Marked Effect of β-Lactoglobulin Polymorphism on the Ratio of Casein to Total Protein in Milk

Marked Effect of b-Lactoglobulin Polymorphism on the Ratio of Casein to Total Protein in Milk ´ N, M. NILSSON, and L. JANSON A. LUNDE Department of Animal Breeding and Genetics, Swedish University of Agricultural Sciences, PO Box 7023, S-750 07 Uppsala, Sweden

ABSTRACT The relationship between genetic variants for milk protein and the composition of milk was analyzed on 4475 repeated milk samples from individual cows; 371 dairy cows of the Swedish Red and White breed and 204 cows of the Swedish Holstein breed were used. The registrations included percentages of casein, protein, fat, and lactose in combination with milk yield and SCC. The genotype of individual cows for as1-CN, b-CN, k-CN, and b-LG was determined by alkaline and acidic PAGE. A mixed animal model was used for the analysis; b-LG and aggregate casein genotypes were included simultaneously as separate fixed effects in the statistical model. The results suggest a positive additive effect of the b-LG B allele on casein content and on the ratio of casein to total protein. For the latter trait, the b-LG genotype accounted for a relatively large part of the phenotypic variance, corresponding to a reduction in residual variance of 11% when included in the model. The corresponding value for casein content was 0.5%. The lack of unfavorable associations between milk protein variants and the traits included in this study makes the b-LG gene an obvious candidate when the breeding objective is improved conversion of milk protein into cheese. ( Key words: milk protein, b-lactoglobulin gene, casein, milk production) Abbreviation key: lnSCC = SCC transformed to the natural logarithm scale, SLB = Swedish Holstein breed, SRB = Swedish Red and White breed. INTRODUCTION Traditionally, fluid milk and other dairy products have been essential components of the human diet in a variety of cultures, mainly because of the high quality of the protein contained in milk and the variety of minerals and vitamins. Today, the dairy indus-

try has the technological ability to produce large quantities of a broad spectrum of dairy products. However, the dairy industry continually strives to improve the quality of its products. It has long been known that the manufacturing properties of milk are influenced by the relative composition of its proteins. The proportion of the protein components in milk show individual variation because of environmental and genetic factors. The genes that encode the major milk proteins are thought of as candidate genes for the observed variation in protein composition. As an example, cheese yield is related to the casein content in milk, particularly the ratio of casein to total protein (casein number). This ratio is partly controlled by the different genetic variants of the polymorphic milk proteins, and b-LG have the largest effect (16, 17, 19, 21, 24). The effect is likely to be associated with observed differences in levels of expression of the b-LG alleles in heterozygous cows (7, 8). However, before the genetic variability of milk proteins can be used in selection programs, it needs to be established that no unfavorable associations exist between the economically interesting milk protein alleles and other traits of importance for milk production. Also, the gene frequencies in the population of interest should be investigated because the prospects of improving a given trait depend on the frequency of the favorable allele. Finally, the size and nature of the gene effects should be firmly established because implementation of poorly defined markers in existing breeding programs is likely to do more harm than good. Presented here are the results from a study aimed at investigating the effects of milk protein genes that segregate in the Swedish Red and White ( SRB) and the Swedish Holstein ( SLB) populations on the yield and composition of milk. Also, observed frequencies of the different milk protein variants are given for the two breeds. MATERIALS AND METHODS

Received September 9, 1996. Accepted May 5, 1997. 1997 J Dairy Sci 80:2996–3005

The present study included 394 SRB cows and 251 SLB cows belonging to four different experimental 2996

β-LACTOGLOBULIN GENES AND CASEIN IN MILK TABLE 1. Mean test day yield and composition of milk from 371 Swedish Red and White cows. Trait

n

X

SD

Milk yield, kg Protein content, % Protein yield, g Casein content, % Casein yield, g Casein number1 Fat content, % Fat yield, g Lactose content, % Lactose yield, g ln SCC2

2934 2921 2921 2945 2923 2945 2921 2921 2921 2921 2919

22.39 3.42 753 2.47 344 73.14 4.71 1041 4.64 1041 3.99

6.63 0.32 193 0.24 90 2.11 0.65 303 0.22 319 1.22

1(Casein/total 2Transformed

protein) × 100. to the natural logarithm scale.

2997

The different sampling occasions were grouped into four sampling periods based on observations from the data on the seasonal variation in casein and protein contents. Parity numbers >5 were pooled with parity number 5. Stage of lactation had a linear relationship with milk components between lactation wk 4 and 51. Therefore, after observations outside this interval were discarded, the effect of stage of lactation was included in the model as a covariate. The phenotype of individual cows for polymorphic milk proteins as1-CN, b-CN, k-CN, and b-LG was determined by PAGE at alkaline and acidic pH, essentially as described by Swaisgood (23). Statistical Methods

herds in Sweden. Two of the herds were pure SLB herds, one was a pure SRB herd, and the fourth herd included cows of both breeds. Separate afternoon and morning milk samples were collected monthly on each individual cow from September 1989 until December 1990. Altogether, the analysis included 5336 test day registrations on percentages of protein, casein, fat, and lactose in combination with milk yield. Based on these data, individual values were calculated for yields of protein, casein, fat, and lactose and for casein number. In parallel, the SCC of the milk samples was analyzed. The milk samples from the different herds were treated with Bronopol (Boots Microcheck, Nottingham, England) and sent for compositional analysis to the milk laboratory of the Department of Animal Breeding and Genetics (Uppsala, Sweden). The SCC of the milk were determined with a cell counter (Fossomatic 90; A/S Foss Electric, Hillerød, Denmark), and the protein content was determined by infrared techniques (Milko Scan 93; A/S Foss Electric). Casein content of the individual morning samples was analyzed by first adding 80 ml of CaCl2 to 20 ml of skimmed milk. After the milk sample was heated to 42°C, 100 ml of rennet were added, and activity continued for 3 min. The coagulum was cut, 3 min were allowed for syneresis, and the whey fraction was separated from the caseins by draining on a steel net (42 mm in diameter). The whey protein content was analyzed with a Milko Scan 93. The casein content was expressed as the difference between total protein and whey protein. Means and standard deviations of the milk production traits for the SRB and SLB breeds are in Tables 1 and 2, respectively. In addition to information on herd, data were collected on individual milk samples for sampling period, lactation number, stage of lactation, and SCC transformed to the natural logarithm scale ( ln SCC) .

To account for the random, additive genetic effect of the polygenic background, a mixed animal model was used to estimate the fixed effects of milk protein genotypes on milk composition. The PEST package ( 9 ) was used for the analyses, and input values of the variance components were estimated using the VARCOMP procedure in the SAS package (20). Relationships among animals from sires, dams, paternal grandsires, and maternal grandsires were considered. In total, 825 SRB and 465 SLB cows were included in the analysis when ancestors without records were also considered. Because data were taken from several observations on each individual, a random environmental effect of cow was added to the model. Because of the strong genetic linkage between the different casein loci (13), the alleles at these loci could not be expected to segregate independently of each other, resulting in a restricted number of allele combinations. Thus, in the analyses, the casein loci were treated jointly as aggregate genotypes, making it possible to also account for variation in intervening

TABLE 2. Mean test day yield and composition of milk from 204 Swedish Holstein cows. Trait

n

X

SD

Milk yield, kg Protein content, % Protein yield, g Casein content, % Casein yield, g Casein number1 Fat content, % Fat yield, g Lactose content, % Lactose yield, g ln SCC2

1534 1522 1522 1530 1530 1530 1522 1522 1522 1522 1520

22.37 3.29 723 2.36 323 72.54 4.08 896 4.63 1045 4.39

7.54 0.32 218 0.24 99 2.13 0.64 292 0.21 370 1.11

1(Casein/total 2Transformed

protein) × 100. to the natural logarithm scale. Journal of Dairy Science Vol. 80, No. 11, 1997

2998

´ N ET AL. LUNDE

sequences that may harbor regulatory elements affecting gene expression, as has been discussed by Lien et al. (12). Because the number of aggregate genotypes that could be analyzed differed considerably between SRB and SLB, the two breeds were analyzed separately. To obtain reasonable numbers in each group, each aggregate genotype that was included in the analysis of genotype effects was based on at least 7 cows for each of the two breeds. The following model was used: yijklmno = hi + spj + lnk + cgl + blgm + b1slijklmno + b2SCCijklmno + cijklmn + aijklmn + eijklmno

typic value of the heterozygote, d, illustrates the effect of dominance within the locus. In addition to the estimations of genotypic values regarding these three loci, contrasts between existing genotypes were performed. The contribution of each of the various fixed effects, relative to the residual variance, was estimated as the relative difference in mean square error between the full model and models that ignored one effect at a time. This measure relates to the proportion of the phenotypic variance accounted for by the respective fixed effects. Also, the contributions from the random effects of animal and permanent environment of cow were estimated.

where yijklmno = observation ijklmno of cow ilmn; hi = the fixed effect of herd i ( i = 1, 2, or 3 for SRB, and i = 1, 2 for SLB); spj = fixed effect of sampling period j ( j = 1, 2, 3, or 4 for September 1989 to December 1989, January 1990 to April 1990, May 1990 to August 1990 and September 1990 to December 1990, respectively); lnk = fixed effect of lactation number k ( k = 1, 2, ... 5; lactation numbers >5 were pooled with lactation numbers of 5); cgl = fixed effect of casein genotype l ( l = 1, 2, ... 6 for SRB, and l = 1, 2 ... 9 for SLB); blgm = fixed effect of b-LG genotype m ( m = 1, 2, or 3); b1 = regression coefficient of the observation on stage of lactation; slijklmno = stage of lactation ( w k ) of observation ijklmno of cow ilmn; b2 = regression coefficient of the observation on lnSCC; SCCijklmno = lnSCC of observation ijklmno of cow ilmn; cilmn = random environmental effect of cow ilmn; ailmn = random, additive, genetic effect of polygenic background for cow ilmn; and eijklmno = residual random term. For b-LG, b-CN, and k-CN (only for SLB), the relevant contrasts were set up to estimate the genotypic values ( 5 ) of the three possible genotypes within each locus. Because the estimation of the genotypic values of the homozygous genotypes, a or –a, does not include heterozygous individuals, a represents purely additive gene effects, and the genoJournal of Dairy Science Vol. 80, No. 11, 1997

RESULTS The gene frequencies, calculated with the gene counting method, are given for the SRB and SLB breeds in Tables 3 and 4, respectively. The genotype frequencies of each locus were tested for possible deviations from the Hardy-Weinberg equilibrium using chi-square calculations. However, no unequal segregation of alleles between genotypes was shown for any of the loci. For the SRB breed, only the b-CN, kCN, and b-LG proteins show genetic variation. For bCN, only the alleles A1 and A2 occurred at frequencies >1%. Six aggregate genotypes with ≥7 cows were found within the SRB breed (Table 5), and nine genotypes were found within the SLB breed (Table 6). Within SRB, as1-CN was shown to be monomorphic BB, and the k-CN BB genotype was only found in 6 individuals, which explains the lower number of aggregate genotypes for this breed. The results from the analysis of effects of milk protein variants on yield and composition of milk are

TABLE 3. Gene frequencies of b-LG, as1-CN, b-CN, and k-CN in 3941 Swedish Red and White cows. Locus

Allele

Frequency

b-LG

A B B A1 A2 A3 B A B

(%) 33.3 66.7 100.0 46.0 53.1 0.1 0.8 83.3 16.7

as1-CN b-CN

k-CN

1Includes genotyped cows with incomplete information on milk traits.

β-LACTOGLOBULIN GENES AND CASEIN IN MILK TABLE 4. Gene frequencies of b-LG, as1-CN, b-CN, and k-CN in 2511 Swedish Holstein cows. Locus b-LG as1-CN b-CN

k-CN

Allele

Frequency

A B B C A1 A2 A3 B A B

(%) 49.8 50.2 85.5 14.5 46.4 52.6 0.2 0.8 80.1 19.9

1Includes genotyped cows with incomplete information on milk traits.

shown in Tables 7 and 8 for SRB and Tables 9 and 10 for SLB. The genotypic value of b-LG BB was significantly ( P < 0.0001) associated with high casein number and casein content ( P < 0.01) in both breeds independently. All three possible b-LG genotypes differed significantly ( P < 0.0001) from one another in casein number, and the heterozygote was intermediate. Based on the genotypic values given in Tables 7 and 9, the difference between the two homozygotes in casein number was estimated to be approximately 3 percentage units for both breeds,

2999

corresponding to 1.5 phenotypic standard deviations for this trait. The as1-CN polymorphism, observed only within the SLB breed, was associated ( P < 0.01) with the content and yield of fat; the BB variant showed a positive relationship. The b-CN A1A1 was related to low yields of milk, casein, protein, and lactose, but these effects were only significant ( P < 0.01) within the SRB breed. No effects of dominance were found within the b-CN locus. The contrasts between the kCN genotypes were not significant for any of the traits within the two breeds. The proportions of the phenotypic variance accounted for by the different fixed and random effects, expressed as the difference in mean square error between the full and reduced models relative to the residual variance, are given for casein content and casein number in Table 11. The presentation is restricted to those traits for which significant effects of milk protein genes were found in both breeds independently. Because effects for these two traits were similar for both SRB and SLB, data for the two breeds were pooled for this analysis. The b-LG genotype accounted for a considerable part of the phenotypic variance in casein number, corresponding to a reduction of the residual variance of 11.1% when included

TABLE 6. Frequencies of milk protein genotypes, both as separate and aggregate as1-CN, b-CN, and k-CN genotypes, in 2041 Swedish Holstein cows included in the analyses of genotype effects on yield and composition of milk. Locus

Genotype

Frequency

TABLE 5. Frequencies of milk protein genotypes, both as separate and aggregate b-CN and k-CN genotypes, in 3711 Swedish Red and White cows included in the analyses of genotype effects on yield and composition of milk.

b-LG

Locus

as1-CN

AA AB BB BB BC A1A1 A1A2 A2A2 AA AB BB

(%) 23.5 52.5 24.0 76.5 23.5 19.6 52.5 27.9 70.1 26.5 3.4

b-LG

as1-CN b-CN

k-CN Casein genotype2

Genotype

Frequency

AA AB BB BB A1A1 A1A2 A2A2 AA AB b-CN b-CN b-CN b-CN b-CN b-CN

(%) 12.1 41.0 46.9 100.0 22.1 49.9 28.0 69.3 30.7 18.3 3.8 35.6 14.3 15.4 12.7

A1A1 A1A1 A1A2 A1A2 A2A2 A2A2

k-CN k-CN k-CN k-CN k-CN k-CN

AA AB AA AB AA AB

1Included here are only cows belonging to casein genotype groups comprising ≥7 cows. 2Aggregate b-CN and k-CN genotype.

b-CN

k-CN

Casein genotype2

as1-CN as1-CN as1-CN as1-CN as1-CN as1-CN as1-CN as1-CN as1-CN

BB BB BB BB BB BB BB BC BC

b-CN b-CN b-CN b-CN b-CN b-CN b-CN b-CN b-CN

A1A1 A1A1 A1A1 A1A2 A1A2 A2A2 A2A2 A1A2 A2A2

k-CN k-CN k-CN k-CN k-CN k-CN k-CN k-CN k-CN

AA AB BB AA AB AA AB AA AA

9.3 6.9 3.4 26.5 15.7 10.8 3.9 10.3 13.2

1Included here are only cows belonging to casein genotype groups comprising ≥7 cows. 2Aggregate a -CN, b-CN, and k-CN genotype. s1

Journal of Dairy Science Vol. 80, No. 11, 1997

´ N ET AL. LUNDE

3000

in the model. Only the effect of permanent environment of cow showed a larger effect (16.1%); the lnSCC of the milk accounted for a reduction of the residual variance of 10.3%. The contribution from the aggregate casein genotype regarding both content and casein number was <0.5%. The same was true for the magnitude of the effect of the b-LG genotype regarding casein content. The corresponding figures for the other effects varied depending on the trait being analyzed; the highest contribution, 39.7%, was from the effect of stage of lactation on casein content. DISCUSSION The present analysis shows a pronounced effect of the b-LG locus on casein number. Because the difference between the two homozygotes corresponds to 1.5

standard deviations, the b-LG could be regarded as a major gene for this trait. The size of the effect is the second largest effect in the model, based on the relative reduction of the residual variance that each of these effects give rise to when included in the model. The genotypic values of the homozygous and heterozygous genotypes at the b-LG locus suggest a positive effect of the B allele on casein number and suggest moreover that the effect is additive, which was reported also by Rahali and Menard (19). Thus, at constant milk protein content, the casein fraction is higher in milk from cows with the B allele, resulting in a larger proportion of the milk protein being converted into cheese, as shown by van den Berg et al. (24). From the results by Schaar et al. (21), the relationship between casein number and cheese yield can be derived. Based on those calculations, the

TABLE 7. Genotypic values of homozygote ( a ) and heterozygote ( d ) b-LG and b-CN genotypes and additional contrasts between homozygotes and heterozygotes within the b-LG, b-CN, and k-CN loci for milk production traits for 371 cows of the Swedish Red and White breed. Contrast Milk protein genotype

Casein content X

b-LG a( AA) = –a( BB) 2 d( AB) 3 AA – AB AB – BB b-CN a( A A ) = –a( A A

2 2)

1 1

d( A

1A2)

5

A1A1 – A1A26 A1A2 – A2A27 k-CN AA – AB8

(%) SE

–0.042** –0.015 –0.027 –0.057*** 4

0.014 0.019 0.028 0.018

Casein yield X –4.5 1.9 –6.4 –2.5

Casein number1

Protein content

(g) SE

X

SE

X

SE

3.9 5.3 7.8 5.2

–1.491*** –0.212 –1.279*** –1.702***

0.154 0.205 0.302 0.201

0.013 –0.013 0.027 –0.000

0.019 0.025 0.037 0.024

Protein yield

(%) X

(g) SE 3.7 3.1 0.6 6.8

8.1 10.8 15.9 10.6

0.010

0.014

–10.7** 4.0

–0.146

0.155

0.018

0.019

–21.3** 8.2

0.020 –0.010 0.031

0.019 0.026 0.020

3.1 5.3 –13.8† 7.4 –7.5 5.7

–0.010 –0.135 –0.156

0.204 0.287 0.221

0.025 –0.007 0.043

0.025 0.035 0.027

11.0 10.8 –32.3* 15.1 –10.3 11.7

0.007

0.021

0.135

0.233

0.003

0.028

–4.1

6.0

–9.5

12.3

protein) × 100. = AA – 1/2(AA + BB). = AB – 1/2(AA + BB). ) = 1/2(b-CN A1A1 k-CN AA + b-CN A1A1 k-CN AB) – 1/4(b-CN A1A1 k-CN AA + b-CN A1A1 k-CN AB + b-CN A2A2 k-CN AA + b-

1(Casein/total 2a ( AA) 3d ( AB) 4a

( A 1A1

CN A2A2 k-CN AB). 5d ( A A ) = 1/2(b-CN A1A2 k-CN AA + b-CN A1A2 k-CN AB) – 1/4(b-CN A1A1 k-CN AA + b-CN A1A1 k-CN AB + b-CN A2A2 k-CN AA + b1 2

CN A2A2 k-CN AB). 6A A – A A = 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB) – 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB). 1 1 1 2 1 1 1 1 1 2 1 2 7A A – A A = 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB) – 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB). 1 2 2 2 1 2 1 2 2 2 2 2 8AA – AB = 1/3(b-CN A A k-CN AA + b-CN A A k-CN AA + b-CN A A k-CN AA) – 1/3(b-CN A A k-CN AB + b-CN A A k-CN AB + 1 1 1 2 2 2 1 1 1 2 b-CN A2A2 k-CN AB). †P < 0.10. **P < 0.01. ***P < 0.001. Journal of Dairy Science Vol. 80, No. 11, 1997

β-LACTOGLOBULIN GENES AND CASEIN IN MILK

casein number in milk from b-LG BB cows was 3% higher than that from b-LG AA cows; this difference corresponds to a 2% increase in cheese yield. With the intermediate frequency of the B allele in the Swedish population, fixation of the B allele would lead to a mean increase of 1% more cheese from a constant volume of milk. The greater number of rare B alleles there are in a population, the more genetic progress can be made by selecting for them. Analyses of gene effects involving several traits and loci, as was done in the present study, almost inevitably results in numerous significance tests, which increases the risks of finding chance associations. However, because the same effect was found in two independent samples of different breeds, this effect is probably not a matter of spurious associations; rather, the b-LG gene itself likely influences the ratio of casein to total protein. McLean et al. ( 1 6 ) suggested that the higher casein number associated with

3001

the B variant was not due to higher casein synthesis but rather to lower synthesis of b-LG, the major whey protein, in the milk from cows with the B allele. Recent results indicate the existence of polymorphisms upstream of the coding part of the milk protein genes, presumably in the regulatory region (1, 3, 22, 26), that might be associated with observed differences in levels of expression of these genes (2, 7, 8, 25). A positive effect of the b-LG B variant was found also for casein content, although the reduction in residual variance was marginal when the effect of bLG genotype was added to the model. Apart from the relationship between b-LG polymorphism and casein number and casein content, no effect of milk protein variants was found that was consistent in both breeds. Thus, no unfavorable association was observed with milk yield or the yields and percentages

TABLE 8. Genotypic values of homozygote ( a ) and heterozygote ( d ) b-LG and b-CN genotypes and additional contrasts between homozygotes and heterozygotes within the b-LG, b-CN, and k-CN loci for milk production traits for 371 cows of the Swedish Red and White breed. Contrast Milk protein genotype

Milk yield X

b-LG a( AA) = –a( BB) 1 d( AB) 2 AA – AB AB – BB b-CN a( A A ) = –a( A A

2 2)

1 1

d( A

1A2)

4

A1A1 – A1A25 A1A2 – A2A26 k-CN AA – AB7 1a ( AA) 2d ( AB) 3a

0.06 0.17 –0.11 0.22 3

(kg) SE 0.27 0.36 0.53 0.35

Fat content X –0.027 0.002 –0.029 –0.025

(%) SE

Fat yield X

(g) SE

Lactose content X

(%) SE

Lactose yield (g) X

SE 12.2 16.3 23.9 15.9

0.045 0.060 0.088 0.059

–4.4 11.0 –15.4 6.5

11.9 16.1 23.6 15.6

–0.026† 0.003 –0.030 –0.023

0.014 0.018 0.027 0.018

–3.3 7.3 –10.7 4.0

–0.71** 0.27

0.074†

0.045

–20.2†

12.1

–0.003

0.014

–34.3** 12.3

0.13 –0.85† –0.58

0.36 0.50 0.39

0.004 0.070 0.079

0.059 0.083 0.064

4.8 –25.0 –15.4

16.1 22.5 17.4

0.021 –0.024 0.018

0.018 0.026 0.020

8.5 –42.8† –25.8

16.2 22.8 17.6

–0.21

0.41

0.047

0.068

–5.2

18.1

0.032

0.021

–4.3

18.4

= AA – 1/2(AA + BB). = AB – 1/2(AA + BB). ) = 1/2(b-CN A1A1 k-CN AA + b-CN A1A1 k-CN AB) – 1/4(b-CN A1A1 k-CN AA + b-CN A1A1 k-CN AB + b-CN A2A2 k-CN AA + b-

( A 1A1

CN A2A2 k-CN AB). 4d ( A A ) = 1/2(b-CN A1A2 k-CN AA + b-CN A1A2 k-CN AB) – 1/4(b-CN A1A1 k-CN AA + b-CN A1A1 k-CN AB + b-CN A2A2 k-CN AA + b1 2

CN A2A2 k-CN AB). 5A A – A A = 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB) – 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB). 1 1 1 2 1 1 1 1 1 2 1 2 6A A – A A = 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB) – 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB). 1 2 2 2 1 2 1 2 2 2 2 2 7AA – AB = 1/3(b-CN A A k-CN AA + b-CN A A k-CN AA + b-CN A A k-CN AA) – 1/3(b-CN A A k-CN AB + b-CN A A k-CN AB + 1 1 1 2 2 2 1 1 1 2 b-CN A2A2 k-CN AB). †P < 0.10. **P < 0.01. ***P < 0.001. Journal of Dairy Science Vol. 80, No. 11, 1997

´ N ET AL. LUNDE

3002

of the major milk components, protein, fat, and lactose. Except for the effect of b-LG genotype on casein number, milk protein genotypes only accounted for a marginal proportion of the phenotypic variation in the

studied traits. The reduction in residual variance from inclusion of the random effect of additive genetics of the polygenic background using an animal model, generally exceeding 5% (results not shown), indicates that genes other than those coding for milk

TABLE 9. Genotypic values of homozygote ( a ) and heterozygote ( d ) b-LG and b-CN genotypes and additional contrasts between homozygotes and heterozygotes within the b-LG, as1-CN, b-CN, and k-CN loci for milk production traits for 204 cows of the Swedish Holstein breed. Contrast Milk production genotype

Casein content X

b-LG a( AA) = –a( BB) 2 d( AB) 3 AA – AB AB – BB as1-CN BB – BC4 b-CN a( A A ) = –a( A A

2 2)

1 1

d( A

1A2)

6

A1A1 – A1A27 A1A2 – A2A28 k-CN AA – AB9 a( AA) = –a( BB) 10 d( A A ) 11 1 2

AA – AB12 AB – BB13

5

(%) SE

Casein yield X

(g) SE

Casein number1

Protein content

Protein yield

(%)

(g)

X

SE

X

SE

X

SE

–1.609*** 0.070 –1.679*** –1.539***

0.134 0.175 0.220 0.222

0.002 0.001 0.001 0.002

0.021 0.027 0.034 0.034

14.7 13.2 1.5 27.9

10.3 14.1 17.4 17.5

0.277

0.036

0.042

27.5

20.0

–0.050** 0.006 –0.056* –0.045†

0.017 0.022 0.027 0.028

0.2 7.7 –7.6 7.9

5.1 6.9 8.6 8.6

0.028

0.034

15.1

10.1

0.068

–0.001

0.022

–6.9

6.8

–0.209

0.174

–0.003

0.027

–9.5

13.7

0.031 –0.032 0.051

0.025 0.034 0.043

–0.7 –6.9 –8.9

7.7 10.5 13.6

–0.041 –0.015 –0.297

0.199 0.272 0.349

0.049 –0.061 0.073

0.031 0.042 0.054

1.4 –14.9 –10.0

15.5 21.1 27.5

–0.015 –0.050

0.031 0.035

10.6 –18.7†

9.8 10.9

0.070 –0.154

0.251 0.283

–0.023 –0.065

0.039 0.044

20.8 –34.0

19.8 22.0

0.036 –0.086 –0.014

0.052 0.055 0.070

–24.1 5.5 –42.8†

16.7 17.5 22.1

–0.119 –0.035 –0.272

0.419 0.444 0.559

0.049 –0.114† –0.015

0.065 0.069 0.087

–37.7 3.6 –71.7

34.0 35.5 45.1

protein) × 100. = AA – 1/2(AA + BB). = AB – 1/2(AA + BB). BC = 1/2(as1-CN BB b-CN A1A2 kCN AA + as1-CN BB b-CN A2A2 k-CN AA) – 1/2(as1-CN BC b-CN A1A2 k-CN AA + as1-CN BC bCN A2A2 k-CN AA). 5a ( A A ) = 1/2(b-CN A1A1 k-CN AA + b-CN A1A1 k-CN AB) – 1/4(b-CN A1A1 k-CN AA + b-CN A1A1 k-CN AB + b-CN A2A2 k-CN AA + b1(Casein/total 2a ( AA) 3d ( AB) 4BB –

1 1

CN A2A2 k-CN AB). 6d ( A A ) = 1/3(as1-CN BB b-CN A1A2 k-CN AA + as1-CN BB b-CN A1A2 k-CN AB + as1-CN BC b-CN A1A2 k-CN AA) – 1 2

1/4(as1-CN BB b-CN A1A1 k-CN AA + as1-CN BB b-CN A1A1 k-CN AB) – 1/6(as1-CN BB b-CN A2A2 k-CN AA + as1-CN BB b-CN A2A2 k-CN AB + as1-CN BC b-CN A2A2 k-CN AA). 7A A – A A = 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB) – 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB). 1 1 1 2 1 1 1 1 1 2 1 2 8A A – A A = 1/3(a -CN BB b-CN A A k-CN AA + a -CN BB b-CN A A k-CN AB + a -CN BC b-CN A A k-CN AA) – 1 2 2 2 s1 1 2 s1 1 2 s1 1 2 1/3(as1-CN BB b-CN A2A2 k-CN AA + as1-CN BB b-CN A2A2 k-CN AB + as1-CN BC b-CN A2A2 k-CN AA). 9AA – AB = 1/3(b-CN A A k-CN AA + b-CN A A k-CN AA + b-CN A A k-CN AA) – 1/3(b-CN A A k-CN AB + b-CN A A k-CN AB + 1 1 1 2 2 2 1 1 1 2 b-CN A2A2 k-CN AB). 10a ( AA) = ( as1-CN BB b-CN A1A1 k-CN AA) – ( as1-CN BB b-CN A1A1 k-CN BB). 11d ( AB) = ( as1-CN BB b-CN A1A1 k-CN AB) – 1/2(as1-CN BB b-CN A1A1 k-CN AA + as1-CN BB b-CN A1A1 k-CN BB). 12AA-AB = ( a -CN BB b-CN A A k-CN AA) – ( a -CN BB b-CN A A k-CN AB). s1 1 1 s1 1 1 13AB-BB = ( a -CN BB b-CN A A k-CN AB) – ( a -CN BB b-CN A A k-CN BB). s1 1 1 s1 1 1 †P < 0.10. **P < 0.01. ***P < 0.001. Journal of Dairy Science Vol. 80, No. 11, 1997

β-LACTOGLOBULIN GENES AND CASEIN IN MILK

proteins influence the traits included in the analysis. For casein number, the effect of the polygenic background is, however, considerably smaller than the effect of b-LG genotype. The choice of the appropriate statistical model is important. The present data were for animals from a

3003

limited number of experimental herds, some of which had previously exchanged genetic material. Thus, in addition to paternal relationships, information regarding maternal relatives needed to be considered. The relationships would otherwise have caused bias in the estimates of the effects of milk protein genes

TABLE 10. Genotypic values of homozygote ( a ) and heterozygote ( d ) b-LG and b-CN genotypes and additional contrasts between homozygotes and heterozygotes within the b-LG, as1-CN, b-CN, and k-CN loci for milk production traits for 204 cows of the Swedish Holstein breed. Contrast Milk protein genotype

Milk yield X

b-LG a( AA) = –a( BB) 2 d( AB) 3 AA – AB AB – BB as1-CN BB – BC4 b-CN a( A A ) = –a( A A

2 2)

1 1

d( A

1A2)

6

A1A1 – A1A27 A1A2 – A2A28 k-CN AA – AB9 a( AA) = –a( BB) 10 d( A A ) 11 1 2

AA – AB12 AB – BB13

5

(kg) SE

0.55† 0.44 0.10 0.99†

0.32 0.43 0.53 0.53

0.51

0.62

–0.21

Fat content X –0.088† –0.037 –0.050 –0.125

(%) SE 0.049 0.064 0.080 0.081

Fat yield X 0.9 1.8 –0.8 2.7

(g) SE

Lactose content X

(%) SE

Lactose yield (g) X

SE

15.5 20.6 25.7 25.9

0.003 0.002 0.002 0.006

0.015 0.020 0.025 0.026

24.1 22.6 1.6 46.7†

15.0 20.4 25.3 25.4

0.291** 0.100

86.7** 31.1

0.048

0.031

30.5

29.5

0.42

0.033

0.064

–6.7

20.3

0.009

0.020

–7.7

20.0

–0.23 0.03 –0.61

0.47 0.65 0.84

0.010 –0.017 0.116

0.072 0.099 0.127

–11.1 –2.6 –13.0

23.1 31.6 40.7

–0.005 0.011 0.021

0.023 0.031 0.040

–14.1 8.2 –27.0

22.7 30.8 40.0

0.73 –0.71

0.60 0.67

0.087 0.022

0.091 0.103

54.0† –27.28

29.3 32.95

0.023 –0.006

0.029 0.033

33.5 –29.2

28.8 32.2

–1.25 0.55 –1.96

1.03 1.08 1.37

0.064 –0.042 0.085

0.154 0.163 0.206

–50.98 23.70 –78.26

49.71 52.33 66.15

–0.083† 0.076 –0.089

0.049 0.051 0.065

–65.6 36.4 –94.8

49.4 51.7 65.5

protein) × 100. = AA – 1/2(AA + BB). = AB – 1/2(AA + BB). BC = 1/2(as1-CN BB b-CN A1A2 kCN AA + as1-CN BB b-CN A2A2 k-CN AA) – 1/2(as1-CN BC b-CN A1A2 k-CN AA + as1-CN BC bCN A2A2 k-CN AA). 5a ( A A ) = 1/2(b-CN A1A1 k-CN AA + b-CN A1A1 k-CN AB) – 1/4(b-CN A1A1 k-CN AA + b-CN A1A1 k-CN AB + b-CN A2A2 k-CN AA + b1(Casein/total 2a ( AA) 3d ( AB) 4BB –

1 1

CN A2A2 k-CN AB). 6d ( A A ) = 1/3(as1-CN BB b-CN A1A2 k-CN AA + as1-CN BB b-CN A1A2 k-CN AB + as1-CN BC b-CN A1A2 k-CN AA) – 1 2

1/4(as1-CN BB b-CN A1A1 k-CN AA + as1-CN BB b-CN A1A1 k-CN AB) – 1/6(as1-CN BB b-CN A2A2 k-CN AA + as1-CN BB b-CN A2A2 k-CN AB + as1-CN BC b-CN A2A2 k-CN AA). 7A A – A A = 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB) – 1/2(b-CN A A k-CN AA + b-CN A A k-CN AB). 1 1 1 2 1 1 1 1 1 2 1 2 8A A – A A = 1/3(a -CN BB b-CN A A k-CN AA + a -CN BB b-CN A A k-CN AB + a -CN BC b-CN A A k-CN AA) – 1 2 2 2 s1 1 2 s1 1 2 s1 1 2 1/3(as1-CN BB b-CN A2A2 k-CN AA + as1-CN BB b-CN A2A2 k-CN AB + as1-CN BC b-CN A2A2 k-CN AA). 9AA – AB = 1/3(b-CN A A k-CN AA + b-CN A A k-CN AA + b-CN A A k-CN AA) – 1/3(b-CN A A k-CN AB + b-CN A A k-CN AB + 1 1 1 2 2 2 1 1 1 2 b-CN A2A2 k-CN AB). 10a ( AA) = ( as1-CN BB b-CN A1A1 k-CN AA) – ( as1-CN BB b-CN A1A1 k-CN BB). 11d ( AB) = ( as1-CN BB b-CN A1A1 k-CN AB) – 1/2(as1-CN BB b-CN A1A1 k-CN AA + as1-CN BB b-CN A1A1 k-CN BB). 12AA-AB = ( a -CN BB b-CN A A k-CN AA) – ( a -CN BB b-CN A A k-CN AB). s1 1 1 s1 1 1 13AB-BB = ( a -CN BB b-CN A A k-CN AB) – ( a -CN BB b-CN A A k-CN BB). s1 1 1 s1 1 1 †P < 0.10. **P < 0.01. ***P < 0.001. Journal of Dairy Science Vol. 80, No. 11, 1997

´ N ET AL. LUNDE

3004

TABLE 11. The contribution of different fixed and random effects, relative to the residual variance, to the phenotypic variation in casein content and casein number,1 analyzed on a material including 575 Swedish Red and White and Swedish Holstein cows. MSE Reduction Parameter

Casein content Casein number

b-LG Genotype Casein genotype3 Stage of lactation Lactation number ln SCC4 Year-season Breed-herd Animal Cow5

0.5 0.1 39.7 0.1 0.0 7.3 1.0 6.9 9.6

(%) 11.1 0.4 1.9 1.5 10.3 4.2 0.1 2.0 16.1

protein) × 100. 2The estimates are given as the relative difference in mean square error (MSE) between the full model and models that ignored one effect at a time: MSE reduction = [(MSE ( reduced model) – MSE( full model) )/MSE ( reduced model) ] × 100. 3Aggregate a -CN, b-CN, and k-CN genotype. s1 4Transformed to the natural logarithm scale. 5Effect of permanent environment of cow. 1(Casein/total

through the nonrandom distribution of background genomes in related individuals in combination with increased likelihood of sharing milk protein genotype (6, 11). Moreover, because repeated measures on the same individual were not treated as independent observations, a statistical package that could handle two random effects was required for the present analyses. To test whether the observed frequencies were likely to be representative for the SRB and SLB populations, the gene frequencies at the b-LG, as1-CN, bCN, and k-CN loci were compared with recent estimates ( 1 0 ) from a random sample of 689 SRB and 304 SLB cows in commercial herds. In general, the frequencies obtained in the experimental herds agreed quite well with those from the commercial herds. However, a comparison of the two studies revealed that, within the present group of SRB cows, the b-LG B variant occurred more frequently (67% compared with 59%), and the frequency of the b-CN A2 allele was 5% units higher; the A1 allele was reduced correspondingly. For the two studies on SLB cows, the only difference in allele frequencies was observed within the k-CN locus, where frequency of the B allele was 7% units higher in the present study. Compared with most other reports in the literature [reviewed by Lin et al. ( 1 5 ) and Bovenhuis (4)], the low frequency of k-CN B within both the SRB and SLB breed is noticeable. However, the value for k-CN Journal of Dairy Science Vol. 80, No. 11, 1997

B in SRB corresponded well with results on related breeds in Norway ( 1 4 ) and Finland (18). The lack of polymorphism within the as1-CN locus for the SRB cows was in accordance with results on Finnish Ayrshire ( 1 8 ) for which, similarly, only the BB genotype was present. Another observed discrepancy concerns the SLB breed for which the as1-CN C allele occurs more frequently than in other Holstein populations. Apart from the exceptions indicated previously, the gene frequencies correspond to previously reported values (4, 15). CONCLUSIONS The effect of b-LG polymorphism on casein number was relatively large, and no unfavorable effects were found on the other major milk constituents. Therefore, this gene appears to be an obvious candidate for use when breeding cows for more efficient production of milk that is suitable for cheese production. ACKNOWLEDGMENTS The authors thank Lena Andersson-Eklund for discussion of this manuscript. This work was supported by the Farmers Research Council for Information and Development. REFERENCES 1 Bleck, G. T., and R. D. Bremel. 1993. Sequence and single-base polymorphisms of the bovine a-lactalbumin 5′-flanking region. Gene 126:213. 2 Bleck, G. T., and R. D. Bremel. 1993. Correlation of the alactalbumin (+15) polymorphism and milk composition of Holsteins. J. Dairy Sci. 76:2292. 3 Bleck, G. T., J. C. Conroy, and M. B. Wheeler. 1996. Polymorphisms in the bovine b-casein 5′ flanking region. J. Dairy Sci. 79:347. 4 Bovenhuis, H. 1992. The relevance of milk protein polymorphisms for dairy cattle breeding. Ph.D. Diss., Wageningen Agric. Univ., Wageningen, The Netherlands. 5 Falconer, D. S. 1989. Introduction to Quantitative Genetics. 3rd ed. Longman, London, United Kingdom. 6 Famula, T. R., and J. F. Medrano. 1994. Estimation of genotype effects for milk proteins with animal and sire transmitting ability models. J. Dairy Sci. 77:3153. 7 Ford, C. A., M. B. Connett, and R. J. Wilkins. 1993. bLactoglobulin expression in bovine mammary tissue. Proc. N.Z. Soc. Anim. Prod. 53:167. 8 Graml, R., G. Weiss, J. Buchberger, and F. Pirchner. 1989. Different rates of synthesis of whey protein and casein by alleles of the b-lactoglobulin and as1-casein locus in cattle. Genet. Sel. Evol. 21:547. 9 Groeneveld, E., M. Kovac, and T. Wang. 1990. PEST, a general purpose BLUP package for multivariate prediction and estimation. Proc. 4th World Congr. Genet. Appl. Livest. Prod., Edinburgh, Scotland XIII:488. 10 Janson, L., A. Lunde´n, A. Andre´n, and T. Allmere. 1993. Genetiska mjo¨lkproteinvarianter och deras betydelse fo¨r produkternas kvalite´. Page 116 in Husdjurskonferensen 1993, Uppsala. SLU Info Rapporter, Allma¨nt 181, Uppsala 1993, ISSN 1101-3761.

β-LACTOGLOBULIN GENES AND CASEIN IN MILK 11 Kennedy, B. W., M. Quinton, and J.A.M. van Arendonk. 1992. Estimation of effects of single genes on quantitative traits. J. Anim. Sci. 70:2000. 12 Lien, S., L. Gomez-Raya, T. Steine, E. Fimland, and S. Rogne. 1995. Associations between casein haplotypes and milk yield traits. J. Dairy Sci. 78:2047. 13 Lien, S., S. Kamin´ski, P. Alestro¨m, and S. Rogne. 1993. A simple and powerful method for linkage analysis by amplification of DNA from single sperm cells. Genomics 16:41. 14 Lien, S., and S. Rogne. 1993. Bovine casein haplotypes: number, frequencies and applicability as genetic markers. Anim. Genet. 24:373. 15 Lin, C. Y., M. P. Sabour, and A. J. Lee. 1992. Direct typing of milk proteins as an aid for genetic improvement of dairy bulls and cows: a review. Anim. Breed. Abstr. 60:1. 16 McLean, D. M., E.R.B. Graham, R. W. Ponzoni, and H. A. McKenzie. 1984. Effects of milk protein genetic variants on milk yield and composition. J. Dairy Res. 51:531. 17 Ng-Kwai-Hang, K. F., J. F. Hayes, J. E. Moxley, and H. G. Monardes. 1986. Relationships between milk protein polymorphisms and major milk constituents in Holstein-Friesian cows. J. Dairy Sci. 69:22. 18 Piironen, T., M. Ojala, T. Niini, E.-L. Syva¨oja, and J. Seta¨la¨. 1992. Effects of milk protein genetic variants and lactation stage on renneting properties of bovine milk. Page 46 in Proc. 43rd Annu. Mtg. Eur. Assoc. Anim. Prod., Madrid, Spain.

3005

19 Rahali, V., and J. L. Menard. 1991. Influence des variants ge´ne´tiques de la b-lactoglobuline et de la k-case´ine sur la composition du lait et son aptitude fromage`re. Lait 71:275. 20 SAS User’s Guide: Statistics, Version 5 Edition. 1985. SAS Inst., Inc., Cary, NC. 21 Schaar, J., B. Hansson, and H.-E. Pettersson. 1985. Effects of genetic variants of k-casein and b-lactoglobulin on cheesemaking. J. Dairy Res. 52:429. 22 Schild, T. A., V. Wagner, and H. Geldermann. 1994. Variants within the 5′-flanking region of bovine milk protein genes: I. kcasein-encoding gene. Theor. Appl. Genet. 89:116. 23 Swaisgood, H. E., ed. 1975. Methods of Gel Electrophoresis of Milk Proteins. ADSA, Champaign, IL (now in Savoy, IL). 24 van den Berg, G., J.T.M. Escher, P. J. de Koning, and H. Bovenhuis. 1992. Genetic polymorphism of k-casein and blactoglobulin in relation to milk composition and processing properties. Neth. Milk Dairy J. 46:145. 25 van Eenennaam, A. L., and J. F. Medrano. 1991. Differences in allelic protein expression in the milk of heterozygous k-casein cows. J. Dairy Sci. 74:1491. 26 Wagner, V. A., T. A. Schild, and H. Geldermann. 1994. DNA variants within the 5′-flanking region of milk-protein encoding genes. II. The b-lactoglobulin-encoding gene. Theor. Appl. Genet. 89:121.

Journal of Dairy Science Vol. 80, No. 11, 1997