(olfaction/olfactory receptor/canine genome/genetics/evolution)

Four members of the canine olfactory receptor gene family were characterized. The predicted proteins shared 40-64% identity with previously identified olfactory receptors. The four subfamilies identified in Southern hybridization experiments had as few as 2 and as many as 20 members. All four genes were expressed exclusively in olfactory epithelium. Expression of multiple members of the larger subfamilies was detected, suggesting that most if not all of the cross-hybridizing bands in genomic Southern blots represented actively transcribed olfactory receptor genes. Analysis of large DNA fragments using Southern blots of pulsed-field gels indicated that subfamily members were clustered together, and that two of the subfamilies were closely linked in the dog genome. Analysis of the four olfactory receptor gene subfamilies in 26 breeds of dog provided evidence that the number of genes per subfamily was stable in spite of differential selection on the basis of olfactory acuity in scent hounds, sight hounds, and toy breeds. One of the first steps in an animal's detection of a volatile compound in its environment is the binding of the odorant to G protein-coupled receptors expressed on the surface of olfactory neurons. The pattern of receptors to which the odorant binds, and thus the pattern of neurons that transmit an action potential to the olfactory bulb in the brain, allow the animal to identify the compound and respond to its presence. The gene family that encodes these odorant receptors was recently identified (1). The family is thought to be quite large, numbering in the hundreds of genes in most mammals (1-3). Representatives of the olfactory receptor gene family have been cloned from rat, human, mouse, catfish, and dog (1-9). Most appear to be expressed in the olfactory neuroepithelium, although expression of individual receptors in the testis (2) and tongue (10) have been reported. Olfactory receptors have seven transmembrane domains and were initially cloned on the basis of their similarity to other membrane-bound G protein-coupled receptors (1). All olfactory receptors identified to date share some characteristic sequence motifs and have a central variable region that corresponds to a putative ligand binding site (1, 11). This variability in the binding site is expected for a set of receptors that must discriminate among thousands of different odorants. The genes that encode the family of olfactory receptors are relatively small ('1 kb in size) and lack introns within the coding region. In human and mouse the genes are often found in linked arrays, with closely related genes found close to one another (3, 4). This arrangement is expected for a gene family whose numbers have increased through the process of unequal exchange (12). The study of olfactory receptor genes from dogs is of special interest for two reasons. First, dogs have extremely sensitive noses. Humans use dogs for hunting, tracking, drug detection, and bomb sniffing because of their olfactory acuity and trainability. Second, there is great variation in the size of the olfactory epithelium in different breeds of dog, and different levels of reliance by the breeds on olfactory cues (13, 14). Thus studying olfactory receptor genes from dogs and comparing the genes in different breeds of dog could provide insight into the evolution of this gene family in response to natural and artificial selection for enhanced olfactory ability. In this study we identified genes encoding four subfamilies of olfactory receptors. We characterized their expression pattern and organization in the genome, and tested whether unequal crossing-over during the selective breeding of dogs has contributed to the present day differences in the olfactory behavior of different breeds. MATERIALS AND METHODS Library Construction and Screening. High-molecular-weight genomic DNA was isolated from canine spleen (15). The DNA was briefly digested with Sau3A and size-selected (35-50 kb) on a sucrose gradient before cloning into the SuperCosl cosmid vector (Stratagene). The cosmid clones were packaged with Gigapack II (Stratagene) and introduced into NM554 bacteria, producing a library of 300,000 recombinants. To create an olfactory receptor-specific probe, degenerate oligonucleotides corresponding to protein sequence PMY(L/ F)FL (primer NL61) and TC(A/G)SHL (primer NL63) were used to amplify a collection of olfactory receptor gene fragments from dog genomic DNA using the polymerase chain reaction (PCR). Primer sequences were obtained from Randall Reed (Johns Hopkins University). The primers contained EcoRI and HindIII restriction sites (shown in boldface type) to aid in cloning, and had the following sequences, 5' to 3': NL61, CGGAATTCCC(GATC)ATGTA(CT)(CT)T(GATC) TT(CT)CT); and NL63, ATAAGCTTAG(GA)TG(GATC) (GC)(TA)(GATC)(GC)C(GA)CA(GATC)GT. The resulting gene fragments were digested with EcoRI and HindIII, subcloned into pBluescript KS+ (Stratagene), and sequenced using a Sequenase 2.0 kit (United States Biochemical). Ten of 16 clones analyzed were sufficiently similar to olfactory receptor genes from other species to be considered presumptive canine olfactory receptor genes. These 10 clones were pooled and 32P-labeled by random priming with the Multiprime DNA Labeling System (Amersham). Both the cosmid library and an EMBL3 phage library of dog genomic DNA (a gift from John Gerlach, Michigan State University) were screened with this complex probe at medium stringency (55°C) and washed in 0.2x standard saline citrate (SSC) and 0.1% SDS by using standard procedures (15). Sequencing from the Genomic Clones. Restriction fragments of the cosmids and phage containing the genes of interest were identified by Southern blot analysis using the 10-clone pool as a probe, and the fragments were subcloned into the vector pMOB (16). Sequences of the candidate olfactory receptor genes were Abbreviation: CHEF, contour-clamped homogeneous field electrophoresis. Data deposition: The sequences reported in this paper have been deposited in the GenBank data base (accession nos. U53679-U53682). *To whom reprint requests should be addressed. 10897 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 10898 Genetics: Issel-Tarver and Rine obtained by y8 transposon-facilitated DNA sequencing (16) using an automated laser fluorescence (ALF) Sequencer (Pharmacia). Four different genes were recovered and designated CfOLF1-4 for Canis familiaris olfactory receptor 1-4. Southern Blot Hybridizations. Dog genomic DNA was digested with restriction enzymes and electrophoretically separated on 0.8% agarose gels. The DNA was transferred to GeneScreen nylon membranes (DuPont) and hybridized in 0.5 M NaHPO4 (pH 7.2), 7% SDS, and 1 mM EDTA at 60°C (17). Washes were done at 60°C in 40 mM NaHPO4 (pH 7.2), 1% SDS, and 1 mM EDTA. Probes specific to each of the four genes were generated by PCR using cosmid or phage templates. The CfOLF1 probe covered nucleotides 11-920 of its open reading frame and was amplified with primers 1-L (5'-AACTACACCTTGGTGACCGAG-3') and 1-R (5'-TTAACCTTACAGCTCTCTTAGC-3'). The CfOLF2 probe covered nucleotides 27-866 of its open reading frame and was amplified with primers 2-L (5'-GAATGAATTCCTTCTCGTGG-3') and 2-R (5'-ATCAGAGGGTTTAGCATGG-3'). The CfOLF3 probe covered nucleotides 9-921 of its open reading frame and was amplified with primers 3-L (5'-AGGTAACCAGACTTGGGT-3') and 3-R (5'-TTGCCCTAATAGTTTCTG-3'). The CfOLF4 probe covered nucleotides 2-870 of its open reading frame and was amplified with primers 4-L (5'-TGGAACTAGAGAATGATACACG-3') and 4-R (5'-TCCTGAGGCTGTAGATGAAG -3'). Probes were labeled as described above. Contour-Clamped Homogeneous Field Electrophoresis (CHEF) Gels. Agarose blocks containing 5 ,ug of canine genomic DNA from MDCK cells were incubated with restriction enzymes NotI, Pacl, PmeI, and Sfi1 (18). Gels were cast with 1% LE agarose (FMC) and placed in 0.5x TBE in a CHEF-DR3 Apparatus (Bio-Rad). The DNA was electrophoretically separated at 6 V/cm. Run times varied from 22 to 28 h, and the gels resolved fragments between 20 and 500 kb. Lambda concatamers were used as size markers. Southern transfer and hybridizations were performed as described above. RNA Extraction and RNase Protection Experiment. Poly(A+) RNAwas extracted from 1 g of quick-frozen fresh tissue using the Fast Track 2.0 system (Invitrogen). To ensure that it was intact, the RNA was inspected after electrophoresis on an agarose gel containing formaldehyde and ethidium bromide. RNase protection experiments were carried out according to standard procedures (17) using an RNase A/Ti mix (Ambion, Austin, TX). Samples consisted of 1-2 ,ug of poly(A+) RNA mixed with 10 ,ug of yeast tRNA (Sigma) and 7 x 105 cpm of a 32P-labeled antisense RNA probe. Probes were made with PCR-amplified gene fragments cloned into a pCRII vector (Invitrogen) and transcribed with SP6 RNA polymerase (Promega). The subclones were sequenced to confirm that no PCR-generated mutations were present. The CfOLF1 probe covered 190 nt (nucleotide positions 221-411 of its open reading frame), CfOLF2 covered 407 nt (positions 393-800), CfOLF3 covered 188 nt (positions 580-768), and CfOLF4 covered 155 nt (positions 684-839). Protected fragments were separated on 6% acrylamide/8 M urea gels.