SNP genotypes of olfactory receptor genes associated with olfactory ability in German Shepherd dogs

SHORT COMMUNICATION doi: 10.1111/age.12389 SNP genotypes of olfactory receptor genes associated with olfactory ability in German Shepherd dogs M. Yang*, G.-J. Geng, W. Zhang, L. Cui, H.-X. Zhang and J.-L. Zheng *Police-dog Technology Department, National Police University of China, Shenyang, Liaoning 110034, China. Technology Department, Shenyang Traffic Police Detachment, Shenyang, Liaoning 110001, China. Forensic Medicine Department, National Police University of China, Shenyang, Liaoning 110854, China. Document Inspection Department, National Police University of China, Shenyang, Liaoning 110854, China. Mark Inspection Department, National Police University of China, Shenyang, Liaoning 110854, China. Summary To find out the relationship between SNP genotypes of canine olfactory receptor genes and olfactory ability, 28 males and 20 females from German Shepherd dogs in police service were scored by odor detection tests and analyzed using the Beckman GenomeLab SNPstream. The representative 22 SNP loci from the exonic regions of 12 olfactory receptor genes were investigated, and three kinds of odor (human, ice drug and trinitrotoluene) were detected. The results showed that the SNP genotypes at the OR10H1-like:c.632C>T, OR10H1-like:c.770A>T, OR2K2-like:c.518G>A, OR4C11-like: c.511t>g and OR4C11-like:c.692G>A loci had a statistically significant effect on the scenting abilities (P < 0.001). The kind of odor influenced the performances of the dogs (P < 0.001). In addition, there were interactions between genotype and the kind of odor at the following loci: OR10H1-like:c.632C>T, OR10H1-like:c.770A>T, OR4C11-like:c.511T>G and OR4C11-like:c.692G>A (P < 0.001). The dogs with genotype CC at the OR10H1-like: c.632c>t, genotype AA at the OR10H1-like:c.770A>T, genotype TT at the OR4C11-like: c.511t>g and genotype at the OR4C11-like:c.692G>A loci did better at detecting the ice drug. We concluded that there was linkage between certain SNP genotypes and the olfactory ability of dogs and that SNP genotypes might be useful in determining dogs scenting potential. Keywords correlation, dog, scenting behavior, genetic markers Certain dog breeds, such as German Shepherd dogs, are very popular in the field of police work because of their sensitive nose. Because of the willingness of these dogs to cooperate with humans, police forces make use of this trait in dogs as tools for detecting a variety of odors such as explosives, drugs, humans and even molds (Lesniak et al. 2008; Pinc et al. 2011; Lippi and Cervellin 2012). Whether the olfactory sensation of dogs is acute or not is often evaluated through the use of many kinds of behavior tests, which have raised questions about the objectivity the tests in the absence of a gold standard (Doty and Kamath 2014). The spread of different phenotypes in dogs has resulted from restricting gene flow and artificially selecting offspring (Spady and Ostrander 2008; Akey et al. 2010; Address for correspondence M. Yang, Police-dog Technology Department, National Police University of China, 4 Baishan Road, Yuhong District, Shenyang, Liaoning 110034, China. E-mail: 1807645913@qq.com Accepted for publication 13 October 2015 Chen et al. 2012; Randi et al. 2014). Gene polymorphisms are associated with behavior of dogs in the fields of social behavior, scenting behavior, activity impulsivity, etc. (Hejjas et al. 2007; Kubinyi et al. 2012; Kis et al. 2014). Over half of the canine olfactory receptor (cor) genes were found to show high polymorphism, some of which were breed specific and some of which might affect individual olfactory perception (Keller and Vosshall 2008; Robin et al. 2009; Quignon et al. 2012; Kim et al. 2012). Therefore, the correlation between the genetic polymorphism of cor genes and olfactory ability was further studied. If a certain allele is likely to endow olfactory receptors with ligand-binding capacity, that might contribute to the olfaction sensitivity of dogs. Lesniak et al. (2008) verified that the specific alleles at two loci, cor9s13: c.592g>a and cor52n9:c.176a>g, were statistically significantly linked to odor recognition capabilities of dogs. However, it is possible that a certain polymorphism at a given locus does not fully determine a dog s odor sensing skills. Understanding the genetic mechanism of olfactory behavior traits could help breeders to breed dogs that cater 240

OR genotype associated with olfactory ability 241 to human requirements and create a genetic method of picking out excellent individuals. An experimental group of 48 German Shepherd dogs (males, n = 28; females, n = 20), 2 3 years old, were selected from the Police Dog Technology School of Ministry of Public Security, China. All the dogs had reached the standard for police dogs and were treated in a humane manner (Lesniak et al. 2008). Experienced handlers recorded the sniffing ability of each dog, which had been trained to detect specific kinds of odor, including the odors of human, ice drug and trinitrotoluene, and were in the last phase of training. The standards for scoring are shown in Table S1. A blood sample was obtained from each dog, and then DNA was isolated. Twenty-two representative SNP sites, of which five were breed specific in German Shepherd dogs, lay in exonic regions of 12 cor genes (Robin et al. 2009) (Table S2). Pairs of specific primers for 12 cor gene fraction amplification and 22 single extensive primers for SNP genotypes were designed using the PRIMER-BLAST and PRIMER PREMIER 5 programs. The primer sequences, fragment sizes and annealing temperatures (Tm) mentioned above are shown in Tables S3 and S4. Polymerase chain reaction (PCR) was conducted using a Veriti thermal cycler (Applied Biosystems, Inc.) in a total volume of 50 ll that included 4 ll of 100 ng of genomic DNA, 25 ll of Premix Taq, 1 lm (final concentration) of each specific primer and up to 50 ll of water. The following conditions for the reaction were applied: 10 s at 98 C, 30 s at 53 C or 58 C, 30 s at 72 C and 32 amplification cycles. PCR products were loaded onto a 2% agarose gel stained with GeneFinder (Biov), and DNA was extracted from the gel. SNP primer extension reaction was performed in a 5-ll reaction system that consisted 2 ll of Premix, 2 ll of the purified PCR products and 1 ll (1 lm) of primer, starting with 25 cycles of denaturation at 96 C for 10 s, annealing at 50 C for 10 s and extension at 72 C for 30 s, followed by an extra extension at 72 C for 5 min. A total of 4.25 ll of product digestion was initiated by adding 0.5 ll of109 buffer and 0.25 ll of1u/ll SAP (TaKaRa) for 1 h in a 37 C water bath and for 15 min in a 80 C water bath. One half microliter (0.5 ll) of the product was added to 0.5 ll of Size Standard 80 and 39 ll of sample loading solution. The capillary electrophoresis was analyzed on a Beckman GeXP Genetic Analysis System according to GenomeLab SNP-Start Primer Extension Kit (Beckman). Results were represented as mean values standard error for sniffing scores. Allele frequency, genotype frequency, H e (genetic heterozygosity) and PIC (polymorphism information content) were counted by EXCEL software. PROC- GLM of SAS 9.2 software followed by the Tukey test was applied to analyze the relationship. The interaction of genotype with the kind of odor on smelling ability was calculated according to the formula: Y ij ¼ l þ G i þ O j þ (GO) ij þ e ij ; where Y ij is the score of the smelling ability, l is the total mean values, G i is the major effect on genotype i level, E j is the major effect on odor j level, (GO) ij is the interaction of genotype with the kind of odor on smelling ability and Ɛ ij is the effect of random error. Overall, each SNP locus had one main effective allele (Table S5) with a frequency between 0.594 and 0.917. In particular, the OR2K2-like:c.72 > C/T, cor52e17: c.791g>a, OR2M5:c.98T>C, OR2M5:c.170C>T, OR2M5: c.862g>a and OR08H10:c.241G>A loci presented as homozygous, reflecting a unique genotype because some alleles are breed specific or rare in the dog population, whereas the other loci presented two or three genotypes (Robin et al. 2009; Kim et al. 2012). According to the allele frequencies, genetic disequilibrium was found at the OR52J3:c.469G>A, OR2K2-like:c.518G>A and OR4C11- like:c.256a>g loci (v 2 > 9.210, P < 0.01). On the contrary, the 13 other loci were in Hardy Weinberg equilibrium (v 2 < 5.991, P > 0.05). Meanwhile, 16 loci showed low polymorphism levels (PIC<0.25), and six loci showed a medium level of polymorphism (0.25 < PIC < 0.50) (Table 1). This phenomenon might be a result of the recruitment system used for police dogs. We also observed that the olfactory sensitivity of dogs varied with certain genotypes. The biggest difference occurred relating to the OR4C11-like:c.511T>G and OR4C11-like:c.692G>A loci, as dogs with a TT and genotype respectively were scored at 26.85 for the detection Table 1 Genetic polymorphism analysis of SNPs. Num Loci v 2 H e PIC 1 OR10H1-like:c.632C>T 0.087 0.202 0.183 2 OR10H1-like:c.770A>T 0.087 0.202 0.183 3 OR52J3:c.469G>A 42.013 0.278 0.239 4 OR2K2-like:c.172>C/T 0 0 5 OR2K2-like:c.518G>A 41.245 0.249 0.217 6 OR8S1:c.658T>C 1.732 0.329 0.275 7 OR1L4:c.214G>C 0.167 0.457 0.353 8 OR51I2-like:c.710T>G 0.034 0.186 0.169 9 OR51I2-like:c.770T>C 0.034 0.186 0.169 10 OR51I2-like:c.778T>A 0.034 0.186 0.169 11 OR49-like:c.147A>G 0.088 0.152 0.141 12 cor52e17:c.214g>a 1.732 0.329 0.275 13 cor52e17:c.791g>a 0 0 14 OR2M5:c.98T>C 0 0 15 OR2M5:c.170C>T 0 0 16 OR2M5:c.862G>A 0 0 17 OR52J3-like:c.106G>A 2.320 0.482 0.366 18 OR4C11-like:c256A>G 36.700 0.152 0.141 19 OR4C11-like:c.511T>G 2.131 0.469 0.359 20 OR4C11-like 692G>A 2.131 0.469 0.359 21 OR08H10:c.241G>A 0 0 22 OR08H10:c.927G>A 0.472 0.249 0.218 PIC, polymorphism information content.

242 Yang et al. Table 2 Scores of smelling behavior of dogs with different SNP genotypes. SNP Odor Genotype (l SE) OR10H1-like:c.632C>T CC CT Human 29.64 0.07 29.85 0.11 Ice drug 26.88 0.45 16.88 0.16 Trinitrotoluene 17.68 0.43 17.95 1.17 OR10H1-like:c.770A>T AA AT Human 29.64 0.07 29.85 0.11 Ice drug 26.88 0.45 16.88 0.16 Trinitrotoluene 17.68 0.43 17.95 1.17 OR52J3:c.469G>A AA Human 29.33 0.05 29.43 0.10 Ice drug 25.58 2.17 20.60 0.60 Trinitrotoluene 18.43 0.94 17.15 0.21 OR2K2-like:c.172 > C/T CC Human 29.10 0.19 Ice drug 23.75 0.84 Trinitrotoluene 18.58 0.37 OR2K2-like:c.18G>A AA Human 28.43 1.11 29.57 0.04 Ice drug 19.13 2.28 25.78 0.60 Trinitrotoluene 15.75 1.10 18.35 0.16 OR8S1:c.658T>C TT TC CC Human 29.76 0.05 29.55 0.26 25.95 0.14 Ice drug 22.80 0.87 24.10 2.29 21.00 0.34 Trinitrotoluene 17.51 0.23 17.63 0.10 13.2 0.79 OR1L4:c.14G>C GC CC Human 28.95 0.39 28.80 0.34 29.93 0.04 Ice drug 27.6 0.12 20.31 1.12 25.88 1.63 Trinitrotoluene 17.75 0.26 17.22 0.48 16.24 0.37 OR51I2-like:c.710T>G TT TG OR51I2-like:c.70T>C TT TC OR51I2-like:c.778T>A AA AT OR49-like:c.147A>G AA AG Human 29.30 0.13 29.85 0.11 Ice drug 22.95 0.62 23.10 0.88 Trinitrotoluene 17.18 0.21 17.20 0.07 cor52e17:c.214g>a GA AA Human 29.59 0.06 28.65 0.78 30.00 0.15 Ice drug 21.68 0.89 24.60 1.44 28.50 0.68 Trinitrotoluene 17.71 0.22 15.60 0.70 17.70 0.23 cor52e17:c.791g>a OR2M5:c.98T>C TT OR2M5:c.170C>T CC

OR genotype associated with olfactory ability 243 Table 2 (Continued) SNP Odor Genotype (l SE) OR2M5:c.862G>A OR52J3-like:c.106G>A GA AA Human 30.00 0.69 29.58 0.10 28.65 0.78 Ice drug 29.40 0.22 20.60 1.06 25.80 1.39 Trinitrotoluene 16.50 0.41 18.22 0.19 14.98 0.80 OR4C11-like:c.256A>G AA Human 29.33 0.13 29.55 0.33 Ice drug 24.62 0.56 15.30 0.61 Trinitrotoluene 16.83 0.19 18.30 0.11 OR4C11-like:c.511T>G TT TG Human 29.74 0.13 29.05 0.27 29.55 0.38 Ice drug 26.85 0.58 23.13 1.14 15.30 0.62 Trinitrotoluene 16.65 0.44 16.94 0.35 18.30 0.23 OR4C11-like:c.692G>A GA AA Human 29.74 0.13 29.05 0.27 29.55 0.38 Ice drug 26.85 0.58 23.13 1.14 15.30 0.62 Trinitrotoluene 16.65 0.44 16.94 0.35 18.30 0.23 OR08H10:c.241G>A OR08H10:c.927G>A GA Human 29.21 0.22 29.65 0.20 Ice drug 23.31 0.95 22.40 1.92 Trinitrotoluene 16.94 0.36 17.38 0.11 of ice drug, whereas those with a and AA genotype respectively were scored at 15.30 (Table 2). To distinguish the dogs olfactory acuity, the relationship between genotypes and phenotypes was studied further. The genotypes at the OR10H1-like:c.632C>T, OR10H1-like:c.770A>T, OR2K2-like:c.518G>A, OR4C11-like:c.511T>G and OR4C11-like:c.692G>A loci had a statistically significant relationship with the scenting ability of police dogs (P < 0.01) (Table S6). Notably, three of five breed-specific SNP loci showed this relationship in contrast to only two of 17 non-breed-specific SNP sites. This verified again that a breed-specific SNP usually gives the strongest association signal for susceptibility to a certain phenotype (Olsson et al. 2011). In particular, the mean values of smelling scores of individuals homozygous for the main effective alleles, heterozygous individuals and those homozygous for the non-main effective alleles varied from high to low, respectively, at the five loci (Table 2). These finding are in accordance with the findings of Lesniak et al. (2008). We again proved that genetic differences in olfactory abilities existed in individuals and that the polymorphisms of cor genes might play an important role in olfactory sensitivity. The allelic transition at these loci were associated with the difference in scores for individuals. The genotype frequencies revealed that individuals homozygous for the main effective alleles comprised the majority of the analyzed group of dogs: 77.1% at the OR10H1-like:c.632C>T and OR10H1-like:c.770A>T loci and 85.4% at the OR2K2-like: c.518g>a locus. This phenomenon might be because the ratings of dogs smelling behavior made by scientists and experienced dog trainers were very similar (Lesniak et al. 2008). However, there was a different situation for the OR4C11-like:c.511T>G and OR10H1-like:c.632C>T loci, where heterozygous individuals represented 58.3% of the total. Maybe experienced dog trainers were incapable of selecting acute sniffing dogs that acted almost as well as or as bad as others. In addition, the kind of odor could have influenced the performances of the dogs (P < 0.01) (Table S6). All the dogs displayed behavioral responses to the odor of human more quickly. As one of the mammalian prey species, dogs may have been more sensitive to a human odor compared to non-prey-associated odors (Nilsson et al. 2014). Furthermore, the interaction between certain genotypes and a certain odor was found for the OR10H1-like:c.632C>T, OR10H1-like:c.770A>T, OR4C11-like:c.511T>G and OR4C11-like:c.692G>A loci (P <0.01) (Table S6). Dogs homozygous for the main effective alleles at those four loci obtained higher scores in detecting ice drug (Fig. 1). Thus, people could utilize this interaction to train German Shepherd dogs with preferable genotypes to carry out an assignment to detect ice drug.

244 Yang et al. (a) (b) Figure 1 The interaction of a genotype and a kind of odor on olfactory abilities of dogs. (a) Interaction with the OR10H1-like:c.632C>T and OR10H1-like:c.770A>T loci. (b) Interaction with the OR4C11-like: c.511t>g and OR4C11-like:c.692G>A loci. In conclusion, there was a correlation between SNP genotypes of cor genes and olfactory ability of dogs. The scenting ability might differ according to the detection of different kinds of odors. Dogs with a certain genotype might be more competent in detecting a specific odor. These preliminary results showed that molecular genetic studies on cor genes might be a valuable tool to improve the selection of sniffing dogs and its working fields. Acknowledgement This work was supported by the Application and Innovation Foundation of Ministry of Public Security of the People s Republic of China (2011YYCXSYJQ165). References Akey J.M., Ruhe A.L., Akey D.T. et al. (2010) Tracking footprints of artificial selection in the dog genome. Proceedings of the National Academy of Sciences of the United States of America 107, 1160 5. Chen R., Irwin D.M., and Zhang Y.P. (2012) Differences in selection drive olfactory receptor genes in different directions in dogs and wolf. Mol Biol Evol, 29, 3475 84. Doty R.L. & Kamath V. (2014) The influences of age on olfaction: a review. Frontiers in Psychology 5, 20. Hejjas K., Vas J., Topal J. et al. (2007) Association of polymorphisms in the dopamine D4 receptor gene and the activity-impulsivity endophenotype in dogs. Animal Genetics 38, 629 33. Keller A. & Vosshall L.B. (2008) Better smelling through genetics: mammalian odor perception. Current Opinion in Neurobiology 18, 364 9. Kis A., Bence M., Lakatos G. et al. (2014) Oxytocin receptor gene polymorphisms are associated with human directed social behavior in dogs (Canis familiaris). PLoS ONE 9, e83993. Kim R.N., Kim D.S., Choi S.H. et al. (2012) Genome analysis of the domestic dog (Korean Jindo) by massively parallel sequencing. DNA Research 19, 275 87. Kubinyi E., Vas J., Hejjas K. et al. (2012) Polymorphism in the tyrosine hydroxylase (TH) gene is associated with activityimpulsivity in German Shepherd dogs. PLoS ONE 7, e30271. Lesniak A., Walczak M., Jezierski T. et al. (2008) Canine olfactory receptor gene polymorphism and its relation to odor detection performance by sniffer dogs. Journal of Heredity 99, 518 27. Lippi G. & Cervellin G. (2012) Canine olfactory detection of cancer versus laboratory testing: myth or opportunity? Clinical Chemistry and Laboratory Medicine 50, 435 9. Nilsson S., Sj oberg J., Amundin M. et al. (2014) Behavioral responses to mammalian blood odor and a blood odor component in four species of large carnivores. PLoS ONE 9, e112694. Olsson M., Meadows J.R., Truve K. et al. (2011) A novel unstable duplication upstream of HAS2 predisposes to a breed-defining skin phenotype and a periodic fever syndrome in Chinese Shar- Pei dogs. PLoS Genetics 7, e1001332. Pinc L., Bartoš L., Reslová A. et al. (2011) Dogs Discriminate Identical Twins. PLoS One, 6, e20704. Quignon P., Rimbault M., Robin S. et al. (2012) Genetics of canine olfaction and receptor diversity. Mammalian Genome 23, 132 43. Randi E., Hulva P., Fabbri E. et al. (2014) Multilocus detection of wolf 9 dog hybridization in Italy, and guidelines for marker selection. PLoS ONE 9, e86409. Robin S., Tacher S., Rimbault M. et al. (2009) Genetic diversity of canine olfactory receptors. BMC Genomics 10, 21. Spady T.C. & Ostrander E.A. (2008) Canine behavioral genetics: pointing out the phenotypes and herding up the genes. American Journal of Human Genetics 82, 10 8. Supporting information Additional supporting information may be found in the online version of this article. Table S1 Standard for olfactory ability. Table S2 Characteristics of the analyzed SNPs. Table S3 Primer sequences of cor genes. Table S4 Single extensive primer sequences of SNPs. Table S5 Allele frequency and genotype frequency of five SNPs. Table S6 Interaction of genotype and a kind of odor on scenting score.