Detection and genetic characterization of ovine CSN1S2⁎B polymorphisms and their associations with milk production traits

Livestock Science 153 (2013) 10–19

Contents lists available at SciVerse ScienceDirect

Livestock Science journal homepage: www.elsevier.com/locate/livsci

Detection and genetic characterization of ovine CSN1S2nB polymorphisms and their associations with milk production traits J.M. Corral a, J.A. Padilla b,n, M. Izquierdo a, M. Martı´nez-Tranco´n b, J.C. Parejo b, J. Salazar b, F.I. Herna´ndez-Garcı´a a a b

´n La Orden-Valdesequera, Junta de Extremadura. A-5, Km 372, Guadajira 06071 Badajoz, Spain Centro de Investigacio ´ceres, Spain Gene´tica y Mejora Animal, Facultad de Veterinaria, Universidad de Extremadura, UEX, Avda. de la Universidad s/n, 10071 Ca

a r t i c l e in f o

abstract

Article history: Received 6 August 2012 Received in revised form 8 January 2013 Accepted 17 January 2013

CSN1S2 gene encodes the aS2-casein, one of the phosphoproteins (aS1, b, aS2 and k) secreted in ruminants milk in form of stable calcium–phosphate micelles. CSN1S2nB variant contains the amino acid substitutions p.Asp75Tyr and p.Ile105Val in the mature protein. Recently, we have shown that two adjacent SNPs, [FN_601350: g.153G4T and 155C4T], located at exon 10, are causers of the p.Asp75Tyr substitution, while the non-synonymous mutation SNP [FN_601350: g.504A4G], in exon 11, led to codon exchange of p.Ile105Val. PCR-RFLP and RT-PCR analyses were used to detect and characterize the single nucleotide polymorphisms involved in the B variant of the CSN1S2 gene and to assess their effects on milk yield and composition traits. Four hundred and thirty samples from different breeds (Berrinchon du cher, Castellana, Ile de France and Merino) were analyzed and population genetic parameters were estimated for each locus and breed. Both polymorphic sites of the SNPs g.[153G4T and 155C4T] are conserved across analyzed breeds. Haplotype allele frequencies from the SNPs g.[153G4T and 155C4T]—GC, GT, TC and TT—varied among the breeds, being GC the most frequent allele in all populations (40.5). Allele G from SNP g.504A4G had the minor allele frequency in all breeds analyzed (ranged from 0.225 to 0.333). The value of genetic differentiation (FST ¼ 0.0419) across breeds and loci indicates that there was no genetic divergence among the analyzed breeds. The haplotype frequencies, the deduced amino acid haplotypes and the occurrence of linkage disequilibrium were estimated in a large population (377 samples) of Merino sheep breed. The eight possible haplotypes of these two loci were identified, and substantial linkage disequilibrium was found. GCA was the haplotype most frequent (470%) and Asp75–Ile105 the most frequent aminoacid haplotype (0.713). Recombinants haplotypes were observed in low frequencies (r0.1). Association study between the ovine CSN1S2nB polymorphisms and the milk traits were estimated in 669 lactations from 218 Merino ewes. The CSN1S2 p.75 and CSN1S2 p.105 polymorphisms affected the non-fat solids and lactose percentages in milk, respectively. The CSN1S2 Tyr75Tyr ewes had the highest levels of non fat solids and the ewes carrying CSN1S2 Val105Val genotype had the smallest lactose percentage. Similar effects of the single markers have been observed for homozygous Tyr75– Val105 ewes on lactose percentage. This work provides the basis to study in further sheep breeds the CSN1S2nB polymorphisms and their effects on milk traits. & 2013 Elsevier B.V. All rights reserved.

Keywords: Sheep aS1-casein B variant SNPs Milk traits

n

Corresponding author. Tel.: þ 34 927257148; fax: þ 34 927257110. E-mail addresses: [email protected] (J.M. Corral), [email protected], [email protected] (J.A. Padilla), [email protected] (M. Izquierdo), [email protected] (M. Martı´nez-Tranco´n), [email protected] (J.C. Parejo), [email protected] (J. Salazar), [email protected] (F.I. Herna´ndez-Garcı´a). 1871-1413/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.livsci.2013.01.008

J.M. Corral et al. / Livestock Science 153 (2013) 10–19

1. Introduction

Caseins are a family of phosphoproteins (aS1, b, aS2 and k) that comprise the major protein component of ruminant milk, functioning as nutritive carriers of both aminoacids and minerals in milk (Ginger and Grigor, 1999). Casein genes are organized in a tightly linked cluster that span the genes for the evolutionarily related caseins, Ca-sensitive caseins aS1, b, aS2 (CSN1S1, CSN2 and CSN1S2, respectively) and also the functionally associated k-casein (CSN3) gene (Rijnkels et al., 2003). Among ruminants, extensive genetic variations in casein genes of the cattle and goat species have been identified and their functional effects on milk composition and cheese-making properties have been characterized (Caroli et al., 2009; Martin et al., 2002) and, in some cases, they are already being included in selection schemes (Manfredi et al., 1995; Ng-Kwai Hang, 1998). However, the knowledge of milk protein genetic variants is more fragmentary in ovine species (Amigo et al. 2000; Moioli et al., 2007) and the genetic control of the observed protein variation was assessed in only a few cases (Chessa et al., 2003). So far, some single nucleotide polymorphisms (SNPs) have been described and characterized in the ovine CSN1S1 (Ceriotti et al., 2004, 2005; Ferranti et al., 1995), CSN2 (Ceriotti et al., 2004; Chessa et al., 2010; Corral et al., 2010) and CSN3 (Ceriotti et al., 2004) casein genes and all of them have resulted in aminoacid exchanges. The ovine aS2-casein presents seven variants recently named from A to G by Giambra and Erhardt (2012). The A, B, C and D variants have been identified by protein isoelectric-focusing in the ovine milk (Chessa et al., 2003; Giambra et al., 2010a). The aS2-CN A and B variants (Chessa et al., 2003) were recently characterized in Italian sheep breeds by structural analysis of the protein by Picariello et al. (2009). The B variant (CSN1S2nB) differs from the most common form A by two aminoacid exchanges, Asp75Tyr and Ile105Val, in the mature protein. The variants C and D were identified in German sheep breeds (Giambra et al., 2010a), in which three aminoacid exchanges were deduced from mRNA sequences: Val45Ile, Ala48Ser (allele C) and Arg46Ser (allele D; Giambra and Erhardt, 2012). Aminoacid substitutions responsible for E and F variants were firstly deduced from mRNA: Asp49Asn and Lys200Asn (Boisnard et al., 1991; Chessa et al., 2010) from the mature protein, respectively. Furthermore, cDNA sequencing led to detection of the Arg161His substitution named as G variant (Giambra and Erhardt, 2012). However, until now, except for the recently PCR-RFLP test designed to detect the G variant—Arg161His (Giambra and Erhardt, 2012), DNA-based methods to quickly and accurately detect and genotype large population samples for the different variants of this gene have not been designed. Recently and for the first time at the DNA level, we have shown that the aminoacid substitutions of the CSN1S2nB variant involve three SNPs (GenBank: FN601350.1, Padilla et al., in press). Two of them are located at exon 10, g.[153G4T and 155C4T], affecting the first and third nucleotides of the codon #75 leading to the Asp75Tyr substitution, and the third one is a non-synonymous mutation g.504A4G, located at exon 11, causing codon

11

exchange of Ile105Val (GenBank Accession number: FN601350). Most association studies between the ovine milk protein polymorphisms and the milk performance traits have assessed mainly the effects of single gene and some controversial results have been revealed (Amigo et al., 2000; Barillet, 2007; Barillet et al., 2005). Casein genes are organized in a cluster and, therefore, the association studies must consider the entire casein haplotype instead of a single locus (Sacchi et al., 2005). Although some casein haplotypes were included in association analyses in a few cases (Corral et al., 2010; Giambra et al., 2010b), current knowledge about the ovine CN haplotype is very scarce, and, therefore, it is necessary to increase our knowledge with new studies about the variability into the casein genes and to design the corresponding genotype tests predicting phenotype differences. In this sense, we have accomplished a method, based on Polymerase Chain Reaction-Restriction Fragment Length Polymorphism (PCR-RFLP) and Real Time-Polymerase Chain Reaction (RT-PCR) techniques, which allow an accurate and reliable identification of the SNPs involved in the B variant of the CSN1S2 gene, in order to (i) detect and characterize all possible CSN1S2nB genotypes in a large Merino population, (ii) verify whether these polymorphisms are conserved in the genome of other sheep breeds and to determine population parameters, and (iii) assess the effects of ovine CSN1S2nB genotypes on milk yield and composition traits. 2. Materials and methods 2.1. Genotyping data Individual blood samples of 430 ewes from six different commercial flocks (159) and one experimental flock (218) from Merino breed (MER) as well as individual blood samples of 53 ewes from three breeds (15 Berrinchon du cher—BER; 18 Castellana—CAS; 20 Ile de France—ILE) were collected. DNA was isolated from whole blood using a standard commercial kit (UltraCleanTM DNA BloodSpin kit, MO BIO, Carlsbad, CA, USA) following the manufacturer’s protocol. After DNA isolation, samples were genotyped in a different way for SNPs in exon 10 (by PCR-RFLP) and exon 11 (by RTPCR), because exon 10 contains multiple polymorphic sites (143G4A, 153G4T, 155C4T and 176A4C, of which solely the SNP 153G4T is a non-synonymous substitution; Padilla et al., 2013) that prevent from using other methods that require a perfect match between probes or primers and the target, as the hybridization method used in exon 11. 2.2. Production data Milk production records included a total of 669 lactations from 218 genotyped Merino ewes (all the animals come from the experimental flock). A daily milk yield of each ewe was recorded fortnightly. The first record was measured at weaning (between 30 and 60 days after lambing). Only lactations longer than 45 days (three controls) were used. Milk composition was analyzed by infrared using the Milko Scan FT (Foss Electric, Hillerød,

12

J.M. Corral et al. / Livestock Science 153 (2013) 10–19

Denmark). Lactation records included milk yield and fat, protein, lactose and non-fat solids percentages of milk standardized to 120 lactation days. Apart from the lactation records, we have also recorded the birth–weaning interval, the year–season of lambing, the type of birth and the age of dam. 2.3. Genotyping for the exon 10 SNPs g.[153G4T and 155C 4T] 2.3.1. PCR reactions Primers were designed on the basis of the available sheep as2-casein (CSN1S2) sequence (GenBank Accession number: FN601350) to amplify the genomic region encoding exon 10, by using Primer3 software (Rozen and Skaletsky 2000). The primer sequences were the following: E10F: 50 CTGAAGTTGCCCCAGAGGTA 30 and E10R: 50 CATTTGGAGAAGAAGCAGTGG 30 . PCR was carried out in a total volume of 20 mL containing 1  PCR buffer, 2.5 mM MgCl2, 200 mM each dNTP, 250 nM each primer, 1 U DNA polymerase (DyNAZymeTM II DNA polymerase, Finnzymes, Oy, Finland) and approximately 100 ng of genomic DNA. PCR amplification was performed using a PTC-200 thermal cycler (MJ Research, Inc., Watertown, MA, USA) with an initial denaturation step at 94 1C for 5 min followed by 35 cycles of 94 1C for 30 s, 55 1C for 30 s and 72 1C for 1 min, with a final extension of 72 1C for 5 min. The size of expected product was 225 bp (GenBank Accession number: FN601350, positions 22–246). 2.3.2. RFLP reactions The SNPs g.[153G 4T;155C 4T] can lead to four codons: GAC, GAT, TAC and TAT. The different haplotypes: GC, GT, TC and TT were identified by three sequential RFLP reactions with RsaI, Hpy8I and BseGI restriction endonucleases (Table 1). Reaction 1. PCR amplicon (225 bp) of each sample was digested with 2 units of the Rsa I (recognition site 50 -GT/ AC-31) overnight at 37 1C. This endonuclease recognizes TAC codon (haplotype TC) and produces three fragments: 132, 71 and (because there is the second restriction enzyme site; base 203–225) 22 base pairs. The haplotypes GC, GT and TT are not recognized by the enzyme Rsa I and they are characterized by two fragments: 203 and 22. Reaction 2. The samples in which genotypes were not identified in the reaction 1 were digested with 6 units of the Hpy8I enzyme (recognition site 50 -GTN/NAC-30 ), overnight at 37 1C. Hpy8I recognizes 155C4T polymorphism that is present in TAC and GAC codons and it produces three fragments: 131, 60 and (second restriction enzyme site; base 191–225) 34 base pairs. Since the TC haplotype was previously identified the GC haplotypes can be deduced. The haplotypes GT and TT are not recognized by the enzyme Hpy8I and they are characterized by two fragments: 191 and 34. Reaction 3. The samples in which genotypes were not identified in the reaction 2 were digested with 6 units of the BseGI (recognition site 50 -GGATG/2-30 ) endonuclease, overnight at 55 1C. The haplotype GT (codon GAT) is recognized by BseGI enzyme and it is represented by

two fragments: 137 and 88 base pairs. The haplotype TT is not recognized by BseGI enzyme and it is characterized by a 225 bp fragment. RFLP reactions were performed with 10 mL of PCR product and digested with the appropriate endonuclease (Fermentas, Hannover, MD), according to the supplier’s directions for buffer conditions. The restriction fragments were separated by electrophoresis in a 4% agarose gel and stained with ethidium bromide. In order to validate the PCR-RFLP reactions, 10 samples of the known genotype by sequencing were included in the test. 2.4. Genotyping for the exon 11 SNP—RT-PCR reactions The single nucleotide polymorphism (SNP) g.504A 4G was genotyped using the TaqMan SNP Genotyping Assay (Applied Biosystems, Foster City, CA, USA). To amplify the genomic region encoding exon 11, primer and probe designs were carried out with Custom Taqmans SNP Genotyping Assay Service of Applied Biosystems on the basis of the available sheep as2-casein (CSN1S2) sequence (GenBank Accession number: FN601350). The primers used were E11-504R-F (CCCCCAGTATCTCCAGTATCTGTAT) and E11-504R-R (GATCAGGTTAAGAGAAATGCTGGC) that amplified a 76 bp fragment of the ovine CSN1S2 gene. The probes were E11-504R-M (CAAGGCCCAATTGTTTT, labeled with the dye VIC) and E11-504R-V (AAGGCCCAGTTGTTTT, labeled with the dye FAM). The PCR reactions were carried out using a 7500 FAST Real-Time PCR System apparatus (Applied Biosystems. Foster City, CA, USA) in a total volume of 20 mL containing 10 mL of TaqMan Universal PCR Master Mix, 0.5 mL of 40  Assay Mix, 5 mL of genomic DNA (20 ng/mL) and Milli-Q water. Real-time PCR cycling conditions consisted of a pre-PCR (Holding stage) of 1 min at 60 1C and initial cycle of 10 min at 95 1C, followed by 40 cycles of 15 sec at 95 1C and 1 min at 60 1C. Data acquisition and analysis of the results were performed using the 7500 Software vers.2.0.1 (Applied Biosystems). Each sample was analyzed in duplicate and the same results were obtained in each one. In each PCR run, a blank without DNA was analyzed. To verify allele identification, three samples, representative of the three genotypes, CSN1S2 g.504A4G-namely, AA, AG, and GG (aminoacid sequences: Ile105Ile, Ile105Val, Val105Val), which had been previously sequenced on a ABI Prism 3130 analyzer (Applied Biosystems, Foster City, Ca, USA), were included as positive controls in each run. 2.5. Statistical analysis Genotype and allele frequencies within each locus and breed were calculated by direct counting. Genotype frequencies were tested for Hardy–Weinberg equilibrium using Guo and Thompson (1992) with 1000 dememorization steps for 100 batches and 10,000 iterations per batch, to estimate without bias the exact P-value of this test. Distribution of gene variability among breeds (and among flocks within Merino population) were analyzed by the analogous Wright’s F-statistics (FIT, FIS and FST; Weir and

J.M. Corral et al. / Livestock Science 153 (2013) 10–19

13

Table 1 Results of the sequential restriction reactions used to identify the different genotypes generated from SNPs g. [153G 4T and 155C4T]. Reaction 1

Reaction 2

Reaction 3

Rsa I (gt/ac)

Hpy8I (gtn/nac)

BseGI (ggatg/2) Genotype

Codon target GAT

n

Codon target TAC

Genotype

Codon target TAC GAC

Genotype

þ/ þ þ/  þ/  þ/ 

TC/TC TC/   TC/   TC/  

þ /þ þ/ þ/

TC/GC TC/  T TC/  T

þ/ /

TC/GTa TC/TT

/ / / / / /

notTC/notTC

þ /þ þ/ þ/ / / /

GC/GC GC/  T GC/-T  T/  T  T/  T  T/  T

þ/ / þ/þ þ/ /

GC/GT GC/TTa GT/GT GT/TT TT/TT

8 84 9 9 226 68 7 11 5 3 430

Note: þ ¼ presence of restriction site;  ¼absence of restriction site. In bold: genotypes identified after each restriction reaction. a GT/TC and GC/TT are the same genotype (GTTC and GTCT in different gametic phase).

Cockerham, 1984). The significances of F-statistics were tested by 1000 permutations and the confidence interval at 95%. An unbiased estimate of expected heterozygosity (He)–gene diversity—for each locus and population was computed by Nei and Roychoudhury equation (1974). The observed heterozygosity (Ho) for each locus and population was estimated as the proportion of heterozygous in a given population. Haplotype frequencies for SNPs g.[153G4T and 155C4T] and g.504A4G, the deduced aminoacid exchanges (Asp75Tyr and Ile105Val) and the occurrence of linkage disequilibrium were estimated for the Merino population using the maximum-likelihood method, implemented via the expectation-maximization (EM) algorithm (Excoffier and Slatkin, 1995; Hawley and Kidd, 1995; Long et al., 1995; Terwilliger and Ott, 1994; Xie and Ott, 1993). The effects of CSN1S2nB Asp75Tyr and Ile105Val polymorphisms on milk yield (first milk recording yield and milk yield standardized to 120 days) and composition (fat, protein, non-fat solids and lactose percentages standardized to 120 days) were estimated using the mixed linear model (procedure REML) from SAS V8e Institute (2007): Y ¼ X b þ Zu þe where Y was the vector of observations for milk yield and composition traits; b was the vector of fixed factors: CSN1S2 p.75 and p.105 genotypes (3 levels each) or the combined genotypes (9 levels), the year–season of lambing (15 different levels), the age of ewe (3 levels: younger than 2.5 years, between 2.5–5 and 5–8.5 years), the type of lambing (2 levels: single or twins) and the lambing– weaning interval (3 levels: less than 50 days, from 50 to 60 and longer than 60 days from lambing); X was the incidence matrix for these effects; u was the vector of random ewe effects; Z was the incidence matrix for random ewe effects, and e was a vector of the residual random errors, assumed to be normally distributed with mean E(e)¼0 and variance V(e)¼R. Linear combinations of parameters values L0 b (L being a vector of coefficients)

and their corresponding variances V (L0 b), estimated with this model, allowed to test differences genotype means using a t distribution. The Bonferroni correction for multiple comparisons was applied. 3. Results 3.1. Genotype identification Fig. 1 shows all the different restriction patterns obtained by the PCR-RFLP tests described in the methodology. These tests allowed us to identify the first and third nucleotide position of the four possible codons (GAC, GAT, TAC and TAT) generated from the SNPs g.[153G 4T and 155C 4T]. Table 1 shows a summary of the genotype identification of the 430 samples analyzed by the sequential PCR-RFLP tests. As it can be observed the 10 expected genotypes have been identified. The genotypes TC/GT and GC/TT contain the same changes in the first (G-T) and third (C-T) positions of the codon and could be considered as the same genotype, but the PCR-RFLP test can discriminate if these changes resulting from the union of GC and TT or GT and TC gametes. Real-time PCR allelic discrimination assay has allowed us to identify the three genotypes from SNP g.504A4G: AA, AG and GG. In Fig. 2, the two-dimensional plot of amplification parameters, which demonstrate the separation of the three genotypes, are shown. 3.2. Genetic variability The frequencies of the different genotypes, the allele frequencies and the population genetic index found in each breed and locus analyzed are shown in Table 2 (only aminoacid exchanges are shown). Protein variations p. Asp75Tyr and p. Ile105Val (D75Y and I105V in aminoacid letter code) were deduced from the observed genotypes of loci g.[153G4 T and 155C 4T] and g.504A 4G, respectively. Our results showed that the occurrence of the different genotypes and alleles vary between breeds.

14

J.M. Corral et al. / Livestock Science 153 (2013) 10–19

Fig. 1. Restriction patterns obtained by PCR–RFLP tests of CSN1S2 SNP g.[153G 4T;155C4T] of exon 10, used to deduce the different genotypes. [M ¼20 bp ladder; n ¼ undigested PCR product (225 bp),þ ¼ presence of restriction site,  ¼ absence of restriction site]. To visualize fragments of low number of bases the samples were separated on 9% polyacrylamide gels (acrylamide/bisacrylamide 19:1) in 1  TBE.

Table 2 Genotype and allele frequencies, expected (unbiased) heterozygosities (Hnb) frequencies, FIS index and the Hardy–Weinberg Equilibrium (HWE) across breeds and loci analyzed. (Only amino acid (letter-code) changes are shown.) Frequencies

Breeds BER n¼ 15

CAS n¼ 18

ILE n¼ 20

Locus:g.[153G 4T;155C 4T] p.75Asp(D) 4 Tyr(Y) Genotypes DD 0.933 0.500 0.800 DY 0.067 0.333 0.150 YY 0.000 0.167 0.050 Alleles GC (D) GT (D) TC (Y) TT (Y)

Fig. 2. Allelic distribution of CSN1S2 SNP g.504A 4G of exon 11.

Thus, in the locus p. Asp75Tyr, the most common genotype was DD (Asp75Asp) in all breeds analyzed and the YY (Tyr75Tyr) was present in very low frequencies and absent in BER sheep breed. In the locus p. Ile105Val the II (Ile105Ile) was the most frequent genotype in all breeds analyzed, with the exception of CAS breed in which the heterozygote IV (Ile105Val) was the most frequent. The VV (Val105Val) genotype was present in lowest frequencies and absent in BER sheep breed. Both polymorphic sites of the SNPs g.[153G4 T and 155C 4T] of the codon #75 were present in all breeds analyzed, indicating that these polymorphic sites are conserved. Allele frequencies of the SNPs g.[153G 4T and 155C 4T] GC, GT, TC and TT varied among the breeds,

MER n ¼377

0.705 0.252 0.043

0.533 0.433 – 0.033

0.639 0.028 0.250 0.083

0.750 0.125 0.100 0.025

0.719 0.113 0.139 0.029

Genetic variability Hnb 0.067 FIS 0.000 HWE; P value –

0.457 0.277 0.310

0.244 0.337 0.247

0.280 0.102 0.062

0.600 0.350 0.050

0.554 0.374 0.072

0.775 0.225

0.741 0.259

0.358 0.022 1.000

0.384 0.026 0.687

Locus:g.-504A4 G p. 105Ile(I)4Val(V) Genotypes II 1.000 0.389 IV 0.000 0.556 VV 0.000 0.055 Alleles A (I) 1.000 0.667 G (V) – 0.333 Genetic variability Hnb 0.000 FIS  HWE; P value 

0.457  0.223 0.597

Breeds: BER, Berrinchon du cher; CAS, Castellana; ILE, Ile de France; MER, Merino.

being GC the most frequent allele in all populations and the minor allele frequency corresponds to TT allele in the ILE sheep (0.025). TC allele was absent in BER sheep breed.

J.M. Corral et al. / Livestock Science 153 (2013) 10–19

( 413%) and GTG (10%). The estimated frequencies of these haplotypes were higher than the expected frequencies. The rest of haplotypes occurred at a very low frequencies (o2.5%). In relation to the aminoacid haplotypes the most frequent haplotype was Asp75–Ile105 ( 471%), followed by Tyr75–Val105 (14%), Asp75–Val105 ( o12%) and Tyr75–Ile105 ( 43%). According to the nomenclature proposed by Picariello et al. (2009) and Giambra and Erhardt (2012), the A allele has Asp and Ile at positions p. 75 and p.105, respectively, and the B allele, Tyr and Val in the same positions. However, in this paper, we found the recombinants haplotypes Asp–Val and Tyr– Ile, which we named as A0 and B0 , respectively. Table 5 shows the observed frequency of each genotype and the corresponding deduced aminoacid in the mature protein for the two loci, as well as the frequencies of combined genotypes in Merino sheep population which was analyzed. In relation to aminoacid exchange Asp75Tyr (D75Y), the observed frequencies of DD, DY and YY genotypes (0.706, 0.252 and 0.042, respectively) were similar to those found by Picariello et al., (2009) in Comisana breed. All (nine) possible combined genotypes have been found in the analyzed Merino samples. The most frequent combined genotype found was DI/DI (0.533), followed by DI/YV (0.207) and DI/DV (0.151), while the rest of genotypes had frequencies lower than 0.05.

The allelic distribution of SNP g.504A4G shows that the allele G has been found as the minor allele frequency in all breed analyzed, ranging between 0.225 and 0.333 in ILE and CAS breeds, respectively. This allele was absent in BER breed. Genetic variability parameters as expected (unbiased) heterozygosity, FIS index and the HW equilibrium for each locus in the different breeds analyzed are also shown in Table 2. All breeds were in HW equilibrium. FIS is a measure of the deviation of genotypic frequencies from panmictic frequencies in terms of heterozygous deficiency or excess. Negative FIS values indicate heterozygote excess and positive values indicate heterozygote deficiency compared with Hardy–Weinberg equilibrium expectations. Distribution of gene variability among breeds were analyzed (Table 3) by the analogous Wright’s F-statistics (FIT, FIS and FST; Weir and Cockerham, 1984). The value of genetic differentiation for the two loci among the breeds (FST ¼0.042) indicates that there was no genetic differentiation among the analyzed population. 3.3. Haplotype and combined genotype frequencies in Merino breed Haplotype frequencies for SNPs g.[153G 4T and 155C4T] and g.504A4G, the deduced aminoacid exchanges (Asp75Tyr and Ile105Val) and the occurrence of linkage disequilibrium found in the Merino population are shown in Table 4. The eight possible haplotypes of these two loci were identified, including recombinant haplotypes (3, 4, 5, and 6). The w2 test was significantly different from the null hypothesis (no association), indicating substantial linkage disequilibrium. The haplotype GCA was the most frequent ( 470%), followed by TCG

3.4. Genotype association with milk traits Effects on milk production and composition traits of the CSN1S2 p. Asp75Tyr, CSN1S2 p. Ile105Val genotypes and the combined genotypes were estimated in an experimental flock of Merino breed (Table 6). Distribution of gene variability among Merino flocks was analyzed by the analogous Wright’s F-statistics (FIT, FIS and FST). The value of genetic differentiation for both loci among the seven flocks (FST ¼0.034) indicate that there was no genetic differentiation within the analyzed population and, consequently, any flock can be representative of the Merino population. No significant differences were found on milk yield among genotypes derived from the CSN1S2nB

Table 3 Multilocus estimate of allele frequency-based correlation (FIS, FST, FIT) among breeds. Locus

Protein position

FIS

FST

FIT

153–155 504 All

75 105

0.1219 0.0143 0.0603

0.0402 0.0432 0.0419

0.1572 0.0569 0.0997

15

Table 4 Haplotype frequencies for SNPs g.[153G 4T;155C4T] and g.504A 4G, and for the deduced aminoacid exchanges in Merino population. Haplotype frequencies in Merino population

Nucleotide position

Protein position

g.153g 4t g.155c4 t g.504a 4g Expa Obsb

p.75 Asp 4Tyr; p.105 Ile 4Val Expa Obsb

1

2

3

4

5

6

7

8

G C A 0.533 0.701

G T A 0.084 0.012

G C G 0.186 0.018

G T G 0.029 0.100

T C A 0.103 0.005

T T A 0.022 0.023

T C G 0.036 0.134

T T G 0.007 0.006

A

A

B

B

Asp(D)–Ile(I) 0.616 0.713

Asp(D)–Val(V) 0.215 0.119

Tyr(Y)–Ile(I) 0.125 0.029

Tyr(Y)–Val(V) 0.043 0.140

Significant (P o0.001) association between 153–155 and 504. a Expected haplotype frequency under independence hypothesis. b Observed haplotype frequency.

16

J.M. Corral et al. / Livestock Science 153 (2013) 10–19

Table 5 Genotypes frequencies found in Merino population. Locus g.[153G 4T;155C 4T]

p.75D 4Y

Locus g.504A4 G;

p.75D 4Y; p.105 I4 V

Genotypes

Freq

Genotypes

Freq

Genotypes p.105 I4 V

Freq

Combined Genotypes

Freq

GC/GC GC/GT GT/GT

0.536 0.146 0.024

DD

0.706

AA

II

0.554

DI/DI DI/DV DV/DV

0.533 0.151 0.021

GC/TC GT/TC, GC/TT GT/TT

0.201 0.043 0.008

DY

0.252

AG

IV

0.374

DI/YI DI/YV DV/YV

0.008 0.207 0.037

TC/TC TC/TT TT/TT

0.019 0.016 0.008

YY

0.042

GG

VV

0.072

YI/YI YI/YV YV/YV

0.013 0.016 0.013

Table 6 Least squares means and standard errors of milk yield and composition traits in Merino sheep with different as2-casein (CSN1S2) SNPs. Genotypes

p.75D 4Y DD DY YY

N

a

p.105I4Va II IV VV Combined genotypesb DI/DI DI/DV DI/YI DI/YV DV/YV YI/YV YV/YV

479 156 29 664 418 222 29 669 398 58 7 131 13 15 7 629

FMRY (Kg)

MY120 (Kg)

Mean7 s.e

Mean7 s.e.

(P¼ 0.3207) 0.489 70.015 0.479 70.023 0.561 70.051

(P ¼ 0.3180) 43.11 71.41 41.48 72.25 49.46 74.93

(P¼ 0.9049) 0.491 70.015 0.495 70.020 0.473 70.048

(P ¼ 0.6673) 43.38 71.47 43.56 71.92 39.33 74.60

(P¼ 0.6241) 0.489 70.016 0.495 70.036 0.554 70.102 0.483 70.025 0.4607 0.079 0.547 70.074 0.475 70.111

(P ¼ 0.3953) 43.007 1.52 44.507 3.51 50.77 79.98 41.85 72.44 37.28 77.63 48.73 77.22 37.73 710.74

N

366 123 22 511 319 169 27 515 303 45 6 99 13 10 5

FP120 (%)

PP120 (%)

SNF120 (%)

Mean7 s.e

Mean7 s.e

Mean7s.e

(P ¼ 0.8806) 7.457 0.10 7.407 0.15 7.327 0.32

(P ¼ 0.5553) 6.597 0.08 6.487 0.12 6.727 0.25

(P ¼0.0467) 11.64 70.06a 11.54 70.09a 12.037 0.18b

(P ¼ 0.2217) 7.447 0.10 7.387 0.13 6.947 0.29

(P ¼ 0.5450) 6.567 0.09 6.467 0.11 6.697 0.25

(P ¼0.4987) 11.68 70.06 11.61 70.08 11.52 70.17

P¼ 0.6374 7.46.7 0.11 7.457 0.23 6.927 0.60 7.467 0.17 6.757 0.46 7.177 0.49 7.367 0.66

P¼ 0.6574 6.617 0.08 6.547 0.18 5.977 0.48 6.527 0.13 6.037 0.36 6.697 0.39 6.727 0.52

P¼ 0.1782 11.66 70.06 11.59 70.14 11.54 70.35 11.53 70.10 11.42 70.27 12.32 70.28 11.607 0.38

481

N

LAC120 (%) Mean 7s.e

310 104 16 430 270 138 22 430 258 35 5 83 12 7 5

(P ¼0.6681) 4.46 7 0.03 4.43 7 0.04 4.40 7 0.09 (P ¼0.0569) 4.46 7 0.028a 4.45 7 0.037 a 4.26 7 0.084 b P ¼ 0.0220 4.46 7 0.03 4.47 7 0.06 4.74 7 0.16 4.43 7 0.04 4.26 7 0.12 4.51 7 0.14 4.08 7 0.17

405

FMRY, first milk recording yield; MY120, 120-day milk yield; FP120%, Fat %; PP120%, Protein %; SNF120%, Solids-non fat %; LAC120 %, Lactose %; N, number of lactations. a Aminoacid (letter-code) position in mature protein. b DV/DV and YI/YI genotypes were excluded for lactations number o 1%.

polymorphisms. The CSN1S2 p. Asp75Tyr had effect on the percentage of non-fat solids in milk, with Tyr75Tyr ewes (12.0370.18%) displaying higher levels than their Asp75Asp (11.6470.06%) and Asp75Tyr (11.5470.09%) counterparts. On the other hand, there were significant differences (Po0.05) on the percentage of lactose of milk among CSN1S2 p. Ile105Ile (4.4670.028%), Ile105Val (4.45 70.037%) and Val105Val (4.26 70.086%) genotypes. Therefore, the smallest lactose percentage was associated with carrier ewes of the Val105Val genotype. For the association studies of the combined genotypes, DV/DV and YI/YI genotypes were excluded since lactations number was o1%. Similar effects to single makers in the lactose and non-fat solids percentages have been

observed when the Bonferroni correction is not applied. Thus, least square means differences have been observed (P¼0.022) among the combined genotypes YV/YV and DV/ YV with the rest of combined genotypes on lactose percentage. 4. Discussion PCR-RFLP and RT-PCR tests used in this work have allowed an accurate identification of the three SNP involved in the aminoacid substitutions of the CSN1S2nB variant, confirming the sequencing results previously obtained by us (Padilla et al., 2013). It was necessary the use of Rsa I, Hpy8I and BseGI restriction endonucleases

J.M. Corral et al. / Livestock Science 153 (2013) 10–19

to genotyping the adjacent SNPs g. [153G 4T and 155C4T] by PCR-RFLP, which makes the technique time-consuming. However, the high polymorphism of the sequence containing these SNPs (exon 10; Padilla et al., 2013) did not allows to use others available methods for the allelic discrimination of these adjacent SNPs. The designed procedure also allows discriminating among all possible genotypes derived from the three SNPs involved in the B variant and to deduce as aminoacid substitutions—Asp75Tyr (allele A) and Ile105Val (B)—as the recombinant haplotypes: Asp75–Val105 (A0 ) and Tyr75–Ile105 (B0 ), which had not been previously studied. So far, the discrimination between aS2-CN A and B phenotypes was performed by protein isoelectricfocusing (IEF) analysis in Italian and German breeds (Chessa et al., 2003; Giambra et al., 2010a), but according to Picariello et al. (2009) only the Asp75Tyr substitution was responsible for the higher isoelectric point of B protein variant, allowing, therefore, its detection by IEF analysis. Moreover, contrarily to IEF genotyping that requires milk samples of the females after lambing, our tests permits to genotype males and females at birth, and this information can be used in breeding programs to increase selection accuracy. Thus, these tests are useful in animal breeding and genetic studies because they are faster and cheaper than the protein and cDNA sequencing methods used until now for the studies of CSN1S2nB (Giambra and Erhardt, 2012; Picariello et al., 2009). Both polymorphic sites of the adjacent SNPs g.[153G 4T and 155C 4T] of the codon #75 are conserved in the genomes of the analyzed breeds. Allele frequencies found for Asp75Tyr (D75Y) substitution were similar to those found in Italian and German breeds (Chessa et al., 2003; Giambra et al., 2010a; Picariello et al., 2009). The locus SNP g. 504A4G, responsible of protein variation p. 105Ile 4Val (I4V), was polymorphic in CAS, ILE and MER breeds and the allele A (Ile) was the more frequent. Future research on the allelic distribution of this locus should include animals from different breeds and countries, in order to characterize this locus better since had not been previously studied. The value of genetic differentiation FST across breeds and loci indicates that there was no genetic divergence among the analyzed breeds, perhaps due to the similar productive purposes of these breeds. Analysis of the effects of casein loci on sheep milk yield and composition traits is a necessary prerequisite before using this information in selection schemes devoted to improve these traits. In this research, the effects of the two aminoacid substitutions that compose the CSN1S2nB variant on milk yield and composition traits have been estimated in an experimental flock of Merino breed. This breed have been used mainly for wool and meat production and it has never been selected for milk production traits, as evidenced by their low milk production (Mean¼40.22 Kg, 120-day milk yield, Corral et al., 2010), although, at present, the Merino milk is also used to produce prestigious artisan regional cheeses in Spain. Moreover, the values of genetic differentiation FST for each

17

locus and between loci performed within the Merino breed indicate that this population is not structured, i.e., there has been no genetic divergence between flocks which comprise the population and, consequently, any flock can be representative of the Merino population. Therefore, we expect that for these loci, any sufficiently large sample from the population (as the experimental flock used in this study) is representative of the gene and the genotype frequencies of the total population. In ruminants, many functional milk protein DNA variants with level expression of milk proteins are known (Ibeagha-Awemu et al., 2008). Major effects were assessed for the aS-caseins in goat. CSN1S1 was observed early as the most variable gene associated with different rates (high, medium, low and null) of milk aS1-casein content (Caroli et al., 2007; Martin et al., 2002; Moioli et al., 2007), and also with cheese yield and milk coagulation properties (Clark and Sherbon, 2000; Remeuf, 1993; Vassal et al., 1994). aS1-Casein genotypes information is included since 1996 in French Alpine and Saanen selection schemes (Barillet, 2007). DNA variants (A–F, 0) of goat CSN1S2 gene were also observed associated with different rates (normal, low and null) of milk aS2-casein synthesis (Ibeagha-Awemu et al., 2008). In sheep, research on association between genetic polymorphisms of caseins and milk yield and composition traits is mainly focused on aS1- and b-caseins loci (Van der Werf, 2007), but to our knowledge, this information has not been incorporated into selection schemes yet as it has been in goat. Different alleles (A–F, H and I) were identified for CSN1S1 gene associated with fat yield, fat, protein and aS1-casein contents and cheesemaking properties in Italian and German breeds (Giambra et al., 2010b; Moioli et al., 2007). In Merino breed an A4G transition in CSN2 gene that produces Met183Val aminoacid exchange has been associated with milk yield (Corral et al., 2010). The highest milk yield was associated with ewes carrying CSN2 GG genotype In relation to aS2-casein (CSN1S2) variants of sheep until now only the B variant has been included into association studies in East Friesian dairy sheep. These studies were performed by IEF and significant influences on protein yield were revealed (Giambra et al., 2010b). Our data showed that Asp75Tyr mutation affects significantly the non-fat solids percentage (SNF%) in milk. The highest SNF content was associated with ewes carrying CSN1S2 Tyr75Tyr genotype. Considering that this substitution is responsible for the B variant detected by IEF (Picariello et al., 2009), then we can relate our results with those that were found by IEF. In this sense, we can think that both results (Giambra’s and ours) are congruent, since the non-fat solids (SNF) consist of protein, lactose and ash (minerals). In Merino sheep, SNF content has a moderate heritability (0.3470.03) and it is genetically correlated with protein and lactose percentages (Corral, 2008) as in dairy cattle (Gaunt et al., 1968; Hossein-Zadeh and Ardalan, 2011) and goat (Kala and Prakash, 1990). Therefore, we can expect that the selection of Tyr75Tyr ewes will increase the SNF content and its components in Merino sheep. In the dairy industry, it is well-known that the greater the amount of fat and SNF in

18

J.M. Corral et al. / Livestock Science 153 (2013) 10–19

raw milk, the greater the yield of cheese will be (FAO, 2009). This genotypic information could be used in the Merino breeding program, whose milk is transformed into cheese after weaning. On the other hand, SNP g. 504A4G p.Ile105Val affects the lactose percentage in milk. The smallest lactose percentages were associated with ewes carrying the CSN1S2 Val105Val genotype. In goat, Badaoui et al. (2007) identified a silent mutation in exon 45 (C5493T) of acetyl-CoA carboxylase alpha (ACACA) gene that was suggestively associated with lactose content. In cattle, the allele E located at promoter of DGAT1 gene had also significant effects on lactose content (Sanders et al., 2006). The rate of lactose synthesis largely determines the rate of water secretion into milk and indirectly the fat and protein content (Miglior et al., 2006). In Merino sheep (Izquierdo et al., 2010) we obtained a moderate estimation of heritability of lactose content (0.2870.05), and a positive genetic correlation between lactose content and milk yield (0.32) and negative with the percentages of fat ( 0.51) and protein ( 0.61). Thus, genetic selection to increase the content of lactose (Ile105Ile and Ile105Val genotypes) may increase milk yield and it may also decrease the protein and fat contents. In relation to combined genotypes, similar effects to single makers on the 120 days lactose percentage were observed (P¼0.022). Thus, the genotypic combinations that carry two copies of 105Val (YV/YV or DV/YV) have the smallest lactose content. In this sense, in the combined genotypes, the SNP g.[153G4T and 155C4T] causer of p.75Asp4Tyr substitution does not appear to influence the single genotypic effect of the SNP g. 504A4G, (p. 105Ile4Val) on milk lactose content. On the contrary, both SNPs appear to interact with one another masking the effect of the SNP Asp75Tyr on non-fat solids percentage. Nevertheless, the small frequencies of some combined genotypes might be a very important factor influencing the significance of the associations, especially when the Bonferroni correction for multiple comparisons was applied. In general, SNPs single and combined effects on the milk yield and composition traits should be confirmed in an increased sample size and in other dairy sheep breeds, prior to be included in ovine breeding programmes, as it could have breed-specific allele effects, similar to those found in casein genes in goats (Caravaca et al., 2011). Moreover, the biological significance of these associations still remains to be evaluated in further studies and negative correlations between milk yield and content traits should be considered, especially in ovine breeds whose milk is often transformed into cheese (Ramos et al., 2009), like Merino sheep. The existence of a genotype test that predicts phenotype differences of functional DNA variant, as described here, is a crucial aspect for the start of effective Marker Assisted Selection programme in sheep breeding. The value of genetic information is basically driven by the increase in selection accuracy resulting from the knowledge of genotypes; in particular the relative increase in selection accuracy of the youngest selection candidates (Van der Werf, 2007).

5. Conclusions In this study, we have designed DNA-based tests which permit to accurately and reliably discriminate among all possible genotypes derived from the three SNPs involved in the B variant of the CSN1S2 gene (153G4T, 155C4T and 504A4G), and to deduce both aminoacid substitutions—Asp75Tyr (allele A) and Ile105Val (allele B)—as the recombinant haplotypes: Asp75–Val105 (A0 ) and Tyr75– Ile105 (B0 ), which had not been previously studied. These polymorphisms are conserved across breed analyzed (Berrinchon du cher Castellana, Ile de France and Merino). Genotype effects on milk composition were detected in Merino population. Ewes with genotype Tyr75Tyr had the highest levels of non-fat solids. In addition, first results regarding the influence of CSN1S2 p.105 polymorphism on sheep milk performance traits are presented. CSN1S2 p.105 polymorphism affects the lactose percentages in milk. The ewes carrying CSN1S2 Val105Val genotype had the smallest lactose percentage. Combined genotypes had similar effects as single markers for lactose percentage. Thus, the genotypic combinations that carry two copies of Val105 (YV/YV or DV/YV) have the smallest lactose content. This research provides the basis to study on further sheep breeds the CSN1S2nB polymorphisms and their effects on milk traits.

Conflict of interest statement There are no known conflicts of interest associated with this publication.

Acknowledgments The authors thank the Junta de Extremadura PRI þDTþI and the FEDER for financial support and to the Spanish Merino Association for providing animal samples. References Amigo, L, Recio, I., Ramos, M., 2000. Genetic polymorphism of ovine milk proteins: its influence on technological properties of milk—a review. Int. Dairy J. 10, 135–149. Badaoui, B., Serradilla, J.M., Tomas, A, Urrutia, B., Ares, J.L., Carrizosa, J.A., Sa´nchez, A., Jordana, J., Amills, M., 2007. Goat acetyl-coenzyme A carboxylase a: molecular characterization, polymorphism, and association with milk traits. J. Dairy Sci. 90, 1039–1043. Barillet, F., 2007. Genetic improvement for dairy production in sheep and goats. Small Ruminant Res. 70, 60–75. Barillet, F., Arranz, J.J., Carta, A., 2005. Mapping quantitative trait loci for milk production and genetic polymorphisms of milk proteins in dairy sheep. Genet. Sel. Evol. 37 (Suppl. 1), S109–S123. Boisnard, M., Hue, D., Bouniol, C., Mercier, J.C., Gaye, P., 1991. Multiple mRNA species code for two non-allelic forms of ovine aS2-casein. Eur. J. Biochem. 201, 633–641. Caravaca, F., Ares, J.L., Carrizosa, J., Urrutia, B., Baena, F., Jordana, J., Badaoui, B., Sa nchez, A., Angiolillo, A., Amills, M., Serradilla, J.M., 2011. Effects of a s1-casein (CSN1S1) and k-casein (CSN3) genotypes on milk coagulation properties in Murciano-Granadina goats. J. Dairy Res. 78, 32–37. Caroli, A.M., Chessa, S., Erhardt, G.J., 2009. Invited review: milk protein polymorphisms in cattle: effect on animal breeding and human nutrition. J. Dairy Sci. 92, 5335–5352.

J.M. Corral et al. / Livestock Science 153 (2013) 10–19

Caroli, A.M., Chiatti, F., Chessa, S., Rignanese, D., Ibeagha-Awemu, E.M., Erhardt, G.J., 2007. Characterizaion of the casein gene complex in West African goats and description of a new aS1-casein polymorphism. J. Dairy Sci. 90, 2989–2996. Ceriotti, G., Chessa, S., Bolla, P., Budelli, E., Bianchi, L., Duranti, E., Caroli, A., 2004. Single nucleotide polymorphisms in the ovine casein genes detected by polymerase chain reaction-single strand conformation polymorphism. J. Dairy Sci. 87, 2606–2613. Ceriotti, G., Chiatti, F., Bolla, R., Martini, M., Caroli, A., 2005. Genetic variability of the ovine as1-casein. Ital. J. Anim. Sci. 4, 64–66. Chessa, S., Bolla, P., Dario, C., Pieragostini, E., Caroli, A., 2003. Genetic milk protein polymorphisms in the Gentile di Puglia ovine breed: monitoring by isoelectric focusing. Sci. Tecnol. Lattiero Casearie 54, 191–198. Chessa, S., Rignanese, D., Berbenni, M., Cerioti, G., Martini, M., Pagnacco, G., Caroli, A., 2010. New genetic polymorphisms within ovine b and aS2-caseins. Small Ruminant Res. 88, 84–88. Clark, S., Sherbon, J.W., 2000. AlphaS1-casein, milk composition and coagulation properties of goat milk. Small Ruminant Res. 38, 123–134. Corral, J.M., 2008. Caracterizacio´n de las variantes gene´ticas de las proteı´nas de la leche en ovinos de raza Merina y su asociacio´n con ˜ o experimental. Ph.D. la produccio´n y calidad de la leche en un reban Thesis. Extremadura University, Ca´ceres, Spain. Corral, J.M., Padilla, J.A., Izquierdo, M., 2010. Associations between milk protein genetic polymorphisms and milk productions traits in Merino sheep breed. Livest. Sci. 129, 73–79. Excoffier, L., Slatkin., M., 1995. Maximum-likelihood estimation of molecular haplotype frequencies in a diploid population. Mol. Biol. Evol. 12, 921–927. FAO, 2009. Milk Testing and payment systems resource book: a practical guide to assist milk producer groups. In: Draaiyer, J., Dugdill, B., Bennett, A., Mounsey, J. (Eds.) Animal Production and Health Division. FAO, Roma. /ftp://ftp.fao.org/docrep/fao/012/i0980e/i0980e00. pdfS. Ferranti, P., Malorni, A., Nitti, G., Laezza, P., Pizzano, R., Chianese, L., Addeo, F., 1995. Primary structure of ovine aS1-caseins: localization of phosphorylation sites and characterization of genetic variants A, C and D. J. Dairy Res. 62, 281–296. Gaunt, S.N., Wilcox, C.J., Farthing, B.R., Thompson, N.R., 1968. Genetic interrelationships of Holstein milk composition and yield. J. Dairy Sci. 51, 1396–1402. ¨ Giambra, I.J., Jager, S., Erhardt, G., 2010a. Isoelectric focusing reveals additional casein variants in German sheep breeds. Small Ruminant Res. 90, 11–17. Giambra, I.J., Stam, E., Atema, F., Schuiling, E., Brandt, H., Erhardt, G., 2010b. Association study between milk protein variants and milk performance traits in East Friesian Dairy sheep. In: Proceedings of the 9th World Congress on Genetics Applied to Livestock Production, p. 36. Giambra, I.J., Erhardt, G., 2012. Molecular genetic characterization of ovine CSN1S2 variants C and D reveal further important variability within CSN1S2. Anim. Genet. 43 (5), 642–645. Ginger, M.R., Grigor, M.R., 1999. Comparative aspects of milk caseins. Comp. Biochem. Physiol. Pt. B: Biochem. Mol. Biol. 124, 133–145. Guo, S.W., Thompson, E.A., 1992. Performing the exact test of Hardy– Weinberg proportion for multiple alleles. Biometrics 48, 361–372. Hawley, M., Kidd, K., 1995. Haplo: a program using the EM algorithm to estimate the frequencies of multi-site haplotypes. J. Hered. 86, 409–411. Hossein-Zadeh, N.G., Ardalan, M., 2011. Estimation of genetic parameters for milk urea nitrogen and its relationship with milk constituents in Iranian Holsteins. Livst. Sci. 135, 274–281. Ibeagha-Awemu, E.M., Kgwatalala, P., Zhao, X., 2008. A critical analysis of production-associated DNA polymorphisms in the genes of cattle, goat, sheep, and pig. Mamm. Genome 19, 591–617. Izquierdo, M. Corral, J.M., Padilla, J.A., Herna´ndez, F.I., 2010. Variance components and genetic correlations of milk production and composition in Merino sheep. In: Proceedings of the 9th World Congress

19

on Genetics Applied to Livestock Production. Leipzig. Germany, pp 2–133. /www.kongressband.de/wcgalp2010/assets/pdf/0746.pdfS. Kala, S.N., Prakash, B., 1990. Genetic and phenotypic parameters of milk yield and milk composition in two Indian goat breeds. Small Ruminant Res. 3, 475–484. Long, J.C., Williams, R.C., Urbanek, M., 1995. An E–M algorithm and testing strategy for multiple-locus haplotypes. Am. J. Hum. Genet. 56, 799–810. Manfredi, E., Ricordeau, G., Barbieri, M.E., Amigues, Y., Bibe´, B., 1995. Ge´notype case´ine as1 et se´lection des boucs sur descendance dans les races Alpine et Saanen. Genet. Sel. Evol. 27, 451–458. Martin, P., Szymanowska, M., Zwierzchowski, L., Leroux, C., 2002. The impact of genetic polymorphisms on the protein composition of ruminant milks. Reprod. Nutr. Dev. 42, 433–459. Miglior, F., Sewalem, A., Jamrozik, J., Kistemaker, G., Lefebvre, D.M., Moore, R.K., 2006. Genetic analysis of MUN and lactose and their relationships with economically important traits in Canadian Holstein cattle. Interbull Bull. 35, 58–63. Moioli, B., D’Andrea, M., Pilla, F., 2007. Candidate genes affecting sheep and goat milk quality. Small Ruminant Res. 68, 179–192. Nei, M., Roychoudhury, A.K., 1974. Sampling variances of heterozygosity and genetic distance. Genetics 76, 379–390. Ng-Kwai Hang, K.F., 1998. Genetic polymorphism of milk proteins: relationships with production traits, milk composition and technological properties. Can. J. Dairy Sci. 74, 4002–4012. Padilla, J.A., Corral, J.M., Quesada, A., Izquierdo, M., Parejo, J.C., Martı´nezTranco´n, M. Single nucleotide polymorphisms in the ovine CSN1S2 gene for as2-casein. Span. J. Agric. 11(1). 2013. http://dx.doi.org/10. 5424/2013111-3309. Picariello, G., Rignanese, D., Chessa, S., Ceriotti, G., Trani, A., Caroli, A., Di Luccia, A., 2009. Characterization and genetic study of the ovine aS2casein (CSN1S2) allele B. Protein J. 28, 333–340. Rozen, S., Skaletsky, H., 2000. Primer3 on the WWW for general users and for biologist programmers. In: Krawetz, S., Misener, S. (Eds.), Bioinformatics Methods and Protocols: Methods in Molecular Biology, Humana Press, Totowa, NJ, pp. 365–386. Ramos, A.M., Matos, C.A.P., Russo-Almeida, P.A., Bettencourt, C.M.V., Matos, J., Martins, A., Pinheiro, C., Rangel-Figueiredo, T., 2009. Candidate genes for milk production traits in Portuguese dairy sheep. Small Ruminant Res. 82, 117–121. Remeuf, F., 1993. Influence of genetic polymorphism of caprine aS1-casein on physicochemical and technological properties of goat’s milk. Lait 73, 549–557. Rijnkels, M., Elnitski, L., Miller, W., Rosen, J.M., 2003. Multispecies comparative analysis of a mammalian specific genomic domain encoding secretory proteins. Genomics 82, 417–432. Sacchi, P., Chessa, S., Budelli, E., Bolla, P., Ceriotti, G., Soglia, D., Rasero, R., Cauvin, E., Caroli, A., 2005. Casein haplotype structure in five Italian goat breeds. J. Dairy Sci. 88, 1561–1568. Sanders, K., Bennewitz, J., Reinsch, N., Thaller, G., Prinzenberg, E.M., Kuhn, C., Kalm, E., 2006. Characterization of the DGAT1 mutations and CSN1S1 promoter in German Angeln dairy cattle population. J. Dairy Sci. 89, 3164–3174. SAS Institute, 2007. SAS/STAT8 User’s Guide: SAS Online Doc. /http:// v8doc.sas.com/sashtml/stat/chap41/index.htmS. Terwilliger, J., Ott, J., 1994. Handbook of Human Genetic Linkage. Johns Hopkins University Press, Baltimore. Van der Werf, J.H.J., 2007. Marker-assisted selection in sheep and goats. In: Guimara~ es, E., Ruane, J., Scherf, B., Sonnino, A., Dargie, J. (Eds.), Marker-assisted Selection: Current status and future perspectives in crops, livestock, forestry and fish, FAO, Rome, Italy, pp. 230–247. Vassal, L., Delacroix-Buchet, A., Bouillon, J., 1994. Effect of AA, EE and FF genetic variants of alpha s1-casein from goat milk on cheese yield and sensory properties of traditional cheeses: preliminary observations. Lait 74, 89–103. Weir, B.S., Cockerham, C.C., 1984. Estimating F-statistics for the analysis of population structure. Evolution 38, 1358–1370. Xie, X., Ott, J., 1993. Testing linkage disequilibrium between a disease gene and marker loci. Am. J. Hum. Genet. 53 (Suppl.), 1107.