 2.90%, 63.49%, and 2.49%, respectively (Figure 9a). The black arrows indicate four changes in the secondary structure of the protein when the sequence of amino acids changed from glycine to arginine at the c.646G>C (G216R) locus (Figure 9b), and the percentages of the alpha helix, beta turn, random coil, and extended strand of CAST changed to 31.26%, 2.49%, 63.62%, and 2.63%, respectively. The black arrows indicate three changes in the secondary structure of the protein when the amino acid sequence changed from arginine to cysteine of at the c.1210C>T (R404C) locus. (Figure 9c), and the percentages of the alpha helix, beta turn, random coil, and extended strand of CAST changed to 31.40%, 3.04%, 62.79%, and 2.77%, respectively.
AlphaFold and SWISS-MODEL were used to predict the tertiary structure of the ovine CAST protein (UniProtKB: Q95208). The calcium-dependent complex between calpain-2 and calpastatin (sequence identity 67.06%) was used as a template for modeling (PDB ID: 3DF0.1.C). It was predicted that the amino acid change at the G216R (c.646G>C) site would cause visible changes in the CAST structure (Figure 9d).
AlphaFold and SWISS-MODEL were used to predict the tertiary structure of the ovine CAST protein (UniProtKB: Q95208). The calcium-dependent complex between calpain-2 and calpastatin (sequence identity 67.06%) was used as a template for modeling (PDB ID: 3DF0.1.C). It was predicted that the amino acid change at the G216R (c.646G>C) site would cause visible changes in the CAST structure (Figure 9d).

Discussion
The CAST gene has been considered a major functional gene related to carcass traits and meat quality in cattle and pig [9][10][11][12][13][14]. However, studies on associations of the expression levels and genetic variants of CAST with FA in sheep are rare. In this study, the results of CAST expression analyses showed that the expression levels of CAST are significantly positive correlated with C18:0 and C18:3n and show a trend of positive correlation with n3, medium-chain fatty acids (MCFA), and long-chain fatty acids (LCFA). In contrast, the expression levels of CAST showed a trend of a negative correlation with n6, short-chain fatty acids (SCFA), and n6/n3. Therefore, the high expression levels of CAST are more desirable for a better FA composition. Sheep meat generally contains higher levels of SFA, which are widely correlated with health problems, such as heart disease, stroke, and obesity [27], so consumers favor leaner meats containing less SFA and high levels of PUFA [28,29]. PUFA, mainly long-chain omega-3 FAs, are considered beneficial for human health as they reduce serum low-density lipoprotein-cholesterol and total cholesterol and modulate immune functions [30]. Additionally, the desirable sensorial characteristic of meat is associated with PUFA and MUFA [31]. Importantly, sheep meat is rich in omega-3 FA, which is beneficial for human health and immunity [32]. Meat production with a higher PUFA and lower SFA content is, therefore, important to improve human health without requiring substantial changes in customers' habits of meat consumption [33]. Thus, based on the results of expression analyses, the CAST gene could be considered an important candidate gene for a better FA composition and class of meat quality in the sheep industry.
The missense c.646G>C (G216R) and c.1210C>T (R404C) mutations, as well as the silent c.1437G>A (479T) and c.2097C>T (699G) mutations, are the first variants in the CAST gene to have a confirmed association with FA in sheep. The amino acid sequence is not altered by synonymous mutations, but they can affect mRNA expression, splicing, stability [34][35][36], and secondary structure [37,38], as well as protein translation, folding [39], and function [40]. In this study, the MFE of the silent c.1437G>A (479T) and c.2097C>T (699G) mutations changed from −16.20 kcal/mol to −11.10 kcal/mol and −13.40 kcal/mol to −11.50 kcal/mol, respectively. As MFE decreases, mRNA's secondary structure could become more stable [37,38]. Additionally, the c.1437G>A (479T) and c.2097C>T (699G) mutations are predicted to change the mRNA secondary structure of CAST. The secondary structure and stability of mRNA can be altered by even single base-pair exchanges, according to some studies [41,42]. According to the results of association studies and bioinformatics analysis, the synonymous c.1437G>A (479T) and c.2097C>T (699G) mutations might be potentially functional mutations in CAST for the meat quality of sheep. They could affect the FA synthesis in sheep by altering the mRNA stability and secondary structure of CAST. However, further experiments are required to test this hypothesis.
Among the two missense mutations (c.646G>C (G216R) and c.1210C>T (R404C)) identified in this study, c.646G>C (G216R) was modeled in the tertiary structure of the CAST protein, and it was predicted that a visible structure change would occur. Both c.646G>C (G216R) and c.1210C>T (R404C) are located within the Calpain_inhib domain (104-227 aa and 372-503 aa, respectively), which is a conserved domain predicted by SMART. CAST plays an important role in the calpain-calpastatin system, which is an endogenous calciumdependent proteinase system that mediates the proteolysis of key myofibrillar proteins during the postmortem storage of carcasses and tenderization [7,8,43]. Calpain can be inhibited by CAST, preventing calpain proteolytic activation, membrane binding, and catalytic activity [44]. This may be confirmed by the predominant isoform of CAST, which contains four inhibitory domains (Figure 6c), each of which may inhibit calpain activity, with one molecule of calpastatin able to inhibit four calpain molecules [45]. Therefore, we hypothesize that the missense mutations c.646G>C (G216R) and c.1210C>T (R404C) might change the structure of two Calpain_inhib domains to influence the interaction with calpain, which results in the dynamic change in FA synthesis and FA metabolism in sheep. Of course, this hypothesis still requires further experiments.
Based on the results of the expression and association analyses in this study, the CAST gene could be considered an important candidate gene for better FA compositions and meat quality of sheep. FA composition is influenced by diet, genetics, breed, sex, and environmental factors [46,47]. Nevertheless, FA composition is the well-defined compounds describing phenotypic traits, which are possible to improve through genetic selection. FA compositions show moderate-to-high heritability ranging from 0.15 to 0.63 [48,49]. Identification of genetic factors controlling FA composition could be implemented in breeding programs to select animals that produce higher PUFA and lower SFA in meat. Hence, the c.646G>C (G216R), c.1210C>T (R404C), c.1437G>A (479T), and c.2097C>T (699G) mutations of CAST identified in this study could be considered useful molecular markers for the purpose of optimizing the composition and class of FA in sheep breeding and industry.

Conclusions
In summary, this study revealed that the high expression levels of the CAST gene are positively correlated with the better FA composition and FA class of grazing Sonid lambs and indicated significant associations between the c.646G>C (G216R), c.1210C>T (R404C), c.1437G>A (479T), and c.2097C>T (699G) mutations of the CAST gene and the FA composition and class of grazing Sonid lambs. These association results give breeders more and higher possibilities to select sheep by different favorable alleles of these markers in the breeding of Sonid sheep. Therefore, these findings suggest that the favorable allele in each mutation of the CAST gene could be a potentially useful genetic marker for breeding programs aimed at improving the FA composition and class of meat quality of Sonid sheep.
Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/ani13020195/s1: Table S1: Descriptive statistics for the studied traits in longissimus thoracis with the number of considered Sonid sheep, the mean value, and the standard deviation (SD); Table S2: Correlation analyses between any two traits of fatty acid compositions and classes; Table S3: Statistical difference of the CAST gene expression between different tissues in Sonid sheep; Table S4: Genotypic frequencies, allelic frequencies, and diversity parameters of four polymorphisms in the Sonid population; Table S5: Linkage disequilibrium as measured by D' and r 2 among four mutations in the CAST gene; Table S6: Associations of CAST variants with fatty acid composition in the longissimus thoracis muscle of Sonid sheep; Table S7: Associations between the haplotypes of CAST and fatty acid composition in the longissimus thoracis muscle of Sonid sheep.