Insulin-like growth factor 2 (IGF2) belongs to the family of structurally related polypeptides, which also include insulin-like growth factor 1 (IGF1), insulin and relaxin (Blundell & Humbel, Reference Blundell and Humbel1980; Dafgård et al. Reference Dafgård, Bajaj, Honegger, Pitts, Wood and Blundell1985). Both IGFs are produced mostly in liver and are secreted into the circulation, where they function as classical endocrine agents (Werner et al. Reference Werner, Adamo, Roberts and Leroith1994). In addition, the IGFs also employ autocrine and paracrine actions.
The mature IGF2 is a 67-amino-acid monomeric protein (O'Dell & Day, Reference O'Dell and Day1998). The IGF2 gene comprises several exons and multiple promoters and therefore has multiple transcripts (Schofield & Tate, Reference Schofield and Tate1987; Hedley et al. Reference Hedley, Dalin and Engström1989; Joujou-Sisic et al. Reference Joujou-Sisic, Granerus, Wetterling, Wikström, Engström, Jeffcott, Schofield and Welin1993; Bäcklin et al. Reference Bäcklin, Gessbo, Forsberg, Shokrai, Rozell and Engström1998).
The bovine IGF2 gene is localized on the telomeric end of the short arm of chromosome 29 (Schmutz et al. Reference Schmutz, Moker, Gallagher, Kappes and Womack1996; MacNeil & Grosz, Reference MacNeil and Grosz2002; Goodall & Schmutz, Reference Goodall and Schmutz2003). It consists of 10 exons and 3 promoters; exons 8, 9 and 10 are the coding region (Engström et al. Reference Engström, Shokrai, Otte, Granérus, Gessbo, Bierke, Madej, Sjölund and Ward1998; Goodall & Schmutz, Reference Goodall and Schmutz2007). Seven IGF2 transcripts are expressed in cattle in a tissue- and developmental stage-specific manner (Goodall & Schmutz, Reference Goodall and Schmutz2007), each containing the same coding exons 8–10 but different leader exons. The full sequence of the bovine INS/IGF2 gene cluster was recently published (Flisikowski & Fries, Reference Flisikowski and Fries2008; GenBank Acc. No. EU518675).
IGF2 is expressed in most bovine fetal tissues only from the paternal allele, with the maternal allele being transcriptionally silent (Dindot et al. Reference Dindot, Farin, Farin, Romano, Walker, Long and Piedrahita2004) but a bi-allelic expression driven by promoters 3 and 4 has been observed in the adult tissues (Curchoe et al. Reference Curchoe, Zhang, Bin, Zhang, Yang, Feng, O'Neill and Cindy2005).
Many experimental data indicate IGF2 as a local regulator of mammogenesis and lactogenesis. In lactating goats, local infusion of IGF2 increased milk synthesis (Prosser et al. Reference Prosser, Davis, Farr, Moore and Gluckman1994). Hovey et al. (Reference Hovey, Harris, Hadsell, Lee, Ormandy and Vonderhaar2003) provided evidence that the effect of prolactin on the mammary gland is mediated by locally secreted IGF2. IGF2 expression in mammary gland increases concurrently with the prolactin receptor and is strongly decreased in mammary tissues from prolactin receptor KO mice. Moreover, IGF2 stimulates alveolar development in mouse mammary gland in whole organ culture, and prolactin induces expression of IGF2 in mouse mammary gland explants, mammary HC11 cells as well as LUC gene transcription from IGF2 promoter 3 in transfected CHO cells in vitro (Hovey et al. Reference Hovey, Harris, Hadsell, Lee, Ormandy and Vonderhaar2003).
Human IGF2 gene polymorphism has been applied as a predictive marker for cancer risk and obesity (Heude et al. Reference Heude, Ong, Luben, Wareham and Sandhu2007; Suzuki et al. Reference Suzuki, Li, Dong, Hassan, Abbruzzese and Li2008). Effects of IGF2 gene polymorphism on meat traits were studied in pigs (Jeon et al. Reference Jeon, Carlborg, Törnsten, Giuffra, Amarger, Chardon, Andersson-Eklund, Andersson, Hansson, Lundstrom and Andersson1999; Nezer et al. Reference Nezer, Moreau, Brouwers, Coppieter, Detilleux, Hanset, Karim, Kvasz, Leroy and Georges1999; Van Laere et al. Reference Van Laere, Nguyen, Braunschweig, Nezer, Colette, Moreau, Archibald, Haley, Buys, Tally, Andersson, Georges and Andersson2003). Only a few nucleotide polymorphisms have been found in the bovine IGF2 gene and only in a few cases have their associations with production traits in cattle been studied. A g.292C>T [position in exon 2 (position 8656 relative to transcription start site)] transition (RFLP-BsrI) was reported by Goodall & Schmutz (Reference Goodall and Schmutz2003), associated with the rib eye area and the carcass fat per cent in beef cattle (Goodall & Schmutz, Reference Goodall and Schmutz2007). Zhao et al. (Reference Zhao, Davis and Hines2002) found a T/G transversion (RFLP-AciI) in exon 9 of the IGF2 gene in Angus cattle; animals with GG genotype tended to have higher body mass and daily weight gain. A single nucleotide ‘C’ deletion/insertion (InDel) polymorphism was found in IGF2 gene exon 6 in European and Zebu cattle and a G/T polymorphism (RFLP-MboII) only in Bos indicus subspecies by Flisikowski et al. (Reference Flisikowski, Adamowicz, Maj, Hiendleder, Pareek, Świtoński and Zwierzchowski2005). The InDel polymorphism appeared associated with the breeding value of Polish HF bulls (Flisikowski et al. Reference Flisikowski, Adamowicz, Strabel, Jankowski, Świtonski and Zwierzchowski2007). The estimated breeding value for protein yield and for milk yield appeared higher in animals without C nucleotide (the −/−genotype) than in C/C and C/− animals. Using combination of mouse mammary gland gene expression and comparative mapping, Ron et al. (Reference Ron, Israeli, Seroussi, Weller, Gregg, Shani and Medrano2007) found IGF2 to be a candidate gene for a QTL affecting milk production traits in cattle.
The aim of this study was to search for novel polymorphisms in the bovine IGF2 gene exon 10 and for associations between two single nucleotide polymorphisms (SNPs), the g.8656C>T in exon 2 (RFLP-BsrI) and g.24507G>T in exon 10 (RFLP-HaeIII), with milk performance of Polish Holstein-Friesian (HF) cows.
Material and Methods
Animal material and phenotype recording
The Polish Holstein-Friesian (HF) dairy cows (of Black-and-White type) were maintained at the Institute of Genetics and Animal Breeding experimental farm, in a herd with an average annual milk yield of 8300 kg. For the present study 238 cows were randomly chosen. The number of animals in the pedigree amounted to 1441. The cows were descendants of 68 sires; 5 cows were descendants of sires with only one daughter. During the whole period the cows were kept in a loose barn with outside run. During the test period animals were fed a complete total mixed ration (TMR) consisting of corn silage, wilted grass silage and concentrates, supplemented with mineral and vitamin mixture, according to the INRA system (Jarrige, Reference Jarrige1988). Water was available ad libitum. Milk samples were taken from each cow once a month, in the course of routine control milking during the whole lactation (lactations 1–5) for 4 years. The average number of lactations per cow was 3. Milk yield and content of the major milk components [fat, protein, lactose, total solids and the somatic cell count (SCC)] were collected. Fat, protein and lactose contents in milk samples were estimated in fresh milk using Milko Scan 104A/B (FOSS A/S, Hillerød, Denmark). Somatic cells were counted by means of Fossomatic (FOSS A/S) apparatus. SCC values were transformed to the natural log scale, LnSCC.
All procedures involving animals were performed in accordance with the Guiding Principals for the Care and Use of Research Animals and were approved by the Local Ethics Commission (Warsaw Agricultural University; Permissions No. 3/2005 and 23/2008).
DNA extraction and genotyping
Blood samples for DNA genotyping were collected from the jugular vein by an authorized veterinarian on K3EDTA (1·6 mg EDTA/ml blood) and stored at −25°C for a few weeks or at −80°C for up to several months. Isolation of DNA from whole blood was performed according to Kanai et al. (Reference Kanai, Fujii, Saito and Yokoyama1994).
The g.8656C>T [positions of mutations analysed in this study are given to the initiation of transcription site of the bovine IGF2 gene according to Flisikowski & Fries, Reference Flisikowski and Fries2008 (GenBank Acc. No. EU518675)] transition was genotyped using RFLP-BsrI with a modification of the method of Goodall & Schmutz (Reference Goodall and Schmutz2003). The 184-bp fragment, from nucleotide (nt) 21 to nt 204 of the bovine IGF2 gene exon 2 (GenBank Acc. No. AY237543) was PCR-amplified with primers: IGF2F – 5′-TTGCCTCCCAGTCAAGCCTG-3′ and IGF2R – 5′-GCTGTGTTGTCTCTGAAGCT-3′, designed using Primer3 program (http://frodo.wi.mit.edu/). PCR was performed in a reaction volume of 10 μl containing approximately 50 ng of bovine genomic DNA, 0·30 μm each primer, 0·2 mm-dNTPs, 2 mm-MgCl2 and 0·8 units of Taq polymerase (PolGen, Poland), 1 μl PCR polymerase buffer. PCR amplification cycles were: 94°C for 1 min, 64°C for 1 min, 72°C for 1 min (30 cycles). The reaction was carried out in MJ Research PTC-225 Thermal Cycler. The amplified DNA fragment was digested at 65°C for 3 h with 5 units of BsrI nuclease (New England Biolabs, Ipswich MA, USA).
The search for polymorphisms in the IGF2 gene exon 10 was carried out using DNA sequencing. DNA was extracted from blood samples of six unrelated HF cows. The 273-bp DNA fragment, from nt 327 to nt 599 of the bovine IGF2 gene exon 10 (GenBank Acc. No. DQ298740) was amplified using the following primers: IGF–F5 –5′-AATCCCTG-TACCGTCCTGTC-3′; IGF–R4 – 5′TTTGCTTTTCTGGTGTTGCT-3′. The sequence of primers was taken from Curchoe et al. (Reference Curchoe, Zhang, Bin, Zhang, Yang, Feng, O'Neill and Cindy2005). PCR was performed in a reaction volume of 50 μl containing approximately 100 ng of bovine genomic DNA, 0·30 μm each primer, 25 μl RedTaq Ready mix (Sigma-Aldrich Inc., St. Louis MO, USA) and H2O. The initial denaturation for 5 min at 94°C was followed by 34 cycles: 1 min at 94°C (denaturation), 1 min at 56°C (annealing) and 1 min at 72°C (elongation). After purification with the GenElute PCR DNA Purification Kit (Sigma), the PCR products were sequenced in an ABI377 sequencer (Applied Biosystems, Foster City CA, USA). Sequencing was done at Institute of Biochemistry and Biophysics, Polish Academy of Sciences (Warsaw). Sequences were analysed using the Chromas program (http://www.technelysium.com.au/chromas.html) and were compared with each other and with the sequence deposited in GenBank (Acc. No. DQ298740) using Blast Program (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Nine putative SNPs were found. The NebCutter 2.0 program (tools.neb.com/NEBcutter/index.php3) was used to find restriction endonucleases, enabling their further identification by the RFLP analysis.
Genotyping of cows at the newly found g.24507G>T transversion was carried out using RFLP-HaeIII. The 273-bp fragment was amplified as described above but in a reaction volume of 10 μl containing approx. 50 ng of bovine genomic DNA. The PCR-amplified DNA fragment was digested at 37°C for 3 h with 5 units of HaeIII nuclease (New England BioLabs, USA). Digestion products were separated on 2% agarose (GIBCO-BRL, Paisley, UK) gels in TRIS-borate-EDTA (TBE) buffer. DNA bands were stained with ethidium bromide, visualized and scanned in FX Molecular Imager (BioRad, Hercules CA, USA).
Statistical procedures
Differences in milk trait levels between IGF2 genotypes were checked using Duncan's test with Bonferroni's adjustment, and those between observed and expected frequencies were estimated according to the Chi-square formula test. Levels of significance were established at P⩽0·001 or at P⩽0·003, depending on number of tested hypotheses. The model included the animal IGF2 haplotypes or individual genotypes, year-season of calving interaction and parity as fixed effects and the animal additive genetic effect, permanent environmental effect of individual cows, and day of the test (with 87 levels) as random effects. Excepting the model for milk yield, for all other traits the fixed regression on milk yield was used.
To determine the effect of year-season, two seasons of calving were established (October–March and April–September). Each calving year between 2002 and 2007 was the separate subclass; there were 12 created classes of year-season of calving. According to parity, the animals were divided into three classes, with the class 3 covering parities >2. Legendre polynomials of standardized days-in-milk (days in lactation) (Brotherstone et al. Reference Brotherstone, White and Meyer2000) were fitted as fixed covariates within each parity subclass, in order to represent changes in considered traits due to the stage of lactation.
Results and Discussion
IGF2 gene polymorphism
Sequencing the 273-bp fragment of the IGF2 gene exon 10 (encoding 3′-UTR) of six HF cows revealed nine possible SNPs. The identity of five of them was confirmed by RFLP with the respective restriction endonucleases (Table 1). Four confirmed SNPs were found as a haplotype in two cows only (out of six genotyped). Therefore, only the g.24507G>T transversion (RFLP-HaeIII) appeared useful for further genotyping in a larger cohort of animals (Fig. 1).
Fig. 1. RFLP-PCR genotyping of two single-nucleotide polymorphisms in the bovine IGF2 gene. (A) g.8656C>T (RFLP-BsrI) in exon 2: M – 501 bp DNA marker (pUC19 cut with MspI, BTL, Poland); ND – non-digested 184 bp PCR product. Digestion of the 184-bp PCR product with the BsrI restriction endonuclease resulted in three DNA bands (118, 58 and 8 bp) for homozygote TT, four bands (176, 118, 58 and 8 bp) for heterozygote, and two bands (176 and 8 bp) for CC homozygote (8 bp band not seen on the electrophorogram). (B) g.24507G>T (RFLP-HaeIII) genotypes in exon 10: M – 11–1444 DNA marker (pUC19 cuts with HaeIII and TaqI, BTL, Poland); ND – non-digested 273 bp PCR product. Digestion with the HaeIII restriction endonuclease resulted in two DNA bands (233 and 40 bp) for homozygote TT, three bands (233, 190, and 40/43 bp) for heterozygote, and two bands (190 and 40/43 bp) for homozygote GG. GG/CC – the genotype with additional T/C transition at position 24524 (HaeIII cutting site); genotype GG and CC at positions 24507 and 24524, resulting in 176, 83 and 14 bp (14 bp band not seen).
Table 1. Single nucleotide polymorphisms found in exon 10 of the bovine IGF2 gene (273-bp fragment)
The g.24507G>T transversion is located in exon 10 of the bovine IGF2 gene at nucleotide position 408 (according to GenBank Acc. No. DQ298740), the same position where the A/G transition was previously reported by Curchoe et al. (Reference Curchoe, Zhang, Bin, Zhang, Yang, Feng, O'Neill and Cindy2005). In study of Curchoe et al., the SSCP method was used to identify IGF2 genotypes while, in the present work, the RFLP-HaeIII was used. This finding is surprising, since usually SNPs are biallelic, although it cannot be ruled out that at this position a triallelic polymorphism exists. We exclude the possibility of an experimental error since in both studies the DNA sequencing was used to confirm the nature of the mutation.
Two-hundred-and-thirty-eight Polish HF cows were genotyped at SNP g.24507G>T in exon 10 (RFLP-HaeIII) and SNP g.8656C>T in exon 2 (RFLP-BsrI). The genotype and allele frequencies of both SNPs and frequency of haplotypes are shown in Table 2. Distribution of genotypes followed the Hardy-Weinberg rule. C and T allele frequencies at g.8656C>T agrees with the results of Goodall & Schmutz (Reference Goodall and Schmutz2007) showing that Holstein cattle had almost equal frequency of the IGF2 alleles C and T. However, in other cattle breeds (Hereford, Simmental, Charolais and Angus) the frequency of allele T was low and varied from 0·02 to 0·28. According to Sherman et al. (Reference Sherman, Nkrumah, Murdoch, Li, Wang, Fu and Moore2008), the frequency of allele T was 0·19 and 0·16 in Angus and Charolais cattle, respectively.
Table 2. Distribution of genotypes, allele and haplotype frequencies in Polish Holstein-Friesian cattle in the bovine IGF2 gene
Out of nine theoretically possible combinations of individual genotypes (32=9), seven were identified in the studied group of animals (Table 2). Missing were haplotypes CC/TT and CT/TT; two haplotypes, CC/GT and TT/TT, appeared very rare. The present results suggest the existence of intragenic recombination in the bovine IGF2 locus.
Association of IGF2 gene polymorphism with milk performance
The association of IGF2 haplotypes, combination of genotypes g.8656C>T in exon 2 and g.24507G>T in exon 10, with dairy traits were estimated in 156 cows, out of 238 genotyped. In total, 4153 observations of their milk yield and composition were collected. For most of the traits under study differences were found between IGF2 haplotypes; the differences appeared significant at P⩽0·001 (Table 3). Only for lactose yield and content were no associations found. The lowest daily milk yield was associated with the CT/GT genotype while the highest milk yield was associated with the genotypes TT/GG and CT/GG. However, no association was shown between daily milk yield and any of the two SNPs analysed separately (Supplementary Tables 1 and 2). The lowest yield of milk constituents, fat, protein and total solids were shown for cows with the haplotype CT/GG. In concordance, the lowest protein content was shown in cows with the CT genotype at g.8656C>T SNP (Supplementary Table 1). Protein content (%) in milk was lowest in cows with the CT/GG haplotype indicating the negative influence of GG genotype at the g.24507G>T polymorphic site (Supplementary Table 2). The highest fat content was found in cows with haplotypes TT/GT and CT/GT, but no effect was shown for both SNPs when analysed separately. The CC/GG genotype was associated with the lowest SCC, in perfect accordance with the effect of CC and GG genotypes analysed separately.
Table 3. Estimates and their standard errors (se) for daily milk yield and major milk constituents as referred to haplotypes of the bovine IGF2 gene polymorphisms: g.8656C>T in exon 2 and g.24507G>T in exon 10Footnote †
† Number of milk samples/animals analysed. Haplotypes CC/GT (n=57/2) and TT/TT (n=18/1) were not included in calculations owing to their very low frequency
A,B,C,D,E Within lines estimates without a common superscript differ significantly at P⩽0·001
Supplementary Table 1. Estimates and their standard errors (se) for daily yield of milk and major milk constituents as referred to IGF2 g.8656C>T genotype in exon 2†
† Number of milk samples/animals analysed
A,B Within lines estimates without a common superscript differ at P⩽0·003
Supplementary Table 2. Estimates and their standard errors (se) for daily yield of milk and major milk constituents as referred to IGF2 g.24507G>T genotype in exon 10Footnote †
† Number of milk samples/animals analysed
A,B Within lines estimates without a common superscript differ at P⩽0·003
To sum up, our results showed that the IGF2 genotype may be predictive for milk yield and composition in Holstein-Friesian cattle. However, in some cases significant associations in haplotype analysis were not confirmed by the analysis of individual SNPs. One possible explanation is that dividing observations into a larger number of classes may lead to more accurate classification. As a result, the variation within groups became less than that between groups, and differences became significant. Similar situations occurred also in our other studies (e.g. Maj et al. Reference Maj, Zwierzchowski, Oprządek, Oprządek and Dymnicki2004). Anyhow, additional studies, performed on more numerous populations of cattle are necessary to confirm such conclusions.
The biological significance of the associations reported in this study is not clear. Both polymorphisms under study are located in un-translated regions: g.8656C>T in leader exon 2 encoding the 5′-UTR, g.24507G>T in exon 10 encoding the 3′-UTR, therefore, both mutations do not change the amino acid sequence of IGF2 protein. Translational control of gene expression may occur through specific cis-regulating sequences located in 5′- or 3′-untranslated regions of mRNAs (de Moor et al. Reference de Moor, Meijer and Lissenden2005; Pickering & Willis, Reference Pickering and Willis2005). However, there are no experimental data showing such regulation of IGF2 gene expression. Therefore, the g.8656C>T and g.24507G>T polymorphisms in the IGF2 gene are probably not causative to the traits under study and might be considered rather as genetic markers possibly linked to other polymorphism(s) located closely in the same or in another gene.
This study was funded by the Ministry of Science and Higher Education grants PBZ-KBN-113/P06/2005 and N N311 312235 and Institute of Genetics and Animal Breeding Polish Academy of Sciences Projects S.I.-2·1.
