Association of novel SNPs in gonadotropin genes with sperm quality traits of Boer goats and Boer crosses

ABSTRACT The present study aimed to investigate single-nucleotide polymorphisms (SNPs) of caprine FSHβ and caprine LHβ genes and to determine effects of the SNPs on fresh and post-thaw semen quality traits of Boer and Boer-crossbred bucks. The single-strand conformation polymorphism analysis and comparative sequencing revealed three SNPs in exon 3 of caprine follicle-stimulating hormone beta (FSHβ), including, three SNPs; 200A > G (FSHβ3-a), 226T > C (FSHβ3-b), and 237A > G (FSHβ3-c), while two of them (FSHβ3-a and FSHβ3-b) were novel. Furthermore, one SNP in exon 2 of luteinizing hormone beta (LHβ) (207T > C) (LHβ2) was detected. The associations of the four detected SNPs with quality traits of fresh and post-thaw semen were evaluated. Analyses of variance revealed significant association of the candidate genes with libido and semen quality traits. The three SNPs of FSHβ3 had significant effect on libido, progressive motility, and abnormality of fresh semen (P < .05), and on the motility, velocity, and viability traits of post-thaw semen (P < .05). LHβ2 polymorphism only showed association with sperm viability of post-thaw semen. The results of the present study suggest that the SNPs in caprine FSHβ and LHβ are associated with semen quality of male goats.


Introduction
Fertility of male goats is important in herd fertility and of importance in the goat industry. With the development of molecular marker techniques and marker assistant selection in animal breeding, the candidate gene method may identify markers that can predict sperm quality traits in male goats. Giesecke et al. (2010) elucidated that hormone and hormone receptors can be proper candidate genes for sperm quality traits due to their importance in male reproductive pathways. Spermatogenesis is controlled initially by the pituitary gland. Interaction among gonadotropin-releasing hormone, follicle-stimulating hormone (FSH), luteinizing hormone (LH), and testosterone are essential for spermatogenesis (Barth et al. 2008). There are some single-nucleotide polymorphisms (SNPs) in FSH and LH genes, which cause a lack of or deficiency in FSH or LH hormones that consequently cause defects in reproduction and fertility of men (Layman et al. 2002;Lee et al. 2003;Mafra et al. 2010).
FSH and LH are carbohydrate-containing proteins, and both of them are composed from two subunits: alpha and beta, which are controlled by two distinct genes. The alpha subunit is reported as common to all members of pituitary glycoprotein hormones and the beta subunit bestows biologic specificity (Fares 2006;Bernard et al. 2010). The FSHβ gene composed of three exons and two introns (Han & Miller 2009). FSH stimulates the germinal cells in the seminiferous tubules of the testis and spermatogenesis up to the secondary spermatocyte stage (Rassoulzadegan et al. 1993;Ohta et al. 2007). It was found that male homozygous FSHβ knockout mice showed normal levels of serum testosterone; they had small testes and oligospermia (Layman 2000). Xing et al. (2003) reported that sperm of FSHimmunized men showed considerable decrease in acrosomal glycoprotein and defective chromatin condensation. Quantitative trait loci (QTL) surveys showed there are QTLs on the 3rd, 8th, and 10th chromosomes of pigs and on the 5th chromosome of beef cattle, associated with concentration of serum FSH (Rohrer et al. 2001;Casas et al. 2004).
LH, which is produced and secreted in the pituitary gland, induces spermatogenesis by regulating the testosterone synthesis and secretion in Leydig cells of the testes (Stouffer 2003). Any abnormality in the function or level of LH hormone that is directly influenced by LH gene sequence and expression may affect spermatogenesis and cause infertility (Ramanujam et al. 2000). They revealed association of an SNP (1502G > A, which caused a change from glycine to serine) of the LHβ gene with male infertility. In addition, some SNPs of LHβ are associated with reproductive diseases, such as azoospermia, prostate cancer, and polycystic ovary syndrome, which cause poor fertility or sterility in men and women (Elkins et al. 2003;Lee et al. 2003;Mafra et al. 2010). The objectives of this study were to investigate SNPs of caprine FSHβ and LHβ genes and to evaluate the association of the SNPs on libido, and fresh and post-thaw semen quality traits of Boer and Boer-crossbred bucks.

Experimental animals, semen collection, and evaluation
Fifty-six mature bucks (38 pure South African Boer and 18 Australian Boer-feral crossbred) aged 2-3 years were used for the experiment. The bucks belonged to a big private farm in 30 km distance of Kuala Lumpur, Malaysia. Three samples of semen were collected from each buck at one-week intervals using artificial vagina, during June to December. Each sample was divided into two parts after measuring the semen volume (VOL). One part was used for evaluation of fresh semen and the other part was frozen in at least six straws; the straws were thawed for each evaluation after six months of freezing. The fresh semen quality traits, which included sperm concentration (SCON), sperm general motility (MOT), sperm progressive motility (PROG), live sperm percentage (LIVE), and abnormal sperm rate (ASR), were evaluated using a light microscope (Maina et al. 2006). Acrosome integrity (ACI) of sperm was evaluated for post-thaw semen. The staining was followed by the method of Lenz et al. (1983) with some modifications of Monterroso and Hansen (1999). The quality of the thawed frozen semen was evaluated using computer-assisted semen analysis (CASA) system (HTM-IVOS, Hamilton-Thorne Biosciences, Beverly, MA, USA). CASA yielded MOT, PROG, percentage of fast motile sperm (FAST), percentage of static sperm (STAT), average path velocity (VAP), straight linear velocity (VSL), curvilinear velocity (VCL), and lateral head displacement (ALH). A libido score system, ranging from 1 to 5, was devised based on the observation of libido behaviour of the bucks at each collection, with modification of the system used by Noran and Mukherjee (1997).

DNA extraction, polymerase chain reaction, and genotyping
Genomic DNA was isolated from either the blood or the semen sample of each animal using a QIAGEN blood and tissue DNA extraction kit. The primer for amplification 5 ′ -UTR and exon 3 of FSHβ gene was taken from Dai et al. (2009). Another two primer pairs were designed based on caprine LHβ sequence in NCBI (GenBank: AM258985.1). The primers and the fragment size are displayed in Table 1.
The PCR was programmed as follows: an initial denaturation step at 95°C for 5 min, followed by 35 cycles of 94°C for 30 s, the specific annealing temperature for 30 s, and 72°C for 60 s. A final extension step was performed at 72°C for 10 min. Electrophoresis of the amplicons was carried out in 1.5% agarose gels containing ethidium bromide in 1× TBE buffer and the gels were visualized under ultraviolet light.
For single-strand conformation polymorphism (SSCP) analysis, 5 µl of the PCR products were mixed with 10 µl of denaturing solution (98% formamide; 20 mM Ethylenediaminetetraacetic acid, pH 8.0; 0.05% bromophenol blue; 0.05% xylene cyanol), and the mixture was denatured at 95°C for 8 min, and chilled on ice for 10 min. Electrophoresis was carried out in non-denaturing 12% polyacrylamide gels in 1× TBE buffer at 4°C and 90 V for 12 h. The gels were subsequently stained using silver staining method (0.1% AgNO 3 ) and scanned using a densitometer (L 800, BIO-RAD). The DNA samples exhibiting different patterns on the SSCP gels were selected for sequencing. The PCR products were purified using PCR purification kits (Fermentas, Cambridge, UK) and sequenced (1st BASE Sequencing Services, Singapore). Nucleotide sequence alignments and comparisons were accomplished using the BioEdit 7.0.9.0 software (Ibis Biosciences, Carlsbad, CA, USA). The SNPs were screened using BioEdit 7.0.9.0 and BLAST (Basic Local Alignment Search Tool, National Library of Medicine, Bethesda, MD, USA) software. The effect of the SNPs on amino acid changes in related proteins was evaluated using ExPASy software (Swiss Institute of Bioinformatics, Geneva, Switzerland).

Statistical analysis
Mixed model analysis was used to test the effects of breed, age, season, genotype, and buck nested in breed, on the traits of the fresh semen, the thawed frozen semen at six months. In the model, the bucks were categorized into two groups based on their age (2-year-old group with animals between 22 and 28 months of age and 3-year-old group with animals between 34 and 40 months of age). The effect of season was not significant on any semen quality traits. Therefore, the effect of season was removed from the final model. Actually, Malaysia is located between the equator and the northern 7°latitude and it has a predominantly equatorial climate with ignorable variation of photoperiod between different months. The monthly mean temperature records in the farm obtained from Malaysian metrological department showed a small range of temperature between different months. Furthermore, the interaction of age and breed was significant only for libido score of bucks. The effects of individual SNPs were evaluated. The genotypes of all SNPs for each gene were considered in the model. The data were analysed using SAS 9.2 software. The following model was finally utilized for association analysis for fresh and post-thaw semen quality traits: where μ was the mean for each trait; A j , B k , and GEN l , were the effects of age, breed, and genotype, respectively; b was the random effect of the buck, which was nested in the breed effect; and e was the random error. Tukey's method was used for means separation.

Results
Two different primer pairs for caprine FSHβ gene were investigated for SNPs. No SNP was detected in FSHβU, which included the 5 ′ -UTR, exon 1, and part of intron 1 regions. However, three SNPs A200 > G (FSHβ3-a), T226 > C (FSHβ3-b), and A237 > G (FSHβ3-c) were detected in exon 3 of the FSHβ gene (GenBank: KF179313-KF179316). Figure 1 shows the polymorphic pattern and sequencing results of FSHβ3 gene. Analysis of the sequences of FSHβ3 of different bucks using the ExPASy software revealed that the FSHβ3-a and FSHβ3-c SNPs caused changes in the amino acid of the related protein 'serine to glycine' and 'glutamine to argenine', repectively, while FSHβ3-b was synonymous.
One SNP T207 > C (LHβ2) was found in exon 2 of LHβ gene in the bucks (Figure 2) (GenBank: KC442863-KC442864). Investigation for SNPs at FSHβU and LHβ3 showed that they were monomorphic. LHβ2 was a synonymous SNP and caused no change in the amino acid sequence of the relevant protein.
Of the four polymorphic loci studied, none showed all three possible genotypes. The highest allele frequencies of the investigated loci for overall population were 0.98 (C in LHβ2). No difference was observed in allele frequencies of FSHβ3-a between Boer and Boer crosses. In LHβ2, only the crosses were polymorph (Table 2).
Statistical analysis showed that the age and breed effects significantly influenced most of the traits, except libido, fresh motility, and some of post-thaw traits ( Table 3). The least square means and P-values of the effects of FSHβ3-a, FSHβ3-b, and FSHβ3-c loci on fresh and post-thaw semen quality traits are given in Table 4. Association analysis revealed that the FSHβ3-c locus significantly affected most of the fresh and post-thaw semen quality traits, including motility traits, sperm viability, and abnormality. However, FSHβ3-a and FSHβ3-b also showed association with some post-thaw semen quality traits, including ALH, BCF, and LIVE (P < .05).
Comparison of the analysis of variance results of individual locus and haplotype effects revealed that FSHβ3-c had the predominant effect on semen quality traits ( Figure 3). Four haplotypes were found for FSHβ3. The GGCTGA haplotype goats had superior semen quality. The goats with GGCTGA haplotype for FSHβ3 showed greater values for LIVE, ACI, motility traits in thawed frozen semen compared to those of other haplotypes (P < .05).
The statistical results of the effect of LHβ2 genotypes on semen quality traits are given as least square means and Pvalues in Table 4. The results showed that the transition mutation in position 207 of LHβ significantly affect LIVE and BCF of thawed frozen semen. The CC genotype goats showed significant superiority in LIVE and BCF traits by 33% and 14%, respectively, compared to those of CG bucks (P < .05).

Discussion
In the present study, most of the fresh and thawed frozen semen quality traits were significantly (P < .05) different between 2-and 3-year-old bucks, except for libido, fresh motility and some of the post-thaw traits. These differences could be probably due to two reasons. The two groups of the younger and the older bucks were fed the same amount of forage and concentrate. The 3year old bucks were heavier and consequently needed more energy, protein, and other nutrients for their maintenance compared to the younger bucks. It was reported that the quality and quantity level of feeding can affect sperm quality and mating behaviour of goats (Zarazaga et al. 2009). On the other hand, the bucks used in this study had never been used for breeding before. Therefore, the older and mature bucks may have been experiencing suppressed reproductive performance more than the younger bucks as they have not been used for mating or semen collection for a long period.
The significant effect of breed type on some fresh and postthaw semen traits, which is observed in this study can be explained by the differences in membrane lipid composition and differences in sperm cryotolerance after freezing in different breeds (Kaeoket et al. 2010).
In this study, no polymorphism was detected in the 5 ′ -UTR, exon 1, and part of intron 1 regions of the caprine FSHβ gene in the present study. Although one SNP in the 5 ′ -UTR region of the bovine FSHβ gene was previously found (Schlee et al. 1995;Dai et al. 2009;Yang et al. 2010), no SNP was reported in the exon 1 region of FSHβ gene. Similar to this study, Zhang et al. (2011) did not find any polymorphisms in exon 1 of FSHβ gene. The sequence results of caprine FSHβ3 gene revealed the amplified fragment to be 310 bp, while Dai et al. (2009) reported the amplicon as 313 bp in cattle. This could result from differences between two different species. Except    Note: FMOT: sperm motility of fresh semen; FPRG: sperm progressive motility of fresh semen; MOT: sperm motility; PRG: sperm progressive motility; RAPD: rapid sperm motility; STAT: static sperm; VAP: average path velocity; VSL: straight-line velocity; VCL: curvilinear velocity; ALH: the amplitude of lateral head displacement; BCF: beat cell frequency; LIVE: live sperm percentage; ASR: abnormal sperm rate; ACI: acrosome integrity of post-thaw semen. a,b,c means for a particular trait (column) not sharing any superscripts were significantly (P < .05) different.
FSHβ3-c, which was reported by Zhang et al. (2011), the SNPs of FSHβ3-a and FSHβ3-b were novel and never reported in goat before. Dai et al. (2009) also reported six SNPs in the same region of the bovine FSHβ gene and only the fourth and sixth SNPs reported were consistent with the FSHβ3-a (200A > G) and FSHβ3-c (237A > G) loci, respectively, in the present study. However, in contrast with the present result, they reported a substitution of cytosine (C) with adenine (A) in fourth SNP of the bovine FSHβ3 gene. FSHβ3-b (T226 > C) was not matched with other SNPs reported. These differences in the positions of SNPs and their frequency may be due to differences in species. Such differences in SNPs and their frequency in different species and/or breeds were reported in previous studies. Ishak et al. (2011) reported the FSHβ3-a SNP in five cattle breeds, Bali, Brahman, Simmental, Limousine, and Holstein. They observed Bali cattle to be monomorphic for this locus. Allele frequencies varied considerably in the different cattle breeds from exclude (Bali cattle) to 89% (Holstein). Therefore, the difference of the allele frequencies between the Boer bucks and Boer crosses may be due to differences in breed type. In addition, differences in allele frequencies and position of some SNPs between the bucks and those reported on bulls can result from differences in the sequence of caprine FSHβ and bovine FSHβ genes. However, similarly, the present study had not observed the CC genotype in FSHβ3-c.
In the present study, one of the homozygote genotypes for each of the three SNPs of FSHβ3 was not detected. This was probably because of the low frequency of these alleles. When the frequency of a particular allele is low, it may be observed mainly in the heterozygote individuals (Falconer & Mackay 1996). It could be that the homozygote genotype with low frequency might have been eliminated because of its extremely poor semen quality or sterility (Dai et al. 2009). Previous literature reported some mutations in the FSHβ gene that could cause azoospermia in homozygous men (Lindstedt et al. 1998;Layman et al. 2002;Simoni et al. 2016), although there is no report of association between SNPs of FSHΒ gene and sterility in Boer or Boer-crosses bucks. The low frequency of the favourable allele of FSHΒ3-c was probably due to the lack of selection for improving semen quality in the population.
Of the two amplicon of LHβ genes (LHβ2 and LHβ3), only one SNP was detected in LHβ2 (intron 1 and exon 2) at position 207T > C. In agreement with the present study, Yuan et al. (2010) did not find any SNP in exon 3 of LHβ in Hu sheep. This is the first attempt to investigate the relationship of the FSHβ loci polymorphism and semen quality traits in goats. Of the three detected FSHβ3 SNPs, FSHβ3-c was the most effective SNP on libido, and fresh and post-thaw semen quality traits. These results were similar to previous report of effect of FSHβ3 SNPs on semen quality of bulls (Dai et al. 2009). Allele A of FSHβ3-a with lower frequency was detected only as a heterozygote genotype. Allele G, as a favourable allele of FSHβ3-a, was associated with higher BCF and lower ALH in post-thaw semen. However, results of investigating the association between FSHβ SNPs and semen quality traits were different in different species. Lin et al. (2006) did not find association between FSHβ gene variations and sperm quality traits that were observed in swine population. Similar to our results, Dai et al. (2009) found a single mutation in the same position (FSHβ3-a) in cattle, but replacing A with C consequently caused the change Ser > Arg. Since variation in protein sequence can change the protein structure and function (Ng & Henikoff 2003), it is expected that the non-synonymous SNP in FSHβ gene consequently influence the phenotype. Similar to the present study, Dai et al. (2009) found that a heterozygote bull in FSHβ3-a had lower semen quality.
Allele C in FSHβ3-b, in higher frequency, was found as the favourable allele for the viability of the post-thaw sperm of the bucks. The transversion in FSHβ3-b was synonymous and may not affect the protein sequence. It was found that silent SNPs may also affect the relevant protein via change in transcription and may influence the accuracy or efficiency of splicing of mRNA or transcriptional control (Cartegni et al. 2002;Komar 2007). There are many reviews describing the potential effect of pre-mRNA splicing on phenotype of the traits (Nissim-Rafinia & Kerem 2002;Faustino & Cooper 2003;Ho et al. 2011). In addition, it has been explained that silent polymorphisms may affect the rate of translation because of changes in codon usage, during producing nascent protein (Kimchi-Sarfaty et al. 2007). Therefore, the influence of FSHβ3-b polymorphism on semen quality traits may occur due to its effect on the level of FSH hormone.
The SNP (237 A > G) in FSHβ3-c caused polymorphism of the FSHβ protein (115 Glu > Arg). SNP in the same position of FSHβ3 in cattle was reported as synonymous (Dai et al. 2009). Allele A was identified as the favourable allele for motility and viability of fresh semen and many post-thaw motility traits. This may be explained by influencing the non-synonymous SNP on altering single amino acids in the encoded proteins that may consequently affect the relevant traits. Investigation of the FSHβ3 haplotype effect on semen quality revealed that the effect of FSHβ3-c on semen quality was dominant compared to those of FSHβ3-a and FSHβ3-b. Based on sequence analysis, the human FSHβ gene is highly conserved, and non-synonymous SNPs, which cause amino acid changes, are extremely rare (Lamminen et al. 2005). Grigorova et al. (2010) suggested an association between the human FSHβ haplotype and fertility.
FSH as a gonadotropin hormone that directly affects Sertoli cells and regulates their functions, inducing expression of androgen-dependent Sertoli cells (Abel et al. 2009). Moreover, FSH stimulates the germinal cells and is responsible for spermatogenesis (Hafez & Hafez 2000). Ohta et al. (2007) elucidated that FSH is required for the initiation and maintenance of the quality and quantity in spermatogenesis. In addition, association between some of the single mutation of FSH with male fertility disorder in human (Layman et al. 2002) shows the role of normal FSH on fertility and semen quality. Layman (2000) also reported that FSHβ knockout mice showed a deficiency of sperm cells in their semen.
The allele C of the detected SNP in exon 2 of LHβ was the favourable allele for the LIVE and BCF of post-thaw semen quality traits. It was reported that mutations in the LH gene may influence the level of LH serum in men (Punab et al. 2015); LH affects Leydig cells, induces testosterone production, and indirectly controls spermatogenesis (McLachlan et al. 1996;França et al. 2005). Therefore, varying LH levels or function may influence semen quality. Ran et al. (2009) also found the same SNP in exon 2 of caprine LHβ gene, and determined that it was associated with litter size. The same SNP was also reported in the Hu breed sheep as associated with prolificacy (Yuan et al. 2010). Similarly, Cheng et al. (2016) found four SNP in intron 1 and exon 2 of water Buffalo's LH gene that were associated with ejaculate volume, sperm abnormalities, and sperm concentration. In contrast to the present study, Lin et al. (2006) did not find any association of LHβ SNP with semen quality traits of pigs. It may be due to differences among the species. The SNP in LHβ was a synonymous mutation. Nevertheless, it has been reported that silent mutations may affect gene expression and consequently affect the phenotype (Cartegni et al. 2002;Kimchi-Sarfaty et al. 2007).
Moreover, previous research has revealed associations among some SNPs of the LHβ gene and reproductive diseases, such as azoospermia, prostate cancer, and polycystic ovary syndrome, which cause poor fertility or sterility in men and women (Elkins et al. 2003;Lee et al. 2003;Mafra et al. 2010). Lower frequency of the T allele and the missing TT genotype for LHβ gene, as reported in the present study, suggests that bucks with this genotype might have been selected against naturally because of poor semen quality and low fertility, or even sterility.

Conclusion
Three SNPs in exon 3 of FSHβ3, and one SNP in exon 2 of the LHβ gene were found. All four SNPs showed association with quality traits of post-thaw semen. Specifically, allele A of FSHβ3-c was largely associated with higher post-thaw semen quality traits of the bucks. Considering the high incidence of associations between the candidate genes and post-thaw semen quality traits, using these markers in selection of bucks for an artificial insemination programme can improve the success rate of AI and reduce rearing of male animals, and consequently improve productivity of the herd. The genes information may also be used for early and more accurate selection of goats with higher genetic merits. Early selection of bucks would decrease the generation interval and increase the rate of genetic response.
In the present study, due to a limited source of Boer and Boercrosses bucks, the sample size was slightly small. Therefore, it is suggested that the effects of informative alleles at the candidate genes be validated using a larger sample size, with a larger pool of heterozygotes for the target SNPs, as well as goats from a different population, plus members of the future generation of the population investigated. It is also recommended that since the SNPs may affect the relevant proteins expression or efficiency, further studies be conducted to investigate the relationship between gene expression and protein expression for the SNPs of the identified genes. Studying the association of the SNPs with bucks' fertility also would be a reasonable next step.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by grants from the Ministry of Higher Education Malaysia [02-11-08-613FR].