The promoter area of the gene has particular importance as it may include the region for various transcription binding sites that control the post-translation modification of the gene. RNA polymerase II is the major enzyme to control the transcription mechanism of the gene, which is further related to the identification and classification of the core promoter region in the sequence (Sandelin et al., Reference Sandelin, Carninci, Lenhard, Ponjavic, Hayashizaki and Hume2007). Cows and buffalo kept for milk production are a core farming activity in many parts of the world, and the choice between the two is often influenced by milk constituents, especially milk fat, as well as the variable incidence of pathogenic mastitis (Bhattarai et al., Reference Bhattarai, Worku, Dad, Rehman, Gong and Zhang2018). The DGAT1 gene (K232A mutation) is considered to be one of the most important candidate genes for encoding acyl-CoA:diacylglycerol acyltransferase which eventually affects the milk fat content (Cases et al., Reference Cases, Smith, Zheng, Myers, Lear, Sande, Novak, Collins, Welch, Lusis, Erickson and Farese1998; Winter et al., Reference Winter, Krämer, Werner, Kollers, Kata, Durstewitz, Buitkamp, Womack, Thaller and Fries2002; Bovenhuis et al., Reference Bovenhuis, Visker, Poulsen, Sehested, van Valenberg, van Arendonk, Larsen and Buitenhuis2016). The majority of previous research has been done in cattle, and there are few studies of DGAT1 gene in buffalo (Yuan et al., Reference Yuan, Zhou, Deng, Hu and Li2007; Cardoso et al., Reference Cardoso, de Souza, Aspilcueta-Borquis, Neto, de Camargo, Hurtado-Lugo, Scalez, de Freitas, Albuquerque and Tonhati2015; Freitas et al., Reference Freitas, de Camargo, Aspilcueta-Borquis, Stafuzza, Venturini, Tanamati, Hurtado-Lugo, Barros and Tonhati2016a; Parikh et al., Reference Parikh, Patel, Patel, Patel, Patel, Patil, Kansara, Jakhesara and Rank2016; Silva et al., Reference Silva, Silva Filho, Matos, Schierholt, Costa, Marques, Costa, Sales, Figueiró and Marques2016). None of these reports mentioned the identification of K232A polymorphism in the buffalo populations that they studied. Since the milk fat content of cow and buffalo differ considerably, DGAT1 gene expression and function in cow and buffalo could be a subject of study. Although few SNPs were identified in the buffalo DGAT1 sequence, as yet no association study has been performed to correlate DGAT1 influence in buffalo milk fat content (Yuan et al., Reference Yuan, Zhou, Deng, Hu and Li2007; Cardoso et al., Reference Cardoso, de Souza, Aspilcueta-Borquis, Neto, de Camargo, Hurtado-Lugo, Scalez, de Freitas, Albuquerque and Tonhati2015; Freitas et al., Reference Freitas, de Camargo, Aspilcueta-Borquis, Stafuzza, Venturini, Tanamati, Hurtado-Lugo, Barros and Tonhati2016a; Silva et al., Reference Silva, Silva Filho, Matos, Schierholt, Costa, Marques, Costa, Sales, Figueiró and Marques2016). Differences in the genomic sequence, especially from coding sequence or transcription factors binding sites in the DGAT1 gene, could help to explain the differences in milk fat quantity among these two species.
Postulating that the DGAT1 gene is a candidate gene for milk fat synthesis, the present study is undertaken to determine the gene structure and prevailing differences in DGAT1 in a large set of tropically adapted animals representing diversified Holstein cow and Murrah buffalo breeds. We specifically investigated the sequences of the 5-flanking region and coding region to find out functional variations.
Materials and methods
Animal sampling, DNA extraction, primer design and SNP study
For the initial mutation screening for exon 8 (K232A), a total of 84 blood samples including 41 cows and 43 buffaloes were used and DNA was extracted as described by Bhattarai et al. (Reference Bhattarai, Chen, Rehman, Hao, Ullah, Dad, Talpur, Kadariya, Cui, Fan and Zhang2017). Information from public database of NCBI (http://www.ncbi.nlm.nih.gov), DGAT1 gene referring to Bos_taurus_UMD_3.1.1 (GCF_000003055.6) for cow and UMD_CASPUR_WB_2.0 (GCF_000471725.1) for buffalo was annotated. Primers were designed to amplify the target sequence of DGAT-1 gene. Primer premier 5.0 software was used to designed all the primers (Premier Biosoft International, Palo Alto, CA, USA), which was synthesized by Sangon Biological Engineering Technology (Shanghai, China). PCR protocol was followed as described previously (Bhattarai et al., Reference Bhattarai, Chen, Rehman, Hao, Ullah, Dad, Talpur, Kadariya, Cui, Fan and Zhang2017).
Bioinformatics analysis and software used
We used http://www.fruitfly.org/seq_tools/promoter.html for finding possible promoter sequence and http://www.cbs.dtu.dk/services/Promoter/ for predicting transcription start sites in the 5′ UTR and 5′ upstream sequence of cow-DGAT1 and Buffalo-DGAT1 gene (Li et al., Reference Li, Wang, He, Xie and Zhang2015). CpG Island was predicted with http://www.urogene.org/methprimer/ and http://www.ebi.ac.uk/Tools/emboss/cpgplot/ software package. Transcription factor (TF) binding sites at predicted sequence was analyzed with a variety of tools (http://www.cbrc.jp/research/db/TFSEARCH.html; http://Zlab.bu.e-du/cluster-buster/cbust.html; http://zlab.bu.edu/cluster-buster/cbust.html).
Promoter cloning and generation of luciferase reporter constructs
In order to clone the bovine DGAT-1 promoter region, we designed gene-specific primers (online Supplementary Table S1) to amplify a 2.0-kb genomic region upstream of the bovine DGAT-1 gene TSS. For the generation of the luciferase reporter construct, the 2.0-kb bovine DGAT-1 promoter fragment was excised simply by digestion with KpnI and BgIII (TaKaRa, Dalian, China) and ligated into the pGL3-basic vector digested with the same restriction enzymes (Fig. 2A).
Cell lines, culture conditions, transient transfection and luciferase reporter assay
We used BMEC (bovine mammary epithelium cell line), HEK293T (human embryonic kidney 293T) and CHO (Chinese hamster ovary cell) cells for the culture experiments. Cells were plated at a density of 1.2 × 105 cells/well in 48-well dishes and 24 h later, they were transfected with the plasmids. Transfection was done in each well of cell plate by adding 450 ng constructs using Lipofectamine 2000 (Invitrogen). Again about, 450 ng recombinant plasmids were co-transfected with 50 ng pRL-TK. We used pGL3-basic as a negative control. Transfection period was 48 h. Finally, the ratio of firefly luciferase light units to Renilla luciferase light units was analyzed which included three independent experiments.
Statistical analysis
One-way ANOVA was used and the unpaired Student's t-test was used to detect significant differences (P < 0.05). Values were represented as the mean ± sd, and statistical significance was indicated as follows: *P < 0.05; **P < 0.01; ***P < 0.001. Data were representative of at least three independent experiments.
Results and discussion
Structure characteristics and difference screening in cow and buffalo sequence of the bovine DGAT1 gene
The gene was characterized individually in cow and buffalo sequence investigating the mRNA sequence, protein sequence and the components of 5′ upstream and 3′ downstream region. The genetic open reading frame (ORF) trans-skeleton was determined in cow and buffalo-DGAT1 sequence (online Supplementary Fig. S1). The mRNA and interacting protein were compared between cow DGAT1 and buffalo DGAT1 gene (online Supplementary Table S2 and Fig. S2). The mRNA at 102 (C>T) changes the amino acid from Alanine (A) to Valine (V); 123 (T>C) changes Valine (V) to Alanine (A); 707 (G>A) and 708 (C>A) changes Lysine (K) to Alanine (A) and finally 1464 (C>T) changes Alanine (A) to Valine (V). These changes could be responsible for the difference in the protein structure or functions of DGAT1 between cow and buffalo which could be helpful for further analysis.
Characterization of the bovine DGAT1 gene 5′-regulatory region
We used −2000 bp upstream sequences (considering ATG as +1) to predict the possible promoter sequence, transcription start sites, transcription binding sites and CpG Island in cow and buffalo-DGAT1 respectively (online Supplementary Figs. S3 and S4). Our results revealed that the sequences between +16 to +65; −225 to −396 and −991 to −1040 are predicted as possible promoter sequence in cow-DGAT1 gene. Similarly, two regions are predicted as possible promoter sequence in buffalo-DGAT1 sequence viz. (−383 to −432), (−1025 to −1074). There were few regions predicted for CpG Island when we considered −2000 upstream bp sequences.
Single nucleotide polymorphism and comparison
Previous results from various authors revealed the SNP at exon 8 of the cow (K232A) is highly significant to milk fat. In order to know if this exists in cow and buffalo population, we tested the mutation targeting exon 8 and found the mutation was present only in Chinese Holstein bovine population but not in Chinese cross Murrah buffalo breed (Fig. 1D). We used 84 animals (43 cows and 41 buffalo) to find the frequency of the distribution of genotype. We did not find the change in nucleotide at the same identical position in buffalo-DGAT1 gene indicating that not all of the buffaloes (Chinese cross Murrah) have lysine at 232 positions, although about 25% of the cattle have lysine at 232 positions (online Supplementary Table S3).
Fig. 1. (A) Identification of the DGAT1 core promoter in cow. (B) Identification of the DGAT1 core promoter in buffalo. (C) Comparison of the identified functional promoter part of cow and buffalo DGAT1 Promoter. Bar represents the relative luciferase activities of the promoter fragments. Values are represented as Mean ± standard deviations. The error bars denote the standard deviation. The unpaired Student's t-test was used to detect significant differences. (D) SNP identification at 707 (G>A) and 708 (C>A) position respectively in cow DGAT1 sequence.
Construction of plasmid vectors, isolation of promoter region and its comparison in cow and buffalo
Based on the predicted positions of promoter regions and availability of suitable primers, 4 reporter vectors were constructed to test the core promoter activity in cow DGAT1 gene (viz. CV1, CV2, CV3, CV4) and 3 reporter vectors were constructed to test the core promoter activity in buffalo DGAT1 gene (viz. BV, BV2, BV3), (Fig. 2A). The transfection of the corresponding luciferase reporter plasmids into BMEC, 293T and CHO cells, and the results of these analyses show −93 to −556 bp was essential for cow-DGAT1and −84 to −590 bp was essential for the buffalo-DGAT1 to maintain the promoter activity (Fig. 1A, 1B). The sequence between −93 to −556 bp region in cow and −84 to −590 bp region in buffalo are highly similar and possess similar transcription binding factors including ATF, AF-2α, NF-1, C-jun, Sp-1 and C-Myb (Fig. 2B). The consensus sequences of ATF, AF-2α, NF-1, C-jun, Sp-1 and C-Myb were predicted by the JASPAR program (jaspar.genereg.net). Though the buffalo DGAT1 promoter activity was higher compared to the cow DGAT1 core promoter, the results were not significantly different. Nevertheless, this difference may go some way to explaining species differences in milk fat, such as the differences in milk fat globule size and composition shown by Ménard et al. (Reference Ménard, Ahmad, Rousseau, Briard-Bion, Gaucheron and Lopez2010).
Fig. 2. (A) Predication of the promoter sequence and designed of the primers and PCR amplified fragments of truncated DGAT1 promoter (cow and buffalo) after digestion by KnpI and BgIII endonuclease. (B) Sequence and identified TFB sites of the proximal minimal promoter of DGAT1 gene in cow (left) and buffalo (right). The Putative TFB sites are underlined and colored. http://alggen.lsi.upc.edu/recerca/menu_recerca.html and http://gene-regulation.com/cgi-bin/pub/programs/alibaba2/webbaba2.cgi software was used to predict the possible transcript factors binding sites.
In the mammary gland, the regular lipogenic processes produce TAG and the final stage in this synthetic pathway is catalyzed by DGAT1 (Cases et al., Reference Cases, Smith, Zheng, Myers, Lear, Sande, Novak, Collins, Welch, Lusis, Erickson and Farese1998). The genetic polymorphism K232A in the DGAT1 gene exhibits higher milk fat percentage for K allele, and our results revealed that all of the 43 buffalo were KK allelic, whilst 11 cows were KK allelic and 32 cows were AA allelic (online Supplementary Table S3). The previous report analyzed the effect of fat content on DGAT1 (K232A) genotype and found the DGAT1 (KK) allelic population has mean 4.95% of fat content while DGAT1 (AA) allelic population has mean 3.99% fat content in Dutch Holstein-Friesian cows (Argov-Argaman et al., Reference Argov-Argaman, Mida, Cohen, Visker and Hettinga2013). Comparison of the milk fat content between Jersey and Holstein revealed higher milk fat in Jersey cow (White et al., Reference White, Bertrand, Wade, Washburn, Green and Jenkins2001) and another report of allele distribution between Jersey and Holstein shows, KK (lysine) allele in most of Jersey and AA (alanine) allele in most of Holstein cows (Winter et al., Reference Winter, Krämer, Werner, Kollers, Kata, Durstewitz, Buitkamp, Womack, Thaller and Fries2002). When we interpret the results of these three studies we can conclude that breeds having KK (lysine) allele have higher fat content in the milk. Allelic determination of buffalo population was done in this study and all the buffaloes were KK allelic in genotype. We can propose that the absence of the alanine allelic population from buffalo explains the higher fat content of buffalo milk and, since segregation of alleles could be important for phenotypic characters in cow and buffalo (Freitas et al., Reference Freitas, de Camargo, Stafuzza, Aspilcueta-Borquis, Venturini, Dias, Cardoso and Tonhati2016b), it should be better incorporated to describe the difference in milk fat content.
To understand this phenomenon, the promoter analysis was done in our study. The prediction of the promoter regions and consequent construction of reporter gene to check the activity of promoter area was performed which revealed that the CV2 (−93 to −556) in cow DGAT1 was highly active and the region BV2 (−84 to −590) in buffalo-DGAT1 was highly active compared to other regions that could contain the core promoter. However, the available of similar transcription binding sites in both the expected core promoter regions (Fig. 2B) and the comparison of the strength of promoter activity of CV2 and BV2 failed to demonstrate a significant difference in these two promoters (Fig. 1C). The potential binding motif for the transcription factor Sp-1 is found in repeats in the DGAT1 gene which are termed as the variable number of tandem repeat (VNTR) and the DGAT1 VNTR also showed a strong association with milk fat percentage (Gautier et al., 2007). The allele with 2 repeats affected fat percentage favorably (Cardoso et al., Reference Cardoso, de Souza, Aspilcueta-Borquis, Neto, de Camargo, Hurtado-Lugo, Scalez, de Freitas, Albuquerque and Tonhati2015). Similar to the previous report we found the presence of the Sp-1 binding site (Fig. 2B) in the promoter region of both cow and buffalo sequence yet we did not find the repeats as mentioned in previous reports. This could be the reason for the non-significant difference between two identified promoters (for cow and buffalo). The causal agent to cause the K232A variation and activity of K allele for higher fat percent remains hidden and calls for future work.
In conclusion, we have demonstrated the K232A mutation as a candidate marker associated with milk fat. Since all the Chinese cross Murrah buffalo had the K allele and the majority (about 75%) of the Chinese Holstein cow had the A alleles it can be concluded that the presence of the lysine allelic population in buffalo explains the higher fat content of buffalo milk. However, the presence of identical promoter regions in cow and buffalo together with the observation of similar transcription factor binding sites suggests a new avenue for further research to elucidate the core reason behind the difference of DGAT1 expression, function and mechanism in cow and buffalo.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0022029920000126.
Acknowledgements
The authors are grateful for financial assistance from the research platform of the Ministry of Agriculture and International Exchange and Cooperation in Science and Technology and EUFP7 projects (PIIFR-GA-2012-912205 and FP7-KBBE-2013-7-613689).
