During the last decade, a large number of quantitative trait loci (QTL) have been identified, particularly for dairy traits such as milk composition (Tellam, Reference Tellam2007). Identifying the genes controlling such traits will provide markers that can be used to improve genetic selection programs and hence, give opportunities to select cows for targeted milk production. Over the past few years, great interest has been taken in the relations between cow genetics and milk characteristics. However, most of the studies focus on major production traits, such as milk, protein and lipid yields and protein and lipid contents (e.g. Kulig et al. Reference Kulig, Kmiec, Kowalewska-Luczak and Andziak2009; Sun et al. Reference Sun, Jia, Ma, Zhang, Wang, Yu and Zhang2009; da Silva et al. Reference da Silva, Sonstegard, Thallman, Connor, Schnabel and Van Tassell2010), and some have also extended their study to milk fatty acid (FA) composition (e.g. Schennink et al. Reference Schennink, Stoop, Visker, Heck, Bovenhuis, van der Poel, van Valenberg and van Arendonk2007, Reference Schennink, Heck, Bovenhuis, Visker, van Valenberg and van Arendonk2008; Conte et al. Reference Conte, Mele, Chessa, Castiglioni, Serra, Pagnacco and Secchiari2010). Nevertheless, also other milk components, such as casein (CN), whey proteins and Ca, are essential and play, among others, an important role in the processing of milk (Wedholm et al. Reference Wedholm, Larsen, Lindmark Månsson, Karlsson and Andrén2006b; Amenu & Deeth, Reference Amenu and Deeth2007; Glantz et al. Reference Glantz, Devold, Vegarud, Lindmark Månsson, Stålhammar and Paulsson2010). Thus, a more comprehensive study on compositional traits will explore the relations with cow genetics more in detail, which could be used for direct selection within practical breeding work.
Polymorphisms in the leptin (LEP), leptin receptor (LEPR) and acyl-CoA:diacylglycerol acyltransferase 1 (DGAT1) gene loci are some examples that have been shown to be associated with milk composition traits. The LEP gene, located on Bos taurus autosome (BTA) 4, encodes the hormone leptin, which is involved in the regulation of feed intake, energy balance, fertility and metabolism (Fruhbeck et al. Reference Fruhbeck, Jebb and Prentice1998). Various polymorphisms in this gene have been shown to be associated with milk (e.g. Liefers et al. Reference Liefers, te Pas, Veerkamp and van der Lende2002; Banos et al. Reference Banos, Woolliams, Woodward, Forbes and Coffey2008; Kulig et al. Reference Kulig, Kmiec, Kowalewska-Luczak and Andziak2009), protein (e.g. Buchanan et al. Reference Buchanan, Van Kessel, Waldner and Christensen2003; Chebel et al. Reference Chebel, Susca and Santos2008; Kulig et al. Reference Kulig, Kmiec, Kowalewska-Luczak and Andziak2009), lipid (e.g. Komisarek et al. Reference Komisarek, Szyda, Michalak and Dorynek2005; Chebel et al. Reference Chebel, Susca and Santos2008; Kulig et al. Reference Kulig, Kmiec, Kowalewska-Luczak and Andziak2009) and lactose yields (Liefers et al. Reference Liefers, te Pas, Veerkamp and van der Lende2002) as well as protein (Liefers et al. Reference Liefers, Veerkamp, Te Pas, Delavaud, Chilliard, Platje and van der Lende2005) and lipid contents (Komisarek et al. Reference Komisarek, Szyda, Michalak and Dorynek2005). Moreover, the LEPR gene, located on BTA3, affects leptin concentrations during pregnancy (Liefers et al. Reference Liefers, Veerkamp, te Pas, Delavaud, Chilliard and van der Lende2004) and the T945M polymorphism in the LEPR gene has been shown to impact protein and lipid contents (Komisarek & Dorynek, Reference Komisarek and Dorynek2006), whereas no associations were found with milk (Komisarek & Dorynek, Reference Komisarek and Dorynek2006; Banos et al. Reference Banos, Woolliams, Woodward, Forbes and Coffey2008), protein or lipid yields (Komisarek & Dorynek, Reference Komisarek and Dorynek2006) for this polymorphism. The DGAT1 gene located on BTA14, on the other hand, has been shown to have a major effect on milk fat composition (Grisart et al. Reference Grisart, Coppieters, Farnir, Karim, Ford, Berzi, Cambisano, Mni, Reid, Simon, Spelman, Georges and Snell2002). It encodes for the enzyme DGAT, which catalyzes the final step of triglyceride synthesis (Cases et al. Reference Cases, Smith, Zheng, Myers, Lear, Sande, Novak, Collins, Welch, Lusis, Erickson and Farese1998). It has been shown that polymorphisms in the DGAT1 gene are associated with lipid yield (e.g. Thaller et al. Reference Thaller, Krämer, Winter, Kaupe, Erhardt and Fries2003; Kaupe et al. Reference Kaupe, Brandt, Prinzenberg and Erhardt2007; da Silva et al. Reference da Silva, Sonstegard, Thallman, Connor, Schnabel and Van Tassell2010) and content (e.g. Kaminski et al. Reference Kaminski, Brym, Rusc, Wójcik, Ahman and Mägi2006; Näslund et al. Reference Näslund, Fikse, Pielberg and Lundén2008; Signorelli et al. Reference Signorelli, Orrù, Napolitano, De Matteis, Scatà, Catillo, Marchitelli and Moioli2009) as well as FA composition (e.g. Schennink et al. Reference Schennink, Stoop, Visker, Heck, Bovenhuis, van der Poel, van Valenberg and van Arendonk2007, Reference Schennink, Heck, Bovenhuis, Visker, van Valenberg and van Arendonk2008; Conte et al. Reference Conte, Mele, Chessa, Castiglioni, Serra, Pagnacco and Secchiari2010), and also have effects on milk yield (e.g. Spelman et al. Reference Spelman, Ford, McElhinney, Gregory and Snell2002; Banos et al. Reference Banos, Woolliams, Woodward, Forbes and Coffey2008; da Silva et al. Reference da Silva, Sonstegard, Thallman, Connor, Schnabel and Van Tassell2010) and protein yield (e.g. Kaupe et al. Reference Kaupe, Brandt, Prinzenberg and Erhardt2007; Sun et al. Reference Sun, Jia, Ma, Zhang, Wang, Yu and Zhang2009; Berry et al. Reference Berry, Howard, O'Boyle, Waters, Kearney and McCabe2010) and content (e.g. Thaller et al. Reference Thaller, Krämer, Winter, Kaupe, Erhardt and Fries2003; Kaminski et al. Reference Kaminski, Brym, Rusc, Wójcik, Ahman and Mägi2006; Näslund et al. Reference Näslund, Fikse, Pielberg and Lundén2008). However, also milk protein genes existing in different genetic variants, e.g. β-CN, κ-CN and β-lactoglobulin (β-LG), could act as markers in genetic selection programs and have been subject to several studies on associations with milk composition traits during the past years (e.g. Ng-Kwai-Hang, Reference Ng-Kwai-Hang1998; Buchberger & Dovc, Reference Buchberger and Dovc2000; Kaminski et al. Reference Kaminski, Brym, Rusc, Wójcik, Ahman and Mägi2006).
Until now, this approach has usually only been used to examine a few compositional parameters. In this study, milk composition was extensively studied with regard to traits in the protein, lipid and mineral profiles as well as common parameters such as lactose, somatic cell count (SCC) and pH. Therefore, we decided to analyse a numerous number of components reflecting the complexity in milk composition and components correlating with each other to increase the knowledge of gene effects. The aim was to determine if polymorphisms in the LEP, LEPR and DGAT1 genes as well as genetic polymorphism of β-CN, κ-CN and β-LG impact these compositional traits in the two most common breeds of dairy cattle in Sweden, and thus could have the possibility to present useful genetic markers in the breeding work in the future.
Materials and Methods
Animals and samples
A total of 48 individual morning milk samples, blood samples and hair root samples were collected at one sampling occasion in September 2008 from cows belonging to the research farm Nötcenter Viken (Falköping, Sweden). This farm has a unique nucleus breeding herd producing a large number of bull dams and an integrated part of the Swedish breeding occurs at this research farm. The cows were selected based on breed, milk production and cell count (SCC ranging from 7500 to 500 000 cells/ml). Twenty-four cows were of the Swedish Holstein (SH) breed and 24 cows were of the Swedish Red (SR) breed, all being feed the same diet and milked 3 times a day. The cows had been fed in the stable for several weeks and been adjusted to the winter feeding regime at the time for sampling. The SH cows included in this study were sired by 13 proven bulls and the SR cows by 10 proven bulls. The cows were in the second lactation and to exclude extremes in the beginning of the lactation, the cows were in lactation week 7–53. Milk yield of the whole sampling day (the total of 3 milkings) was recorded for each cow. Whole and skim milk samples, defatted by centrifugation at 2000 g for 30 min, were analyzed directly or stored at −20°C.
Milk composition
Samples of fresh milk were analyzed for contents of protein, lipid, lactose and urea as well as for freezing point by using an infrared technique and for somatic cells by using flow cytometry, as described by Glantz et al. (Reference Glantz, Lindmark Månsson, Stålhammar, Bårström, Fröjelin, Knutsson, Teluk and Paulsson2009). For the 10 cows with the lowest SCC of each breed, a more extended study of milk composition traits was made in addition. The protein profile, including contents of CN, whey proteins and non-protein nitrogen, was analyzed by using the Kjeldahl method, the lipid profile, including FA composition and free FAs, was analyzed with gas chromatography and a colorimetric technique, respectively, and the mineral profile, including contents of total Ca and P, was analyzed with inductively coupled plasma mass spectrometry, all in milk samples stored at −20°C. The measurements were performed according to Glantz et al. (Reference Glantz, Lindmark Månsson, Stålhammar, Bårström, Fröjelin, Knutsson, Teluk and Paulsson2009) and the analyses were conducted by a certified dairy analysis laboratory (Eurofins Steins Laboratory, Jönköping, Sweden and Holstebro, Denmark). Yields per cow per day were calculated by multiplying each percentage by the milk yield of the sampling day. The mineral profile also included analyses of free Ca2+ concentrations at 25 and 32°C using an Orion 97–20 Ionplus Calcium Electrode (Thermo Electron Corporation, Beverly, MA, USA), as described previously (Glantz et al. Reference Glantz, Lindmark Månsson, Stålhammar and Paulsson2011). Additionally, pH of fresh milk samples was measured. The samples were analyzed at least in duplicate for all the above-mentioned traits.
Genetic analyses
Genotypes and alleles were determined on 4 single nucleotide polymorphisms (SNP) within the LEP gene, 1 SNP within the LEPR gene and 1 SNP within the DGAT1 gene for the 48 cows. The cows were genotyped for the A1457G and C963T polymorphisms in the LEP promoter region, the A252T and A59V polymorphisms in the LEP exon 2 and 3 regions, respectively, and the T945M polymorphism in the LEPR exon 20 region. The SNP determined at the DGAT1 gene locus was Nt984+8(A-G) in exon 12. Mutations were chosen on the basis of available knowledge about their associations with economically important traits (Liefers et al. Reference Liefers, Veerkamp, Te Pas, Delavaud, Chilliard, Platje and van der Lende2005; Komisarek & Dorynek, Reference Komisarek and Dorynek2006; Banos et al. Reference Banos, Woolliams, Woodward, Forbes and Coffey2008), commercially available (Igenity, Harlow, UK) and on potential important SNPs for milk composition. The analyses were performed by Igenity (Harlow, UK) and in brief, DNA was extracted from hair root samples and the genotyping was carried out with allele specific primer extension reactions analyzed with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry.
Genetic variants of β-CN (A1, A2), κ-CN (A, B, E) and β-LG (A, B) were determined for the 10 cows with the lowest SCC of each breed in blood samples with polymerase chain reaction based methods. The analyses were conducted by Dr Van Haeringen Laboratorium b.v. (Wageningen, the Netherlands).
Statistical analyses
To estimate the effects of polymorphisms and genetic variants of milk proteins on milk composition traits, univariate models were used. The general linear model procedure from Minitab (version 14, Minitab Ltd, Coventry, UK) was used to calculate least squares means and se and for pairwise testing of significant differences in alleles. For the LEP, LEPR and DGAT1 genes, data was analyzed with the following model, based on the selection criteria:
where y ijk(l)=milk trait; μ=general mean; breedi=fixed effect of breed (i=SR or SH); milk volumej=fixed effect of milk production (j=10–20 kg, 21–30 kg or >30 kg); gene allelek=fixed effect of gene allele (k=LEP A1457G A or G; LEP A252T A or T; LEP A59V C or T; LEP C963T C or T; LEPR T945M C or T; DGAT1 Nt984+8(A-G) A or G); cell countl=fixed effect of cell count (l=<50 000 cells/ml, 51–100 000 cells/ml, 101–200 000 cells/ml or >200 000 cells/ml) for the study including 48 cows to account for the selection based on this parameter in this data set; εijk(l)=random residual effect. A fixed effect of lactation stage as well as interactions have been analyzed but were not significant and were therefore excluded from the statistical model.
Due to the close linkage between the CN loci (Bovenhuis et al. Reference Bovenhuis, van Arendonk and Korver1992), a multigene model was used for the genetic variants of milk proteins:
where y ijklm=milk trait; μ=general mean; breedi=fixed effect of breed (i=SR or SH); milk volumej=fixed effect of milk production (j=10–20 kg, 21–30 kg or >30 kg); β-CN allelek=fixed effect of β-CN allele (k=A1 or A2); κ-CN allelel=fixed effect of κ-CN allele (l=A, B or E); β-LG allelem=fixed effect of β-LG allele (m=A or B); εijklm=random residual effect. A fixed effect of lactation stage as well as interactions have been analyzed but were not significant and were therefore excluded from the statistical model.
Results and Discussion
Means and variation of analyzed milk traits
The mean values and sd of the analyzed milk traits are presented in Tables 1 & 2. In spite of the high amount of milk produced by the cows in this study, the contents of milk components are average and in the range previously found for individual milk samples from SR and SH cows (Lundén et al. Reference Lundén, Nilsson and Janson1997; Wedholm et al. Reference Wedholm, Hallén, Larsen, Lindmark Månsson, Karlsson and Allmere2006a; Hallén et al. Reference Hallén, Wedholm, Andrén and Lundén2008; Näslund et al. Reference Näslund, Fikse, Pielberg and Lundén2008). The mean values obtained in this study for protein and lipid contents are higher for both SR and SH cows, except for lipid content of SH cows, compared with data from the national cow database (protein content: 3·53 g/100 g and 3·36 g/100 g for SR and SH, respectively, and lipid content: 4·35 g/100 g and 4·07 g/100 g for SR and SH, respectively). Only a minor breed difference is found for components in the protein and mineral profiles, with SR cows having a higher urea content, lower ionic Ca at 25°C and higher milk pH compared with the SH cows. However, for the FA composition, the differences are larger (P=0·000–0·046). SR cows are shown to have a higher amount of saturated FA, whereas SH cows have higher amounts of monounsaturated FA, polyunsaturated FA, C20 and >C20 and n−3 FA. Short and medium chain FAs (C4 to C14 and part of C16) are produced in the lactating cell (Walstra & Jenness, Reference Walstra and Jenness1984). From Table 2 it can be derived that SR cows produce higher amounts of short (C4–C6) and medium chain (C8–C12) FAs than SH cows. This observation is very interesting. Although the cows are being fed the same diet, there are possibilities to change FA composition through the use of genetic selection.
Table 1. Unadjusted means±sd of protein, non-protein nitrogen (NPN) and mineral profiles as well as lactose, somatic cell count (SCC), freezing point and pH in milk from the Swedish Red and Swedish Holstein breeds
† n=24 for each breed
‡ n=10 for each breed
Table 2. Unadjusted means±sd of lipid profile and fatty acid (FA) composition† in milk from the Swedish Red and Swedish Holstein breeds (n=10 for each breed)
† Mean fatty acid as a proportion (wt/wt) of the total fat fraction of 100%
‡ n=24 for each breed
The data in the current study are based on one morning milk sample per cow. However, the samples were taken at the same milking, in a research herd, in the beginning of the winter feeding season and in the middle of the lactation from healthy cows with a high production, which makes this study design appropriate. It has been shown that both yield and content of milk components vary according to stage of lactation (Auldist et al. Reference Auldist, Walsh and Thomson1998; Kay et al. Reference Kay, Weber, Moore, Bauman, Hansen, Chester-Jones, Crooker and Baumgard2005; Kgwatalala et al. Reference Kgwatalala, Ibeagha-Awemu, Mustafa and Zhao2009). However, correlations exist between milk composition data of the morning milk sampling and the cows' 305-days lactation (data not shown; Glantz et al. 2012), justifying that the morning milk samples in this study can be used to represent the entire lactation.
Allele and genotype frequencies
The allele and genotype frequencies estimated for the individual SNP loci within the LEP, LEPR and DGAT1 genes are presented in Table 3. It can be seen that frequencies differ somewhat between the SH and SR cows, although the same tendency of minor allele frequency is found for LEP A252T, LEP A59V, LEPR T945M and DGAT1 Nt984+8(A-G) for both the breeds. The cows included in this study belong to a research farm and are selected by this farm based for high Nordic total merit, which is a combined value of production, functional and conformation traits. The observed distribution of all genotypes is consistent with expected values obtained according to the Hardy-Weinberg law, except for LEP A252T for SR cows and DGAT1 Nt984+8(A-G) for SH cows. This might be explained by the limited number of cows included in the study. The higher frequency of the A allele of LEP A1457G compared with the G allele (56% vs. 44%) and the higher frequency of the C allele of LEP C963T compared with the T allele (61% vs. 39%) for the SH cows are in agreement with previous studies on Holstein-Friesian heifers (Liefers et al. Reference Liefers, Veerkamp, Te Pas, Delavaud, Chilliard, Platje and van der Lende2005) and UK Holstein cows (Banos et al. Reference Banos, Woolliams, Woodward, Forbes and Coffey2008). The same tendency was also reported in Polish Holstein-Friesian cattle regarding LEP C963T (Komisarek, Reference Komisarek2010). Furthermore, Banos et al. (Reference Banos, Woolliams, Woodward, Forbes and Coffey2008) reported a higher frequency of the LEP A252T A variant compared with the LEP A252T T variant in UK Holstein cows, which confirms our results for SH cows. Also consistent with Banos et al. (Reference Banos, Woolliams, Woodward, Forbes and Coffey2008), we found no TT genotype of LEP A252T for either SH or SR cows. For LEPR T945M, a higher frequency of the C variant compared with the T variant (88% vs. 12%) is found for the SH cows, and the same tendency has been reported earlier in Holstein-Friesian cows (Liefers et al. Reference Liefers, Veerkamp, te Pas, Delavaud, Chilliard and van der Lende2004), UK Holstein cows (Banos et al. Reference Banos, Woolliams, Woodward, Forbes and Coffey2008) and Polish Holstein-Friesian cattle (Komisarek, Reference Komisarek2010). In addition, none or a minor frequency (1%) of the TT genotype of LEPR T945M have been reported previously (Liefers et al. Reference Liefers, Veerkamp, te Pas, Delavaud, Chilliard and van der Lende2004; Banos et al. Reference Banos, Woolliams, Woodward, Forbes and Coffey2008), which is in agreement with our results for both the breeds. Thus, despite the limited sample of cows, the frequencies for SH cows seem valid. However, no comparisons with earlier studies are available for SR cows regarding the studied SNPs since, to our knowledge, this is the first time these SNPs have been studied for SR cows.
Table 3. Allele and genotype frequencies in the leptin, leptin receptor and acyl-CoA:diacylglycerol acyltransferase 1 (DGAT1) genes in the Swedish Red (SR; n=24) and Swedish Holstein (SH; n=24) breeds
† A=adenine, G=guanine, T=thymine, C=cytosine
Table 4 shows the allele and genotype frequencies for β-CN, κ-CN and β-LG. For all of the analyzed milk protein genes, the same tendency of minor allele frequency is found for both the breeds. The allele frequencies observed for β- and κ-CN are consistent with the trends found previously for SR and SH cows (Lundén et al. Reference Lundén, Nilsson and Janson1997; Hallén et al. Reference Hallén, Allmere, Näslund, Andrén and Lundén2007) and Holstein-Friesian cattle (Bobe et al. Reference Bobe, Beitz, Freeman and Lindberg1999; Heck et al. Reference Heck, Schennink, van Valenberg, Bovenhuis, Visker, van Arendonk and van Hooijdonk2009). In the present data set, the results indicate a higher frequency of the κ-CN E allele, which has been shown to be associated with poor milk coagulation properties (Ikonen et al. Reference Ikonen, Ahlfors, Kempe, Ojala and Ruottinen1999a; Wedholm et al. Reference Wedholm, Larsen, Lindmark Månsson, Karlsson and Andrén2006b; Hallén et al. Reference Hallén, Allmere, Näslund, Andrén and Lundén2007), in SR cows (30%) compared with SH cows (10%), which also has been reported earlier (Ikonen et al. Reference Ikonen, Ojala and Syväoja1997; Lien et al. Reference Lien, Kantanen, Olsaker, Holm, Eytorsdottir, Sandberg, Dalsgard and Adalsteinsson1999; Hallén et al. Reference Hallén, Wedholm, Andrén and Lundén2008). The higher frequency of the A variant of β-LG compared with the B variant for both SR and SH (60% vs. 40% and 70% vs. 30%, respectively) is in agreement with the results reported in Holstein-Friesian cattle (Heck et al. Reference Heck, Schennink, van Valenberg, Bovenhuis, Visker, van Arendonk and van Hooijdonk2009), but contradictory to the results reported earlier for SR and SH cows (Lundén et al. Reference Lundén, Nilsson and Janson1997; Hallén et al. Reference Hallén, Allmere, Näslund, Andrén and Lundén2007).
Table 4. Allele and genotype frequencies of β-lactoglobulin (β-LG), β-casein (β-CN) and κ-casein (κ-CN) in the Swedish Red (SR; n=10) and Swedish Holstein (SH; n=10) breeds
Effects of polymorphisms in the LEP, LEPR and DGAT1 genes on milk composition
Table 5 shows effects of polymorphisms in the LEP, LEPR and DGAT1 genes on milk composition traits. Results are shown only for (1) SNPs that had an effect (P<0·10) on at least one compositional trait and (2) components that were significantly (P<0·10) influenced by the studied SNPs. The level of significance was set at P<0·10 due to the low sample number in an attempt to indicate interesting results that may be of merit and importance. However many results were significant at P<0·05 and these have been highlighted in Tables 5 & 6. The results demonstrate that there are associations between LEP A1457G, LEP A59V, LEP C963T, LEPR T945M as well as DGAT1 Nt984+8(A-G) and milk composition in Swedish dairy cattle. The LEP A1457G polymorphism is shown to have an effect on protein content, with the G allele giving a higher content. The same observation was made by Liefers et al. (Reference Liefers, Veerkamp, Te Pas, Delavaud, Chilliard, Platje and van der Lende2005) in a study on Holstein-Friesian heifers. In the same study, however, no association was observed with lactose content (Liefers et al. Reference Liefers, Veerkamp, Te Pas, Delavaud, Chilliard, Platje and van der Lende2005), whereas our study indicates a favourable effect of the G allele on this component. Furthermore, the results show an effect on concentration of ionic Ca, milk pH and some FA. Also consistent with our results, Liefers et al. (Reference Liefers, Veerkamp, Te Pas, Delavaud, Chilliard, Platje and van der Lende2005) found no effect of this polymorphism on milk, protein, lipid or lactose yields as well as lipid content, whereas others have reported higher milk yield in G allele cows (Banos et al. Reference Banos, Woolliams, Woodward, Forbes and Coffey2008). For the LEP A1457G polymorphism, the reliability of the estimates for the analyzed traits increased with 2–11%.
Table 5. Effects of leptin (LEP), leptin receptor (LEPR) and acyl-CoA:diacylglycerol acyltransferase 1 (DGAT1) alleles on milk composition traits.† Results are shown only for components with P<0·10 (results with P<0·05 are shown in bold)
† Values are expressed as least squares means±se. Mean fatty acid as a proportion (wt/wt) of the total fat fraction of 100%
‡ FA=fatty acid
Table 6. Effects of β-lactoglobulin, β-casein and κ-casein alleles on milk composition traits.† Results are shown only for components with P<0·10 (results with P<0·05 are shown in bold)
† Values are expressed as least squares means±se. Mean fatty acid as a proportion (wt/wt) of the total fat fraction of 100%
ab Mean values within a row with different superscripts differ (P<0·10)
The C allele of LEP A59V is shown to give higher protein and total Ca contents, total Ca and P yields, milk pH and SCC. Also FA composition is shown to be affected by this polymorphism. The reliability of the estimates for these traits increased with 1–11%. However, earlier studies on Jersey cows have shown no effect of this polymorphism on protein content (Kulig et al. Reference Kulig, Kmiec, Kowalewska-Luczak and Andziak2009) or SCC (Kulig et al. Reference Kulig, Kmiec and Wojdak-Maksymiec2010) and Komisarek et al. (Reference Komisarek, Szyda, Michalak and Dorynek2005) found no association between the LEP A59V polymorphism and breeding value for protein content in Holstein–Friesian bulls. On the other hand, associations were found with breeding values for lipid yield and content in the same study (Komisarek et al. Reference Komisarek, Szyda, Michalak and Dorynek2005), which might support the associations between LEP A59V and FA composition found in this study.
The LEP C963T polymorphism is shown to be associated with milk yield, protein and lactose contents, milk pH and concentration of ionic Ca. The C allele cows produce more milk than T allele cows, but decreased contents of protein and lactose. These results are in contrast to earlier studies on Holstein-Friesian heifers showing no effect of LEP C963T on milk yield or protein and lactose contents (Liefers et al. Reference Liefers, Veerkamp, Te Pas, Delavaud, Chilliard, Platje and van der Lende2005) and on UK Holstein cows showing no effect on milk yield for this polymorphism (Banos et al. Reference Banos, Woolliams, Woodward, Forbes and Coffey2008). For the LEP C963T polymorphism, the reliability of the estimates for the analyzed traits increased with 2–9%.
No significant (P<0·10) effect of the LEP A252T polymorphism on any of the analyzed traits was found in this study. Banos et al. (Reference Banos, Woolliams, Woodward, Forbes and Coffey2008) have earlier reported on a favourable effect of the T allele on milk yield for this polymorphism in UK Holstein cows.
The T allele of LEPR T945M is shown to be associated with increased concentration of ionic Ca, but decreased milk pH. The reliability of the estimates for these traits increased with 7–16%. Komisarek & Dorynek (Reference Komisarek and Dorynek2006) have previously reported on an effect of this polymorphism on protein and lipid contents in Jersey cows, however, this can not be confirmed in this study. Other studies have analyzed milk yield (Komisarek & Dorynek, Reference Komisarek and Dorynek2006; Banos et al. Reference Banos, Woolliams, Woodward, Forbes and Coffey2008), protein and lipid yields (Komisarek & Dorynek, Reference Komisarek and Dorynek2006) and SCC (Komisarek, Reference Komisarek2010), but found no effect on these traits, which is in agreement with the results in our study.
In Table 5 it is shown that the LEP polymorphisms have an effect on protein content, total or ionic Ca and milk pH, and also the LEPR polymorphism is associated with the latter traits. All of these parameters have been shown to affect milk gelation (Amenu & Deeth, Reference Amenu and Deeth2007; Glantz et al. Reference Glantz, Devold, Vegarud, Lindmark Månsson, Stålhammar and Paulsson2010). Considering that more than 30% of the milk produced in Sweden is being processed into cheese (Swedish Dairy Association, 2009), it would be valuable to include these traits in the breeding objective. By doing this, an indirect selection of cheese characteristics would be possible. However, it could be of interest to examine the effects of the LEP and LEPR polymorphisms on technological properties important for cheese production, in order to have a direct selection of cheese characteristics.
The DGAT1 gene has been shown to have major effect on milk fat composition (Grisart et al. Reference Grisart, Coppieters, Farnir, Karim, Ford, Berzi, Cambisano, Mni, Reid, Simon, Spelman, Georges and Snell2002), but also on milk yield and protein yield and content (Kaupe et al. Reference Kaupe, Brandt, Prinzenberg and Erhardt2007; Berry et al. Reference Berry, Howard, O'Boyle, Waters, Kearney and McCabe2010; da Silva et al. Reference da Silva, Sonstegard, Thallman, Connor, Schnabel and Van Tassell2010). The K232A polymorphism has been shown to be the most likely polymorphism that causes this QTL effect (Grisart et al. Reference Grisart, Coppieters, Farnir, Karim, Ford, Berzi, Cambisano, Mni, Reid, Simon, Spelman, Georges and Snell2002). However, also other polymorphisms within the DGAT1 gene have been revealed (Grisart et al. Reference Grisart, Coppieters, Farnir, Karim, Ford, Berzi, Cambisano, Mni, Reid, Simon, Spelman, Georges and Snell2002), which suggests that other compositional traits could be influenced by these polymorphisms. In this study, the Nt984+8(A-G) polymorphism was analyzed. As can be seen in Table 5, associations are only found between DGAT1 Nt984+8(A-G) and some FA and SCC. The K232A polymorphism has, however, been shown to impact most of the FAs in milk (Schennink et al. Reference Schennink, Stoop, Visker, Heck, Bovenhuis, van der Poel, van Valenberg and van Arendonk2007, Reference Schennink, Heck, Bovenhuis, Visker, van Valenberg and van Arendonk2008; Conte et al. Reference Conte, Mele, Chessa, Castiglioni, Serra, Pagnacco and Secchiari2010). Thus, since no other associations were found between the Nt984+8(A-G) polymorphism and other compositional traits in our study, it is suggested that this polymorphism is not a desirable marker within genetic selection programs.
Effects of alleles of β-CN, κ-CN and β-LG on milk composition
Effects of β-CN, κ-CN and β-LG alleles on milk composition traits are presented in Table 6. Results are shown only for components that were significantly (P<0·10) influenced by the studied milk protein genes. As for polymorphisms in the LEP, LEPR and DGAT1 genes, genetic variants of the studied milk protein genes also show an effect on milk composition. The A allele of β-LG is shown to give more whey proteins and increase the concentration of ionic Ca compared with the B allele, whereas the B allele is shown to increase the amount of some FA. For β-LG, the reliability of the estimates for these parameters increased by 5–8%. Earlier studies on effects of β-LG on milk composition involves mainly major milk production traits, such as milk yield as well as yields and contents of protein, casein, lipid and lactose, however, the results are often contradictory (Lundén et al. Reference Lundén, Nilsson and Janson1997; Ikonen et al. Reference Ikonen, Ojala and Ruottinen1999b; Buchberger & Dovc, Reference Buchberger and Dovc2000). Despite this, we found no effect of β-LG on any of these traits in this study. The A2 variant of β-CN is shown to give higher yields of milk, protein, casein, whey protein, non-protein nitrogen, lactose, total Ca and P as well as higher contents of lactose and total Ca compared with the A1 variant. The results found are in agreement with previous studies on milk and protein yields (Bovenhuis et al. Reference Bovenhuis, van Arendonk and Korver1992; Lundén et al. Reference Lundén, Nilsson and Janson1997; Ikonen et al. Reference Ikonen, Ojala and Ruottinen1999b; Heck et al. Reference Heck, Schennink, van Valenberg, Bovenhuis, Visker, van Arendonk and van Hooijdonk2009) as well as casein and lactose yields (Lundén et al. Reference Lundén, Nilsson and Janson1997). However, Lundén et al. (Reference Lundén, Nilsson and Janson1997) found no effect of β-CN on lactose content. β-CN is shown to increase the reliability of the estimates by 2–10%. κ-CN is shown to influence the yields of milk, protein, casein, lactose, total Ca and P as well as whey protein content, lactose content and milk pH. The results indicate that the yields and contents as well as pH increase with allele A>B>E, except for whey protein content showing the opposite tendency. For κ-CN, the reliability of the estimates for the analyzed traits increased by 2–26%. The result obtained for protein yield is consistent with the study by Aleandri et al. (Reference Aleandri, Buttazzoni, Schneider, Caroli and Davoli1990), however, other studies have found no effect (Lundén et al. Reference Lundén, Nilsson and Janson1997) or effects on other compositional traits (Ng-Kwai-Hang, Reference Ng-Kwai-Hang1998; Buchberger & Dovc, Reference Buchberger and Dovc2000; Kaminski et al. Reference Kaminski, Brym, Rusc, Wójcik, Ahman and Mägi2006).
Markers in genetic selection programs
The composition of milk is crucial for both the nutritional value as well as technological properties of milk. Depending on the application, different components are of importance, thus making animal breeding an efficient tool in optimizing milk quality traits. Thus far, however, production traits considered in the Nordic breeding objective are only protein and lipid yields as well as a negative weight for milk yield, which thus favour animals with higher contents (Nordic Cattle Genetic Evaluation, 2011). By including more traits within breeding programs, a direct selection of important traits will be possible. Despite the limited material, the results have demonstrated several interesting and important associations between alleles of LEP A1457G, LEP A59V, LEP C963T, LEPR T945M, β-CN, κ-CN as well as β-LG and milk composition traits in Swedish dairy cattle. Thus, this study shows possibilities of using these polymorphisms as markers within genetic selection programs to improve and adjust several parameters. However, when comparing different studies, results are often contradictory. Thus, it is essential to verify these results as well as include more cows in the study before breeding practices are being adjusted for direct selection of more compositional traits. Using genetic markers to select for specific parameters will accelerate the present breeding work and thus increase the economical output, however, a genetic marker will most probably not influence only one trait. Selecting for a specific trait may result in another one being unfavourably altered. Hence, it is important to control the selection so that milk composition does not negatively influence the nutritional value and processability of milk. In conclusion, the results of this study show several interesting associations and hence, the future importance of this rapidly developing and challenging research field.
The authors wish to thank the staff at Nötcenter Viken for invaluable help during milking and the collection of data as well as Mia Davidsson and Johanna Ryssnäs at the Department of Food Technology, Engineering and Nutrition, Lund University, Sweden, for their help with the ionic Ca measurements. The financial support from Sparbanksstiftelsen Färs & Frosta, Sjöbo, Sweden, and the Swedish Farmer's Foundation for Agricultural Research (SLF), Stockholm, Sweden, are gratefully acknowledged.
