healthmedicinet

Animal ethics statement

All animals were housed at an approved commercial AI center in Örnsro, Sweden. Semen
samples were collected by employees of the AI center as part of their regular breeding
and reproductive measures in cattle industry. Bulls with the tail stump sperm defect
were slaughtered because their semen was not suitable for artificial insemination.
The decision to slaughter the bulls was made solely by the owner (i.e., AI center) of the bulls. None of the authors of the present study was involved in
the decision to slaughter the bulls. Testicles of an affected bull were collected
after slaughter. Consent from the owner of the bulls was obtained to use the semen
and tissue samples for this study. No ethical approval was required for this study.

Animals

Three bulls of the Swedish Red cattle breed born between 2008 and 2012 with a sterilizing
tail stump sperm defect were included in the study together with 18 unaffected fertile
male half-sibs. The bulls were housed in an AI bull center in Örnsro, Sweden. The
age of the bulls during semen collection ranged from 11 to 16 months. Employees from
the AI center collected semen approximately twice a week as part of their regular
practice.

Sperm motility, morphology and testicular histology

We examined ten ejaculates per bull. Aliquots of fresh semen were put into vials to
measure sperm concentration using a photometric method and a haemocytometer (Bürker
chamber). A drop of semen (approximately 7 ?l was put on a pre-warmed slide to evaluate
sperm morphology. Head and sperm tail morphology of 200 spermatozoa was assessed from
slides stained with the Williams stain (bright field microscopy) and from a wet mount
formol-saline sample using a phase contrast microscope with 1000× magnification, respectively.
Moreover, sperm head morphology was assessed in dry smears stained with carbol fuchsin
according to Williams 35] and Lagerlöf 36]. Testicles from an affected bull were collected after slaughter. Histological specimens
were taken from the testicles, fixed in Bouin’s solution and embedded in paraffin.
Sections (5 ?m) were cut and stained with haematoxylin and eosin.

Genotyping of affected and unaffected animals

Twenty-one bulls (three affected, 18 unaffected) of the Swedish Red cattle breed were
genotyped using the Illumina BovineSNP50 Bead chip (Illumina, Inc., San Diego, CA,
USA). The chromosomal position of the SNPs corresponded to the UMD3.1 assembly of
the bovine genome 37]. Mitochondrial, X-chromosomal, Y-chromosomal SNPs and SNPs with unknown chromosomal
position were not considered for further analyses. After quality control (per SNP
and per individual call-rate higher than 90 %, no deviation from the Hardy-Weinberg
equilibrium (P??0.0001)), 46,035 SNPs were retained for further analyses. Beagle genetic analysis software 38] was used to impute sporadically missing genotypes and to infer haplotypes.

Homozygosity mapping

Segments of extended homozygosity were identified in three affected bulls using the
homozyg-function implemented in the whole genome association analysis toolset PLINK39], 40]. Due to the relatively sparse genome coverage of the genotype data (1 SNP per 56 kb),
we restricted our analysis to runs of homozygosity (ROH) with a minimum number of
20 contiguous homozygous SNPs and a minimum length of 500 kb.

Generation of sequence data

Genomic DNA of an affected bull was prepared from a semen sample following standard
protocols using proteinase K digestion and phenol-chloroform extraction. A gDNA sequencing
library with 420 bp insert size was prepared using the TruSeq DNA Sample Preparation
Kit (Illumina inc., San Diego, CA, USA). The sample was sequenced on an Illumina HiSeq2500
system using TruSeq SBS v3 chemistry (Illumina inc., San Diego, CA, USA) and the 2×100
bp paired-end read module. The fastq-files were generated with the CASAVA bcl2fastq conversion software (version 1.8.3, Illumina inc., San Diego, CA, USA). The alignment
of the reads to the University of Maryland reference sequence (UMD3.1, 37]) was performed with the Burrows-Wheeler Aligner41]. The resulting SAM file was converted into a BAM file with SAMtools42]. Duplicate reads were identified and marked with the MarkDuplicates command of Picard-tools 43].

Identification of candidate causal variants

Single nucleotide and short insertion and deletion polymorphisms were genotyped in
the affected bull together with 300 previously sequenced animals from eleven cattle
breeds (Gelbvieh (?=?12), Nordic Finncattle (?=?6), Fleckvieh (?=?153), Original Simmental (?=?15), Holstein-Friesian (?=?31), Brown Swiss (?=?50), Murnau-Werdenfelser (?=?2), Ayrshire (?=?2), Red-Holstein (?=?21), Original Braunvieh (?=?8)) other than Swedish Red 44] using the multi-sample approach implemented in the mpileup function of SAMtools42] and a variant calling pipeline as detailed by Jansen et al. 25]. Larger insertions and deletions and structural rearrangements were identified in
the affected animal and 226 sequenced control animals with an average genome coverage
above 8-fold using the Pindel software package 45]. To identify mutations compatible with recessive inheritance, all polymorphic sites
were filtered for variants that were homozygous for the alternate allele in the affected
bull and homozygous for the reference allele in 300 sequenced control animals. Candidate
causal variants were annotated using the Variant Effect Predictor tool 46], 47]. Additionally, sequence variants of 1147 animals from 29 breeds that were sequenced
for the 1000 bull genomes project 17] were analyzed to obtain genotypes of compatible variants in a larger cohort. The
animals of the 1000 bull genomes project were mostly influential sires that had been
widely used for artificial insemination.

Validation of the ss1815612719 polymorphism

PCR primers TTCAGTGCCAGGTTCATTGC and TTGGCTGGATGAGGTCAGTT were designed with Primer
3 48] to scrutinize the ss1815612719 polymorphism by Sanger sequencing in two affected
bulls and 97 unaffected artificial insemination bulls of the Swedish Red cattle breed.
DNA was extracted from semen samples following standard protocols using proteinase
K digestion and phenol-chloroform extraction. Genomic PCR products were sequenced
using a 3730×1 DNA Analyzer (Applied Biosystems) and data were analyzed with the Variant
Reporter v1.0 program (Applied Biosystems).

Bioinformatic analysis of ARMC3

The ARMC3 protein sequence was obtained from ensembl (ENSBTAT00000061467) and the
ClustalW2 tool 49] was used for multiple species alignment. The annotation of ARMC3 protein domains
was carried out using the Simple Modular Architecture Research Tool50].