1. Introduction
The dairy industry is a crucial cog in the agricultural economy of many countries. Therefore, milk yield and its components are among the major goals targeted by animal geneticists [
1]. Recently, increasing interest in sheep milk has led to extensive studies aiming to explore the local sheep breed’s genetic potential in milk production [
2]. Awassi is the most common breed of sheep in Middle East countries where its products, including meat, milk, and wool, play an important socioeconomic role in Jordan. Awassi lambs are profitable and the milk of Awassi ewes is a commodity valued by farmers for its ability to cover annual production costs. For this reason, Awassi daily milk production has been enhanced via genetic selection from 0.5 kg to 1.2 kg [
3]. Meanwhile, milk production and composition in Mediterranean countries was observed to fluctuate within and between flocks and countries as a result of different environmental factors and variations within and between Awassi genotypes [
4,
5,
6,
7,
8,
9]. In the past, animal breeders have made effective efforts to improve the performance of livestock by artificial selection and breeding to increase the frequency of certain desirable traits. This only involved the direct use of phenotypic observations without any knowledge of the underlying molecular information. Therefore, there is a need to adopt methods of selection based on genomic queues [
10].
Milk protein polymorphisms are of great importance in the dairy industry and animal-breeding programs due to its association with quantitative and qualitative milk traits and its potential use in genetic selection programs of dairy sheep breeds [
11]. Many studies with diverse breeds of dairy sheep have indicated Beta-lactoglobulin (
β-LG), Prolactin (
PRL) and Kappa casein (
CSN3) as promising candidate genes for milk production and composition [
2,
12,
13,
14,
15,
16].
Beta-lactoglobulin, coded by the
β-LG gene, is synthesized by the secreting cells of the mammary gland.
β-LG is the primary whey protein in ruminant’s milk and accounts for approximately 17 to 22% of total milk proteins [
15]. The
β-LG gene is located on ovine chromosome 3 [
17], and exon number 2 of
β-LG revealed three allelic polymorphisms (A, B and C) based on different amino acid changes. The alleles A and B (Tyr/His) differ at the amino acid position 20 [
18] and the genetic variant C differs from variant A by an amino acid exchange at position 148 (Arg/Gln) (GenBank accession No. X12817). The most common genetic variants detected in all studied sheep breeds are A and B, while the variant C is regarded as rare, and only found with low frequencies in Carranzana, Black Merino and White Merino breeds [
15]. Subsequent studies have demonstrated the
β-LG gene’s polymorphic effect on milk components including yield, protein, fat, and lactose content [
14,
19,
20] as well as the impact of protein genetic variants on the properties of milk [
2].
Prolactin, coded by the
PRL gene, is a lactogenic hormone found in many species. The
PRL gene plays a key role in the development of the mammary gland and milk secretion; its depletion in sheep provokes a severe reduction in milk production [
21], suggesting that
PRL is a functional candidate gene contributing to variations in milk production. Many researchers suggested the association of
PRL gene polymorphisms with milk yield and composition traits in different dairy sheep breeds [
13,
14,
16,
22]. The
PRL gene is in a region of the ovine chromosome 20 with putative quantitative trait locus (QTL) for milk yield and composition [
23]. Thus, PRL is primarily responsible for the synthesis of fat, proteins, and all other major components of milk [
24]. These characteristics make
PRL a strong candidate gene for milk traits, and potentially be used as a positional marker gene associated with milk yield and composition.
The four caseins (
αS1-, αS2-, β- and κ-casein) are the major proteins in sheep milk, accounting for about 80% of total protein in milk [
24]. Among the 4 caseins,
κ-casein (
CSN3) accounts for approximately 15% of total casein, and thus represents one of the most important proteins due to its essential role in micelle formation and stabilization [
2], and thus determines the manufacturing properties of milk. Therefore,
CSN3 could be used as a positional marker gene associated with high milk production and milk composition improvement. Molecular analyses of ovine
CSN3, mainly of exon 4, revealed synonymous and non-synonymous sequence differences [
12,
25,
26]. However, limited information about the association of this marker with milk yield and composition in dairy sheep could be found in the literature. One positive association between
CSN3 gene polymorphism and lactose percentage has been reported in Zel sheep [
20]. Staiger et al. [
16] and Gras et al. [
12] reported that the polymorphisms found in the
CSN3 gene did not have any significant effect on milk yield and composition traits. Hence, substantiating phenotypic data with molecular information to improve milk and its composition is critical for the future of dairy sheep industry [
16].
The present study screens some Awassi genetic loci for possible variants in the β-LG, PRL, and CSN3 genes and establishes their frequency in commercial Awassi sheep flocks in the ultimate objective of investigating the effect of these genotypes and their interaction on milk production and composition traits. To our great knowledge, this is the first study that investigates the association of some molecular information and milk traits in the commercial Awassi flocks in Jordan.
2. Materials and Methods
2.1. Animal and Sample Collection
All experimental protocols involving animals were approved by the Animal Care and Use committee (ACUC), Jordan University and Science Technology (JUST), approval Number 16/3/3/578.
A participatory animal-breeding program was performed in the South, North, and middle of Jordan; this program was supported by Livestock and Range land research directorate of the National center for agricultural research and Extension (NCARE) (recently renamed to National Agricultural Research Center (NARC)) staff. Specific farmers were appointed before starting the program and several meetings were performed with the livestock owners. A total number of 928 ewes that belong to 9 flocks (three in each region) were targeted through the participatory animal-breeding program.
After specifying the collaborating farms, semi-structural questionnaires were completed for each animal in the flock covering parity and age of dam, sex and type of birth of the newborns and available sires and dams in addition to all related environmental factors. All animals (n = 928) were monitored and milk data were recorded directly by three specialized teams in each region (North, middle, and South). The exhaustive milk and pedigree data however was collected from two commercial Awassi flocks, selected based on the availability and ability for collecting performance data, and located in the northern part of Jordan. Data were collected from the two farms during 5 lambing and milking seasons. The team of NARC collected the data by identifying the sires in the breeding seasons, recording lamb information (birth and weaning weights, date of birth and weaning, dam weight at lambing, milk tests as described below), after which the trained farmers continued data collection under the supervision of the livestock research director. The farmers were supplied by special separation cages. They measured milk yield with graded milk cylinders and recorded its weight with small portable digital weighing scales (up to 30 kg) or portable digital mobile weighing scales (up to 200 kg).
Blood samples for DNA harvesting were carefully collected from the jugular vein of the 167 Awassi ewes (2 to 6 years old) that were born to 31 sires using vacuum tubes treated with 0.25% Ethylene Diamine Tetraacetic Acid (EDTA) (BD Vacutainer Systems, Plymouth, UK) and stored at −20 °C until DNA isolation.
2.2. Milk Samples and Analysis
A total of 391 full lactation records on 167 ewes were collected from 2007 and 2011 from the two commercial farms identified. Milk was collected manually via hand milking by skilled workers. Through the pre-weaning period, ewes were milked once a day followed by analyzing milk amount of each ewe twice a day (morning and evening), at biweekly intervals. To measure the amount of milk produced during the suckling period (60 days), lambs were isolated from their dams 12 h before morning milking. Ewes were dried off when their milk production was reduced to less than 100 mL. The average lactation length, including suckling and milking periods was 115 days. Through the suckling period, ewes were milked once, keeping half of the milk for the lamb’s consumption, so the milk amount was multiplied by 2. While in the post-weaning period ewes were milked twice a day (morning and evening), milk amounts were averaged and multiplied by the interval between the two successive tests period then summed to obtain the total milk yield (TMY). Test-Day Milk yield (TDM) was calculated by dividing TMY by Lactation Period of each ewe. Approximately 50 mL of milk sample was collected from the morning milk of each ewe to determine basic composition of fat%, protein%, lactose%, solids-not-fat% (SNF%) and milk density g/cm2. Milk composition (n = 986 milk samples) was analyzed using a Milko Scan (Minor Type 78100, FOSS Electric, Hillerød, Denmark) available at JUST university.
2.3. Genomic DNA Extraction and Polymerase Chain Reaction (PCR)
Genomic DNA was extracted using Wizard Genomic DNA Extraction Kit (OMGA-Bio-Tek, Inc., Madison., WI, USA). DNA quality was tested using 1.5% agarose gel electrophoresis. Polymerase chain reaction (PCR) was used for amplifying the studied genes; using primers targeting exon II of the
β-LG, intron 2 of PRL as shown in
Table 1. Primers for CSN3 were designed, using primer 3 (
http://frodo.wi.mit.edu/primer3/), to target part of the intron 3 and the totality of exon 4 and part of intron 4, using the available nucleotide sequence (Accession No.: 443394) on the NCBI GenBank database. PCR mix (HOT FIREPol DNA Polymerase; Solis BioDyne, Estonia) was carried out in a total volume of 20 μLcontaining10μL of nuclease-free water, 2 μL of genomic DNA (100 ng/ μL) as a template, 2 μL of each primer, and 4 μL (5U/µL) of Taq DNA polymerase (Eppendorf AG, Hamburg, Germany). Primer sequences, annealing temperature, and restriction enzymes used for genotyping are shown in
Table 1. The PCR reaction was carried out in the following conditions of 95 °C for 5 min for initial denaturation followed by 33 cycles at 95 °C for 30 s of denaturation, 40 s annealing (
Table 1) and extension each at 72 °C, and a final extension step at 72 °C for 7 min.
2.4. Restriction Fragment Length Polymorphism (RFLP) Analysis
β-LG and PRL genes variants were identified by the PCR-RFLP method. The amplified β-LG gene fragment (301 bp) was digested by RasI restriction enzyme for 2 h at 37 °C. Restriction products were separated in a 2% agarose gel with ethidium bromide and visualized under ultraviolet (UV) light. The amplified PRL gene fragment (1209 bp) was restricted with HaeIII endonuclease at 37 °C for 3 h. The RFLP profile was visualized by the same way as for β-LG.
2.5. Sequencing Analysis
Sequences were obtained using the same primers used for PCR amplification as shown in
Table 1. The PCR products of the different genotype patterns of the
CSN3 gene were purified and sequenced by Macrogen Incorporation (Seoul, South Korea) to identify the single nucleotide polymorphisms (SNPs) found in these different genotype patterns. Ten randomly chosen PCR samples for each genotype of
β-LG were sequenced from both directions to confirm the results obtained by PCR-RFLP technique. Sequence analysis and alignments were carried out using BioEdit program version 5.0.6. [
29].
2.6. Statistical Analysis
The genotype and allelic frequencies of the
β-LG,
PRL, and
CSN3 loci were calculated using Pop-Gene 32 package version 1.31 program [
30]. A chi-square (χ2) test was performed to test the goodness of fit to Hardy-Weinberg equilibrium expectations for the distribution of genotypes. The effects of genotypes of
β-LG,
PRL, and
CSN3, and their interactions on the traits studied were analyzed using the least-squares method as applied in a mixed-model procedure of SAS/ STAT
® software (SAS Institute Inc., Cary, NC, USA, v9.1). Two statistical models were used as described below.
The first model used for the milk production traits analysis was:
where:
- -
Yijklnmop = the studied traits;
- -
Μ = overall mean of the total milk yield or test-day milk yield;
- -
BLGi = fixed effect of the ith genotype at β-LG locus (I = AA, AB and BB);
- -
PRLj = fixed effect of the jth genotype at PRL locus (j = AA, AB and BB);
- -
CSNk = fixed effect of the kth genotype at CSN3 locus (k = TT and TC);
- -
Pl = fixed effect of the lth parity or number of lambing (l = 1, 2, 3, 4, 5 and 6);
- -
Sm= random effect of mth sires (m = 1, 2 to 31);
- -
SYn = fixed effect of the nth year-season of lambing (n = 2007 to 2011);
- -
Bo = linear regression coefficient dam weight at lambing;
- -
DWo = dam weight at lambing as covariate;
- -
(BLG × PRL)ij = interaction between β-LG genotypes and PRL genotypes (ij = AAAA, AAAB, AABB, ABAA, ABAB, ABBB, BBAA, BBAB, and BBBB);
- -
(BLG × CSN)ik = interaction between β-LG genotypes and CSN3 genotypes (ik = AATT, AATC, ABTT, ABTC, BBTT, and BBTC);
- -
(PRL × CSN) jk = interaction between PRL genotypes and CSN3 genotypes (jk = AATT, AATC, ABTT, ABTC, BBTT, BBTC);
- -
Eijklmnop = random errors with the assumption of N (0, σ2).
The second model used for the milk composition traits analysis:
where:
- -
Yijklnmo = the studied traits;
- -
Μ = overall mean of Fat%; protein%, SNF%, Total solids, lactose%, and density (g/cm2)
- -
BLGi = fixed effect of the ith genotype at β-LG locus (I = AA, AB and BB);
- -
PRLj = fixed effect of the jth genotype at PRL locus (j = AA, AB and BB);
- -
CSNk = fixed effect of the kth genotype at CSN3 locus (k = TT and TC);
- -
Pl = fixed effect of the lth parity or number of lambing (l = 1, 2, 3, 4, 5 and 6);
- -
Sm = random effect of mth sires (m = 1, 2, …, 31);
- -
Bn = linear regression coefficient TDM.
- -
TDMn = TDM covariant.
- -
(BLG × PRL)ij = interaction between β-LG genotypes and PRL genotypes (ij = AAAA, AAAB, AABB, ABAA, ABAB, ABBB, BBAA, BBAB, and BBBB);
- -
(BLG × CSN)ik = interaction between β-LG genotypes and CSN3 genotypes (ik = AATT, AATC, ABTT, ABTC, BBTT, and BBTC);
- -
(PRL × CSN)jk = interaction between PRL genotypes and CSN3 genotypes (jk = AATT, AATC, ABTT, ABTC, BBTT, and BBTC);
- -
Eijklmno = random errors with the assumption of N (0, σ2).
The three-ways interaction effects among the three genes and the flock effect were removed from the model as it was not significant. For all statistical comparisons, a probability level of p < 0.05 was considered to be statistically significant.
4. Discussion
This study reports the association between
β-LG,
PRL, and
CSN3 genes and milk production and composition of Awassi sheep. These genes were chosen because of their direct involvement in the growth and development of the mammary gland, maintenance of milk secretion, and synthesis of milk [
2,
12,
14,
15,
22,
31]. The genes are also located in the region of QTL influencing milk quantity and quality [
12,
14,
15,
22,
23]. This study revealed that in Awassi sheep, these genes present different allelic frequencies and genotypes (
Table 2).
β-LG locus showed a higher frequency of allele B (0.58) than allele A (0.42) in Awassi sheep. Similar results were found in Chios sheep breed [
19], Racka Sheep [
32], Rusty Tsigai breed [
33], Zel breed [
20], Polish Merino breed [
34], and Awassi breed [
3,
35]. No evidence was found for C allele in our study, which is considered a rare variant detected only in a few breeds such as Turcana, Racka, Tsigai, Karakul of Botosani, Transylvanian Merino, Merinoland, and Hungarian Merino [
15,
36].
In agreement with other studies, the frequency of
PRL A allele (0.82) was more than B allele (0.18); Allele frequency was found at 0.64 in Spanish Merino sheep [
14], 0.53 in Black Head sheep breed [
12], 0.75 in Awassi sheep [
37], 0.64 in Serra da Estrela, 0.57 in White Merino, and 0.72 in Black Merino [
38].
However, the gene frequency for the CSN3 locus in Awassi sheep reported in this study differed from East Friesian sheep [
16] and Black Head sheep [
12] where the T and C alleles were of equal frequencies.
The findings presented in this paper indicated that
β-LG gene polymorphisms are not associated with milk production traits in Awassi sheep (
Table 4). These results are in agreement with previous studies reported by Padilla et al. [
14] in Spanish Merino sheep, Triantaphyllopoulos et al. [
19] in Karagouniko and Chios sheep breeds, Giambra et al. [
39] in East Friesian Dairy and Lacaune sheep, Kaweka and Radko [
34] in Polish Mountain, Polish Merino, East Friesian and Bergschaf, and Staiger et al. [
16] in East Friesian sheep where the effect of
β-LG variants on milk yield was not indicated. Although
β-LG gene did not show any associations with high milk yield, inverse response over milk composition must be considered in marker assisted selection (MAS) strategy. We found that the AA and BB genotypes of
β-LG gene were associated with the highest fat percentage compared to AB genotype. Some previous studies reported significant effects of
β-LG on fat content in milk of different sheep breeds. Positive effects on milk fat content varied depending on sheep breed and
β-LG genotype;
β-LG BB genotype had a significant effect on milk fat percentage in Awassi sheep [
40], AA and AB genotypes in Italian Altamurana and Leccese sheep [
41], AA genotype in Merino sheep [
42] and East Friesian Dairy sheep [
39], AB genotype in Zel breed [
20]. A recent study conducted by Padilla et al. [
14] on the Spanish Merino sheep showed that
β-LG A allele had positive significant effects on fat percentage.
We also found that the BB variant of the
β-LG gene had a very large positive effect on protein%, SNF%, lactose%, and milk density compared to AB and AA genotypes (
Table 4). These results were consistent with many previous studies where a strong association of the BB variant with protein percentage [
43,
44] and lactose content [
19,
20,
45], but inconsistent in regard to SNF% [
34,
41,
44,
46] and milk density [
20].
When analyzing the PRL genotypes effects on milk production traits, the AA genotype was associated with the highest milk production (
Table 5). This result is consistent with some published results [
12,
13,
16] who reported a significant difference confirming the superiority of the
PRL AA genotype in milk yield. However, fat percentage was not significantly affected by the
PRL genotype (
Table 6) consistent with some published reports [
14,
38]. Moreover, contrary to available findings of BB genotype’s superiority for protein percentage [
12,
14,
38], our study found a stronger correlation with the AA genotype of the
PRL gene compared to the AB and BB genotypes in Awassi ewes. Another study using Sakiz ewes, Ozmen and Kul [
13] found that heterozygous AB ewes produced the highest milk protein percentage when compared to homozygous AA and BB genotypes animals, while there were no significant differences in protein percentage according to different genotypes in Akkaraman and Awassi ewes.
PRL AA and BB genoty [
13] pe also showed positive effects on SNF% and milk density (
Table 6); however, according to the literature, no other studies examined this association. This study also showed that the BB genotype of the
PRL gene particularly was associated with higher production of lactose in milk compared with the AA and AB genotypes. In contrast to these results, Ozmen and Kul [
13] reported lack of association between lactose content and PRL genotypes in Sakiz, Akkaraman, and Awassi ewes.
Our results did not show any significant effect of the
CSN3 variants on milk production traits as seen in the milk yield of Black Head sheep [
12] and East Friesian sheep [
16]. We found no significant effect of
CSN3 variants on fat, protein, lactose content, and milk density, whereas SNF content was positively affected by only the TT genotype (
Table 6). Gras et al. [
12] also reported no associations when his team examined the influence of
CSN3 polymorphism on fat and protein content in Black Head sheep. On the other hand, Yousefi et al. [
20] reported that the k1 pattern of the
CSN3 locus affected only lactose percentage in milk and milk density in Zel sheep.
The interesting portion of this project was the combined genotype effect on milk production and composition; genotype combination reflects the interactions of multiple genes effects in a certain quantitative trait [
47]. Our study only found a significant impact of the interaction of
β-LG*PRL genotypes on milk production (
Table 6), while PRL
× CSN3 and
β-LG ×
CSN3 genotypes showed minimal combined effects. (
p < 0.05). Mile production was highest in Awassi ewes of AA × BB (
β-LG × PRL) combined genotypes compared to the ABxBB genotype. While milk composition only improved in the
β-LG × PRL and PRL × CSN3 genotypes compared to
β-LG × CSN3 (
p < 0.05). The statistical results show that ewes of AA × BB genotype combination had the highest fat%, and BB × BB genotype combination had the highest SNF%, protein%, lactose% and milk density compared to other genotypes (
Table 6). The highest fat% and SNF% were recorded for the BB × TT genotype (PRL ×
CSN3) compared to the other genotypes.