A Single Nucleotide Polymorphism within the Interferon Gamma Receptor 2 Gene Perfectly Coincides with Polledness in Holstein Cattle

Polledness is a high impact trait in modern milk and beef production to meet the demands of animal welfare and work safety. Previous studies have mapped the polled-locus to the proximal region of the bovine chromosome 1 (BTA1) and narrowed it down to approximately 1 Mb. Sequencing of the positional candidate genes within the 1 Mb polled region and whole genome sequencing of Holsteins revealed a single nucleotide polymorphism (SNP) AC000158: g.1390292G>A within intron 3 of the interferon gamma receptor 2 gene (IFNGR2) in perfect co-segregation with polledness in Holsteins. This complete association was validated in 443 animals of the same breed. This SNP allows reliable genotyping of horned, heterozygous and homozygous polled Holsteins, even in animals that could not be resolved using the previously published haplotype for Holstein.


Introduction
Breeding genetically polled cattle offers many advantages for modern beef and milk production. Polled animals are usually easier to handle, safer to work with and are less aggressive among each other. The risk of injuries causing depression in milk production, insufficiency in beef-quality and faults in leather are decreased. In Germany, dehorning of calves without anaesthetics is allowed until the age of 6 weeks. Nevertheless, this procedure is connected with pain for the animal and can lead to extensive damage if there are complications. Due to animal welfare affairs, a prohibition of dehorning in Germany is foreseeable. Furthermore, dehorning is rendered unnecessary when breeding genetically polled cattle and thus, the farmer can save time and money.
As well as rumination and cloven hooves, pneumatised horns are a typical characteristic of the bovine species. However, the appearance of polled cattle centuries ago is proven by prehistoric carvings in Scotland from before recorded history and Egyptian sculptures and paintings from around 3000 before Christ [1].
A window of nine SNPs from 488 kb to 1,338 kb on BTA1 showed the greatest composite log likelihood (CLL) for the polled condition comparing a subset holding significant numbers of polled individuals to the breeds in the International Bovine HapMap study (IBHM). The SNP featuring the largest difference in allelic frequency between polled and horned animals was located within the interferon gamma receptor 2 gene (IFNGR2) [12].
Genotyping 285 animals of different cattle breeds with the bovine SNP50 BeadChip (Illumina) displayed a region of 381 kb on BTA1 where all 101 polled animals were homozygous for all genotyped SNPs [13]. This region harbours the three genes histone cluster 1, H4i like (HIST1H4C), oligodendrocyte transcription factor 1 (OLIG1) and chromosome 1 open reading frame, human C21orf62 (C1H21orf62) and the two pseudogenes oligodendrocyte transcription factor 2 like (OLIG2 like) and 40S ribosomal protein S13 like (LOC782947). Using the Agilent 44k bovine array, expression levels from tissuesamples of the horn-forming region of polled and horned animals were examined. Differences in relative expression levels among horned and polled cattle could not be found for the genes located in the 1-Mb-polled-region on BTA1 [14].
A 202-bp-indel located between the genes IFNAR2 and OLIG1 is highly associated with polledness in several European cattle breeds, but not in Holsteins [15]. For Holsteins only a haplotype of seven intergenic polymorphisms was detected spanning an interval of 260 kb including the genes HIST1H4C, OLIG1 and C1H21orf62 and the pseudogenes OLIG2 like and LOC782947 [15]. To our knowledge, no reliable single polymorphism differentiating polled and horned Holsteins has yet been found.

Mutation analysis and genotyping
Highest association with polledness in Holsteins was identified for polymorphisms within the interferon gamma receptor 2 gene (IFNGR2). In total, 41 SNPs and two indel mutations were detected within IFNGR2 by sequencing of PCR products (Table S1 Figure S1). Using next generation sequencing technique, further 18 SNPs and five indels were discovered within IFNGR2 (Table S2 Figure S1). The SNPs AC000158: g.1390505T>C, AC000158: g.1390528T>C, AC000158: g.1376880T>A, AC000158: g.1376884T>C, AC000158: g.1376932A>G and AC000158: g.1376977A>G were localised within the coding sequence of exon 3 and 7 of IFNGR2, respectively ( Figure 1). AC000158: g.1376520T>C was located within the 3' UTR ( Figure 1). All other polymorphisms were intronic. The indels AC000158: g. 1394768_771delCTGC and AC000158: g.1400851_852delCT were not observed in breeds other than Holstein. We genotyped all SNPs and indels with high association to polledness in further 11 homozygous polled, 61 heterozygous polled and 336 horned Holsteins. Although most Holsteins were born in Germany or male animals had a German herdbook number, their genetic origin was highly diverse ( Figure S2), as crossing with animals from different countries is common in Holstein breeding to improve milk yield. The ancestry is known for 82 of the 86 polled animals. Five of these animals were female. Two of these female Holsteins were born in Germany and Spain, respectively, and one female Holstein was born in the United Kingdom. Of the male Holsteins, 57% were born in Germany, 14% in the United States of America, 14% in the Netherlands, 5% in Canada, 4% in the Czech Republic, 3% in the United Kingdom and 1% in Spain and Belgium, respectively ( Figure S3). Of these animals, 87% have ancestors from other than their native country. Parents came from Germany, the United States of America, the Netherlands, Canada or France and grandparents from Germany, the United States of America, the Netherlands, Canada, Italy or Austria. This high diversity of ancestry assures representation of many important polled lineages including animals from Aggravation, Burket-Falls, Hickorymea, Dansire, Sandy-Valley, West Port Holsteins, Arron Doon Holsteins, Golden Oaks, Baldus Holsteins and Hollysprings.
Polymorphisms that showed association in all Holsteins were genotyped in further 21 homozygous polled, 43 heterozygous polled and 28 horned German Fleckvieh and five homozygous polled, nine heterozygous polled and eight horned Charolais.

Association analysis
The SNP AC000158: g.1390292G>A, within intron 3 showed a highly significant association with the polled genotype (χ ² =423, -log 10 P=91.85) ( Table 1 Figure 1) and perfect cosegregation in all 443 Holsteins. All horned Holsteins showed the genotype G/G, all heterozygous polled Holsteins the genotype G/A, and all homozygous polled Holsteins the genotype A/A. However, association could not be confirmed for the tested animals of the breeds Charolais and German Fleckvieh, as they had the genotype G/G in either case, whether they were horned or polled.
There was one progeny-proven pure-bred heterozygous polled Holstein sire used in artificial insemination with a polled mother and polled grandmother that was heterozygous for the SNP AC000158: g.1390292G>A. This animal carried neither the previously published 202-bp-indel associated with polledness in cattle [15] nor the polled-associated alleles from three tested markers (P5ID, PC1768587A and P80kbID) belonging to the previously published polled-associated haplotype of seven polymorphisms [15]. Thus, this sire would not have been identified as genetically polled without the SNP AC000158: g.1390292G>A. Furthermore, there was one homozygous polled pure-bred Holstein that was a compound heterozygote for the polled and horned alleles at both the seven polymorphism haplotype and 202 bp indel loci predictive of polled [15]. However, this animal was homozygous A/A for the SNP AC000158: g.1390292G>A, and was therefore identified reliably as homozygous polled by only genotyping this SNP.
The SNP AC000158: g.1385927G>A in intron 4 was also significantly associated with the horned and polled genotypes in the initially analysed animals (χ ² =13, -log 10 P=2.82). Association of this SNP could not be confirmed analysing an extended number of Holsteins. All three genotypes were found in the analysed homozygous polled animals. This SNP was thus not genotyped in all horned animals.
The indels AC000158: g.1394768_771delCTGC and AC000158: g.1400851_852delCT showed significant association (Table 1 Figure 1) but since they did not perfectly co-segregate with the horned and polled phenotype as did the SNP AC000158: g.1390292G>A, they were genotyped in all polled but not all of the horned animals.
The r 2 -values indicating correlation among mutated alleles showed that the three polymorphisms AC000158: g. 1390292G>A, AC000158: g.1394768_771delCTGC and AC000158: g.1400851_852delCT are in a strong linkage disequilibrium (LD) (Figure 2). Although these three mutations are located over 200 kb proximal on BTA1 to the three markers P5ID, PC1768587A and P80kbID [15], all six polymorphisms form a haplotype block of 519 kb length ( Figure 2).
IFNGR2 encodes the non-ligand-binding beta chain of the gamma interferon receptor which is involved in immunological pathways. There are currently no known functions of the IFNGR2 gene that could play a role in the growth of horns. The SNP AC000158: g.1390292G>A is unlikely to have a causal function for polledness in Holsteins, as it is located in a noncoding region of the IFNGR2 gene (intron 3) and effects on gene regulating elements were not found. The other associated polymorphisms (AC000158: g.1394768_771delCTGC and AC000158: g.1400851_852delCT) are also located within introns. The indel AC000158: g.1400851_852delCT in intron 1 was associated with polledness and disrupts a transcription factor binding site (Table 1) but the affected factor ICSBP has no distinguishable role in the development of horns and the distribution of alleles does not discriminate the polled genotype in the analysed animals. AC000158: g.1394768_771delCTGC and AC000158: g.1400851_852delCT, therefore, do not seem to have a causal function, either.
SNPs within exon 3 of IFNGR2 (AC000158: g.1390505T>C and AC000158: g.1390528T>C) led to amino acid changes but were predicted to be benign. Only a SNP AC000158: g. 1376880T>A causing an amino acid change from threonine to serine on cDNA position 338 was evaluated as probably damaging to protein function using PolyPhen-2 but failed the significance for association with polledness.
The SNP AC000158: g.1376520T>C within the 3' UTR disrupts a binding site for the factor AP-1 (Table 1), which is involved in cell transformation, inflammation, innate immune response and cell death. However, this SNP was not associated with polledness. AC000158: g.1390292G>A, AC000158: g. 1394768_771delCTGC and AC000158: g.1400851_852delCT are newly discovered polymorphisms located in the IFNGR2 gene. P5ID, PC1768587A and P80kbID represent three markers of the polled-haplotype previously published for Holstein [15]. LD coefficients (r 2 ) between pairs of polymorphisms are indicated as black letters in red fields. All pictured polymorphisms share one common haplotype block of 519 kb length.

IFNGR2 Polymorphism and Polledness
PLOS ONE | www.plosone.org within the encoded protein shows no significant associations with polledness in cattle. Thus, the coding sequence of IFNGR2 is unlikely to exert an influence on the development of horns. MicroRNAs located within the IFNGR2-gene were not found in the human and mouse database miRBase (http:// www.mirbase.org/). High conservation of microRNA-regions in various species is often observed [16]. Multiple sequence alignment including the species cow, mouse, human and sheep was performed using the program ClustalW2 (http:// www.ebi.ac.uk/Tools/msa/clustalw2/). For intron 3 of IFNGR2 harbouring the highly associated SNP AC000158: g. 1390292G>A identity-values from 59% between the species human and cow to 90% between the species cow and sheep were found. However, there could be mutations within currently unknown regulatory regions nearby or within the IFNGR2 gene or mutations within the 1-Mb-polled-region that interact with the IFNGR2 gene that may cause polledness in Holsteins. Thus, the IFNGR2 gene cannot be excluded as a potential candidate gene for polledness in cattle.
Nevertheless, the SNP AC000158: g.1390292G>A showed perfect co-segregation with polledness in Holsteins. Nearly 90% of the examined Holsteins have a multinational background with ancestors from at least one of 11 different countries. Despite of that their polled genotype could be determined reliably by analysing the SNP AC000158: g. 1390292G>A. The genotype of one heterozygous polled Holstein sire could be confirmed using this SNP, while this animal neither carried the previously published haplotype block nor the 202-bp-indel associated with polledness in cattle [15]. One homozygous polled Holstein that was a compound heterozygous for the haplotype block of seven polymorphisms and the 202-bp-indel [15] was revealed to be homozygous polled by the SNP AC000158: g.1390292G>A. Thus, the SNP AC000158: g.1390292G>A can be used as single polymorphism for testing of the polled genotype of pure-bred Holsteins, independent of their origin.
The IFNGR2 gene is located within the 1-Mb-polled-region but more than 200 kb proximal to the published homozygosity region [13]. The high association of the SNP AC000158: g. 1390292G>A within intron 3 of IFNGR2 indicates that the causal polled mutation is not necessarily located within this homozygous region of 381 kb. Performing homozygosity mapping, small associated regions can be missed due to the distribution of SNPs examined using the BovineSNP50 BeadChip (Illumina) if these small regions are located inbetween the genotyped SNPs. Furthermore, homozygosity mapping across breeds would not be successful if the polled mutation is non-allelic among the different polled breeds. The candidate region for polled therefore still contains the entire non-recombinant region of approximately 1 Mb between the microsatellites RP42-218J17_MS1 (661 kb) and BM6438 (2,025 kb) [9].

Conclusions
Herein, we demonstrate a single SNP as a useful and reliable marker for polledness in Holsteins, independent of their origin. We furthermore propose that the whole well-established candidate interval that showed no recombination in polled animals from 661 kb to 2,025 kb on BTA1 should be scanned to find the causal polled mutation.

Ethics statement
All animal work has been conducted according to the national and international guidelines for animal welfare. The Lower Saxony state veterinary office at the Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit, Oldenburg, Germany, was the responsible Institutional Animal Care and Use Committee (IACUC) for this specific study. The EDTA-blood and semen sampling for the present study had been approved by the IACUC of Lower Saxony, the state veterinary office at the Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit, Oldenburg, Germany (registration number 02A/101).

Animals, phenotypic data and DNA preparation
EDTA preserved blood samples of three male and one female homozygous polled, five male and five female heterozygous polled and two male and 19 female horned, not closely related pure-bred Holsteins were used for the detection of new polymorphisms. The genotypes of the homozygous polled animals were proved by progeny-testing. The heterozygous polled animals were each offspring of a polled sire and a horned dam.
For genotyping of polymorphisms with highest association with polledness we used further EDTA preserved blood samples and semen samples of 11 homozygous polled, 61 heterozygous and 336 horned unrelated Holsteins of highly diverse origin (Figures S2-S3).
In addition, we used EDTA preserved blood samples of 21 homozygous polled, 43 heterozygous polled and 28 horned German Fleckvieh and five homozygous polled, nine heterozygous polled and eight horned Charolais. Bovine genomic DNA was isolated from EDTA preserved blood using the NucleoSpin 96 Blood Quick Pure Kit (Macherey Nagel, Düren, Germany) or from EDTA preserved blood samples or semen samples following a standard protocol with ethanolprecipitation.
The amplicons were purified using the MinElute 96 UF Plate (Qiagen, Hilden, Germany) or ExoSAP-IT (Affymetrix, München, Germany) and directly sequenced with the DYEnamic ET Terminator Cycle Sequencing kit (GE Healthcare, Freiburg, Germany) on a MegaBACE 1000 capillary sequencer (GE Healthcare) or an ABI Prism 3500 capillary sequencer (Applied Biosystems, Darmstadt, Germany). Determined sequence data was analyzed using the Sequencher 4.8 program (GeneCodes, Ann Arbor, MI) and screened for polymorphisms.
Additionally, the genomes of one homozygous polled and three horned Holsteins and two DNA-pools, respectively, with six horned Holsteins each were sequenced using paired-end sequencing on the Illumina HiSeq 2000 (LGC Genomics, Berlin, Germany). On the average, an eight-fold coverage per sample was accomplished.
Indel mutations were amplified by using DY-682-tagged primers (Eurofins MWG Operon, Ebersberg, Germany). PCR reactions were performed in a total volume of 12 μl using 2 μl (~ 20 ng/µl) genomic DNA, 1.2 μl 10x PCR buffer, 0.4 μl DMSO, 0.2 μl dNTPs (10mM each), 0.5 μl (10 pmol/μl) of each primer and 0.2 μl Taq Polymerase (MP Biomedicals). The PCR conditions were as follows: 5 min initial denaturation at 94°C, followed by 36 cycles at 94°C for 30 sec, 58°C to 60°C (optimal annealing temperature according to the primer) for 1 min and extension at 72°C for 30 sec. The PCR was completed with a final cooling at 4°C for 10 min. Afterwards the PCR products were analysed by gel electrophoresis on the automated sequencer LI-COR 4300 (LI-COR, Bad Homburg, Germany) using 4% to 6% polyacrylamide denaturing gels (Rotiphorese Gel40, Carl Roth, Karlsruhe, Germany). Differences in allele sizes caused by deleted or inserted bases were defined by visual examination.

Statistical analyses
Association analysis with polledness in cattle was carried out for the detected polymorphisms. The markers have been tested for their polymorphism information content, Hardy-Weinberg-Equilibrium (HWE) and association with the polled phenotype and the polled genotype using the procedures ALLELE and CASECONTROL of SAS/Genetics, Version 9.3 (Statistical Analysis System, Cary, NC, 2012). Statistical calculation of pairwise linkage disequilibrium (LD) was performed and pictured using HAPLOVIEW 4.2 (http:// www.broad.mit.edu/mpg/haploview/). For detection of SNPs with strong LD among the alleles we used the tagger algorithm r 2 ≥0.8 [18]. .

In silico screening for binding sites of gene-regulating elements
Polymorphisms associated with polledness and mutations located within the UTR-regions were screened for effects on binding sites for gene regulating elements using AliBaba 2.1 (http://www.gene-regulation.com/pub/programs/alibaba2/ index.html).
SNPs localised in coding sequence were examined for causing amino acid changes. Discovered changes in amino acid order were analysed using PolyPhen-2 (http:// genetics.bwh.harvard.edu/pph2) with regard to their influence on protein function.
The expression of genes can also be regulated posttranscriptionally by microRNAs (small non-coding RNAs) [19]. In vertebrates 40-70% of microRNAs are expressed from introns [20]. The database miRBase (http://www.mirbase.org/) was therefore scanned for microRNAs contained inside the IFNGR2-gene in human and mouse. Figure S1. Gene structure of IFNGR2 with PCR-products and all polymorphisms. (a) Rectangles indicate exons. Coding sequence is filled with black, the untranslated region is filled with white. The size of exons and introns is specified in number of base pairs (bp). The black lines below illustrate the sections analysed in Holsteins. Each PCR-product is pictured separately with start and ending position. Position of start and stop codon is stated in bp. All polymorphisms in Holsteins are given with name and position. Each position is given with accordance to Bos taurus assembly UMD3.