The animals used in this study were four adult (24 mouth old) female breeds from China: Qinchuan cattle (QC, n=176, Fufeng, Shaanxi Province, China), Pinan cattle (PN, n=136, Xinye County, Henan Province, China), Xianan cattle (XN, n=160, Biyang County, Henan Province, China) and Jiaxian Red cattle (JX, n=144; Jia County, Pingdingshan City, Henan Province). Animal care and study protocols were in accordance with the Animal Care Commission of the College of Veterinary Medicine, Northwest A&F University. Then, the cattle samples were collected from the same breeding farm and the same batch. The body size data of growth traits are accurately measured and recorded according to the same criteria. They are in a state of uniform feeding, and there is no blood relationship between the three generations. And all breeds were feed under the same conditions.

The samples obtained from the blood belong to each individual. Genomic DNA extracted by phenol chloroform method. The concentration and purity of the genomic DNA were tested and diluted to a uniform final concentration of 25 ngµL-1 for subsequent amplification assays.

The expression test sample is from XN adult female cattle (n=3). Samples included heart, liver, kidney, lung and muscle tissues. In order to research the relationship between SNP and gene transcription expression level, muscle tissues (n=12) were collected. The total RNA of all tissues was determined using the Trizol method of extraction following the manufacturer's instructions. The cDNA which we collected via RNA was reverse-transcribed using PrimeScript^™ RT Reagent Kit with gDNA Eraser (Clontech, TaKaRa).

SNP primers were designed using the ARMS-primer method (http://primer1.soton.ac.uk/primer1.html, last access: 3 March 2019) based on the DNA sequence of the SPARC gene searched in the NCBI (NC_037334.1) and linked to the re-sequencing results. And the RNA expression test primers were designed based on SPARC gene mRNA sequence (NM_174464.2). All primer sequences are shown in Table 1.

The same amount of genomic DNA from 50 different individuals of cattle was mixed into a DNA pool for amplification of the gene of interest, and the volume of PCR reaction was 10 µL: the reaction system included 25 ng of genomic DNA (25 ngµL-1), 0.5 µL of internal and external primers (10 pmolµL-1), 5 µL 2 × Taq PCR Master Mix (GeneStar, Beijing, China), and 2 µL ddH2O. The procedure is as follows: the reactants are held at 95 ∘C for 5 min, at 94 ∘C for 30 s, down to 60 ∘C for 30 s, up to 72 ∘C for 25 s, the above steps for 40 cycles, and finally at 72 ∘C for 10 min. The PCR product is then sequenced to find mutations. The determination is typically made by sequencing the product in both forward and reverse directions (Sangon, Shanghai, China).

Fluorescence quantitative detection of SPARC expression levels in different tissues was performed using a CFX 96TM real-time quantitative RCR instrument (Bio-Rad, Hercules, CA, USA). The beta-actin gene stably expressed in bovine tissues was corrected as an internal reference.

The SNP phenotype of SPARC gene was tested via the T-ARMS-PCR method, and the phenotype showed in 3.0 % agarose gel. The genotype frequency and allele frequency of SPARC gene SNPs were statistically analyzed using Excel software. In addition, SPSS V19.0 was used for the correlation analysis of SNP with various growth traits. According to the experimental design, a simplified model processing analysis was performed: Yijk = μ+Tj+Eijk, where Yijk is the individual phenotypic record, μ the population mean, Tj the genotype effect, and Eijk the random error. The SPARC gene expression level was quantified by using an optimized method of comparing Ct (ΔΔCt) values (commonly referred to as 2-ΔΔCt). All samples were guaranteed at least three technical replicates, and the intensity ratio ± SD average was obtained therefrom.

The bovine SPARC gene was mapped to chromosome 7, and the SNP (g.12454T > C, fourth exon) of the SPARC gene was found in the Chinese cattle genome (Fig. 1). The SNP is a synonymous mutation (59Ala > Ala). It is tested by PCR amplification using the ARMS-primer method. Thus, three genotypes were found at the SNP site (g.12454T > C). The length of target band showed in the gel of TT genotype was run into 306 and 168 bp, the length of band of CT genotype was 306, 168 and 136 bp, and the CC genotypic band was 306 and 136 bp (Fig. 2).

Statistical analysis of 616 cattle SPARC gene SNP typing found that in QC, PN, XN and JX, the frequency of allele T is much greater than the frequency of allele C. We further calculated He (gene heterozygosity), Ne (the number of effective alleles; the reciprocal of homozygotes), Ho (gene homozygous) and PIC (polymorphism information content) with the POPGENE software; the methods are as follows (Nei, 1973; Botstein et al., 1980): Ho=∑i=1nPi2He=1-∑i=1nPi2Ne=1/∑i=1nPi2PIC=1-∑i=1mPi2-∑i=1m-1∑j=i+1m2Pi2Pj2. The diversity parameter of PIC is 0.210, 0.325, 0.313 and 0.159 in QC, PN, XN and JX cattle, respectively. The PIC value of QC and JX < 0.25 indicates low genetic diversity. The PIC values of PN and XN > 0.25 and <0.5 indicate intermediate genetic diversity (Table 2).

The result showed the transcription level of SPARC in five different tissues of adult bovine (Fig. 3). It was expressed in all tissues which we tested (heart, kidney, liver, muscle and lung), and the result showed differences in the tissue expression of the SPARC gene in adult cattle. The SPARC gene was most highly expressed in muscle tissues compared to others (relative to the kidney). On the contrary, it was most lowly expressed in the kidney compared other tissues.

To find out whether the influence of SNP affects the mRNA expression level of the SPARC gene, we analyzed the correlation of SPARC SNP with mRNA expression levels in skeletal muscles from 12 adult cattle (Fig. 4). The results are shown in Fig. 4: the CT gene phenotype (n=6) has a high expression level of the SPARC gene compared to the TT gene phenotype (n=6) (p<0.05).

The results show that the SNP in QC and PN cattle has a significant influence on the rump length and hip width (P<0.05); moreover, the mutation also had a significant influence on the rump length and body length of JX cattle (P<0.05). In XN cattle, the SNP has a trend to affect the traits regarding the body height and cannon bone circumference. It shows an effect of CC type > CT type > TT type but not significantly (P>0.05). To summarize, these data results found that the SNP had a positive influence on the growth of skeletal muscle in the hindquarters of cattle, such as body length, hip length and hip width (Table 3).