Effects of ESR 1 , FSHB and RBP 4 genes on litter size in a Large White and a Landrace Herd *

The polymorphisms of ESR1, FSHB and RBP4 genes were detected by PCR-SSCP, PCR and PCR-RFLP in a Large White and a Landrace herd in Beijing, China and the influence of ESR1, FSHB and RBP4 on litter size traits was analyzed using three models. We found polymorphisms for the three genes in Large White besides for ESR1 and RBP4 genes in Landrace. The results showed that the most genotype effects are of ESR1 among these three genes.


Introduction
Litter size is one of the most important production traits in pig industry (ROTHSCHILD, 1998).In recent years, because of rapid progress of molecular biology and molecular genetics, research to detect major genes or molecular markers influencing litter size had became a hotspot, and had proved some major genes or molecular markers, including the estrogen receptor 1 (ESR1) (ROTHSCHILD et al., 1996;SHORT, 1997), the follicle-stimulating hormone beta subunit (FSHB) (ZHAO et al., 1999), the prolactin receptor (PRLR) (VINCENT et al., 1998) and the retinalbinding protein 4 (RBP4) (MESSER et al., 1996;DROEGEMUELLER et al., 1999;ROTHSCHILD et al., 2000).These investigations generally employed a candidate gene approach and studied chiefly effects of individual genes controlling reproduction.But how these genes influencing reproductive traits involve interaction effects and how much the impact of each beneficial genotype is led to no positive conclusion.The purpose of this study is to provide reliable information how more genes correlate with reproductive traits by dealing with these questions.

Animals
The breeds used in this study are French Large White and Landrace, all came from the first and the second seedstock farm of Beijing Huadu Swine Breeding Company, LTD.

Statistical Analysis
Data of litter size, which included total number born (TNB) and number born alive (NBA), were collected from the breeding farms.Three linear models were established to analyse the genotype effects of ESR1, FSHB and RBP4.Fixed effects involving herd-year-seasons, parity number, genotype and their interaction effects.Breed effect was also considered in these models, but the Large White and Landrace were analysed apart because of large difference between the two breeds.The following linear models were used: where, y ijkl, y ijklmn is the observed value; µ is the population mean value; HYS i is herdyear-season effects, which showed combined effects of breeding farms, years and seasons for farrowing; P j is parity number effects.For Landrace, it is the effects of records of all parities of a sow, and for Large White, a separate analysis was performed for the records of the first, the second and the third and above parities; G k is individual gene effects of ESR1, FSHB or RBP4; ESR1 k , FSHB l , RBP4 m is gene effects of ESR1, FSHB and RBP4, respectively; (ESR1×FSHB×RBP4) klm is interaction effects among ESR1, FSHB and RBP4 (including 2-way and 3-way interactions); e ijkl, e ijklmn is the random residual effect.GLM (General Linear Models) of SAS (8.2 version) was used for processing the data owing to unbalanced data.

Results of genotyping of ESR1, FSHB and RBP4
Banding patterns of ESR1 and FSHB were consistent with what were previously reported, But RBP4 was not entirely consistent with that of ROTHSCHILD et al (2000).Which can be classified into AA (190bp, 154bp and 136bp), BB (190bp,136bp and 108bp) and AB (190bp, 154bp, 136bp and 108bp

Polymorphism distributions of ESR1, FSHB and RBP4
In Landrace, the distributions of ESR1 and FSHB deviated much from each other, the A allele frequency of ESR1 was considerably higher than the other allele.On the contrary, the A allele frequency of FSHB was considerably lower than B allele, no homozygous AA animal was detected and only one heterozygous AB.The two alleles frequencies of RBP4 are close to each other (Table 1).In Large White, the allele frequencies of ESR1 and RBP4 are also close to each other.Frequency distribution of FSHB was similar to that of Landrace, most frequent genotype is also the BB genotype.The genotype frequency distributions at every locus were all in Hardy-Weinberg equilibrium, except for ESR1 in Landrace (using a chisquare test) (Table 1).Association of the ESR1, FSHB and RBP4 genotypes with litter size traits in Landrace In Landrace, we did not consider FSHB gene because of lack of polymorphism, the sample sizes are small again, and results of analysis showed the differences between different parities were not significant.Therefore, total parities were considered together.Genotype effects of ESR1, FSHB and RBP4 were analyzed with three models respectively, which were analysis of individual gene (model I), two genes included no interaction (model II) and two genes and their interaction (model III) (Table 3).The analysis result of model I indicated that least squares means of different genotypes for ESR1 had all the same trends, AB>AA, for TNB and NBA, and reached a significant level (P<0.05), the AB sows had an advantage of 1.08 pigs in TNB and of 1.9 pigs in NBA per litter across all parities over the BB sows, respectively; But no significant differences were found among the three RBP4 genotypes for litter size.
Using model II, difference between genotypes of ESR1 for TNB and NBA all reached a significant level, the AB sows had an advantage of 1.29 piglets in TNB and of 1.22 piglets in NBA per litter over the AA sows, respectively; However, genotype effects of RBP4 did not reach a significant level.Using model III, the sows with AB genotype of ESR1 had an advantage of 1.57 pigs in TNB and of 1.61 pigs in NBA per litter over the AA sows, respectively, and reached a significant and very significant level (P<0.01);The sows with BB genotype of RBP4 had an advantage of 1.60 pigs per litter over the AB, although the difference was not significant, BB>AA>AB, and showed the same trend in NBA, the BB sows had an advantage of 1.66 pigs per litter in NBA over the AB, and reached a significant level.The analysis of individual gene (model I) showed that only the genotype effects of ESR1 of these three genes reached a significant or very significant level at parity 3 and above.For TNB, BB>AA>AB, the sows with BB and AA genotype produced 0.79 and 0.53 piglets per litter more than those of AB genotype, respectively; the traits of NBA were similar to the TNB, the BB sows produced 0.75 piglets per litter more than those of AB genotype, but these differences among three genotypes were not significant.However, the genotype effects of FSHB and RBP4 were not significant.Analysis using model II, ESR1 genotype effects for TNB and NBA were significant at parity 3 and over, BB>AA>AB, the BB sows respectively produced 0.76 of TNB and 0.74 of NBA per litter more than those of AB; FSHB genotype effects for TNB reached a significant level at parity 2, the BB sows produced 1.13 piglets per litter more than those of AB; and the RBP4 genotype effects were not significant.Analysis using model III, effects differences between different genotypes for each gene were further widen, and the interaction effects between ESR1 and RBP4, among ESR1, FSHB and RBP4 reached a significant level at different parities (1,2 and ≥ 3).

Discussion
Previous research showed that favourable alleles of ESR1, FSHB and RBP4 are B, B and A, respectively (ROTHSCHILD et al., 1996;SHORT, 1997;ZHAO et al., 1997;ROTHSCHILD et al., 2000).In this study, favourable alleles of ESR1 and FSHB consist with what were reported previous, but favourable allele of RBP4 is B instead of A, this difference is probably result from different genotyping, which may as a result of mutation of MspI locus, but needs to further study.Differences of the genotype effects of each gene using different models were rather large (Table 3 and Table 4).In Landrace, the genotype effects of ESR1, of more significant difference, increased according to three models.For TNB and NBA, the sows with AB genotype respectively produced 1.08, 1.29, 1.57 and 1.09, 1.22, 1.61 piglets per litter more than those of AA genotype.Although the genotype effects of RBP4 were obviously lower than that of ESR1, the most effects were also observed using model III, additive effect of favourable allele B was 0.40 piglets in TNB and 0.45 piglets in NBA per litter, respectively, this result is close to that of OLLIVIER et al. (1997), but over that of ROTHSCHILD et al. (2000), while the genotype effects of RBP4 were not significant in the analysis using the former two models.The trends of varieties of ESR1, FSHB and RBP4 in Large White are the same as that of Landrace.The most genotype effects of ESR1 were observed upward of parity 3 and with model III, for TNB and NBA, the sows with BB genotype respectively produced 1.23 and 1.16 piglets per litter more than those of AB genotype.The genotype effects of the other two genes were not such significant, but the difference of genotype effects are also widen using model I, II and III.The current study had confirmed that there was interaction effects among genes that are significant associated with litter size traits, which will be further discussed in the second manuscript.Therefore, the interaction effects must be considered when performing analysis of effects of multi genes on the trait.The results of this study were showed the impact of parities on the litter size traits.No significant difference among the first six parities, but the difference is significantly between parity 7~9 and the other parities.This result disagrees with that of some researchers (CHEN et al., 2000;2001), who usually consider that the first parity have more significant effect on litter size than others parities.What needs pointing out is that the data recording of this experiment were up to parity 11, in contrast to other reports were usually no more than 7 parities on recorded.In addition, the effects of herd-year-season on the litter size traits were significant, so the effects of the parities and herd-year-season must be considered when performing statistical analysis.
Figure: Banding patterns of RBP4

Table 1
The polymorphous distributions of ESR1, FSHB and RBP4 and HW equilibrium detection in Landrace and Large White

Association of the ESR1, FSHB and RBP4 genotypes with litter size
Reproductive performance of Landrace and Large White Reproductive performances of different parities in Landrace and Large White are shown in Table2.We have observed that mean values of TNB and NBA of parity 3 and above were higher than the former two parities, mean values of total parities have all reached 11 piglets born and 10 piglets born alive and over, and Large White has produced nearly 1 more piglet than Landrace, it showed that reproductive performance of this study population was very well, especially of Large White.

Table 2
Reproductive performance of the different parities in Landrace and Large White (mean ±S.D.)

Table 3
Least squares means and standard errors for different genotypes of ESR1 and RBP4 for litter size traits with three models in Landrace

Table 4
Least squares means and standard errors for different genotypes of ESR1, FSHB and RBP4 for litter size traits with three models in Large White