The goat αs1-casein (CSN1S1) locus is one of the few major genes identified in goats with evident effects on milk composition (Martin et al. Reference Martin, Szymanowska, Zwierzchowski and Leroux2002). This gene encodes a protein of 199 amino acids (Brignon et al. Reference Brignon, Mahé, Grosclaude and Ribadeau-Dumas1989) and it has been completely sequenced by Ramunno et al. (Reference Ramunno, Cosenza, Rando, Illario, Gallo, Di Berardino and Masina2004). One of the main features of the CSN1S1 locus is its high level of polymorphism (Martin et al. Reference Martin, Szymanowska, Zwierzchowski and Leroux2002; Moioli et al. Reference Moioli, D'Andrea and Pilla2007). In this way, CSN1S1 alleles can be classified, according to the levels of CSN1S1 synthesis to which they are associated with, into four groups (Grosclaude et al. Reference Grosclaude, Mahé, Brignon, Di Stasio and Jeunet1987; reviewed in Martin et al. Reference Martin, Szymanowska, Zwierzchowski and Leroux2002; Moioli et al. Reference Moioli, D'Andrea and Pilla2007; Caroli et al. Reference Caroli, Chiatti, Chessa, Rignanese, Ibeagha-Awemu and Erhardt2007): high (A, B 1, B 2, B 3, B 4, B′, C, H, L, and M), medium (E and I alleles), low (D, F and G alleles) and null (01, 02 and N alleles). The effects of this polymorphism on milk yield, milk composition and cheese yield have been studied in depth in Saanen and Alpine French breeds. According to Mahé et al. (Reference Mahé, Manfredi, Ricordeau, Piacère and Grosclaude1994) and Barbieri et al. (Reference Barbieri, Manfredi, Elsen, Ricordeau, Bouillon, Grosclaude, Mahé and Bibé1995), high content alleles are associated with higher protein, fat and casein contents, whereas there is not any effect on milk production. Moreover, Vassal et al. (Reference Vassal, Delacroix-Buchet and Bouillon1994) demonstrated that cheese yield of milk from AA individuals is 8·2% and 17·7% higher than that of obtained from EE and FF individuals, respectively.
Although the influence of CSN1S1 genotype on milk composition (protein, fat and casein contents) has been sufficiently demonstrated in various European breeds (Trujillo et al. Reference Trujillo, Jordana, Guamis, Serradilla and Amills1998; Martin et al. Reference Martin, Szymanowska, Zwierzchowski and Leroux2002; Moioli et al. Reference Moioli, D'Andrea and Pilla2007) its effect on the synthesis rate of CSN1S1 has been only quantified in a few populations. In French breeds, Grosclaude et al. (Reference Grosclaude, Mahé, Brignon, Di Stasio and Jeunet1987) and Martin et al. (Reference Martin, Ollivier-Bousquet and Grosclaude1999) reported values of 0·0, 0·9–1·2, 2·2–3·2 and 7–7·2 g/l CSN1S1 for null, low, medium and high homozygous genotypes. Moreover, Gómez-Ruiz et al. (Reference Gómez-Ruiz, Miralles, Agüera and Amigo2004) reported values of approximately 9·42, 6·11, 3·35 and 1·45 g/l CSN1S1 for BB, BF, EE and FF CSN1S1 genotypes in a sample of 18 goats of unknown origin. This scarcity of data concerning the effect of CSN1S1 genotype on CSN1S1 milk content in European goat breeds is probably motivated by the complexity of measuring this trait, which is, therefore, not routinely recorded in milk recording schemes. However, the rate of synthesis of CSN1S1 is one of the major factors influencing the efficiency with which caseins are transported from the endoplasmic reticulum to the Golgi apparatus and, in consequence, it has profound effects on milk quality (Chanat et al. Reference Chanat, Martin and Ollivier-Bousquet1999). The main objective of our work was to investigate whether B, E and F CSN1S1 alleles, the most common ones in Spanish breeds (Jordana et al. Reference Jordana, Amills, Díaz, Angulo, Serradilla and Sánchez1996), have a differential effect on milk CSN1S1 content in Malagueña and Murciano-Granadina goats. The B and E alleles differ, amongst other features, by the insertion of a long interspersed nucleotide element at exon 19 which might diminish mRNA stability (Pérez et al. Reference Pérez, Leroux, Bonastre and Martin1994), whereas the F allele results from the alternative splicing of exons 9, 10 and 11 (Leroux et al. Reference Leroux, Mazure and Martin1993; Martin et al. Reference Martin, Szymanowska, Zwierzchowski and Leroux2002). The skipping of these exons involves the loss of 37 amino acids, including a phosphorylation site, and it might alter protein secretion (Chanat et al. Reference Chanat, Martin and Ollivier-Bousquet1999; Martin et al. Reference Martin, Szymanowska, Zwierzchowski and Leroux2002). The comparative analysis of the effects of these genetic polymorphisms on CSN1S1 synthesis rate in Spanish v. French breeds would be helpful in understanding how population-specific factors modulate, on a qualitative and quantitative basis, the allelic expression of casein genes.
Materials and Methods
Three-hundred-and-five and four-hundred-and-sixty milk samples were taken from 89 Malagueña and 138 Murciano-Granadina goats distributed in eight and three herds, respectively (Table 1). Full sibs were not included in the experimental sample. Herds were reproductively disconnected and, therefore, goats in different herds were not genetically related. At least three samples were taken in each milking period from each goat (except for the FF genotype in the Malagueña breed where the mean number of records was 2·8).
Levels of αs1-casein (CSN1S1) synthesis in Malagueña and Murciano-Granadina goats with different CSN1S1 genotypes (genotypic means that are significantly different at P⩽0·05 are shown with different superscript letters)
1 N: Number of analysed goats per genotype
2 NR: total number of records for each genotype
A near infrared spectrophotometer (NIRS) previously calibrated (Agüera et al. Reference Agüera, Urrutia, Sánchez, Ares, Amigo, Serradilla, Davies and Garrido-Varo2004) was used to analyse milk CSN1S1 content. NIRS calibrations were obtained by using, as a standard, samples (n=200) where CSN1S1 content was measured by capillary electrophoresis (Gómez-Ruiz et al. Reference Gómez-Ruiz, Miralles, Agüera and Amigo2004). Sample preparation involved the utilization of spectroscopy by infrared reflection (DESIR) method. This method consists of drying a glass fibre filter previously soaked with the liquid under test in an oven at 40°C for 24 h. For NIR spectra collection a Foss NIRSystems 6500 SY-I, equipped with a spinning module, connected to a computer controlled with the Software ISI NIRS 3 version 3.11 (Infrasoft International, Port Matilda PA, USA) was used. Samples (glass fibre filters) were placed in a small ring cup for solid products. Spectra were obtained collecting reflectance measurements of monochromatic light in the 1100–2500 nm region with 2-nm intervals. Software WINISI II version 1.04 (Infrasoft International) was used for spectra treatment.
Genotyping of the goat CSN1S1 gene
Genomic DNA was isolated from blood samples. Four-hundred μl of whole blood was repeatedly washed with 0·5 ml TE buffer (10 mm-Tris, 1 mm-EDTA) by centrifuging at 13 000 g for 1 min. This process was repeated 4–5 times until a white pellet was obtained. Leucocytes were subsequently resuspended in 0·4 ml lysis buffer (50 mm-KCl, 10 mm-Tris, 2·5 mm-MgCl2, 0·5% Tween 20, pH 8·3) plus 10 μl proteinase K (10 mg/ml) and incubated at 56°C for 4–5 h. Proteins were removed by chloroform (0·4 ml) extraction and genomic DNA was purified by precipitation with 0·8 ml of 95% ethanol plus 40 μl of 2m-NaCl. After a 13 000 g for 5 min centrifugation step, the DNA pellet was washed with 1 ml of 70% ethanol and resuspended in 100–300 μl ultrapure water. The E allele was identified by amplifying exon 19 of the CSN1S1 gene with the following primers: forward, 5′-CTA TCA TGT CAA ACC ATT CTA TCC-3′ and reverse, 5′-CAA TTT CAC TTA AGG ATG TTA CAC-3′. This region contains a truncated long interspersed nucleotide element in the case of allele E (Pérez et al. Reference Pérez, Leroux, Bonastre and Martin1994). In consequence, allele E yields a PCR product of approximately 0·59 kb whereas the remaining alleles generate amplicons with a size of about 0·13 kb. The PCR composition was 1·5 mm-MgCl2, 200 μm of each dNTP, 0·5 μm of each primer, 2 ng/μl genomic DNA and 0·025 U/μl Taq DNA polymerase (Promega, Barcelona, Spain). The thermal profile consisted of 30 cycles of 94°C for 30 s, 57°C for 30 s and 72°C for 2 min plus an extension step of 72°C for 10 min. The resulting amplicons were electrophoresed in 1·8% high resolution agarose gels stained with ethidium bromide at 100 V for 20 min. The F and B alleles were identified by employing a method reported by Ramunno et al. (Reference Ramunno, Cosenza, Pappalardo, Pastore, Gallo, Di Gregorio and Masina2000). Primers were as follows: forward, 5′-TTC TAA AAG TCT CAG AGG CAG-3′ and reverse, 5′-GGG TTG ATA GCC TTG TAT GT-3′. The PCR composition was as previously described except for the dNTP concentration (100 μm). The thermal profile consisted of 1 cycle of 94°C for 3 min plus 35 cycles of 94°C for 30 s, 64°C for 30 s and 72°C for 2 min plus an extension step of 72°C for 10 min. Amplicons were digested with 5 U XmnI enzyme (Fermentas, Madrid, Spain) at 37°C overnight. Digestion products were run in 2·5% high resolution agarose gels stained with ethidium bromide for 3 h at 140 V. According to Ramunno et al. (Reference Ramunno, Cosenza, Pappalardo, Pastore, Gallo, Di Gregorio and Masina2000) this protocol allows to distinguish the F allele (223 bp), the B/E allelic pair (161/63 bp) and the A/0 allelic pair (150/63 bp), whereas the remaining alleles yield a 212 bp fragment. Since these two methods (Pérez et al. Reference Pérez, Leroux, Bonastre and Martin1994; Ramunno et al. Reference Ramunno, Cosenza, Pappalardo, Pastore, Gallo, Di Gregorio and Masina2000) have a limited resolution, we only considered for the experimental work goats carrying alleles that can be determined unambiguously (B, E and F alleles). The precision of this method was assessed by the simultaneous genotyping of a number of samples with a CSN1S1 PCR-RFLP genotyping protocol developed at the Laboratoire de Génétique Biochimique et de Cytogénétique (Institut National de la Recherche Agronomique) in collaboration with the Labogena company at Jouy-en-Josas (http://www.labogena.fr; Leroux et al. Reference Leroux, Mazure and Martin1993). Both procedures yielded identical results with regard to the identification of the B, E and F alleles.
Statistical analyses
Statistical analyses were carried out with MIXED procedure of SAS Statistical Software (SAS Institute, 2007). The model employed was:
where Y is the vector of recorded data, β is the vector of fixed factors including CSN1S1 genotype, herd-year-season, age-parity number; X is the incidence matrix that relates these effects with the data vector; Z 1 and Z 2 are the incidence matrices corresponding to the random part of the model (U i) describing the change through lactation within animal of CSN1S1 content by means of the covariate number of months elapsed from kidding to milk sampling (NM), which accounts for the covariance between repeated records of the same goat, with an element for the intersection (U1) and another for the coefficients of first and second grade terms (U 2) of the polynomial regression equation; e is a vector of residual random errors.
Results and Discussion
The results obtained in this experiment are fairly consistent, at a qualitative level, with previous data concerning the effects of the CSN1S1 polymorphism on milk CSN1S1 content in French goats (Grosclaude et al. Reference Grosclaude, Mahé, Brignon, Di Stasio and Jeunet1987) and in a goat population of unknown origin (Gómez-Ruiz et al. Reference Gómez-Ruiz, Miralles, Agüera and Amigo2004). Our data show that in the Malagueña breed, the BB, EE and FF genotypes are associated with a high, intermediate and low CSN1S1 content (Table 1), respectively. In the Murciano-Granadina breed, the situation is more complex because only the comparison between the BB genotype and the remaining ones yields a significant difference (P⩽0·05). In this way, milk CSN1S1 levels are fairly similar between EE, BF and EF genotypes. Although our results are substantially similar to the ones obtained by Grosclaude et al. (Reference Grosclaude, Mahé, Brignon, Di Stasio and Jeunet1987) and Gómez-Ruiz et al. (Reference Gómez-Ruiz, Miralles, Agüera and Amigo2004) in other goat populations, they are not completely coincident from a quantitative perspective. This lack of coincidence amongst different studies might be explained by technical and biological factors. From an experimental perspective, sample size, number of records per individual and the method of choice for measuring CSN1S1 levels might have a strong impact on the precision with which phenotypic values are obtained. In this way, we have used near infrared spectroscopy to quantitatively analyse CSN1S1, whereas Grosclaude et al. (Reference Grosclaude, Mahé, Brignon, Di Stasio and Jeunet1987) and Gómez-Ruiz et al. (Reference Gómez-Ruiz, Miralles, Agüera and Amigo2004) employed rocket immunoelectrophoresis and capillary electrophoresis, respectively. Moreover, and of outmost importance, we have taken several registers in each individual to ensure that CSN1S1 levels are representative of the whole lactation and not of a particular time point. This is relevant because there is evidence that the stage of lactation has a deep influence on milk CSN1S1 concentration (Benradi, Reference Benradi2007). From a biological point of view, breed-specific environmental and genetic factors might play a prominent role on the allelic expression of the goat CSN1S1 gene thus explaining the differences observed in Spanish and French populations. Alpine and Saanen goats are generally raised in France under rather intensive production systems. On the contrary, Malagueña goats are almost exclusively bred under grazing systems with a supplementary stall feeding of variable magnitude, whereas Murciano-Granadina goats are bred under systems with varied levels of intensification. The relative importance of environmental and genetic non-additive factors in determining milk CSN1S1 concentration has been consistently evidenced by estimating the heritability of this phenotype in goats (h2=0·10, Benradi, Reference Benradi2007) and sheep (h2=0·21, Othmane et al. Reference Othmane, De La Fuente, Carriedo and San Primitivo2002). Moreover, the existence of a significant influence of nutrition on CSN1S1 genotype expression, in the Malagueña breed, has been recently evidenced by De la Torre et al. (Reference De la Torre, Serradilla, Ares, Rodríguez Osorio and Sanz Sampelayo2007), convincingly illustrating the notable complexity of CSN1S1 synthesis regulation.
Differential allelic expression seems to be very pervasive in mammalian genomes and highly context-specific, being modulated by a myriad of genetic and environmental factors that determine, in the end, which allele is transcribed or translated more efficiently (Knight, Reference Knight2004). In the present work, we have shown that the effects of the CSN1S1 polymorphism on the rate of synthesis of the corresponding protein show similar trends in Spanish and French breeds although population-specific differences exist with regard to the magnitude of genotype differences. Gene expression is a very complex process under the influence of many environmental and genetic factors, a feature which pleads for the convenience of validating the effects of major genes in many diverse populations before using them as a source of information in selection schemes.
This work was funded with grants from the Spanish Ministry of Education and Science (1FD1997-1052-C02-01 and AGL2002-04304-C03-02-GAN). Many thanks to Patrice Martin, Christine Leroux and Yves Amigues for providing details of the CSN1S1 genotyping technique developed at the Laboratoire de Génétique Biochimique et de Cytogénétique (Institut National de la Recherche Agronomique) in collaboration with the Labogena company at Jouy-en-Josas (http://www.labogena.fr).
