Reporter gene expression in the olfactory epithelium of H-lacZ6 transgenic mice mimics the cell-selective expression pattern known for some odorant receptor genes. The transgene construct in these mice consists of the lacZ coding region, driven by the proximal olfactory marker protein (OMP) gene promoter, and shows expression in a zonally confined subpopulation of olfactory neurons. To address mechanisms underlying the odorant receptor-like expression pattern of thelacZ construct, we analyzed the transgene-flanking region and identified OR-Z6, the first cloned odorant receptor gene that maps to mouse chromosome 6. OR-Z6bears the highest sequence similarity (85%) to a human odorant receptor gene at the syntenic location on human chromosome 7. We analyzed the expression pattern of OR-Z6 in olfactory tissues of H-lacZ6 mice and show that it bears strong similarities to that mapped for β-galactosidase. Expression of both genes in olfactory neurons is primarily restricted to the same medial subregion of the olfactory epithelium. Axons from both neuronal subpopulations project to the same ventromedial aspect of the anterior olfactory bulbs. Furthermore, colocalization analyses in H-lacZ6 mice demonstrate thatOR-Z6-reactive glomeruli receive axonal input fromlacZ-positive neurons as well. These results suggest that the expression of both genes is coordinated and that transgene expression in H-lacZ6 mice is regulated by locus-dependent mechanisms.
- olfactory neurons
- olfactory epithelium
- olfactory bulb
- expression patterns
- in situ hybridization
- locus dependence
- OMP promoter
The vertebrate olfactory system is remarkable in its ability to detect and discriminate among thousands of different odorants. Tremendous progress has been made in understanding odorant detection that is mediated by odorant receptor (OR) proteins located in the ciliary membrane of the olfactory receptor neurons (ORNs), a bipolar neuron population that resides in the olfactory epithelium (OE) of the vertebrate nose. Vertebrate ORNs are unusual in that they can be replaced throughout life via progenitor cell proliferation and differentiation (Graziadei and Graziadei, 1979). Each ORN seems to express a single OR gene from a family that is estimated to number as many as 1000 different genes in rodents (Buck and Axel, 1991). ORNs expressing a given OR gene exhibit a mosaic-like distribution within one of four suggested zones of the neuroepithelium (Ressler et al., 1993; Strotmann et al., 1994a,b; Vassar et al., 1994), and their axons project onto defined glomeruli in the olfactory bulb (OB) (Ressler et al., 1994; Vassar et al., 1994; Mombaerts et al., 1996). The mechanisms underlying these expression patterns are still elusive but likely depend on a complex interplay including allelic exclusion, locus control elements, and zonally active transcription factors during early development (Ressler et al., 1993; Chess et al., 1994; Sullivan et al., 1996; Ebrahimi et al., 2000). Chromosomal localization studies showed that OR genes are present on several chromosomes where they often exist as clusters (Ben-Arie et al., 1994;Asai et al., 1996; Sullivan et al., 1996; Trask et al., 1998). Recent studies revealed that highly homologous genes from a cluster exhibit expression in identical zones (Strotmann et al., 1999; Tsuboi et al., 1999; Serizawa et al., 2000). However, the relation between locus and zonal confinement is probably more complex because many receptors expressed in the same zone reside at different loci and vice versa (Sullivan et al., 1996).
To address the molecular and cellular bases of OR gene expression, we analyzed the transgenic mouse line H-lacZ6 (Walters et al., 1996a,b) in which transgene expression mimics the expression pattern demonstrated for some OR genes (Ressler et al., 1993; Vassar et al., 1993; Strotmann et al., 1994a,b). Transgene expression in H-lacZ6 mice is restricted to a subpopulation of ORNs primarily located on the tips of endoturbinate-II and -III and ectoturbinate-3 (Treloar et al., 1996; Walters et al., 1996a), the axons of which terminate in closely associated glomeruli in the ventromedial aspects of the anterior OBs (Treloar et al., 1996;Cummings et al., 2000). This pattern prompted us to hypothesize that the lacZ transgene has inserted under the control of an OR gene locus, and we analyzed the region flanking the insertion site for the presence of OR genes. Our hypothesis was supported by the subsequent identification of the new odorant receptor geneOR-Z6 that not only resides in proximity to the transgene locus on mouse chromosome 6 but also exhibits an expression pattern similar to that of the transgene in H-lacZ6 mice. Our results strongly suggest that the expression of both genes in H-lacZ6 mice is subject to similar locus-dependent regulatory mechanisms.
MATERIALS AND METHODS
H-lacZ6 mouse line. The generation of the transgenic mouse line H-lacZ6 was reported previously (Walters et al., 1996a). In brief, the transgene construct used [H-olfactory marker protein (OMP)–lacZ] consists of a truncated OMP promoter (−239 to +55 bp of the rat OMP gene that carry the proximal olf-1, O/E1 binding site) fused to the β-galactosidase coding region and the Simian virus 40 (SV40) polyadenylation signal (Kudrycki et al., 1993). Experiments were performed on homozygous mice of the transgenic H-lacZ6 line. Procedures involving the use of animals were approved by the Institutional Animal Care and Use Committee of the University of Maryland School of Medicine.
Cloning of transgene-flanking sequences. A genomic library in λDashII (Stratagene, La Jolla, CA) was prepared from homozygous H-lacZ6 mice, and plaque lifts from ∼1× 106 phages were screened with an [α-32P]dCTP-labeled lacZprobe (1.3 × 106 cpm/ml) according to standard procedures (final washes in 2× SSC and 0.5% SDS at 60°C). Random primed labeling of a gel-purified 1891 bpEcoRV–EcoRI restriction fragment of thelacZ coding region [pMC1871 lacZ-plasmid (Amersham Pharmacia Biotech, Piscataway, NJ)] was performed using [α-32P]dCTP (3000 Ci/mmol) and the Ready-To-Go DNA-labeling kit (Amersham Pharmacia Biotech). DNA from two different phages, obtained after three rounds of screening, was purified (λ-Magic Minipreps; Promega, Madison, WI) and mapped by restriction digestion, partial sequence analysis, and Southern blot hybridization using 32P-labeled probes for the OMP promoter (296 bp of the proximal region) and the SV40 sequence that hybridize to the 5′-end and 3′-end of the transgene construct, respectively. Probe labeling was as described. Phage λ8 contained a nontruncated copy of the transgene construct in the correct order of OMP–lacZ–SV40 plus 12 kb of genomic downstream sequence and was further analyzed by restriction digestion and Southern blot hybridization using the α-32P-labeled SV40 DNA probe. An 874 bpXbaI hybridization product was excised from the gel, purified (Qiagen, Valencia, CA), and subcloned into anXbaI-linearized vector (pBluescript; Stratagene). Sequence analysis using T3 and T7 vector primers showed that the fragment contained the transition between the 3′-end of the transgene construct (246 bp of SV40 sequence) and downstream flanking genomic DNA (629 bp). The 629 bp flanking fragment was PCR-amplified using primers EW1 and EW20; the amplicon was subjected to restriction digestion usingBamHI and XbaI enzymes (the BamHI site had been introduced by primer EW1) and directionally subcloned into pBluescript (clone pBSKS3′-flank).
Interspecific mouse backcross mapping. Interspecific backcross progeny were generated by mating (C57BL/6J ×Mus spretus) F1 females and C57BL/6J males as described (Copeland and Jenkins, 1991). A total of 205 N2 mice were used to map the Hlz6 locus. DNA isolation, restriction enzyme digestion, agarose gel electrophoresis, Southern blot transfer (Hybond-N+ nylon membrane; Amersham Pharmacia Biotech), and hybridization were as described (Jenkins et al., 1982). The probe, the 629 bp 3′-flanking fragment of mouse genomic DNA, was labeled with [α-32P]dCTP using a nick translation labeling kit (Boehringer Mannheim, Indianapolis, IN); washing was done to a final stringency of 0.8× SSC-phosphate and 0.1% SDS, 65°C. A fragment of 6.1 kb was detected in SphI-digested C57BL/6J DNA, and a fragment of 8.4 kb was detected in SphI-digested M. spretus DNA. The presence or absence of the 8.4 kb SphIM. spretus-specific fragment was followed in backcross mice. A description of the probes and restriction fragment length polymorphisms (RFLPs) for the loci linked to Hlz6 includingPtn, Tcrb, and Hoxa5 has been reported previously (Siracusa et al., 1991; Li et al., 1992). Recombination distances were calculated using Map Manager, version 2.6.5. Gene order was determined by minimizing the number of recombination events required to explain the allele distribution patterns.
P1 clones. Two primers generated from the 3′-flanking fragment (clone pBSKS3′-flank) were used to screen a wild-type mouse genomic library in P1 phagemids by PCR (Genome Systems, St. Louis, MO). Primers ps2 and ps4 yielded a 160 bp amplicon. The four P1 clones obtained (P1 clones 6386–6389) were confirmed to map to theHlz6 locus by PCR using the same primers. PCR standard conditions were 10 mm Tris-HCl, pH 9, 50 mm KCl, 0.1% Triton X-100, 0.2 mm each dNTP, 2 mmMgCl2, 2.5 U of Taq polymerase (Promega), and 0.2 μm each primer. Cycling parameters were 5 min at 95°C, followed by 35 cycles of 1 min at 60°C, 1.5 min at 72°C, and 1 min at 96°C and a final extension for 20 min at 72°C.
Cloning of OR-Z6. P1 clones 6386–6389 (cre−-bacterial host strain NS3516; Genome Systems) were grown in LB medium (Life Technologies, Rockville, MD) supplemented with 20 μg/ml Kanamycin (Sigma, St. Louis, MO), and P1 DNA from 5 hr cultures was purified according to the manufacturer's recommendations. Each 50 ng of P1 DNA was subjected to PCR using the degenerate primers p26 and p27 each at 2 μm. Cycling parameters were 5 min at 95°C, followed by 45 cycles of 2 min at 50°C, 3 min at 72°C, and 1 min at 96°C and a final extension for 20 min at 72°C. Controls contained either 100 ng of mouse genomic DNA or no template. PCR products were resolved on 1% agarose gels. The 390 bp amplicon deriving from P1 6386 DNA was gel-purified (Qiagen) and labeled with dUTP-conjugated digoxigenin (Boehringer Mannheim) by PCR according to the manufacturer's procedure and was used as a probe. For Southern blot hybridization, each 800 ng of restriction-digested P1 DNA was size-fractionated on 0.6% agarose gels, denatured (0.5m NaOH and 1.5 m NaCl), and neutralized (3 m NaCl and 0.5m Tris-HCl, pH 7.5) before passive overnight transfer onto nylon membranes (Nytran; Bio-Rad, Hercules, CA). The membranes were UV-cross-linked (120 mJ/cm2; UV Stratalinker 2400; Stratagene), prehybridized for 2 hr at 62°C in 5× SSC, 0.1%N-lauroylsarkosine, 0.02% SDS, and 1% blocking reagent (Boehringer Mannheim), and hybridized for 16 hr at 62°C in the same buffer containing 25 ng of digoxigenin-labeled DNA probe (see above). Posthybridization washes were 10 min in 2× SSC at room temperature and twice for 30 min each in 0.2× SSC containing 0.1% SDS at 65°C. Detection of digoxigenin-labeled hybridization products was performed using the manufacturer's protocol (DIG-Genius Users Guide; Boehringer Mannheim). The 3 kb XbaI hybridization fragment was subcloned into pGEM5zf (Promega) and subjected to double-strand sequencing. The OR-Z6 sequence was deposited in GenBank (accession number AF320347).
Genomic Southern blot hybridization. Genomic DNA was extracted from liver tissue of H-lacZ6 and wild-type mice (129S3), subjected to restriction digestion, and size-fractionated on 0.6% agarose gels. Southern blotting and hybridization conditions were as described. The DNA probe, the digoxigenin-dUTP-labeled 899 bp fragment of the OR-Z6 coding region, was generated by PCR using primers N and C.
Bacterial artificial chromosome clones.The sequence of the 629 bp flanking fragment (see above) was used to screen The Institute for Genomic Research (TIGR) database that contains the sequenced ends of EcoRI-digested mouse genomic DNA [C57BL/6J female mouse bacterial artificial chromosome (BAC) end sequence (BES)] generated by Osoegawa et al. (2000). BAC clones RPCI-23-282I16 (AQ932615) and RPCI-23-323H21 (AQ988378) were obtained as stab cultures (CHORI; BACPAC Resources, Oakland, CA) and amplified in LB medium containing 20 μg/ml chloramphenicol (Sigma). Purification of BAC DNA and subsequent conditions for PCR using the degenerate primers NL61 and NL63 were as described for P1 DNA using annealing temperatures of 50°C (high stringency) or 40°C (low stringency). PCR products were size-fractionated on agarose gels, gel-purified, and subcloned into the pGEM-T vector system (Promega). Generated subclones were analyzed by restriction digestion and sequencing.
Primers used (5′ to 3′ orientations). The following primers were used: EW1, TGTCTGGATCCAGATGGAGG; EW20, CGATAAGCTCGATATCGA; ps2, CAGTCCTGCAGTAGTGAATTCTC; ps4, CAGTGTTTTGTATTCCTTCTCAG; p26, GCITA(T/C)GA(T/C)CGITA(T/C)GTIGCIATITG; p27, ACIACIGAIAG(A/G)TGIGAI(C/G)C(A/G)CAIGT; NL61, CGGAATTCCC(G/A/T/C)ATGTA(C/T) (C/T)T(G/A/T/C)TT(C/T)CT; NL63, ATAAGCTTAG(G/A)TG(G/A/T/C)(G/C)(T/A)(G/A/T/C)(G/C) C(G/A)CA(G/A/T/C)GT; and OR-Z6 gene-specific primers (coding region) N, CCTGGATCCCAAGAGCTACAC (48–69 bp, sense), and C, CTGAGGAGCAGACAGCAACGC (918–938 bp, antisense).
Reverse transcription-PCR forOR-Z6. Four-week-old 129S3/SvImJ mice (129S3; The Jackson Laboratory, Bar Harbor, ME) were decapitated; their tissues (OE, OBs, and liver) were dissected and snap-frozen in liquid nitrogen. Total RNA was extracted with RNAzol-B (TEL-TEST, Friendswood, TX). Contaminating genomic DNA was removed by DNase I digestion (2 U of DNase I/μg of RNA; Boehringer Mannheim), followed by acidic phenol/chloroform extraction (Ambion, Austin, TX). Total RNA was reverse transcribed (Superscript RTII; random hexamer primers, 20 pmol/μg of RNA; Life Technologies), and 500 ng of cDNA was subjected to PCR. To control for PCR products deriving from amplification of residual genomic DNA, an aliquot of RNA from each tissue was subjected to PCR without preceding cDNA synthesis. OR-Z6 cDNA was selectively amplified using the gene-specific primers N and C that generate an 899 bp fragment of the OR-Z6 coding region. PCR products were resolved on 1% agarose gels containing 0.5 μg/ml ethidium bromide, visualized under UV light, and photographed (Polaroid, Cambridge, MA). PCR products were gel-purified (Qiagen) and sequenced using primers N and C.
Tissue preparation for cryostat sections. Three- to 6-week-old 129S3 mice (The Jackson Laboratory) or H-lacZ6 mice were anesthetized using ketamine at 165 mg/kg of body weight (Animal Health, Fort Dodge, IO) and xylazine at 11 mg/kg of body weight (Bayer, Shawnee, KS). Mice were transcardially perfused with ice-cold buffer-1 (0.1 m PIPES, 5 mmMgCl2, and 5 mm EGTA, pH 6.9) (Emson et al., 1990) followed by fixative [4% (w/v) paraformaldehyde in buffer-1]. OE and OBs were dissected, cryoprotected in 30% sucrose overnight at 4°C, embedded in O.C.T. (Tissue-Tek, Torrance, CA), and frozen in a dry ice/acetone bath. Serial cryosections (10–18 μm) of the OE and of the OBs were thaw-mounted onto slides (Superfrost-Plus; Fisher Scientific, Pittsburgh, PA) and air-dried. All analyses were performed on coronal sections (frontal and perpendicular to the anterior-to-posterior axis of the OBs).
Detection of β-galactosidase expression.Staining of perfused hemiheads or cryosections using the chromogenic substrate 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-gal) and β-galactosidase immunohistochemistry were as described inCummings et al. (2000).
In situ hybridization. Serial coronal cryosections of the OE and OB were fixed for 10 min in 4% paraformaldehyde in PBS at 4°C, rinsed twice in PBS, and incubated for 10 min at room temperature in 0.25% acetic anhydride and 0.1 mtriethanolamine, pH 8.0. Sections were rinsed in PBS and prehybridized in standard hybridization solution (50% formamide, 5× SSC, 5× Denhardt's, 0.2% SDS, 0.5 mg/ml salmon sperm DNA, and 0.25 mg/ml yeast tRNA) for 4 hr at 60°C. OR-Z6 riboprobes were generated from a BamHI subclone in pGEM5zf(+) (Promega) containing a 573 bp fragment of the OR-Z6 coding region. Digoxigenin-labeled OR-Z6 sense and antisense riboprobes were prepared from 1 μg of the linearized subclone (XbaI and EcoRI, respectively) using the T7/Sp6 RNA transcription system (DIG-Genius Users Guide; Boehringer Mannheim). OR37Eriboprobes were prepared similarly; the plasmid (generously provided by J. Strotmann) was first linearized using PstI orApaI restriction enzymes. Blunt ends were obtained by incubation with Klenow enzyme (Promega), and sense and antisense riboprobes were generated using the kit indicated above. Riboprobes for the OB were prepared accordingly, but CTP was replaced by 12 μm [α-32P]CTP (10 mCi/ml; >3000 Ci/mmol; Amersham Pharmacia Biotech). Hybridization of OE sections was performed at 60°C for 16 hr in standard hybridization solution containing 500 ng/ml digoxigenin-labeled sense or antisense RNA probes. Hybridization of OB sections was performed similarly using 1 × 108 cpm/ml [α-32P]CTP-labeled riboprobe containing 10% dextran sulfate (molecular weight, 500,000). After hybridization, sections were washed twice for 5 min each in 4× SSC at room temperature and treated with RNase A (500 mmNaCl, 10 mm Tris-HCl, 1 mmEDTA, and 5 μg/ml RNase A, pH 7.5) for 30 min at 37°C. Sections were washed twice for 10 min each in 2× SSC at room temperature, twice for 30 min each in 0.2× SSC at 60°C, and once for 30 min in 0.1× SSC at 60°C. Sections hybridized with the [α-32P]CTP-labeled riboprobe were dehydrated in increasing percentages of ethanol, dipped into NTB-3 photo emulsion (Kodak, New Haven, CT), and air-dried. After a 4 week exposure in the dark at 4°C, sections were developed in Kodak-D19 (2 min at 16°C), rinsed in water, fixed in Kodak fixer, counterstained with hematoxylin (Sigma), dehydrated in ethanol, and cleared in xylene, and coverslips were mounted with DPX (Aldrich, Milwaukee, WI,). Sections hybridized with a digoxigenin-labeled riboprobe were blocked in TBS (0.1m Tris-HCl and 0.15 m NaCl, pH 7.5) containing 4% goat serum, and hybridization products were detected using alkaline phosphatase-conjugated sheep anti-digoxigenin antibody (Boehringer Mannheim; 1:5000 in TBS; overnight at 4°C) and washed twice for 15 min each in TBS. Bound antibody was visualized using 5-bromo-4-chloro-3-indolylphosphate/nitroblue tetrazolium (Boehringer Mannheim) as substrate, yielding a cytosolic purple precipitate. Endogenous alkaline phosphatase activity was blocked with 0.24 mg/ml Levamisol (Sigma). Color development was stopped in TE buffer (20 mm Tris-HCl and 5 mm EDTA, pH 8.0), and the sections were coverslipped in Dako mounting medium (Dako, Carpinteria, CA).
Zonal tracing. Adjacent coronal cryosections from P19 wild-type mice (129S3; The Jackson Laboratory) were collected on separate sets of slides, and each set was subjected to in situ hybridization using sense and antisense riboprobes either forOR-Z6 or for one of the OR genes shown previously to represent a single rostrocaudal zone of the OE. Clones for generating riboprobes for K21 (zone 1), K20 (zone 2), andL45 (zone 3) were kindly provided by L. Buck (Ressler et al., 1993, 1994). Digoxigenin-labeled riboprobes were prepared as described above. In situ hybridization signals for each OR mRNA were subsequently traced using a camera lucida attached to a Leitz Orthoplan microscope (Leitz, Wetzlar, Germany). The individual maps obtained for OR-Z6, K20, K21, andL45 were color-coded and scanned, and the images were stacked to facilitate comparison.
Preparation of photographs. Photographs of olfactory turbinates and OB whole mounts stained with X-gal were taken on a Leica WILD M3Z dissecting photomicroscope (Leica, Deerfield, IL) using Ektachrome 160T slide film (Kodak). Images were digitized using a Nikon LS-1000 slide scanner (Nikon, Melville, NY). Light photomicrographs were scanned on a Leica DMRX microscope attached to a Phase One Power Phase digital camera (Phase One, Copenhagen, Denmark). Photomicrographs of radioactive in situ hybridization were taken with a dark-field condenser using a Nikon Optiphot-2 microscope (Nikon) fitted onto a Polaroid DMC1 digital camera (Polaroid). Digitized images were minimally processed (adjustments included only small changes in contrast and brightness) using Adobe Photoshop 5.0 (Adobe Systems, Mountain View, CA), and photomicrographs were printed on a Fujix Pictrography 3000 (Fuji, Carlstadt, NJ).
The transgene expression pattern in the OE of H-lacZ6 mice (Walters et al., 1996a) was analyzed by staining whole-mount nose preparations using the chromogenic substrate X-gal. As seen for thein situ hybridization patterns reported for some OR genes (Ressler et al., 1993; Vassar et al., 1993; Strotmann et al., 1994a,b), transgene expression in H-lacZ6 mice was limited to a subset of ORNs (Fig. 1 A). In agreement with our previous studies (Treloar et al., 1996; Walters et al., 1996a), β-galactosidase-positive (β-gal+) neurons appear in a zonal pattern along the anterior-to-posterior axis of the OE and exhibit a high density on endoturbinate-II. On the basis of this expression pattern, we hypothesized that the lacZ transgene had integrated at a locus that directs the expression of nearby OR genes. To address this and to obtain information about genomic sequences flanking the transgene locus, we screened an H-lacZ6 mouse genomic library in λDashII using a32P-labeled lacZ DNA probe. Analyses of the positive phages enabled us to identify and subclone a genomic flanking fragment of 629 bp (clone pBSKS3′-flank) that resides immediately downstream of the transgene construct. This 3′-flanking fragment was subsequently used to determine the chromosomal localization of the transgene as well as to isolate the corresponding genomic region from a wild-type mouse genomic library in P1 phagemids.
The lacZ transgene maps to mouse chromosome 6
As a first step toward understanding the transgene locus (Hlz6), we mapped the chromosomal location of the insertion site by interspecific backcross analysis using a panel that has been typed for over 2900 loci that are well distributed among all autosomes and the X chromosome (Copeland and Jenkins, 1991). DNA from C57BL/6J and M. spretus was analyzed for informative RFLPs using the 629 bp genomic 3′-flanking fragment as a probe. The mapping results indicated that the Hlz6 locus is located in the proximal region of mouse chromosome 6 and linked to the chromosomal markers Ptn, Tcrb, and Hoxa5. We analyzed 174 mice for every marker (Fig. 1 B) and up to 193 mice for some pairs of markers. Each locus was analyzed in pair-wise combinations for recombination frequencies using the additional data. The ratios of the total number of mice exhibiting recombinant chromosomes to the total number of mice analyzed for each pair of loci showed that the most likely gene order is the following: centromere–Ptn–7/181–Hlz6–0/186–Tcrb–9/193–Hoxa5. The recombination frequencies [expressed as genetic distances in centimorgans (cM) ± the SE] are the following:Ptn–3.9 ± 1.4 cM–Hlz6,Tcrb–4.7 ± 1.5 cM–Hoxa5. No recombinants were detected between Hlz6 and Tcrb in 186 animals typed in common, suggesting that the two loci are within 1.6 cM of each other. We compared our interspecific map of chromosome 6 with a composite mouse linkage map that reports the location of many uncloned mouse mutations (Mouse Genome Database; The Jackson Laboratory).Hlz6 mapped in a region that lacks mouse mutations with a phenotype that might be expected for an alteration in this locus (data not shown). The proximal region of mouse chromosome 6 shares regions of homology with both the long and short arms of human chromosome 7 (Fig.1 B). In particular, Tcrb has been mapped to 7q35. The close linkage between Hlz6 and Tcrbin mouse suggests that the human homolog of the Hlz6 locus will map to 7q as well.
Cloning of the new mouse OR gene OR-Z6
To search the vicinity of the Hlz6 locus for the presence of OR genes, we first isolated the genomic region surrounding the Hlz6 locus from a wild-type mouse genomic library in P1 phagemids by PCR, using primers ps2 and ps4 that derived from the sequence of the 3′-flanking fragment. Four genomic P1 clones (P1 clones 6386–6389; Genome Systems) with an average insert size of 80 kb were obtained and confirmed to correspond to the transgene integration site by PCR using the same primers (Fig.2 A). The existence of OR genes at the Hlz6 locus was assessed by subjecting the four P1 clones to PCR using the degenerate oligonucleotides p26 and p27, shown previously to amplify the region between transmembrane domains 3 and 6 of human and mouse OR genes (nucleotide sequences for p26 and p27 were provided by L. Buck) (Ngai et al., 1993). One of the clones, P1 6386, gave rise to a PCR product of the expected size of 390 bp (Fig. 2 A). Assuming that this amplicon represents a mixture of different OR sequences, the PCR product was labeled with digoxigenin and used as a probe in Southern blot hybridization of restriction-digested P1 6386 DNA (Fig. 2 B). The number of detected hybridization products per digestion indicated that one potential OR gene was obtained by PCR. Subcloning and sequence analysis of a 3 kb hybridizing XbaI fragment (Fig.2 B) led indeed to the identification of a new mouse OR gene that we named OR-Z6, recalling its identification in the genomic context of the mouse line H-lacZ6. The intronless open-reading frame of OR-Z6 encodes a typical G-protein-coupled, seven transmembrane domain protein and contains many conserved amino acid motifs and single residues in specific positions characteristic of the large family of OR genes (Fig.3 A). This analysis demonstrated that the transgene construct in H-lacZ6 mice had indeed inserted close to an OR gene. OR-Z6 is the first cloned OR gene that maps to mouse chromosome 6.
OR-Z6 orthologs on human chromosome 7
Subsequent homology searches in the European Molecular Biology Laboratory database using the deduced OR-Z6 protein showed sequence similarities up to 66% with other OR genes but did not reveal any OR-Z6 homologs. Interestingly, when searching the High Throughput Genomic Sequences human genome database, we identified OR genes on human chromosome 7 that have a higher homology to OR-Z6 than to any other OR sequence yet reported and most likely represent human OR-Z6 orthologs. Human clone RP11-707F14 (AC073647) carries three OR-Z6 orthologs, the coding regions of which show 81–94% nucleotide sequence identity among themselves and 78–81% identity with the mouse OR-Z6gene. One of the human ORs shares 85% amino acid similarity withOR-Z6 (Fig. 3 A), whereas the deduced protein sequences of the other two OR-Z6 orthologs are interrupted by frame shifts and stop codons that may reflect pseudogenes or the draft quality of the sequence released. Consistent with previous reports (Ben-Arie et al., 1994; Asai et al., 1996; Sullivan et al., 1996; Trask et al., 1998; Strotmann et al., 1999; Tsuboi et al., 1999), the human OR-Z6 orthologs exist in a tandem array and are separated by 31 and 24 kb spacers.
OR-Z6 is a single OR gene at theHlz6 locus
To identify mouse OR-Z6 homologs, we performed high-stringency genomic Southern blot hybridization using a digoxigenin-labeled DNA probe of the OR-Z6 coding region and genomic DNA from both wild-type mice (129S3) and H-lacZ6 mice. It has been reported that members of a subfamily share at least 80% identity and cross-hybridize with one another (Lancet and Ben-Arie, 1993). The identical hybridization patterns obtained from both DNA types (Fig. 3 B) implied that the integration of thelacZ transgene did not delete or grossly modify theOR-Z6 locus in H-lacZ6 mice. Furthermore, the number of the hybridizing fragments obtained per digestion showed thatOR-Z6 in mouse is most likely a single gene in its subfamily but does not exclude the possibility that OR genes from a different subfamily reside proximal to the OR-Z6 gene. To address this question, we analyzed two BAC clones that extend ∼190 kb each upstream and downstream of the Hlz6 locus. The terms “upstream” and “downstream” refer to the 3′-flanking fragment that maps downstream of the transgene construct in H-lacZ6 mice and were only defined for the purpose of orientation. A computerized search in an end-sequenced mouse genomic BAC library (TIGR database) using the 3′-flanking fragment as a query yielded two BAC clones. Clones RPCI-23-282I16 (AQ932615) and RPCI-23-323H21 (AQ988378) showed 96% identity with the first 348 bp and the last 250 bp of the 3′-flanking fragment, respectively. Thus, the sequence of our transgene 3′-flanking fragment overlaps the two BAC clones and links them into a single contig. PCR analyses of both BAC clones, using the degenerate primers NL61 and NL63 that amplify the region between transmembrane domains 2 and 6 of OR genes (primer sequences were obtained from R. Reed, The Johns Hopkins University), yielded no other OR genes in addition to OR-Z6. Low-stringency PCR yielded PCR products from both BAC clones that were subcloned. A total of 55 subclones derived from both BAC clones were examined by restriction digestion, and 20 of these clones were subjected to sequence analysis (data not shown). No OR genes were identified on BAC PCI-23-323H21 that extends downstream of the Hlz6locus. All sequenced subclones deriving from BAC RPCI-23-282I16 that extends upstream of the Hlz6 locus were identical to theOR-Z6 sequence. Thus, OR-Z6 in mouse may be a single OR gene at the Hlz6 locus or the last member of a cluster that resides even farther upstream in a region not covered by BAC RPCI-23-282I16.
OR-Z6 expression in the OE
The tissue-specific expression of OR-Z6 was addressed by reverse transcription (RT)-PCR in three different tissues using the gene-specific primers N and C (Fig. 3 A) that span 899 bp of the OR-Z6 coding region. Amplicons of the expected size were only obtained after RT-PCR using RNA from OE and OB (Fig.3 C), demonstrating that OR-Z6 is an expressed OR gene and that its mRNA is present in olfactory tissue. No products were obtained from liver RNA or control reactions that omit reverse transcriptase during cDNA synthesis. Sequencing of the RT-PCR products showed that they were identical to the OR-Z6 sequence, initially derived from the genomic XbaI subclone.
To analyze the distribution of OR-Z6 mRNA in the OE, we performed in situ hybridization of serial coronal cryosections along the anterior-to-posterior axis using digoxigenin-labeled antisense and sense riboprobes. Riboprobes were generated from a subcloned BamHI fragment that spans the region between transmembrane domains 1 and 5 of the OR-Z6coding region (Fig. 3 A, top,restriction map). Hybridizing ORNs were predominantly located in the mid-to-caudal aspects of the OE. Interestingly,OR-Z6 + ORNs were strongly restricted to the medial epithelial recess, consisting of the tips of endoturbinate-II and -III and ectoturbinate-3, all of which are exposed at the central lumen of the nasal cavity (Fig.4, left column). Within this zone, OR-Z6 + neurons were randomly distributed as indicated by hybridization of single neurons and small clusters of approximately two to three neurons.OR-Z6 + neurons were evident throughout the depth of the OE and seemed only occasionally to segregate in the more apical OE. Few labeled neurons were found in other turbinates, the septum, or the vomeronasal organ (VNO) (data not shown). The VNO, an accessory olfactory system, mediates the detection of pheromones via specialized classes of pheromone receptors (Dulac and Axel, 1995; Herrada and Dulac, 1997; Matsunami and Buck, 1997; Ryba and Tirindelli, 1997; Leinders-Zufall et al., 2000). Although these receptors differ from the main ORs, occasional cross-hybridization of VNO neurons with OR sequences has been reported (Dulac and Axel, 1995;Ebrahimi et al., 2000). We never detected hybridization signals in sustentacular cells, in olfactory stem cells, or in control hybridizations using the OR-Z6 sense probe. TheOR-Z6 hybridization pattern was essentially bilaterally symmetrical between the left and right nose and reproducible among the mice investigated (n = 10 129S3 mice). Comparison of the OR-Z6 mRNA pattern with that of three ORs, each known to represent a single rostrocaudal zone, further illustrated the divergentOR-Z6 expression pattern (clones K21,K20, and L45 corresponding to zones 1, 2, and 3 were provided by L. Buck). In situ hybridization for each of the four probes in adjacent cryosections, followed by manual tracing of the individual mRNA patterns (data not shown), demonstrated thatOR-Z6 was not only confined to the medial aspect of zone 2 (according to the Buck nomenclature) (Ressler et al., 1993, 1994) but also overlapped with the ventral region of zone 1 and the dorsal aspect of zone 3. A similar restricted type of expression was reported for members of the OR37 subfamily in rat and mouse (Strotmann et al., 1992, 1999).
Expression of OR-Z6 and β-galactosidase in H-lacZ6 mice
The OR-Z6 expression pattern was strikingly similar to that of β-galactosidase in H-lacZ6 mice. This became apparent when we compared the X-gal staining pattern in H-lacZ6 mice with the OR-Z6 mRNA pattern in an age-matched wild-type mouse. Although β-gal+ ORNs in H-lacZ6 mice were more numerous than OR-Z6 + ORNs in wild-type mice, both patterns exhibited a high density of labeled neurons at the tips of the central turbinates (Fig. 4).LacZ + neurons were evident throughout the depth of the OE and did not exhibit obvious laminar segregation. As seen for OR-Z6, a few β-gal+ neurons were present in other turbinates, the septum, and the VNO (data not shown). Analyses of the rostrocaudal distribution in H-lacZ6 mice further demonstrated the related expression patterns of β-galactosidase andOR-Z6 (Fig. 5). Adjacent coronal cryosections from H-lacZ6 mice were collected as separate sets of every 10th section and subjected to either X-gal staining or OR-Z6 in situ hybridization before the labeled cells were counted. On the basis of the similar expression pattern, we included in situ hybridization for OR37E on one set of sections (the OR37E subclone was kindly provided by J. Strotmann). The distribution ofOR-Z6 + neurons in H-lacZ6 mice was the same as that determined in wild-type mice, implying that the insertion of the reporter gene, close to the OR-Z6locus, apparently does not affect the OR-Z6 expression pattern. Furthermore, despite the different total numbers of labeled neurons per reaction, the highest numbers of ORNs labeled for β-galactosidase, OR-Z6, or mOR37E were found at ∼2.40–2.75 mm posterior to the organ of Masera (a small patch ofOMP-positive ORNs separate from the main OE on both sides of the nasal septum) (Giannetti et al., 1995). This result confirmed and extended our previous findings that OR-Z6, β-galactosidase, and OR37 share a similar unusual spatial distribution pattern in the OE.
Neuronal colocalization of OR-Z6and lacZ
Because of the close chromosomal localization of OR-Z6and the lacZ transgene, together with their related expression patterns, we determined whether both genes are expressed in the same neurons. Double-labeling experiments in the OE of H-lacZ6 mice were performed by subjecting coronal cryosections to OR-Z6 in situ hybridization before anti-β-galactosidase immunohistochemistry (IHC). In agreement with the result obtained from single-labeling analyses,OR-Z6 + and β-gal+ neurons cosegregated at the tips of the central turbinates in all H-lacZ6 mice analyzed (n = 5). Surprisingly, only a small fraction of ORNs exhibited coexpression (OR-Z6 + and β-gal+) (Fig.6), whereas the majority of neurons was either OR-Z6-positive or β-galactosidase-positive. Among 2038 β-gal+ and 1070OR-Z6 + neurons counted in a P19 H-lacZ6 mouse, 15 neurons showed double-labeling (0.74% of the β-gal+ population). Similar results were obtained from a P12 H-lacZ6 mouse in which out of 1035 β-gal+ and 618OR-Z6 + neurons, 6 ORNs showed double-labeling (0.58% of the β-gal+population). The numbers presented derive from conservative counts, discarding any ambiguously double-labeled neurons. Furthermore, we noted that in situ hybridization before anti-β-galactosidase IHC decreases the number of immunolabeled neurons by 10–20%. Thus, the actual number of double-labeled cells detected could be slightly higher. Reasoning that the population oflacZ + neurons in H-lacZ6 mice coexpresses other OR genes in addition to OR-Z6, we performed double-labeling for β-galactosidase and OR37Ethat exhibits a similar zonal restriction. Interestingly, among 6605OR37E + neurons and 2226 β-gal+ neurons counted in a P35 H-lacZ6 mouse, 36 neurons exhibited coexpression (1.6% of the β-gal+ population). Despite the small number of H-lacZ6 mice investigated (n= 3), this analysis demonstrates that the expression of β-galactosidase does not exclude the coexpression of eitherOR-Z6 or OR37E in the same neuron.
OR-Z6+ ORNs project to a single glomerulus in each OB
The bulb projections of OR-Z6 +neurons were assessed by in situ hybridization of serial coronal sections of the entire OBs using32P-labeled antisense and sense riboprobes. To compare the position of hybridization signals among different mice, we defined the sections as percentages along the anterior-to-posterior axes of OBs. The first anterior section and the last posterior section containing glomeruli were set as 0 and 100%, respectively. Analyses of four 2-week-old wild-type mice (129S3) and four 4-week-old H-lacZ6 mice showed that axonal projections of OR-Z6 + neurons preferentially converge onto a single glomerulus in each OB (Fig.7).OR-Z6 + glomeruli were reproducibly located in the ventromedial portion of the anterior OBs, at approximately the rostral tip of the subventricular zone. Hybridization signals in the left and right bulbs of individual mice were nearly bilaterally symmetrical (Fig. 7). The positions ofOR-Z6-labeled glomeruli across all mice investigated varied between 19 and 26% of the bulb length, equaling a mouse-to-mouse variation of 250 μm along the anterior-to-posterior axis (two to three glomerulus widths). Furthermore, in each one of the two mouse strains investigated, we detected an additionalOR-Z6 + glomerulus in one of the bulbs that was either directly adjacent to and in the same coronal plane as the primary one (129S3 line) or 300 μm (three to four glomerulus widths) posterior to the primary one (H-lacZ6 line). The fact that these additional glomeruli were closely associated with the primary OR-Z6 + glomerulus in the ventromedial OB is consistent with the general idea that OR gene expression plays a role in guiding axons to specific target sites in the OB (Mombaerts et al., 1996; Wang et al., 1998). In contrast to other reports (Mombaerts et al., 1996), we have never observedOR-Z6 + glomeruli in the lateral OB in any mice investigated.
OR-Z6+ glomeruli in H-lacZ6 mice receive input fromlacZ+ axons
Previous reports showed that β-gal+ORNs in H-lacZ6 mice project to many ventromedially located glomeruli of the anterior OBs (Treloar et al., 1996; Walters et al., 1996a; Cummings et al., 2000). These glomeruli exhibit different degrees of lacZ expression because of the different number of lacZ + axons converging to each glomerulus. Of these, two to three glomeruli are heavily innervated by axons expressing the lacZ transgene, but it remains to be shown whether all axons entering these glomeruli expresslacZ. Because the positions we determined forOR-Z6 + glomeruli (between 19 and 26% of the bulb length) coincide closely with those of thelacZ + glomeruli mapped recently in H-lacZ6 mice (Cummings et al., 2000), we next asked whether axons of both ORN phenotypes also project to the same glomeruli. Alternate coronal cryosections (15 μm) of the entire OBs from young adult H-lacZ6 mice were collected as two sets and separately subjected either to in situ hybridization using a32P-labeled OR-Z6 antisense riboprobe or to anti-β-galactosidase IHC. Analysis of two H-lacZ6 mice showed that the singleOR-Z6 + glomerulus was always strongly labeled for β-galactosidase (Fig.8), coinciding with one of the two to three heavily labeled lacZ +glomeruli. Consistently, the OR-Z6 extra glomerulus detected in one of the H-lacZ6 bulbs (data not shown) exhibitedlacZ labeling as well. The β-gal+ fibers projecting to theOR-Z6 + glomerulus most likely represent axonal projections of the double-labeled neurons detected in the OE. Thus, this analysis demonstrates thatOR-Z6 + glomeruli receive axonal input from at least two phenotypically distinct ORN populations (OR-Z6 + and β-gal+;OR-Z6 + and β-gal−).
We demonstrate that locus-dependent mechanisms probably contribute to the zonal confinement of the OMP–lacZreporter in H-lacZ6 mice (Walters et al., 1996a,b) that mimics the cell-selective expression pattern known for some OR genes. Analysis of the transgene insertion site in H-lacZ6 mice enabled us to identify and characterize the new OR geneOR-Z6, the expression pattern of which resembles that of thelacZ reporter in H-lacZ6 mice.
OR-Z6 is a typical OR gene the deduced protein sequence of which shares many features with the large multigene OR family. Sequence similarities to other ORs were as high as 66% but without selectivity to any known OR, indicating that OR-Z6 defines a new subfamily. We mapped OR-Z6 and the lacZ transgene to the proximal region of mouse chromosome 6, close to Tcrb. Because this region shows synteny to human chromosome 7, whereTcrb maps to 7q35, we predicted that human OR-Z6orthologs map to 7q as well. Analyzing a recent release of the human genome database, we identified three OR-Z6 orthologs on human chromosome 7 that share ∼80% nucleotide sequence identity with mouse OR-Z6. In contrast, mouse OR-Z6 is likely a single OR gene in its subfamily, because we did not detect homologs by Southern blot analysis. The strong mouse-to-human homology suggests that OR-Z6 is the first member of a distinct class of ORs that have been conserved during mammalian evolution. Interestingly, a recent hybridization study indicated that a mouse ortholog of the canine OR subfamily CfOLF3 (Issel-Tarver and Rine, 1997) resides proximal to the marker Tbxas1 on mouse chromosome 6 (Carver et al., 1998). No ORs were assigned to this region, and the canine receptor is only poorly homologous toOR-Z6. Thus, OR-Z6 is the first cloned OR gene identified on mouse chromosome 6.
OR-Z6 expression is restricted to the medial OE
Any given OR gene is expressed by an ORN subpopulation confined to one of four rostrocaudal zones of the mouse OE (for review, seeMombaerts, 1999). By contrast, our in situ data demonstrate that OR-Z6 neuron distribution differs from this zonal rule. Although primarily confined to zone 2, OR-Z6expression overlaps with neighboring zones and exhibits an unusual medial restriction by occupying the tips of the central turbinates only. Within this small subzone, OR-Z6 neurons are randomly distributed, as implied by the coexistence of single-labeled neurons and small labeled ORN clusters. The overall pattern is bilaterally symmetrical between left and right nostrils and is present in both wild-type and H-lacZ6 mice. A similar distribution was reported for the mOR37 genes (Strotmann et al., 1999), the only other exception to the canonical OR expression patterns. In contrast to OR-Z6, mOR37 members share a six amino acid extension in the third extracellular domain (Strotmann et al., 1999; Hoppe et al., 2000) and map to mouse chromosome 4, near to two other OR genes that exhibit related expression patterns, but lack the loop extension. OR-Z6, on mouse chromosome 6, has no significant sequence identity with the mOR37 genes, despite the related expression pattern. Our results demonstrate that the zonal confinement of OR expression is more complex than suggested by the arbitrary division into four rostrocaudal zones. We provide evidence of a medially restricted compartment having OR-Z6 and the genes of the mOR37 locus as the first members. This location exposes OR-Z6 and mOR37 to the main nasal airflow, implying a specific biological significance for these receptors. Thus, it would be of interest to explore the ligand chemistry of OR-Z6 and mOR37.
OR-Z6+ neurons project exclusively to the medial OB
The unusual epithelial distribution is also reflected in OBs where axonal projections of OR-Z6 + neurons converge preferentially onto a single ventromedial glomerulus, as demonstrated by in situ hybridization. We occasionally identified secondary OR-Z6 +glomeruli (see also Royal and Key, 1999; Gogos et al., 2000;Strotmann et al., 2000). Mapping analyses of the entire bulbs showed that OR-Z6 + glomeruli always reside ventromedial and within 19–26% of the rostrocaudal axis of the OBs. The close association of primary and secondary glomeruli reflects the topographic organization of the epithelium-to-bulb projections ofOR-Z6 + neurons, supporting the idea that OR proteins contribute to selecting specific target sites in the OBs (Mombaerts et al., 1996; Wang et al., 1998). Although their physiological significance is unknown, the presence or absence of extra glomeruli clearly demonstrates interindividual variation in olfactory coding. In contrast to reports indicating that axonal projections of ORNs expressing the same receptor each converge onto one lateral and one medial glomerulus per OB (Ressler et al., 1994; Vassar et al., 1994; Mombaerts et al., 1996), we have never observed lateralOR-Z6 + glomeruli in the OBs of any mice. As for OR-Z6, ORNs expressing members of the mOR37 family (Strotmann et al., 2000) also project to single ventromedial glomeruli. In analogy to tracing studies investigating the patterns of epithelium-to-bulb projections (Astic and Saucier, 1986; Schoenfeld et al., 1994), we conclude that the ventromedial position for OR-Z6 + andmOR37 + glomeruli is consistent with the medial restriction of the corresponding neurons in the OE.
The OMP–lacZ-transgene mimics theOR-Z6 expression pattern
The expression patterns of OR-Z6 and the transgene in H-lacZ6 mice exhibit strong similarities, are confined to the tips of the central turbinates, and exist as separate, but overlapping, ORN populations. WhereasOR-Z6 + neurons project primarily to a single glomerulus per OB, lacZ +neurons, which are twice as numerous, converge to a variety of ventromedial glomeruli, exhibiting different degrees of lacZlabeling. The overall distribution ofOR-Z6 + neurons and their bulbar projections are indistinguishable in H-lacZ6 and wild-type mice, demonstrating that insertion of theOMP–lacZ construct close to the OR-Z6locus has little effect on the OR-Z6 expression pattern. Despite their similar zonal confinement, the small number of double-labeled ORNs detected in H-lacZ6 mice (1–2%) shows that coexpression of lacZ and either OR-Z6 ormOR37E is independent and that transgene expression does not interfere with OR gene expression.OR-Z6 + glomeruli in the OBs of H-lacZ6 mice are always positive for β-galactosidase. Because the number of lacZ + fibers in OR-Z6 + glomeruli approximates the number of double-labeled ORN cell bodies in the OE, we conclude that these glomeruli receive input from two differentOR-Z6 + ORN phenotypes (OR-Z6 + andlacZ −;OR-Z6 + andlacZ +), suggesting that singlylacZ-labeled neurons (OR-Z6 − andlacZ +) that converge onto neighboring glomeruli express other ORs. This idea is consistent with reports demonstrating the importance of OR expression for precise axonal convergence (Mombaerts et al., 1996; Wang et al., 1998).LacZ + neurons in H-lacZ6 mice target a large number of glomeruli the overall topography of which is reestablished with remarkable accuracy after chemical deafferentation (Cummings et al., 2000). Thus, we propose that glomerular targeting of lacZ +neurons in H-lacZ6 mice relies on coexpression of OR genes, preferentially directed to the small medial subzone, of whichOR-Z6 and mOR37E are the only two known examples. Furthermore, we speculate that the number oflacZ + glomeruli corresponds to the number of different OR genes expressed by thelacZ + neuron population. The different levels of lacZ labeling in H-lacZ6 glomeruli may reflect the different degrees to whichlacZ + neurons coexpress other OR genes.
Locus-dependent transgene expression in H-lacZ6 mice
The similar spatial distribution oflacZ + andOR-Z6 + neurons implies that the expression of both genes is linked. Because the transgene construct bears no OR gene identity, the question arises as to what molecular mechanisms lead to this pattern. The H-OMP–lacZconstruct consists of a truncated OMP promoter containing the proximal olf-1 (O/E-1) binding site (Kudrycki et al., 1993; Wang and Reed, 1993), the lacZ coding region, and an SV40 polyadenylation sequence. Transgene expression in H-lacZ3 mice, another mouse line generated with this construct, showed that the truncated OMP promoter serves as a minimal promoter, providing temporally and spatially correct expression in all mature ORNs (Walters et al., 1996a,b). In contrast, the expression pattern observed in H-lacZ6 mice implies the involvement of hierarchical mechanisms, capable of overriding the properties of theOMP promoter in the lacZ transgene and restricting lacZ expression not only to a specific zone but also to an ORN subpopulation within this zone, as opposed to the global ORN expression observed in H-lacZ3 mice. Zonally active transcription factors, environmental cues, or lineage predetermination possibly contribute to this patterning. However, theOR-Z6-like transgene expression in H-lacZ6 mice is likely caused by the site of transgene integration that occurred close to OR-Z6. So-called position effects are well known in transgenic mice (Festenstein et al., 1996; Milot et al., 1996). This idea is consistent with studies showing that OR genes closely linked at specific loci exhibit similar expression patterns (Strotmann et al., 1999; Tsuboi et al., 1999; Serizawa et al., 2000). A transgenic approach demonstrated that the zonal affiliation of an OR promoter-driven reporter construct was dependent on the chromosomal insertion site (Qasba and Reed, 1998). Thus, we propose that the transgene pattern in H-lacZ6 mice is locus-predetermined and that the OMP–lacZ construct that randomly inserted near OR-Z6 reports on its genomic environment. Similar locus-dependent mechanisms may apply to the mOR37genes on mouse chromosome 4. In conclusion, our study strongly suggests that the expression of OR-Z6 and theOMP–lacZ transgene is regulated via common, locus-dependent mechanisms that exemplify the locus-dependent regulation of OR gene expression.
This research was supported in part by National Institutes of Health Grants DCD03112 (F.L.M.) and U54NS39407 (E.W.) and by the National Cancer Institute, Department of Health and Human Services (N.G.C., N.A.J., and D.J.G.). We thank Drs. A. Puche and O. Buiakova for helpful discussions and S. Moussavi and Deborah B. Householder for excellent technical assistance.
Correspondence should be addressed to Dr. F. L. Margolis, University of Maryland at Baltimore, School of Medicine, Department of Anatomy and Neurobiology, 685 West Baltimore Street, Health Science Facility 273, Baltimore, MD 21201. E-mail:.