Fat-related traits are of economic interest for both dairy and beef cattle production. Meat and milk nutritional and organoleptic characteristics are determined, at least in part, by the amount and quality of fat (Berner Reference Berner1993; Wood et al. Reference Wood, Richardson, Nute, Fisher, Campo, Kasapidou, Sheard and Enser2004). A certain amount of fat is necessary to improve meat flavour and tenderness whereas low fat content foods are preferred by consumers. On the other hand, milk fat content has also a very important effect on milk organoleptic and nutritional characteristics, just as on cheese production yield (Berner Reference Berner1993; Lucey et al. Reference Lucey, Johnson and Horne2003).
In recent years there has been a growing interest in the detection and characterization of markers associated with fat production traits in order to apply marker-assisted selection (MAS) which could improve them. To gain insight, investigations carried out in this field have identified several genes such as acyl-CoA:diacylglycerol acyltransferase (DGAT1) (Winter et al. Reference Winter, Kramer, Werner, Kollers, Kata, Durstewitz, Buitkamp, Womack, Thaller and Fries2002), leptin (Buchanan et al. Reference Buchanan, Fitzsimmons, Van Kessel, Thue, Winkelman-Sim and Schmutz2002) or fatty acid synthase (FASN) (Roy et al. Reference Roy, Ordovas, Zaragoza, Romero, Moreno, Altarriba and Rodellar2006) that are significantly associated with fat-related traits in cattle.
The solute carrier family 27 member 1 (SLC27A1) is the first described member of the fatty acid transport protein family (FATP) (Schaffer & Lodish Reference Schaffer and Lodish1994). These proteins are proposed to mediate long-chain fatty acids (LCFA) trafficking across the plasma membrane and so have the potential to regulate local and systemic LCFA concentrations and metabolism (Schaffer & Lodish, Reference Schaffer and Lodish1994; Kim et al. Reference Kim, Gimeno, Higashimori, Kim, Choi, Punreddy, Mozell, Tan, Stricker-Krongrad, Hirsch, Fillmore, Liu, Dong, Cline, Stahl, Lodish and Shulman2004; Wu et al. Reference Wu, Ortegon, Tsang, Doege, Feingold and Stahl2006; Gimeno, Reference Gimeno2007).
SLC27A1 is expressed in tissues exhibiting rapid fatty acid metabolism such as heart, muscle and adipose tissues in different species (Hui et al. Reference Hui, Frohnert, Smith, Schaffer and Bernlohr1998; Martin et al. Reference Martin, Nemoto, Gelman, Geffroy, Najib, Fruchart, Roevens, de Martinville, Deeb and Auwerx2000; Ordovas et al. Reference Ordovas, Roy, Zaragoza and Rodellar2006). Adipocytes control flux of fatty acids (FA) to peripheral tissues by storing and hydrolysing triglycerides under hormonal control. Insulin regulates in part this process by promoting membrane translocation of intracellular SLC27A1 to the plasma membrane (Stahl et al. Reference Stahl, Evans, Pattel, Hirsch and Lodish2002; Wu et al. Reference Wu, Ortegon, Tsang, Doege, Feingold and Stahl2006) which suggests an important role for this protein in the control of energy homeostasis.
In addition to this important role in lipid metabolism, SLC27A1 has been related also to traits such as diet-induced obesity (Wu et al. Reference Wu, Ortegon, Tsang, Doege, Feingold and Stahl2006). Moreover, a polymorphism in intron 8 has been associated with increased plasma triglyceride levels in man (Meirhaeghe et al. Reference Meirhaeghe, Martin, Nemoto, Deeb, Cottel, Auwerx, Amouyel and Helbecque2000) and elevated postprandial lipaemia and alterations in LDL particle size distribution (Gertow et al. Reference Gertow, Skoglund-Andersson, Eriksson, Boquist, Orth-Gomer, Schenck-Gustafsson, Hamsten and Fisher2003). In addition, expression of SLC27A1 in heart caused lipotoxic cardiomyopathy in transgenic mice (Chiu et al. Reference Chiu, Kovacs, Blanton, Han, Courtois, Weinheimer, Yamada, Brunet, Xu, Nerbonne, Welch, Fettig, Sharp, Sambandam, Olson, Ory and Schaffer2005).
In previous work our group reported the isolation and characterization of the bovine SLC27A1 gene both structurally and functionally. In this way we established that it is organized in 13 exons extending over more than 40 kb of genomic DNA (Ordovas et al. Reference Ordovas, Roy, Zaragoza and Rodellar2006) and that the gene maps in BTA 7 (Ordovas et al. Reference Ordovas, Roy, Zaragoza, Hayes, Eggen and Rodellar2005) where several quantitative trait loci (QTL) for fat related traits have been found (Casas et al. Reference Casas, Shackelford, Keele, Koohmaraie, Smith and Stone2003; Ron et al. Reference Ron, Feldmesser, Golik, Tager-Cohen, Kliger, Reiss, Domochovsky, Alus, Seroussi, Ezra and Weller2004).
Since genetic variation in the SLC27A1 gene may affect lipid metabolism, the present study was conducted to search for variability in the bovine gene and analyse its potential association with milk fat content in dairy cattle.
Materials and Methods
Sample collection
Two groups of samples were collected. The first consisted of 211 Holstein-Friesian animals with highest [tail H, n=117 estimated breeding values (EBV)=45·19, reliability=54%] and lowest [tail L, n=94 estimated breeding values (EBV)=−24·45, reliability=58%] milk fat yield. EBVs for milk fat yield were evaluated by the National Confederation of the Spanish Holstein-Friesian Associations (CONAFE). These animals were selected from 7631 cows from connected herds located in Aragon (a region in the north-east of Spain), with an EBV reliability >47%. Each tail was constituted by the 1% of the population with the animals having the most extreme EBVs. Table 3 shows the genetic parameters for milk fat yield of the whole Aragon Holstein-Friesian animals and the selected animals for high and low tails.
The second group of samples included 101 unrelated animals belonging to two different Spanish breeds: a beef breed, Asturiana de los Valles (n=50) and a dairy breed, Menorquina (n=51).
Identification of variation
Screening for variation in the bovine SLC27A1 gene was performed by PCR amplification and direct sequencing of most of the coding region in ten animals of each Holstein-Friesian, Asturiana de los Valles and Menorquina breeds. A set of primers was designed to amplify each exon (Table 1) using Primer Express software (Applied Biosystems) and the bovine genomic DNA sequence (GenBank Accesssion number AAFC03051286). All exons were amplified with the exception of exon 13 for which several sets of primers were tested but no amplification was obtained.
Table 1. Set of primers used for PCR amplification. The sets 1, 2, 3, 5, 6, 7, 8, 9, 11 and 12 were used for PCR and sequencing of the SLC27A1 exons. For each one the MgCl2 concentration, annealing temperature and number of cycles used in PCR are indicated. The shaded primers were used for multiplex PCR and subsequent minisequencing reactions and the newly designed are in bold letters. The final concentration of each set in the multiplex PCR is indicated
PCR amplification was performed in an ABI2700 thermocycler in 25 μl using 50 ng of bovine genomic DNA, standard PCR buffer, different amounts of MgCl2 (Table 1), 100 μm each dNTP, 150 nm each primer and 1·25 U Taq DNA polymerase (Invitrogen). The PCR profile included an initial denaturation step of 95°C for 5 min and a final extension step of 72°C for 5 min. Cycling conditions were 95°C for 30 s, specific annealing temperature for each set of primers (Table 1) for 30 s and 72°C for 30 s for different number of cycles depending on the set of primers (Table 1). PCR products were enzymically purified with ExoSAP-IT (Amersham) according to manufacturer's instructions and bi-directionally sequenced.
Sequencing reactions were done in 5 μl using 1 μl of the BigDye Terminator V3.1 Cycle Sequencing Kit (Applied Biosystems), 250 nm each primer and 3–5 ng of PCR purified product. Samples were analysed in an ABI Prism 3130 Genetic Analyzer (Applied Biosystems). ClustalW multiple sequence alignment software (http://www.ebi.ac.uk/clustalw/) was used to analyse the sequences. Prediction of cis elements was done using the Patch software (http://www.gene-regulation.com/cgi-bin/pub/programs/patch/bin/patch.cgi).
SNP genotyping by primer extension assay
SNP genotyping was carried out by minisequencing (Pastinen et al. Reference Pastinen, Partanen and Syvanen1996) analysis using SNaPshot chemistry (Applied Biosystems). To optimize the multiplex PCR of fragments containing all SNPs, new sets of primers were designed when necesssary (Table 1). Multiplex PCRs were carried out in 5 μl using the Multiplex PCR Kit (QIAGEN) according to the manufacturer's instructions, the concentrations of primers showed in Table 1 and 15 ng of genomic DNA as template. The PCR profile included an initial denaturation step of 95°C for 15 min and a final extension step of 72°C for 5 min. Cycling conditions consisted of 40 cycles of 95°C for 30 s, 60°C for 1 min and 72°C for 30 s. Multiplex PCR was enzymically purified with ExoSAP-IT (Amersham) as previously described.
Minisequencing reactions were done using 0·5 μl of SnaPshot Multiplex Kit (Applied Biosystems), 1 μl of purified PCR and 0·5 μl of the 12X extension primer mix (Table 2) in a final volume of 6 μl. Reactions were then purified with 1 unit of Shrimp Alkaline Phosphatase (SAP) (Sigma) according to the manufacturer's instructions. Minisequencing products (1 μl) were mixed with 10 μl of Hi-DiTM formamide and 0·25 μl of GeneScan-120LIZ size standard (Applied Biosystems) and analysed in an ABI Prism 3130 Genetic Analyzer (Applied Biosystems).
Table 2. Extension primers for minisequencing genotyping. The number (n), name of the SNPs and the region of localization in the gene are indicated as well as the size, sequence, direction of the primer according to the sense coding sequence and the concentration in the 12X primer mix for each extension primer. The length of the 5′ poli-T tail used to modify the size of each primer is indicated in brackets
Table 3. Mean, sd and range of the EBVs (maximun and minimum) for milk fat yield (kg) in the Holstein-Friesian population and in the animals of the high and low tails of the milk fat content EBV's distribution
Data analysis
Allelic frequencies were calculated by direct counting in both groups of samples. In the analysis of the potential association with milk fat content in Holstein-Friesian animals, the allele frequency comparison was tested using binomial contrast of proportions using χ2 test. To take into account multiple testing, the Bonferroni correction was applied. A comparison-wise error rate, or type I error P value was calculated using standard statistical procedures (P<0·05).
Results and Discussion
Using the sets of primers shown in Table 1, we searched for variability in the bovine SLC27A1 gene by PCR amplification and direct sequencing of all exons and flanking introns in animals from different breeds. Overall, we examined ∼4 kb of genomic DNA of which 47·5% corresponded to exonic regions. In this way we identified 14 SNPs (Fig. 1). One of them was in the promoter region (SNP 1), seven in intronic regions (SNPs 2, 3, 4, 8, 9, 12 and 14) and the other six in coding exons (SNPs 5, 6, 7, 10, 11 and 13). This gives one polymorphic position every 268 and 323 nucleotides sequenced in intronic and exonic regions respectively. The SNPs were annotated in the reference sequence used for the search of variability taking into consideration that the SLC27A1 gene is coded in the complementary sense (Table 2 and Fig. 1).
Fig. 1. Newly identified SNPs in the bovine SLC27A1 gene by sequencing all the exons and flanking introns. Each SNP was assigned to its position in the reference sequence (AAFC03051286) taking into consideration that the gene is coded in the complementary sense. Also a correlative number was given to each of them.
The analysis of the promoter sequence where SNP 1 is located predicted different cis elements depending on the allele. For example, when C allele was present GATA-1 or GATA-6 elements were predicted while GR or En-1 elements were predicted with T allele. In any case, none of them have been involved in the control of SLC27A1 gene expression in other species. However, given the big repercussion that the SNP 1 could have over the whole gene expression, further studies such as band shift assay should be carried out to determine the real transcription factor binding activity of the region including the SNP.
Intronic polymorphisms are mainly A>G transitions (71%) while exonic polymorphism are mainly C>T transitions (60%). Exonic SNPs are in the third 3′ nucleotide of the codons, all of them constituting silent (synonymous) SNPs. So, with the exception of the SNP of the promoter, all the variation identified is intronic or synonymous. Nevertheless, although both are assumed to be not functional, there are SNPs that can affect intronic elements of regulation (Van Laere et al. Reference Van Laere, Nguyen, Braunschweig, Nezer, Collette, Moreau, Archibald, Haley, Buys, Tally, Andersson, Georges and Andersson2003), consensus sequences of splicing (Aretz et al. Reference Aretz, Uhlhaas, Sun, Pagenstecher, Mangold, Caspari, Moslein, Schulmann, Propping and Friedl2004), the mRNA stability (Capon et al. Reference Capon, Allen, Ameen, Burden, Tillman, Barker and Trembath2004) or the co-translational folding of the protein (Komar et al. Reference Komar, Lesnik and Reiss1999). For example, an intronic regulatory polymorphism in the IGF2 gene causes a major QTL effect on muscle growth in pigs (Van Laere et al. Reference Van Laere, Nguyen, Braunschweig, Nezer, Collette, Moreau, Archibald, Haley, Buys, Tally, Andersson, Georges and Andersson2003) and a synonymous SNP changes the protein activity of the MDR1 protein by altering the co-translational folding of the protein (Kimchi-Sarfaty et al. Reference Kimchi-Sarfaty, Oh, Kim, Sauna, Calcagno, Ambudkar and Gottesman2007). In this regard, SNPs 7 and 13 are in codons of amino acids included in the functional motifs 1 and 2 of the protein respectively. Motif 1 is essential for the transport function of the protein since mutation within it abolishes its function (Stuhlsatz-Krouper et al. Reference Stuhlsatz-Krouper, Bennett and Schaffer1998; Stuhlsatz-Krouper et al. Reference Stuhlsatz-Krouper, Bennett and Schaffer1999). On the other hand, motif 2 has been proposed as a fatty acid binding motif (Black et al. Reference Black, Zhang, Weimar and DiRusso1997; Hui et al. Reference Hui, Frohnert, Smith, Schaffer and Bernlohr1998). So, if the fact that synonymous substitutions can affect the co-translational folding of the protein is taken into consideration (Kimchi-Sarfaty et al. Reference Kimchi-Sarfaty, Oh, Kim, Sauna, Calcagno, Ambudkar and Gottesman2007; Komar et al. Reference Komar, Lesnik and Reiss1999), the SNPs 7 and 13 located within the functional motifs of the protein might exert a variation in the protein activity.
Once the variation was identified, a single reaction genotyping procedure based on primer extension analysis was developed. Firstly, the conditions for amplifying in a multiple reaction the fragments containing all the SNPs were established (Table 1). Then, extension primers for each polymorphism were designed (in the sense or antisense direction of the coding sequence) and the conditions for each primer in the multiple minisequencing reaction were defined (Table 2). As result, a fast and reliable method to genotype simultaneously the 14 SNPs of an individual was obtained.
The expression pattern of the SLC27A1 gene has been studied in different tissues and species including cattle (Hui et al. Reference Hui, Frohnert, Smith, Schaffer and Bernlohr1998; Martin et al. Reference Martin, Nemoto, Gelman, Geffroy, Najib, Fruchart, Roevens, de Martinville, Deeb and Auwerx2000; Ordovas et al. Reference Ordovas, Roy, Zaragoza and Rodellar2006) but no data exist for the expression profile of this gene in the bovine mammary gland. In this regard, our group has preliminary evidence demonstrating that the bovine SLC27A1 gene is expressed in mammary gland (data not shown).
The potential association of the SLC27A1 gene with milk fat content was evaluated by estimation of the allele frequencies in two tails of Holstein-Friesian animals with extreme EBVs for this trait (Fig. 2). All the SNPs, with the exception of SNP 11 and SNP 12, presented mean frequencies higher than 5% of the minor allele. Analysis of the data obtained showed no significant differences between groups for any of them (Fig. 1). These results could indicate that the SLC27A1 gene was not associated with milk fat content. Taking into account several considerations as the distance between tails of the EBV's distribution (each one contains the 1% of the animals with the most extremes EBVs and, so, the 98% of the initial population of 7631 individuals was rejected), the large size of the sample (n=211) and the use of a commercial population (in which the use of the same sires is not done and consanguinity is systematically avoided) any strong association should have been detected if it had existed. In effect, this approach allowed us to describe recently the association of two fatty acid synthase (FASN) polymorphisms with milk fat content in dairy cattle (Roy et al. Reference Roy, Ordovas, Zaragoza, Romero, Moreno, Altarriba and Rodellar2006).
Allele frequency estimation in Asturiana de los Valles and Menorquina breeds (Fig. 2) showed that all the positions were polymorphic with the exception of SNP 1 and SNP 8 which presented the allele C fixated in both of them. Major allele frequency differences between breeds were observed for SNPs 2–7 while the rest of the SNPs presented more similar frequencies. This trend seemed to be maintained also in Holstein-Friesian groups.
Fig. 2. Allele frequencies of the SLC27A1 SNPs (numbered from 1 to 14 according to Fig. 1) in the tails of Holstein-Friesian animals with highest (H) and lowest (L) EBV for milk fat content. For each SNP, bars correspond to frequencies of allele 1 (white) and allele 2 (black). Also the value of allele 1 frequency (p) is indicated below for the high and low Holstein-Friesian groups and the Asturiana de los Valles (pA) and Menorquina (pM) breeds.
Taken overall, the results suggested that the SLC27A1 gene could have a poor effect on milk fat content, or so small as not to be detected in this study. However, further studies should be carried out to analyse in depth the role of the bovine SLC27A1 gene in milk fat content. Indeed, it could also be studied as a candidate gene for other fat-related traits. In this regard, the newly identified SNPs and the genotyping method that we report here provide a very useful tool for such studies.
In addition, although no significant association of the SLC27A1 gene with milk fat content was found, the present results provide new data relevant to dairy science because this is the first time that this fatty acid transporter has been studied in relation to milk traits in cattle or in any other species.
