Associations between maternal milk protein genotypes with preweaning calf growth traits in beef cattle

Abstract

The objectives of this study were to investigate milk casein polymorphisms in dams and to determine the impacts of maternal casein genotypes on growth traits of their sucking calves. Milk samples from 433 dams of the breeds German Angus (GA) and German Simmental (GS) were typed at the milk protein loci α _s1-casein (α_s1-CN), β-casein (β-CN), α _s2-casein (α_s2-CN), and κ-casein (κ-CN) via isoelectric focusing. Associations between casein genotypes in maternal milk with growth traits of their 1,872 calves were analyzed until the age of weaning using linear mixed models, considering either genotypes of individual casein loci (model 1) or composite α _s1-β-α _s2-κ-CN genotypes within the casein cluster (model 2). Besides environmental effects such as sex, age of the dam, and calving year-season, genetic effects (breed group and maternal and paternal effects) were considered in statistical models. The composite casein genotype BBǀA2A2ǀAAǀAB (order of genes on bovine chromosome 6: α _s1-ǀβ-ǀα _s2-ǀκ-CN) was associated with greater average daily weight gains (ADG) and heavier age-adjusted weaning weights (WW) of calves (P < 0.05). The effects of composite genotypes on birth weight of calves were similar (P > 0.05; model 2). With regard to individual casein loci, greater ADG and WW were observed for calves from dams with the genotypes κ-CN BB and α _s1-CN BB, respectively (P < 0.05; model 1). Age-adjusted WW was largest for calves from dams carrying the κ-CN genotype BB (215 kg) compared with calves representing the maternal AB and AA genotypes (both 204 kg). Results from the present study suggested selectable casein genotypes due to their nutritional value of milk (value in terms of offspring performances), offering new perspectives for breeding strategies in beef cattle to improve preweaning calf performance.

Introduction

Bovine milk and colostrum are the most important sources of well-balanced nutrients and are essential for the development of newborn mammals. Proteins contained in milk, especially caseins, influence a wide range of nutritional, functional, and biological activities (Pihlanto and Korhonen, 2003; Caroli et al., 2009). The casein genes CSN1S1, CNS2, CSN1S2, and CSN3, encoding the milk proteins α _s1-casein (α_s1-CN), β-casein (β-CN), α _s2-casein (α_s2-CN), and κ-casein (κ-CN), are tightly linked in the given order in a 250-kilobase (kb) cluster on bovine chromosome 6 (BTA6) (Ferretti et al., 1990; Hayes et al., 1993) (Figure 1). Several single-nucleotide polymorphisms (SNP) of the casein genes change their protein sequences, implying different casein variants. In Bos genus, 10 variants for α _s1-CN (A, B, C, D, E, F, G, H, I, and J), 15 for β-CN (A1, A2, A3, B, C, D, E, F, G, H¹, H², I, J, K, and L), 5 for α _s2-CN (A, B, C, D, and E), and 14 for κ-CN (A, A1, B, B2, C, D, E, F¹, F², G¹, G², H, I, and J) have been reported, displaying differences in frequencies among breeds (Caroli et al., 2009; Gallinat et al., 2013). Casein polymorphisms were associated with milk production traits (Ng-Kwai-Hang et al., 1986; Grosclaude, 1988) and with the technological properties of milk (Marziali and Ng-Kwai-Hang, 1986). For example, Ng-Kwai-Hang et al. (1984) identified relationships between the BB genotypes of α _s1-CN and κ-CN with protein and casein percentages of milk. Bovenhuis et al. (1992) focused on the κ-CN BB genotype and indicated its important role in cheese-making abilities by providing colloidal stability to the casein micelle. Furthermore, Ng-Kwai-Hang et al. (1986) and Grosclaude (1988) identified the superior rennet coagulation properties of κ-CN BB in comparison to the genotypes AA and AB.

Figure 1.

Structural organization of the casein genes CSN1S1, CSN2 (β-casein), CSN1S2, and CSN3 on BTA6 (modified from Martin et al., 2002). The genes encode the four caseins α _s1-, β-, α _s2-, and κ-casein. Sizes of genes are given in kilobases (kb).

To date, research addressing such values of milk proteins has been conducted entirely in dairy cattle populations (i.e., analyzing the effects of milk protein genotypes on milk constituents). However, strong practical relevance is given for the effect of milk protein genotypes on preweaning calf performances in beef cattle, as beef cattle calves are suckling milk from their biological dams. The known participation of κ-CN in the composition of milk protein (Hoogendoorn et al., 1969) stimulated first association studies with calf growth traits. Weston (1971) identified positive effects of the κ-CN BB genotype on weaning weight (WW; P < 0.05) and 180-d gain (P < 0.1). In contrast, Ziehe et al. (1993) reported increased daily weight gains and WWs for calves from AA dams, when compared with the dam group representing the AB genotype. A similar observation was made by Henderson and Marshall (1996). Oppositely, Faria et al. (1999) found no differences in average daily weight gains (ADGs) when contrasting AA and AB genotypes. However, as the individual genotypes for the κ-CN locus were considered in the model as a fixed effect (i.e., Ziehe et al., 1993; Faria et al., 1999), the estimates for the κ-CN genotypes might be biased, due to the possible impact of other linked casein loci (α _s1-, β-, and α _s2-CN) and interactions among single loci (i.e., epistatic effects). As an alternative, individual genotypes of all caseins should be considered simultaneously, as separate fixed effects or by considering all casein loci as composite α _s1-β-α _s2-κ-CN genotypes (Ojala et al., 1997). To our knowledge, no previous study investigated the effects of all four casein loci on growth traits in preweaning calves. Preweaning calf performance is of special importance, because the growth rate is a proper indicator for, e.g., mature live weight and reproductive efficiency (Martins et al., 2000; Corrêa et al., 2006).

The aim of the present study was to investigate milk casein polymorphisms in dams and to infer an association between casein genotypes in maternal milk with growth traits of their sucking calves in a large-scale beef cattle research farm. From a practical breeding perspective, identified valuable casein genotypes can be directly used for early marker-assisted selection strategies in female calves or heifers, when aiming at selection strategies to improve early life nutrition of calves.

Materials and Methods

The Animal Care and Use Committee approval was not obtained for the present study as data were obtained from an existing database.

Experimental location and herd description

Beef cattle data from German Angus (GA) and German Simmental (GS) were collected during the period from 1998 to 2009 at the experimental farm Rudlos of the Institute of Animal Breeding and Genetics, University of Giessen. The beef cattle breed GA was bred in the 1950s in Germany by crossing imported Aberdeen Angus from the United Kingdom with various native German dual-purpose cattle breeds (i.e., German Black Pied and German Red Pied). Rudlos is located in the federal state Hesse in Germany at an altitude of 360 to 450 m above the sea level, with an average daily temperature of 7.5 °C and a total annual rainfall of 850 L/m². The herd included about 280 dams with natural service (1 sire per 30 cows). Half of the herd were GA and half were GS cows. The calving season spanned the period from January to May, with approximately two-thirds of the calvings in February and March. In the winter season, the dams were kept in a deep-litter housing system. Grazing on pasture in groups (mixed breeds) comprising 30 dams (plus calves) started in April/May until October/November. Calves were weaned at an average age of approximately 220 d at the end of the grazing season.

Animals and traits

Milk samples from dams of the two beef cattle breeds GA (n = 222) and GS (n = 211) were collected and used for the determination of α _s1-CN, β-CN, α _s2-CN, and κ-CN fractions. From the natural matings of these dams with 31 sires (GA sires, n = 16; GS sires, n = 15), 812 GA, 636 GS, 174 GA × GS, and 250 GS × GA crossbred calves were recorded for birth weight (BW) and WW at the average age of 216 d (SD = 17.1). ADG from birth to weaning was calculated as (WW − BW)/(age at weaning). For ongoing analyses, WW was age-adjusted at 180 d of life (named WW₁₈₀). Means and standard deviations for BW, WW₁₈₀, ADG, and for the suckling period (SP) of calves are given in Table 1. The average BW of GS calves (44 kg) was larger than the BW of GA calves (37 kg), whereas the average BW of their crossbreeds was in between with 43 kg. The SP varied from 139 to 279 d, with a mean of 216 d over all breeds. The average WW₁₈₀ of crossbred calves (GA × GS) was 245 kg, and lower for GA and GS with 205 and 237 kg, respectively. ADG was greatest for the crossbreed GA × GS (1.12 kg/d), followed by GS (1.07 kg/d), GS × GA (1.04 kg/d), and GA (0.93 kg/d).

Table 1.

Descriptive statistics of ADG, SP, BW, and WW₁₈₀ for the calves of the breeds GA and GS and their crosses

		ADG, kg/d		SP, d		BW, kg		WW180, kg
Breed	n	mean	SD	mean	SD	mean	SD	mean	SD
GA	812	0.93	0.16	218.10	18.29	37.28	5.19	204.67	31.83
GA × GS	174	1.12	0.17	213.62	17.29	43.04	5.44	245.12	32.75
GS × GA	250	1.04	0.16	217.40	14.47	42.04	5.26	229.74	31.02
GS	636	1.07	0.20	215.33	18.48	44.37	6.09	237.46	39.41

Milk protein typing

In total, 433 skimmed milk samples were analyzed for milk protein polymorphisms of the caseins α _s1-CN, β-CN, α _s2-CN, and κ-CN by isoelectric focusing in 0.3-mm-thin polyacrylamide gels (according to Erhardt, 1989, 1993). This method describes the separation of the known α _s1-CN, β-CN, α _s2-CN, and κ-CN variants according to their isoelectric point. Subsequent identification of the genetic variants was done according to Seibert et al. (1985) and Erhardt (1989). Isoelectric focusing allows the simultaneous separation of all polymorphic protein fractions in bovine milk in one single run, and considers variants, which cannot be detected via commercial SNP chip applications.

Allele and genotype frequencies for each casein locus (α _s1-CN, β-CN, α _s2-CN, and κ-CN) were determined via direct counting. A χ ² test was carried out to evaluate if the populations were in Hardy–Weinberg equilibrium. As the casein component of milk represents the product of four genes carrying two alleles (Ginger and Grigor, 1999), composite α _s1-β-α _s2-κ-CN genotypes (in order according to their location on BTA6) were determined for subsequent analyses. Composite genotypes allow consideration of potential interactions between loci within the casein cluster and rare or missing genotypes.

Statistical analyses

Individual caseins

The effects of individual α _s1-CN, β-CN, α _s2-CN, and κ-CN genotypes on BW, ADG, and WW₁₈₀ were considered simultaneously applying the following linear mixed model (individual loci model) 1:

$${\displaystyle \begin{array}{l}{\displaystyle \begin{array}{lll}& {\displaystyle {\mathrm{y}}_{\mathrm{i}\mathrm{j}\mathrm{k}\mathrm{l}\mathrm{m}\mathrm{n}\mathrm{o}\mathrm{p}\mathrm{q}\mathrm{r}\mathrm{s}\mathrm{t}\mathrm{u}}=}& {\alpha }_{\mathrm{s}1}C{N}_{\mathrm{i}}+\beta C{N}_{\mathrm{j}}+{\alpha }_{\mathrm{s}2}C{N}_{\mathrm{k}}+\kappa C{N}_{\mathrm{l}}\\ & & +CY{S}_{\mathrm{m}}+Se{x}_{\mathrm{n}}+Age{D}_{\mathrm{o}}+{\displaystyle Breedgrou{p}_{\mathrm{p}}}\\ & & +B{T}_{\mathrm{q}}+AD{G}_{\mathrm{r}}+B{W}_{\mathrm{s}}+D_{\mathrm{t}}+S_{\mathrm{u}}+{\mathrm{e}}_{\mathrm{i}\mathrm{j}\mathrm{k}\mathrm{l}\mathrm{m}\mathrm{n}\mathrm{o}\mathrm{p}\mathrm{q}\mathrm{r}\mathrm{s}\mathrm{t}\mathrm{u}}\end{array}}\end{array}}$$ (1)

where y_{ijklmnopqrstu} = observations for BW, ADG, and WW₁₈₀: α_s1CN_i = fixed effect of α _s1-CN loci, βCN_j = fixed effect of β-CN loci, α_s2CN_k = fixed effect of α _s2-CN loci, κCN_l = fixed effect of κ-CN loci, CYS_m = fixed effect of calving year-season (CYS), Sex_n= fixed effect of sex of the calf, AgeD_o = fixed effect of age of the dam, Breed group_p = fixed effect of the breed group of the calf, BT_q = fixed effect of birth type (BT) (single or twin), ADG_r = linear regression on ADG for the trait WW₁₈₀, BW_s = linear regression on BW for the traits ADG and WW₁₈₀, D_t = random effect of the dam, S_u = random effect of the sire, and e_{ijklmnopqrstu} = random residual effect. Interactions among breed group and BT were initially tested for each trait. The interaction component was statistically significant (P < 0.05) for ADG and WW₁₈₀. Hence, a fixed combination effect of breed group and BT (= Breed group_BT) was included in the final model 1 for the traits ADG and WW₁₈₀.

The average number of offspring was 4.3 per dam and 55.1 per sire. The age of the dam considered 10 classes from the age of 2 yr until 11 yr. The effect of breed group included the four classes for GA, GS, and the respective crossbreeds GA × GS and GS × GA. BT consisted of two classes for calves born as singles or twins. The combination effect Breed group_BT in model 1 for ADG and WW₁₈₀ comprised eight classes, e.g., the four breed groups combined with the two BT. Calving seasons were defined from December to February, March to May, June to August, and September to November, implying 31 classes for CYS due to the consideration of calving years from 1998 to 2009. The casein β-CN comprised the 10 genotypes: A1A1, A1A2, A1B, A1C, A2A2, A2A3, A2B, A2C, BB, and BC. The κ-CN fraction included the six genotypes: AA, AB, AE, BB, BE, and EE, and α _s2-CN consisted of the two genotypes: AA and AD. The casein α _s1-CN included the three genotypes for BB, BC, and CC. Additive and dominance effects for α _s1-CN were estimated according to Falconer and Mackay (1996).

Casein cluster

The influence of the whole casein cluster including combined α _s1-β-α _s2-κ-CN genotypes on BW, ADG, and WW₁₈₀ was analyzed applying the linear mixed model (composite loci model) 2:

$${\displaystyle \begin{array}{l}{\displaystyle \begin{array}{lll}& {\displaystyle {\mathrm{y}}_{\mathrm{i}\mathrm{j}\mathrm{k}\mathrm{l}\mathrm{m}\mathrm{n}\mathrm{o}\mathrm{p}\mathrm{q}\mathrm{r}}=}& CNcluste{r}_{\mathrm{i}}+CY{S}_{\mathrm{j}}+Se{x}_{\mathrm{k}}+Age{D}_{\mathrm{l}}+Breedgrou{p}_{\mathrm{m}}+\\ & & {\displaystyle B{T}_{\mathrm{n}}+AD{G}_{\mathrm{o}}+B{W}_{\mathrm{p}}+D_{\mathrm{q}}+S_{\mathrm{r}}+e_{\mathrm{i}\mathrm{j}\mathrm{k}\mathrm{l}\mathrm{m}\mathrm{n}\mathrm{o}\mathrm{p}\mathrm{q}\mathrm{r}}}\end{array}}\end{array}}$$ (2)

where y_ijklmnopqr = observations for BW, ADG, and WW₁₈₀: CNcluster_i = fixed effect of the composite α _s1-β-α _s2-κ-CN genotypes. The remaining effects for the composite loci model (2) were defined in analogy to the individual loci model (1). Although 43 different composite casein genotypes were observed, only composite genotypes with frequencies > 5% were considered, implying five classes.

Statistical analyses were conducted using the software R 2.14.2 (R Core Team, 2019), applying the packages lmer, lme4 (Bates et al., 2015), and lmerTest (Kuznetsova et al., 2017) for fitting mixed-effects models. The package emmeans (Russell, 2016) was used for the calculation of least squares means. Statistical significances were declared at P < 0.05.

Results and Discussion

Allele and genotype frequencies at the casein loci

Allele and genotype frequencies for α _s1-, β-, α _s2-, and κ-CN as well as tests for Hardy–Weinberg equilibrium in the studied breeds are summarized in Tables 2 and 3. The genotype frequencies observed in both breeds were within the Hardy–Weinberg expectations for all casein loci (Table 2), which is in agreement with previous studies (e.g., Caroli et al., 2004). For the casein κ-CN, the most frequent genotypes across the breeds were AA and AB with frequencies of 44.8% and 41.5%, respectively. The frequency for the homozygote genotype BB was 8.3%. Falaki et al. (1997) reported a frequency of 6.8% for the κ-CN BB genotype in Italian Simmental cows, which is similar to the observed frequency of 7.3% in GS. Due to similar breeding goals for Simmental beef and Simmental dual-purpose (Perišić et al., 2009), both selection lines did not differ in κ-CN genotype frequencies. With regard to β-CN, there was a predominance of A1 (24.4%) and A2 (69.5%) alleles as shown by the presence of A1A1, A1A2, and A2A2 genotypes in about 89.0% of the milk samples. The rare variants A3, B, and C only occurred in genotypes with frequencies lower than 5.0% across the breeds, apart from the genotype A2B with an average frequency of exactly 5.0%. The genotype A2A2 was the most frequent with 61.9% for the breed GA, which is in agreement with results by Meier et al. (2019) reporting the A2 variant as the predominant allele for β-CN in Angus cattle. Curik et al. (1997) showed similar frequencies for β-CN in Simmental cattle when compared with the frequencies in the present study. For α _s1-CN, the two alleles B and C were observed in both breeds with average genotype frequencies of 70.4% for BB, 23.1% for BC, and 6.5% for CC. The low appearance of the α _s1-CN C allele in GS with a minor allele frequency (MAF) of 4.7% is in agreement with allele frequencies in Croatian Simmental (Curik et al., 1997). For the casein α _s2-CN, an almost monomorphic genotype frequency for the genotype AA with 98.6% was found. The heterozygous genotype AD (1.4%) was detected in 6 of the 433 milk samples. This is the case for most cattle breeds of Western Europe. The alleles B and C did not occur, because the alleles B and C are zebu- and yak-specific (Grosclaude et al., 1979; Caroli et al., 2009). Only the allele D, which is rather common in French breeds (e.g., Montbéliarde; Grosclaude et al., 1979) and which was also described in GS (Erhardt, 1993), was observed with a MAF less than 1% (MAF = 0.7%) for both breeds.

Table 2.

Allele frequencies (in %) and test for Hardy–Weinberg equilibrium (χ ² and corresponding P values) for α _s1-, β-, α _s2-, and κ-casein in the studied breeds

Milk protein	Genotype	GA (n = 222)	GS (n = 211)	All breeds (n = 433)
α s1-casein	B	69.2	95.3	82.3
	C	30.8	4.7	17.7
	χ 2	2.668	5.415
	P	0.116	0.067
β-casein	A1	18.4	30.3	24.4
	A2	80.2	58.8	69.5
	A3	0.2	0.0	0.1
	B	0.9	8.5	4.7
	C	0.2	2.4	1.3
	χ 2	5.801	3.344
	P	0.205	0.567
α s2-casein	A	99.8	98.8	99.3
	D	0.2	1.2	0.7
	χ 2	0.001	0.03
	P	1	1
κ-casein	A	64.3	70.6	67.4
	B	30.9	28.9	29.9
	E	4.8	0.5	2.7
	χ 2	0.649	0.844
	P	0.720	0.543

Table 3.

Genotype frequencies (%) for α _s1-, β-, α _s2-, and κ-casein loci in the studied breeds

Milk protein	Genotype	GA (n = 222)	GS (n = 211)	All breeds (n = 433)
α s1-casein	BB	50.5	91.4	70.4
	BC	37.8	7.6	23.1
	CC	11.7	1.0	6.5
β-casein	A2A2	61.9	35.7	49.2
	A1A2	33.9	34.3	34.1
	A1A1	1.4	10.0	5.6
	A2B	1.8	8.1	5.0
	A1B	—	5.7	2.7
	A2C	0.5	3.3	1.9
	BB	-	1.4	0.6
	A1C	-	1.0	0.5
	A2A3	0.5	—	0.2
	BC	—	0.5	0.2
α s2-casein	AA	99.5	97.6	98.6
	AD	0.5	2.4	1.4
κ-casein	AA	40.8	49.0	44.8
	AB	40.8	42.2	41.5
	BB	9.2	7.3	8.3
	AE	5.9	0.5	3.4
	BE	2.8	0.5	1.7
	EE	0.5	—	0.3

Within the whole casein cluster and across all breeds, 43 composite genotypes were identified. The most common composite genotypes with frequencies larger than 10% were BBǀA1A2ǀAAǀAA (in order to α _s1-ǀ β-ǀ α _s2-ǀ κ-CN) (14.0%), BBǀA2A2ǀAAǀAA (12.6%), and BBǀA1A2ǀAAǀAB (10.9%) (Table 4). During the past decades, selection of cattle for improved beef production traits (e.g., fertile cows with high milk yield for calf growth, high-quality carcass; Campbell and Marshall, 2016) associated with casein genes may explain the expected imbalances in composite genotype frequencies. Positive selection increases frequencies for favorable alleles and neutral loci linked (Kaplan et al., 1989; Kreitman, 2000). Oppositely, negative selection hinders the spread of unfavorable alleles, causing the decreasing frequencies up to its complete loss in the respective population (Kreitman, 2000). Such underlying principles of selection for individual casein loci may explain the quite small number of composite genotypes with high frequencies and a larger number of composite genotypes with low frequencies.

Table 4.

Frequencies (%) of the composite casein genotypes within the casein cluster

Composite casein genotypes1,2	GA (n = 222)	GS (n = 211)	All breeds (n = 433)
BBǀA1A2ǀAAǀAA	14.3	13.7	14.0
BBǀA2A2ǀAAǀAA	14.7	10.2	12.6
BBǀA1A2ǀAAǀAB	5.1	17.1	10.9
BBǀA2A2ǀAAǀAB	6.9	13.2	10.0
BCǀA2A2ǀAAǀAB	11.5	2.0	6.9

¹In order to their location on BTA6 (α _s1-ǀβ-ǀα _s2-ǀκ-CN).

²Composite casein genotypes with average frequencies less than 5% for all breeds are excluded.

Effects of individual casein genotypes on growth traits of preweaning calves

Individual loci model

Table 5 displays the results from variance analyses (i.e., test of significance) and least squares means for preweaning calf performances with respect to casein genotypes. The fixed effects such as sex, age of the dam, CYS, and BW and the combination effect Breed group_BT as well as the casein fractions κ-CN and α _s1-CN (P < 0.05) influenced ADG and WW₁₈₀ (results from model 1).

Table 5.

Least squares means with corresponding standard errors (± SE) for BW, WW₁₈₀, and ADG with respect to the individual casein genotypes

		Growth traits
Casein genotype	n	BW, kg	WW180, kg	ADG, kg/d
α s1-casein
BB	1,301	36.8 ± 1.4a	215 ± 6.1a	1.00 ± 0.04a
BC	448	36.8 ± 1.4a	211 ± 6.4a	0.97 ± 0.05b
CC	115	36.9 ± 1.6a	208 ± 7.2a	0.95 ± 0.04b
P-value1		0.999 ns	0.038*	0.039*
β-casein
A2A2	920	36.9 ± 1.3ab	215 ± 5.5a	0.96 ± 0.03a
A1A2	660	37.5 ± 1.4b	213 ± 5.8a	0.96 ± 0.03a
A1A1	99	36.2 ± 1.5ab	207 ± 6.5a	0.92 ± 0.04a
A2B	81	36.3 ± 1.6ab	216 ± 7.0a	0.95 ± 0.04a
A1B	49	37.1 ± 1.7ab	219 ± 7.7a	0.99 ± 0.04a
A2C	26	38.0 ± 1.8ab	207 ± 8.6a	0.99 ± 0.09a
BB	10	31.0 ± 2.4a	202 ± 11.8a	0.99 ± 0.04a
A1C	7	39.4 ± 1.3ab	199 ± 12.9a	0.99 ± 0.04a
A2A3	5	40.2 ± 3.2ab	224 ± 15.9a	1.02 ± 0.05a
BC	5	36.1 ± 3.2ab	210 ± 16.0a	1.06 ± 0.09a
P-value1		0.039*	0.350ns	0.384ns
α s2-casein
AA	1,836	37.9 ± 1.0a	215 ± 4.8a	0.97 ± 0.03a
AD	19	37.0 ± 1.9a	207 ± 9.4a	0.92 ± 0.05a
P-value1		0.570ns	0.294ns	0.297ns
κ-casein
AA	856	37.1 ± 1.3a	204 ± 5.7a	0.93 ± 0.04a
AB	703	36.3 ± 1.3a	204 ± 5.5a	0.93 ± 0.04a
BB	145	36.0 ± 1.4a	215 ± 6.0b	0.99 ± 0.06b
AE	93	36.0 ± 1.6a	202 ± 7.0ab	0.93 ± 0.05ab
BE	27	36.6 ± 1.9a	214 ± 8.9ab	0.99 ± 0.09ab
EE	9	38.6 ± 3.0a	227 ± 14.7ab	1.06 ± 0.09ab
P-value1		0.250ns	0.011*	0.011*

¹Statistical significances were declared at P < 0.05; *P < 0.05; ns, not significant.

^a,bMeans in the same column with different superscripts are different (P < 0.05).

Least squares means for ADG of calves representing the ĸ-CN BB genotype (1.00 kg/d) were greater (P < 0.05) than ADG of calves representing the genotype AA (0.93 kg/d) and the genotype AB (0.93 kg/d). Least squares means for WW₁₈₀ were significantly larger for calves from dams with BB genotypes (215 kg) compared with calves from dams with AA and AB genotypes (204 kg, respectively). The favorable effect of κ-CN BB genotype on increasing ADG and WW₁₈₀ might be an effect of increased maternal milk production, especially increased protein yield (Ng-Kwai-Hang et al., 1986). In addition, greater ADG as well as WW₁₈₀ of calves consuming milk with BB genotype might be due to the smaller micelle size of the κ-CN allele B, causing positive physical and chemical properties (Hristov et al., 2014). Such differences in ĸ-CN genotypes are in agreement with the results by Weston (1971), who investigated the effects of casein genotypes on calf performance, including WW, adjusted WW₁₈₀, and 180-d gain. Weston (1971) identified significant effects of ĸ-CN on WW (P < 0.05) and on 180-d gain (P < 0.1). Least squares means for WW of calves from dams with genotype κ-CN BB were larger (213 kg) compared with WW of calves from dams with genotypes AA and AB (200 and 195 kg, respectively). The same genotype κ-CN BB contributed to an increased weight gain from birth to an age of 180 d (166 kg, P < 0.1), followed by the genotypes AA with 154 kg and AB with 151 kg 180-d weight gain. In contrast, Ziehe et al. (1993) identified favorable associations between the κ-CN genotype AA with calf ADG and WW. Faria et al. (1999) found no differences in ADG, WW, and BW between AA and AB genotypes for the casein κ-CN. In both latter studies, only the genotypes AA and AB were detected for the casein κ-CN, resulting in lacking information about the possible effects of κ-CN BB on growth traits. Henderson and Marshall (1996), who considered the κ-CN BB genotype besides AA and AB, reported that calves from dams with genotype AA grew faster and were heavier at weaning compared with calves from dams with genotype BB. Disagreements in genotype effects from different studies might be due to the different applied methodologies to investigate the relationship between κ-CN polymorphism and preweaning calf growth. The mentioned studies (i.e., Henderson and Marshall, 1996; Faria et al., 1999) only investigated individual κ-CN effects, but neglected the polygenetic effects of the whole casein gene cluster.

At the α _s1-CN locus, the BB genotype was associated with the greatest ADG (1.00 kg/d) and heaviest WW₁₈₀ (215 kg), whereas the genotype CC was associated with the lowest ADG (0.95 kg/d) and the lowest WW₁₈₀ (208 kg). The heterozygous genotype BC behaved intermediately, with corresponding least squares means of 0.97 kg/d for ADG and of 211 kg for WW₁₈₀ (Table 5). In addition, positive additive effects were observed for the two growth traits (3.8 kg for WW₁₈₀ and 0.02 kg for ADG) at the α _s1-CN locus (P < 0.05 for multiple pairwise genotype contrasts). Similarly, Weston (1971) observed a favorable impact of the α _s1-CN BB genotype on WW. Nevertheless, effects might be biased due to the almost fixation of the α _s1-CN B allele with a frequency of 93.7% in the studied breeds Hereford, Angus, Charolaise, Brown Swiss, and their crossbreeds. In consequence, the α _s1-CN alleles A and C had low frequencies of 0.3% and 6.0%, respectively. The almost monomorphic α _s1-CN locus is a result of long-term selection for high milk production in dairy cattle because the B allele favored selection in this regard (Ng-Kwai-Hang et al., 1986). In the present study, average frequencies for the genotypes BC and CC within the α _s1-CN locus were 23.1% and 6.5%, respectively (Table 3). Furthermore, Ng-Kwai-Hang et al. (1986) reported the lowest milk protein contents for α _s1-CN genotypes containing one C allele. In such a context, our study showed that the genotype with at least one α _s1-CN C allele led to lower ADG (0.95 kg/d) and WW₁₈₀ (211 kg) (Table 5). The influence of the α _s1-CN genotypes on the protein content of milk (Ng-Kwai-Hang et al., 1984) might explain the differences in ADG and WW₁₈₀ with respect to the different α _s1-CN genotypes.

The α _s2-CN locus indicated similar effects on growth traits of calves (P > 0.05; Table 5). The low frequency of the α _s2-CN AD genotype (1.4%) across the breeds, compared with the reference genotype AA (98.6%), may explain the absence of significant effects on BW, WW₁₈₀, and ADG. In genome-wide association studies, markers containing alleles with a MAF less than 5% are mostly excluded from the analysis due to their low heterozygosity (Fadista et al., 2016). The unacceptable MAF of 0.7% for α _s2-CN D allele in both breeds suggested to neglect the almost monomorphic α _s2-CN locus (Table 2). Actual effects of the α _s2-CN AD genotype on growth traits should be studied in breeds carrying higher frequencies for the α _s2-CN D allele (e.g., in Montbéliarde).

The β-CN locus affected neither ADG nor WW₁₈₀ of calves (P > 0.05). The β-CN genotypes with at least one A2 allele (A1A2, A2A2, A2B, and A2A3) contributed to increasing ADG and WW₁₈₀ (Table 5). Several studies (i.e., Tailford et al., 2003; Kamiński et al., 2007; Ul Haq et al., 2014) associated the β-CN A2 allele with benefits for human and animal health, whereas the A1 allele releases the opioid peptide beta-casomorphin-7 (β-CM7). Hence, there is an increasing interest in breeding cows producing pure A2 milk. However, controversial results concerning the influence of A1- vs. A2-milk consumption on health parameters were reported (Gödert et al., 2017). The European Food Safety Authority (EFSA) mentioned the insufficient data to determine a causal relationship between β-CM7 exposure (A1 milk) and noncommunicable diseases (EFSA, 2009).

The β-CN locus significantly affected BW of the calves (P < 0.05). Least squares means for BW indicated that calves from dams with genotype BB weighed less at birth than newborn calves from A1A2 dams (−6.5 kg; P < 0.05). As BW of calves is generally independent of milk consumption, the effect of β-CN on BW suggested that the maternal casein loci may be linked to genomic segments being associated with fetal growth characteristics. In this regard, Eberlein et al. (2009) identified a highly significant quantitative trait locus (QTL) affecting fetal growth on BTA6 in a unique cattle resource population. The authors suggested a nonsynonymous polymorphism in the non-SMC condensing I complex, subunit G (NCAPG) gene, NCAPG: c.1326T > G, as a functional mutation for bovine fetal growth. In the same context, Gutiérrez-Gil et al. (2009) mapped a highly significant QTL for BW and body length at birth on BTA6 in a cross from Charolais and Holsteins. The reported confidence interval from position 27 to 65 cM (Gutiérrez-Gil et al., 2009) reflects the given confidence interval for a QTL with an impact on bovine fetal growth (Maltecca et al., 2009). Osteopontin (gene: SSP1 mapped on BTA6) is a phosphorylated acidic glycoprotein that functions as an immune modulator and affects bone mineralization and blood regulation. Johnson et al. (2014) found the impact of SSP1 on the development of the epitheliochorial placenta, which utilizes placental areolae for histotrophic support of the growing fetuses. Echternkamp (1993) identified associations between the placental development and calf BW in beef cattle, explaining the possible link between the caseins and SSP1. Interestingly, in a functional genomics approach, Sheehy et al. (2009) suggested SPP1 as an important regulator of bovine milk protein gene expressions.

Composite loci model

Table 6 shows the estimates of the effects of the composite casein loci on calf growth traits from model 2. BWs of calves were not affected by the common composite casein genotypes (P > 0.05). As shown in the results from model 1, the β-CN BB genotype was associated with lower BW (−6.5 kg) in comparison to the A1A2 genotype (P < 0.05). The appropriate composite genotype including β-CN BB (e.g., AA|BB|BB|AA) was not included in model 2, due to its frequency lower than 5%.

Table 6.

Least squares means with corresponding standard errors (± SE) for BW, WW₁₈₀, and ADG with respect to the composite α _s1-β-α _s2-κ-CN genotypes

		Growth traits
Composite casein genotype1	n	BW, kg	WW180, kg	ADG, kg/d
BB|A1A2|AA|AA	280	39.4 ± 0.7a	217 ± 3.1ab	0.97 ± 0.02ab
BB|A1A2|AA|AB	188	38.8 ± 0.7a	214 ± 3.5ab	0.95 ± 0.02ab
BB|A2A2|AA|AA	251	39.1 ± 0.7a	218 ± 3.3ab	0.97 ± 0.02ab
BB|A2A2|AA|AB	157	38.1 ± 0.8a	223 ± 3.6b	1.00 ± 0.02b
BC|A2A2|AA|AB	123	37.7 ± 0.8a	211 ± 4.0a	0.94 ± 0.02a
P-value2		0.121ns	0.034*	0.029*

¹Composite genotypes coded in order to their location on BTA6: α _s1-ǀβ-ǀα _s2-ǀκ-CN.

²Statistical significances were declared at P < 0.05; *P < 0.05; ns, not significant.

^a,bMeans in the same column with different superscripts are different (P < 0.05).

The composite casein genotypes significantly influenced ADG and WW₁₈₀ (P < 0.05). Calves grew faster when fed with milk containing the genotype BB|A2A2|AA|AB compared with calves consuming BC|A2A2|AA|AB (1.0 vs. 0.93 kg/d; P < 0.05; Figure 2). Least squares means for WW₁₈₀ were significantly greater with respect to the composite genotype BB|A2A2|AA|AB (223 kg) than for BC|A2A2|AA|AB (211 kg; P < 0.05). Calves from dams with the composite genotype BB|A2A2|AA|AB weighed at an age of 180 d were 12 kg more than calves from dams with the composite genotype BC|A2A2|AA|AB (Figure 3). The only difference between these two composite genotypes is the C allele substitution in α _s1-CN. Previous studies estimating the effects of milk protein genotypes on milk production traits (e.g., Lin et al., 1986) indicated that, among all caseins, α _s1-CN had the greatest influence. Lin et al. (1986) hypothesized that the superior effect of α _s1-CN is due to the largest content of the proteins in milk with 10.6 g/L. The unfavorable effect of the α _s1-CN C allele on growth traits is explained with the detrimental impact on milk production traits (Ng-Kwai-Hang et al., 1984), indicating maternal deficiency in beef cattle. The single β-CN A2A2 genotype effect had no significant effect on ADG and WW₁₈₀ (P > 0.05; model 1), but the joint effect of β-CN A2A2 and κ-CN AB within the composite genotype BB|A2A2|AA|AB was favorably associated (P < 0.05) with both growth traits. Epistatic effects between the κ-CN and β-CN loci, as postulated by Graml and Pirchner (2003), might be an explanation in this regard.

Figure 2.

Least squares means of ADG with respect to composite genotypes within the casein cluster with 95% confidence limits. ¹Composite casein genotypes coded in order to their location on BTA6: α _s1-ǀβ-ǀα _s2-ǀκ-CN. ^a,bThe letters a and b indicate significant differences between casein genotypes at P < 0.05. Red and black bars show standard errors and upper-/ lower confidence level, respectively.

Figure 3.

Least squares means of WW₁₈₀ with respect to composite genotypes within the casein cluster with 95% confidence limits. ¹Composite casein genotypes coded in order to their location on BTA6: α _s1-ǀβ-ǀα _s2-ǀκ-CN. ^a,bThe letters a and b indicate significant differences between casein genotypes at P < 0.05. Red and black bars show standard errors and upper-/ lower confidence level, respectively.

The favorable effect of the composite genotype BB|A2A2|AA|AB on growth traits (ADG and WW₁₈₀) may be caused by other QTL for body weight, being in linkage disequilibrium with the casein genotype. In this regard, genome-wide association studies (e.g., Setoguchi et al., 2009; Liu et al., 2015) detected further QTL for body and carcass weight within the NCAPG gene, suggesting a potential mechanism through bone or muscle growth and/or lipid deposition. Furthermore, the gene SPP1 on BTA6 was associated with body weight of growing Polish Holstein-Friesian cattle (Pareek et al., 2008). Han et al. (2012) observed similar relationships between the porcine SPP1 gene and growth and carcass weight in Korean black pig populations. Han et al. (2012) assumed the impact of SPP1 not only on the early embryo development but also long-term effects on tissue development for bones and muscles.

In conclusion, our results indicate that, in addition to well-known environmental factors, milk protein genotypes influence the preweaning growth of calves. Such a pronounced genetic impact offers new perspectives for breeding strategies aiming at the improvements of the calf development due to the nutritional value of colostrum or milk. Growth of the calf is generally defined as the most important factor for meat productivity (Corrêa et al., 2006). Hence, the selection on favorably associated casein genotypes will contribute to economic efficiency through a shortened fattening period and reduced feeding costs. The present study emphasizes the need to monitor the distribution of milk protein variants in beef cattle.

Glossary

Abbreviations

αs1-CN

alpha-S1-casein

αs2-CN

alpha-S2-casein

β-CM7

beta-casomorphin-7

β-CN

beta-casein

κ-CN

kappa-casein

ADG

average daily weight gain

BT

birth type

BTA6

bovine chromosome 6

BW

birth weight

CYS

calving year-season

EFSA

European Food Safety Authority

GA

German Angus

GS

German Simmental

MAF

minor allele frequency

NCAPG

Non-SMC Condensin I Complex Subunit G

QTL

quantitative trait loci

SNP

single-nucleotide polymorphism

SP

suckling period

WW

weaning weight

WW₁₈₀

weaning weight at day 180

Conflict of interest statement

The authors declare no conflict of interest.