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 GG 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 Chen et al. 2012; Randi et al. 2014). Gene polymorphisms very popular in the field of police work because of their are associated with behavior of dogs in the fields of social sensitive nose. Because of the willingness of these dogs to behavior, scenting behavior, activity–impulsivity, etc. cooperate with humans, police forces make use of this trait (Hejjas et al. 2007; Kubinyi et al. 2012; Kis et al. 2014). in dogs as tools for detecting a variety of odors such as Over half of the canine olfactory receptor (cOR) genes explosives, drugs, humans and even molds (Lesniak et al. were found to show high polymorphism, some of which 2008; Pinc et al. 2011; Lippi and Cervellin 2012). Whether were breed specific and some of which might affect the olfactory sensation of dogs is acute or not is often individual olfactory perception (Keller and Vosshall 2008; evaluated through the use of many kinds of behavior tests, Robin et al. 2009; Quignon et al. 2012; Kim et al. 2012). which have raised questions about the objectivity the tests Therefore, the correlation between the genetic polymor- in the absence of a ‘gold standard’ (Doty and Kamath phism of cOR genes and olfactory ability was further 2014). The spread of different phenotypes in dogs has studied. If a certain allele is likely to endow olfactory resulted from restricting gene flow and artificially selecting receptors with ligand-binding capacity, that might con- offspring (Spady and Ostrander 2008; Akey et al. 2010; tribute to the olfaction sensitivity of dogs. Lesniak et al. (2008) verified that the specific alleles at two loci, cOR9S13: Address for correspondence c.592G>A and cOR52N9:c.176A>G, were statistically sig- nificantly linked to odor recognition capabilities of dogs. M. Yang, Police-dog Technology Department, National Police University of China, 4 Baishan Road, Yuhong District, Shenyang, However, it is possible that a certain polymorphism at a Liaoning 110034, China. given locus does not fully determine a dog’s odor sensing E-mail: [email protected] skills. Understanding the genetic mechanism of olfactory Accepted for publication 13 October 2015 behavior traits could help breeders to breed dogs that cater

to human requirements and create a genetic method of Yij ¼ l þ Gi þ Oj þ (GO)ij þ eij; picking out excellent individuals. l An experimental group of 48 German Shepherd dogs where Yij is the score of the smelling ability, is the total (males, n = 28; females, n = 20), 2–3 years old, were mean values, Gi is the major effect on genotype i level, Ej is selected from the Police Dog Technology School of Ministry the major effect on odor j level, (GO)ij is the interaction of Ɛ of Public Security, China. All the dogs had reached the genotype with the kind of odor on smelling ability and ij is standard for police dogs and were treated in a humane the effect of random error. manner (Lesniak et al. 2008). Experienced handlers Overall, each SNP locus had one main effective allele recorded the sniffing ability of each dog, which had been (Table S5) with a frequency between 0.594 and 0.917. In > trained to detect specific kinds of odor, including the odors particular, the OR2K2-like:c.72 C/T, cOR52E17: > > > of human, ice drug and trinitrotoluene, and were in the last c.791G A, OR2M5:c.98T C, OR2M5:c.170C T, OR2M5: > > phase of training. The standards for scoring are shown in c.862G A and OR08H10:c.241G A loci presented as Table S1. homozygous, reflecting a unique genotype because some A blood sample was obtained from each dog, and then alleles are breed specific or rare in the dog population, DNA was isolated. Twenty-two representative SNP sites, of whereas the other loci presented two or three genotypes which five were breed specific in German Shepherd dogs, lay (Robin et al. 2009; Kim et al. 2012). According to the allele in exonic regions of 12 cOR genes (Robin et al. 2009) frequencies, genetic disequilibrium was found at the > > (Table S2). Pairs of specific primers for 12 cOR gene fraction OR52J3:c.469G A, OR2K2-like:c.518G A and OR4C11- > v2 > < amplification and 22 single extensive primers for SNP like:c.256A G loci ( 9.210, P 0.01). On the contrary, the 13 other loci were in Hardy–Weinberg equilibrium genotypes were designed using the PRIMER-BLAST and PRIMER (v2 < 5.991, P > 0.05). Meanwhile, 16 loci showed low PREMIER 5 programs. The primer sequences, fragment sizes < and annealing temperatures (Tm) mentioned above are polymorphism levels (PIC 0.25), and six loci showed a < < shown in Tables S3 and S4. medium level of polymorphism (0.25 PIC 0.50) Polymerase chain reaction (PCR) was conducted using a (Table 1). This phenomenon might be a result of the Veriti thermal cycler (Applied Biosystems, Inc.) in a total recruitment system used for police dogs. volume of 50 ll that included 4 ll of 100 ng of genomic We also observed that the olfactory sensitivity of dogs varied with certain genotypes. The biggest difference DNA, 25 ll of Premix Taq, 1 lM (final concentration) of > each specific primer and up to 50 ll of water. The following occurred relating to the OR4C11-like:c.511T G and > conditions for the reaction were applied: 10 s at 98 °C, 30 s OR4C11-like:c.692G A loci, as dogs with a TT and GG at 53 °Cor58°C, 30 s at 72 °C and 32 amplification genotype respectively were scored at 26.85 for the detection 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 Table 1 Genetic polymorphism analysis of SNPs. in a 5-ll reaction system that consisted 2 ll of Premix, 2 ll 2 Num Loci v He PIC of the purified PCR products and 1 ll(1lM) of primer, starting with 25 cycles of denaturation at 96 °C for 10 s, 1 OR10H1-like:c.632C>T 0.087 0.202 0.183 OR10H1-like: > annealing at 50 °C for 10 s and extension at 72 °C for 30 s, 2 c.770A T 0.087 0.202 0.183 3 OR52J3:c.469G>A 42.013 0.278 0.239 followed by an extra extension at 72 °C for 5 min. A total 4 OR2K2-like:c.172>C/T – 00 of 4.25 ll of product digestion was initiated by adding 5 OR2K2-like:c.518G>A 41.245 0.249 0.217 0.5 llof109 buffer and 0.25 llof1U/ll SAP (TaKaRa) 6 OR8S1:c.658T>C 1.732 0.329 0.275 for 1 h in a 37 °C water bath and for 15 min in a 80 °C 7 OR1L4:c.214G>C 0.167 0.457 0.353 OR51I2-like: > water bath. One half microliter (0.5 ll) of the product was 8 c.710T G 0.034 0.186 0.169 9 OR51I2-like:c.770T>C 0.034 0.186 0.169 added to 0.5 ll of Size Standard 80 and 39 ll of sample 10 OR51I2-like:c.778T>A 0.034 0.186 0.169 loading solution. The capillary electrophoresis was ana- 11 OR49-like:c.147A>G 0.088 0.152 0.141 lyzed on a Beckman GeXP Genetic Analysis System 12 cOR52E17:c.214G>A 1.732 0.329 0.275 according to GenomeLab SNP-Start Primer Extension Kit 13 cOR52E17:c.791G>A – 00 OR2M5: > – (Beckman). 14 c.98T C 00 15 OR2M5:c.170C>T – 00 Results were represented as mean values standard 16 OR2M5:c.862G>A – 00 error for sniffing scores. Allele frequency, genotype fre- 17 OR52J3-like:c.106G>A 2.320 0.482 0.366 quency, He (genetic heterozygosity) and PIC (polymorphism 18 OR4C11-like:c256A>G 36.700 0.152 0.141 OR4C11-like: > information content) were counted by EXCEL software. PROC- 19 c.511T G 2.131 0.469 0.359 20 OR4C11-like 692G>A 2.131 0.469 0.359 GLM of SAS 9.2 software followed by the Tukey test was 21 OR08H10:c.241G>A – 00 applied to analyze the relationship. The interaction of 22 OR08H10:c.927G>A 0.472 0.249 0.218 genotype with the kind of odor on smelling ability was calculated according to the formula: PIC, polymorphism information content.

Table 2 Scores of smelling behavior of dogs with different SNP genotypes.

SNP Odor Genotype (l SE)

OR10H1-like:c.632C>TCCCT 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>TAAAT 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>AAAGG 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>AAAGG 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>CTTTCCC 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>CGGGCCC 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>GTTTG Human 29.08 0.17 30.00 0.01 Ice drug 22.46 0.87 25.65 1.59 Trinitrotoluene 17.50 0.26 15.58 1.08 OR51I2-like:c.70T>CTTTC Human 29.08 0.17 30.00 0.01 Ice drug 22.46 0.87 25.65 1.59 Trinitrotoluene 17.50 0.26 15.58 1.08 OR51I2-like:c.778T>AAAAT Human 29.08 0.17 30.00 0.01 Ice drug 22.46 0.87 25.65 1.59 Trinitrotoluene 17.50 0.26 15.58 1.08 OR49-like:c.147A>GAAAG 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>AGGGAAA 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>AGG Human 29.39 0.10 Ice drug 22.98 0.53 Trinitrotoluene 17.18 0.16 OR2M5:c.98T>CTT Human 29.39 0.10 Ice drug 22.98 0.53 Trinitrotoluene 17.18 0.16 OR2M5:c.170C>TCC Human 29.39 0.10 Ice drug 22.98 0.53 Trinitrotoluene 17.18 0.16

Table 2 (Continued)

SNP Odor Genotype (l SE)

OR2M5:c.862G>AGG Human 29.39 0.10 Ice drug 22.98 0.53 Trinitrotoluene 17.18 0.16 OR52J3-like:c.106G>AGGGAAA 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>GAAGG 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>GTTTGGG 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>AGGGAAA 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>AGG Human 29.39 0.10 Ice drug 22.98 0.53 Trinitrotoluene 17.18 0.16 OR08H10:c.927G>AGGGA 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 GG and AA genotype group of dogs: 77.1% at the OR10H1-like:c.632C>T and respectively were scored at 15.30 (Table 2). To distinguish OR10H1-like:c.770A>T loci and 85.4% at the OR2K2-like: the dogs’ olfactory acuity, the relationship between geno- c.518G>A locus. This phenomenon might be because the types and phenotypes was studied further. The genotypes at ratings of dogs’ smelling behavior made by scientists and the OR10H1-like:c.632C>T, OR10H1-like:c.770A>T, experienced dog trainers were very similar (Lesniak et al. OR2K2-like:c.518G>A, OR4C11-like:c.511T>G and 2008). However, there was a different situation for the OR4C11-like:c.692G>A loci had a statistically significant OR4C11-like:c.511T>G and OR10H1-like:c.632C>T loci, relationship with the scenting ability of police dogs where heterozygous individuals represented 58.3% of the (P < 0.01) (Table S6). Notably, three of five breed-specific total. Maybe experienced dog trainers were incapable of SNP loci showed this relationship in contrast to only two of selecting acute sniffing dogs that acted almost as well as or 17 non-breed-specific SNP sites. This verified again that a as bad as others. breed-specific SNP usually gives the strongest association In addition, the kind of odor could have influenced the signal for susceptibility to a certain phenotype (Olsson et al. performances of the dogs (P < 0.01) (Table S6). All the dogs 2011). In particular, the mean values of smelling scores of displayed behavioral responses to the odor of human more individuals homozygous for the main effective alleles, quickly. As one of the mammalian prey species, dogs may heterozygous individuals and those homozygous for the have been more sensitive to a human odor compared to non-main effective alleles varied from high to low, respec- non-prey-associated odors (Nilsson et al. 2014). Further- tively, at the five loci (Table 2). These finding are in more, the interaction between certain genotypes and a accordance with the findings of Lesniak et al. (2008). We certain odor was found for the OR10H1-like:c.632C>T, again proved that genetic differences in olfactory abilities OR10H1-like:c.770A>T, OR4C11-like:c.511T>G and existed in individuals and that the polymorphisms of cOR OR4C11-like:c.692G>A loci (P<0.01) (Table S6). Dogs genes might play an important role in olfactory sensitivity. homozygous for the main effective alleles at those four loci The allelic transition at these loci were associated with the obtained higher scores in detecting ice drug (Fig. 1). Thus, difference in scores for individuals. The genotype frequen- people could utilize this interaction to train German cies revealed that individuals homozygous for the main Shepherd dogs with preferable genotypes to carry out an effective alleles comprised the majority of the analyzed assignment to detect ice drug.

(a) 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. 19 – (b) DNA Research , 275 87. Kubinyi E., Vas J., Hejjas K. et al. (2012) Polymorphism in the tyrosine hydroxylase (TH) gene is associated with activity- impulsivity 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., Sjoberg€ 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., TruveK.et al. (2011) A novel unstable Figure 1 The interaction of a genotype and a kind of odor on olfactory duplication upstream of HAS2 predisposes to a breed-defining abilities of dogs. (a) Interaction with the OR10H1-like:c.632C>T and skin phenotype and a periodic fever syndrome in Chinese Shar- OR10H1-like:c.770A>T loci. (b) Interaction with the OR4C11-like: Pei dogs. PLoS Genetics 7, e1001332. c.511T>G and OR4C11-like:c.692G>A loci. Pinc L., Bartosˇ L., Reslova´ A. et al. (2011) Dogs Discriminate Identical Twins. PLoS One, 6, e20704. In conclusion, there was a correlation between SNP Quignon P., Rimbault M., Robin S. et al. (2012) Genetics of canine genotypes of cOR genes and olfactory ability of dogs. The olfaction and receptor diversity. Mammalian Genome 23, 132–43. scenting ability might differ according to the detection of Randi E., Hulva P., Fabbri E. et al. (2014) Multilocus detection of different kinds of odors. Dogs with a certain genotype might wolf 9 dog hybridization in Italy, and guidelines for marker be more competent in detecting a specific odor. These selection. PLoS ONE 9, e86409. preliminary results showed that molecular genetic studies Robin S., Tacher S., Rimbault M. et al. (2009) Genetic diversity of on cOR genes might be a valuable tool to improve the canine olfactory receptors. BMC Genomics 10, 21. selection of sniffing dogs and its working fields. 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. Acknowledgement This work was supported by the Application and Innovation Supporting information Foundation of Ministry of Public Security of the People’s Additional supporting information may be found in the Republic of China (2011YYCXSYJQ165). online version of this article. Table S1 Standard for olfactory ability. References Table S2 Characteristics of the analyzed SNPs. Table S3 Primer sequences of cOR genes. Akey J.M., Ruhe A.L., Akey D.T. et al. (2010) Tracking footprints of Table S4 Single extensive primer sequences of SNPs. artificial selection in the dog genome. Proceedings of the National Table S5 Academy of Sciences of the United States of America 107, 1160–5. Allele frequency and genotype frequency of five Chen R., Irwin D.M., and Zhang Y.P. (2012) Differences in selection SNPs. drive olfactory receptor genes in different directions in dogs and Table S6 Interaction of genotype and a kind of odor on wolf. Mol Biol Evol, 29, 3475–84. scenting score.