Association of the estrogen receptor gene Pvu II restriction polymorphism with expected progeny differences for reproductive and performance traits in swine herds in Brazil

Estrogen has an important function in swine reproduction and growth. A Pvu II restriction enzyme polymorphism has been proven to be an important genetic variation in the estrogen receptor gene (ESR) and may be considered as a candidate gene for use in pig production but there is no data regarding the prevalence of this polymorphism in the Brazilian pig population. We used DNA samples from the following three purebred pig breeds: Large White (336 females and 26 males), Landrace (304 females and 27 males) and Pietrain (125 females and 11 males). The ESR genotyping was performed using PCR-RFLP. For each breed, genotypes for the ESR gene were compared independently for expected progeny differences (EPD) in litter size (LS), average daily weight gain (DWG) (g/day) and back fat thickness (BT) as measured in mm by ultrasound. In the Large White breed, but not the other breeds, the ESR genotype was significantly (p < 0.05) associated to LS, DWG and BT. Large Whites genotyped as AA or AB had higher EPD values for the LS and BT traits compared to BB Large Whites, while AA Large Whites had higher DWG EPD values than BB Large Whites. Our results for the Large White population showed that the A allele has a beneficial effect on LS, DWG and BT expected progeny differences.


Introduction
Reproduction is an important area that influences swine production performance, but has low heritability (0.1-0.3) for reproductive characteristics (Holm et al., 2004;Rydhmer, 2000).This hampers direct selection of reproductive characteristics by conventional selection mating and this means that selection for litter size in pigs has variable results (Bolet et al., 1989).The genetic components of litter size have been studied and their application permits higher selection indices and consistent results on reproduction improvement (Bennett and Leymaster, 1990).Recently, mammalian genome sequencing projects, construction of gene maps and the discovery of gene polymorphisms by molecular genetics have been adding important information about gene control of physiological pathways of reproductive traits.
Estrogen has an important function on reproduction and growth, estradiol regulating the growth, differentiation and physiology of the reproductive process through the estrogen receptor.Rothschild et al. (1991) identified a Pvu II restriction enzyme polymorphism in the estrogen receptor gene in three families of Chinese breeds (Meishan, Fengjing and Minzhu).This polymorphism was associated to reproductive and performance traits in Meishan, Large White and Yorkshire breeds (Rothschild et al., 1996;Short et al., 1997).The polymorphism has been proven to be an important genetic variation in the estrogen receptor gene to be studied as a candidate gene for pig production, however, there is no data on the prevalence of this polymorphism in the Brazilian pig population.
The aim of this study was to assess the allele frequency of the estrogen receptor gene (ESR) Pvu II polymorphism and evaluate the effect of genotypes on the expected progeny differences (EPD) for litter size (LS), average daily weight gain (DWG) and back fat thickness (BT) in Brazilian Large White, Landrace and Pietrain herds.

Swine populations and laboratory analysis
Three purebred populations from a Brazilian pig farm (Granja Resende S/A Company, Brazil), were used in this study: Large White (336 females and 26 males), Landrace (304 females and 27 males) and Pietrain (125 females and 11 males).Blood samples were collected from each animal in sterile tubes containing EDTA anticoagulant and DNA was extracted as described by Borges (1997).A blood aliquot of 0.5 mL was added to 0.5 mL of lysis buffer (Tris-HCL 20 mM, sucrose 640 mM, MgCl 2 10 mM and 4% Triton X-100) in a microcentrifuge tube (2 mL) and the mixture centrifuged for 1 min at 4,000 x g, the leukocytes decanted and the pellet washed three times in the same buffer.After washing, 200-µL of lysis buffer (Tris-HCl 100 mM (pH 7.5) and 5M guanidine) was added to the nucleated white cell pellet and the mixture incubated at 60 °C for 30 min, after which the DNA in the supernatant was precipitated with a mixture of 100 µL of 150 mM Tris-HCL (pH 7.5), 1.2 M ammonium acetate and 300 µL of 100% isopropanol and then centrifuged for 5 min at 4,000 x g.After centrifugation the DNA pellet was washed twice in 1 mL of aqueous ethanol (70%), eluted in 300-500 µL of TE buffer (Tris 10mM; EDTA 1mM) and stored at -20 °C until used.
The estrogen receptor genotype was identified using the polymerase chain reaction -fragment length polymorphism (PCR-RFLP) method as reported by Short et al. (1997).The primers used for the PCR reactions were 5-CCTGTTTTTACAGTGACTTTTACAGAG-3 and 5-CACTTCGAGGGTCAGTCCAATTAG-3 and the PCR conditions were: 100ng of genomic DNA, 1x PCR buffer (Invitrogen), 1.5 mM MgCl 2 , 5 µM of each primer, 100 µM of each dNTP and 1U of Taq DNA polymerase (Invitrogen) in a final volume of 20 µL.The reactions were performed in a MJ Research PTC-100 thermocycler by heating the tubes at 95 °C for 5 min followed by 35 cycles of 95 °C for 45 s, 55 °C for 40 s and 72 °C for 45 s, followed by a final extension of 72 °C for 5 min.The amplicons (120 bp) were di-gested using the Proteus vulgaris Pvu II restriction enzyme at 37 °C for 15 h and then submitted to electrophoresis on 3.5% (w/v) agarose gel containing 50 µg mL -1 of ethidium bromide (Sambrook et al., 2001) and visualized under UV light.The genotypes, described in detail by Short et al. (1997) were AA (120bp), AB (120bp, 65bp, 55bp) and BB  (65bp, 55bp).

Statistical analysis
Population analysis was performed using the Genepop software (Raymond and Rousset, 1995) applying the default parameters.Genotypic and allelic frequencies were calculated independently for each breed (Weir and Cockerham, 1984) and allelic frequencies were compared between each breed pair using the Chi-squared (χ 2 ) test and applying a contingency table (Raymond and Rousset, 1995).
Expected progeny differences (EPD) for litter size (LS) for all parities, average daily weight gain (DWG , g d -1 ) between 63 and 140 days, and back fat thickness (BT) in mm obtained by ultrasound measurements were estimated and kindely provided by personnel at Granja Resende S/A which maintains a selection program based on the breeding values of their animals estimated by a complete animal model applying best linear unbiased prediction (BLUP) (Boldman et al., 1995).For each breed, ESR gene genotypes were independently compared for each trait using analysis of variance (ANOVA) and the EPD means were compared among the genotypes using Duncan's Multiple Range Test.Both tests were performed using the Statistical Analysis Software (SAS Institute Inc., 2001).The linear model fitted for ANOVA was: Y ij : µ + G i + e ij , where Y ij is the EPD for each trait of the ij th animal, µ is the general mean of each trait, G i is the fixed effect of the i th ESR (Pvu II) genotype and e ij is the random error effect associated to the ij th observation.Results were considered statistically significant when p < 0.05.

Results
Allele and genotypic frequencies calculated for each breed are presented in Table 1.The allelic distribution was different across the three breeds (p < 0.001).
Due to the fixed allele A in Pietrain animals, analysis of variance and the mean contrast test were performed only for Landrace and Large White data (Table 2).Although Landrace pigs genotyped as BB had higher EPD values than AA or AB animals for the LS, DWG and BT traits, the differences were not statistically significant (p > 0.05).However, in the Large White breed the estrogen receptor genotype was significantly associated (p < 0.05) with litter size, average daily weight gain and back fat thickness.Animals genotyped as AA or AB had higher EPD values compared to the BB animals for LS and BT traits, and AA animals had higher EPD values than BB animals for DWG (Table 2).

Discussion
The estrogen receptor gene has been studied as an important marker for genetic improvement for reproductive traits in swine breeding programs, especially in Large White, Meishan and Yorkshire purebreds or derived crossbreds.
There is no data reporting the ESR (Pvu II) restriction polymorphism in Pietrain swine pigs.In this study the A alleles were fixed in Pietrain, but not in the Landrace and Large White swine populations.
In a study of 91 German Landrace boars for the Pvu II polymorphism Drogemuller et al (2001) found that all the boars were AA homozygous, so the fact that the Brazilian Landrace population studied by us was polymorphic for the Pvu II site (AA, AB and BB) may be due to the different genetic background between populations.Despite the presence of the polymorphic Pvu II site in the Brazilian Landrace animals, there were no differences among ESR genotypes in the expected progeny differences (EPD) for litter size (LS), average daily weight gain (DWG) and back fat thickness (BT) traits.
Breeding programs have selected females for reproductive efficiency and the controlled matings to improve the number of pigs born may have interfered with the estrogen receptor (ESR) allele frequency because it seems to have a significant effect on litter size (Rothschild, 1996).
The ESR allele frequencies in the Brazilian Large White population was A = 0.68 and was different from that reported for other populations, i.e.A = 0.59 (Rothschild et al., 1996), A = 0.59 to 0.43 (Short et al., 1997) and A = 0.60 (Isler et al., 2002), although this difference was statistically significant (p = 0.004) only when our data were compared to the lower frequency (A = 0.43) reported by Short et al. (1997).
The ESR gene has been analyzed and associated to reproductive or performance traits, with Rothschild et al. (1996) having demonstrated that the B allele in Meishan and Large White synthetic lines had an additive effect for total number of pigs born and for the number of pigs born alive in the first parity and considering all parities.For Meishan and Large White, there were no genotype effects on performance traits (average daily gain, back fat thickness, and functional nipples).Short et al. (1997) studied the ESR polymorphism in four Large White-based commercial lines and found that the B allele had an additive effect for the total number of pigs born and number of pigs born alive in the first parity and in later parities and there were also additive effects for B allele for back fat and for test average daily feed consumed.In different Meishan x Large White F 2 populations analyzed by Gibson et al. (2002) no additive B allele effects were found for nine reproductive traits.However, in another study with a synthetic breed (50% Meishan/50% Landrace) van Rens et al. (2002) found that the AB genotype produced a higher total number of pigs born and number of pigs born alive than the BB genotype.
ESR polymorphism and swine traits 275  Isler et al. (2002) genotyped a population of Yorkshire, Large White and crossbred animals and found that the AA genotype had higher values for total weight of animals born (kg) and total weight of animals born alive (kg) than had the AB or BB genotypes.Goliasova and Wolf (2004) analyzed a Large White population for the ESR polymorphism and found that the AA genotype had better performance over all parities for number of piglets born alive, total number of piglets born per litter, number of piglets weaned per litter and litter weight at weaning.
All the published results were performed with real data collected from each analyzed trait.In our study we analyzed the expected progeny difference (EPD) estimated from real data and compared among ESR genotypes.The EPD is based on the estimated value of the additive genetic component of the analyzed trait; therefore, EPDs are effective estimates for additive genetic variation in the pigs (Boldman et al., 1995).In our study of the first Brazilian Large White pig population ever analyzed for the ESR Pvu II polymorphism the AA genotyped animals had statistically higher (p < 0.05) EPD values for litter size (LS), average daily weight gain (DWG) and back fat thickness (BT) in comparison to BB pigs from the same population (Table 2).Rothschild et al. (1996) have postulated that the influence of the ESR genotype on the different number of piglets born was associated with embryo survival, probably because of the important role of estrogen in the maternal recognition of the conceptus (Spencer and Bazer, 2004).However, pre-ovulatory hormones (luteinizing hormone (LH), estradiol and progesterone), number of embryos at day 35/36 of gestation, weight of embryos and viable embryos did not differ between AA and AB ESR genotypes in Meishan/Landrace animals (van Rens et al., 2000).Moreover, there was no significant B allele additive effect on ovulation rate in a Landrade/Large White population studied by Linville et al. (2001).These results produced by various authors suggest that the effect of the estrogen receptor gene could be associated to percentage fetal survival and/or may be also associated to the success of embryo implantation (Damario et al., 2001).
Analyses of the ESR locus have shown that there is a divergence between the beneficial allele that could be associated with the improvement over reproductive or performance traits in pigs.Rothschild et al. (1996) constructed a linkage map near the estrogen receptor gene and found that the ESR gene polymorphism seems to be responsible for the phenotypic variation in beneficial traits, suggesting that the genetic background of breeds might affect the control of gene expression.Whether the ESR gene is linked to other important major gene or whether it is the major gene associated to litter size and performance traits, and the mechanism by which this gene influences these traits is still unknown.
Our analysis in this Brazilian Large White population showed that the A allele is a favorable allele for expected progeny differences in litter size, average daily weight gain and back fat thickness.Although the average difference of -0.06 pig observed in the BB genotyped animals was statistically significant in the contrast test, the pig industry must study how cost-effective genotyping herds would be.Moreover, studies with a larger data set should be carried out to obtain information about the genetic background of Landrace and Large White swine breeds in Brazil in order to confirm whether or not the estrogen receptor gene Pvu II restriction polymorphism could act as a good marker to help improve reproductive and performance traits in Brazilian swine production.

Table 1 -
Allele and genotype frequencies for the estrogen receptor gene (ESR) in Landrace, Large White and Pietrain swine breeds.

Table 2 -
Number of observations, expected progeny difference (EPD) mean and standard deviation values for litter size (LS), daily weight gain (DWG) and back fat thickness (BT) traits in 362 Landrace and 331 Large White breeds genotyped for the estrogen receptor gene.
*LS: litter size, DWG: average daily weight gain (gram/day), BT: back fat thickness (mm).ab Means within a row with different superscript letters are different (p < 0.05).