Pure Holstein bulls were selected based on their fertility as previously described by Peddinti et al. [34
]. Briefly, the progeny test program from Alta Genetics Inc. (Alta Advantage®
program) involving approximately 180 farms milking an average of 850 cows each was used to select the bulls (Alta Genetics Inc; Calgary, Alberta, Canada). This program provides certain benefits, including DNA verification of the paternity of offspring and pregnancy diagnoses by veterinary palpation, instead of relying solely on non-return rates 60–90 days after breeding. This depth of data allows an accurate determination of both male and female fertility traits. In addition, this program provides unique research materials to identify molecular markers associated with fertility.
Definition of low- and high-fertility bulls
The fertility of bulls (Holstein), drawn from a total population of 874 bulls, is based on an average of 788 breeding, ranging from 101 to 11,997. We used the model described by Zwald et al. [35
], which takes in account the breeding event, environmental factors and herd management factors that influence fertility performance of sires (i.e., effects of herd-year-month, parity, cow, days in milk, sire proven status). All these factors were adjusted using threshold models similar to the previously published models [35
]. The fertility prediction of each bull was obtained using the Probit.F90 software [36
] and expressed as the percent deviation (Table ) of its conception from the average conception of all bulls. For the QTL analysis, the fertility was converted to a Z-score.
Oligonucleotide probe sequences for single-nucleotide polymorphism (SNP) markers
Isolation of pure sperm cells and DNA extraction
Alta Genetics Inc. (Watertown, WI) provided frozen semen straws of selected bulls. Thawed spermatozoa were then purified through a Percoll gradient, washed, counted, and pelleted for DNA isolation [34
]. DNA was extracted from a pool of three different ejaculates of spermatozoa using the Puregene DNA isolation kit (Qiagen, Valencia, CA), with minor modifications. Spermatozoa were homogenized in the lysis buffer (containing 60 mM DTT and 60 μg proteinase K), incubated for 60 minutes at 55°C, and treated with RNase A (12 μg). Proteins were sedimented and DNA was subsequently precipitated using isopropanol. DNA was washed in ethanol, dissolved in TE buffer (pH 8.0), and quantified using the NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies). DNA integrity was verified on an electrophoresis gel stained-agarose. DNA samples with high purity (A260/A280
≥ 1.8) and no degradation were used for the Phase I and Phase II studies.
SNP genotyping (Phase I study)
The 10 K SNP Bovine Gene Chip (Affymetrix/ParAllele GeneChip; Affymetrix Inc., Santa Clara, CA) was used to genotype DNA samples (250 ng/μl) of 10 low-fertility and 10 high-fertility bulls. The experiment was carried out at Baylor College of Medicine (Houston, TX), and assays utilized molecular inversion probe (MIP) technology, allowing the multiplex detection of single base variants using a 4-color array hybridization assay [37
]. The hybridization, washing, staining, and chip scanning procedures were performed using the standard protocol recommended by the manufacturer (Affymetrix Inc., Santa Clara, CA) of 9,919 SNPs analyzed in all DNA samples, the pass and completeness rates were 94.04% and 98.2%, respectively.
Allelic discrimination analysis (Phase II study)
The allelic discrimination technique, based on the TaqMan technology (ABI Prism 7000 Sequence Detection System, Applied Biosystems, Foster City, CA) was performed by scientists at SeqWright, Inc. seqwright.com on DNA samples (4 ng/μl) of 101 low-fertility and 100 high-fertility bulls to validate the Phase I association study. Primer and probe sets were designed on a sequence of 501 nucleotides containing the SNP (250 nucleotides downstream and upstream of the SNP allele) using the Primer Express software (Applied Biosystems, Foster City, CA). The probes were designed and labeled with FAM (6-carboxy-fluorescein) or VIC fluorescent dyes (Applied Biosystems, Foster City, CA) to match perfectly either one of the alleles (Table ).
PCR reactions were carried out in a total volume of 25 μl as recommended by the manufacturer. Each reaction consisted of 5 μl of sperm DNA in 20 μl Master Mix solution (Applied Biosystems, TaqMan Universal PCR Master Mix) containing primers (900 nM) and probes (200 nM). DNA samples were amplified by 40 times (2 min-50°C, 10 min-95°C, 15 sec-92°C, and 1 min 60°C). For each PCR run, negative (no-template) and positive (oligos) controls were added, and each run was preceded or followed by 1-minute incubation at 60°C to determine the background, or final levels of fluorescence, in each reaction. The base calls were made by examining all samples on the allelic discrimination graph, and the Sequence Detection Software (Applied Biosystems, Foster City, CA) was used to determine the homozygosis or heterozygosis of alleles. All samples were run in triplicates for each SNP allele.
The major repository of SNP data in the National Center for Biotechnology Information (NCBI) database (dbSNP), combined with Entrez Genome (NCBI) as well as the Ensembl automatic analysis pipeline, were used to characterize and annotate the markers (SNPs and genes) based on the cattle genome assembly version 3.1. For each SNP, a sequence of 501 nucleotides consisting of the 250 nucleotides immediately upstream and downstream of the SNP allele, were extracted and the best hit in the Bovine 3.1 genome identified using BLASTN. We considered only the hits (i) found on the same chromosome as the query, (ii) with E-values equal or close to 0.00 and (iii) presenting at least 94% alignment with the full length of the query sequence. Furthermore, the same query sequences were used to search for putative transcription factor binding sites through the Transfac-blastX (TRANSFAC Database 7.0 for searching eukaryotic transcription factors).
Functional analysis of integrin beta 5 (ITGB5)
Sperm preparation and treatment
Thawed spermatozoa were purified through a Percoll gradient as previously reported [39
]. Motile spermatozoa were resuspended in the fertilization medium containing Heparin and PHE. The sperm concentration was adjusted to 50 × 106
/ml and incubated with or without the integrin beta 5 antibody (ITGB5; sc-5401: 5 and 20 μg/ml). As a control for structural effects of the antibody, spermatozoa (50 × 106
/ml) were also incubated with a nonmammalian protein (BIT) antibody (sc-33757: 20 μg/ml). After a 2-hour incubation at 38°C under 5% CO2
in air, spermatozoa were washed twice in fertilization medium and used to fertilize the oocytes. The motility of spermatozoa was comparable before and after the incubation period.
Oocyte maturation and in vitro fertilization
Bovine oocytes were purchased from Bomed Inc. (Madison, WI), and maturation took place during the transportation. These oocytes were washed and fertilized with treated and untreated spermatozoa at a final concentration of 106 spermatozoa/ml. After 18 hours of co-incubation, oocytes were collected, denuded, washed, fixed, and placed on slides for staining with 0.1% aceto-orcein on slides. The nuclear status of oocytes was observed under a microscope, and the fertilized oocytes (two extruded polar bodies and/or two pronuclei) were counted to assess the fertilization rates.
In the Phase I study, the 20 bulls of varied fertility were typed for 9,919 SNP using Affymetrix 10 K Xba 142 2.0 array. A total of 1,712 markers fixed for a single genotype in this sample were excluded from further analysis. Markers were tested for Hardy-Weinberg Equilibrium using a 1 degree of freedom chi-square test. Markers with a p value < 0.05 were flagged as potentially out of HWE, but were used for association analysis.
The fertility of the bulls was converted to Z-scores and used as the quantitative trait for the association analysis. Analysis was conducted only at the markers' positions; no attempt at interval mapping was made. At each marker, the mean fertility, scored as a Z-score, was compared between genotypes using single marker regression. If the n for a genotype was 1, the samples were collapsed into the heterozygote. Benjamini- Hochberg FDR was calculated at each marker, adjusting 8,207 tests. A p-value less than 0.01 was set as the threshold of a significant association between the SNP marker and bull fertility; this corresponds to an FDR of 84%. Similar analyses were conducted in Phase II data.