Physiol. Genomics Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Physiol. Genomics 16: 67-81, 2003. First published October 21, 2003; doi:10.1152/physiolgenomics.00117.2003
1094-8341/03 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
16/1/67    most recent
00117.2003v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (17)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Genter, M. B.
Right arrow Articles by Aronow, B. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Genter, M. B.
Right arrow Articles by Aronow, B. J.
Received 17 June 2003; accepted in final form 8 October 2003.
Physiological Genomics 16:67-81 (2003)
1094-8341/03 $5.00 © 2003 American Physiological Society

Microarray-based discovery of highly expressed olfactory mucosal genes: potential roles in the various functions of the olfactory system

Mary Beth Genter1, Paul P. Van Veldhoven2, Anil G. Jegga3, Bhuvana Sakthivel3, Sue Kong3, Kristin Stanley3, David P. Witte3, Catherine L. Ebert3 and Bruce J. Aronow3

1 Department of Environmental Health, University of Cincinnati, Cincinnati, Ohio 45267-0056
2 Katholieke Universiteit Leuven, Department of Molecular and Cellular Biology, Pharmacology Division, B-3000 Leuven, Belgium
3 Children’s Hospital Medical Center, Cincinnati, Ohio 45229


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
We sought to gain a global view of tissue-specific gene expression in the olfactory mucosa (OM), the major site of neurogenesis and neuroregeneration in adult vertebrates, by examination of its overexpressed genes relative to that in 81 other developing and adult mouse tissues. We used a combination of statistical and fold-difference criteria to identify the top 269 cloned cDNAs from an array of 8,734 mouse cDNA elements on the Incyte Mouse GEM1 array. These clones, representing known and poorly characterized gene transcripts, were grouped according to their relative expression patterns across the other tissues and then further examined with respect to gene ontology categories. Approximately one-third of the 269 genes were also highly expressed in developing and/or adult central nervous system tissues. Several of these have been suggested or demonstrated to play roles in neurogenesis, neuronal differentiation, and/or neuronal migration, further suggesting that many of the unknown genes that share this expression pattern may play similar roles. Highly OM-specific genes included a palate, lung, and nasal epithelium carcinoma-associated gene (Plunc); sphingosine phosphate lyase (Sgpl1), and paraoxonase 1 (Pon1). Cell-type-specific expression within OM was established using in situ hybridization for several representative expression pattern clusters. Using the ENSEMBL-assembled mouse genome and comparative genomics analyses to the human genome, we assigned many of the unknown expressed sequence tags (ESTs) and poorly characterized genes to either novel or known gene products and provided predictive classification. Further exploration of this database will provide additional insights into genes and pathways critical for olfactory neurogenesis, neuronal differentiation, olfaction, and mucosal defense.

cDNA microarray; gene expression profiles; gene discovery; bioinformatics; neurodevelopment; Plunc; sphingosine phosphate lyase; paraoxonase


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
THE OLFACTORY MUCOSA (OM) is positioned at the junction of the respiratory tract and nervous system and performs functions integral to both organ systems. Olfactory sensory neurons are responsible for odorant transduction. In addition, inhaled air can be detoxified of toxic agents by mucociliary mechanisms and metabolic enzymes in the nasal respiratory and olfactory tissues. The OM is unique with respect to the rest of the nervous system, as it has a basal cell population that provides a continuous supply of new olfactory neurons. Thus neurons lost to injury or toxic insult can be replaced (13, 14, 27, 33). At the present time, we lack an understanding of the overall genomic expression patterns that allow for this to be accomplished.

We have used the power of microarray-based genomics to examine gene expression in the mouse OM relative to that of many other tissues. For example, other studies utilizing the University of Cincinnati-Children’s Hospital Medical Center (UC-CHMCC) Mouse Tissue Specific Gene Expression Database described gene expression in liver development and regeneration (42) and in various anatomical segments of the gastrointestinal tract (6). In the present study we are specifically interested in understanding gene expression patterns that distinguish the OM, with its capacity to support continuous neurogenesis, from more quiescent regions of the brain, as well as from adjacent tissue in the nasal cavity, namely, the nasal respiratory mucosa. We hypothesized that the molecular definition of these different, but closely related, tissues could be performed by comparing the expression and function of large numbers of genes in OM, nasal respiratory mucosa, and brain. This approach revealed a group of genes with expression highly restricted to the OM. Other clusters of genes were highly expressed in OM and other tissues, including in 1) OM and olfactory bulb; 2) OM and lung; 3) OM and liver; 4) OM and respiratory mucosa; and 5) an OM and skin gene clusters. We also demonstrated the utility of this approach for gene discovery with the detection of novel genes likely to be relevant in the life cycle of olfactory neurons.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Animals, tissue samples, and RNA preparation.
Adult OM, olfactory bulb, and nasal respiratory mucosa were obtained from 6- to 8-wk-old male C57BL/6 mice (Jackson Laboratories) as part of an institutional consortium effort to develop a generalized mouse gene expression database. This database includes a wide range of developing, normal adult, and diseased adult mouse tissues, including multiple brain subregions, dorsal root ganglion, heart, lung, kidney, liver, reproductive and endocrine organs, immune organs, muscle, and skin. The ethmoid turbinates and caudal one-third of the nasal septum served as the source of OM RNA, and the nasoturbinates, maxilloturbinates, and anterior two-thirds of the nasal septum were the source of respiratory mucosal RNA (for anatomy, see Ref. 87). Mice were killed by carbon dioxide inhalation according to institutional and national animal care guidelines, and tissues were rapidly dissected and snap frozen in liquid nitrogen. Frozen tissues, pooled from eight mice, were homogenized on ice with a stainless steel tissue grinder, and total RNA was prepared using TRI Reagent (Molecular Research Center, Cincinnati, OH), following the manufacturer’s protocol. Total RNA was reprecipitated with ethanol/sodium acetate and resuspended in DEPC-treated water. Polyadenylated [poly(A)+] RNA was prepared from total RNA using Oligotex (Qiagen, Valencia, CA). Poly(A)+ RNA was quantitated using RiboGreen dye (Molecular Probes, Eugene, OR) and checked for RNase degradation using an Agilent Bioanalyzer 2100 (Agilent Technologies, Palo Alto, CA). A total of 600 ng of poly(A)+ RNA (50 ng/µl) for each tissue was submitted for cDNA labeling and microarray hybridization.

Fluorescent probe preparation and microarray hybridization.
Probe preparation and microarray hybridization were performed by Incyte Genomics (Palo Alto, CA) using the GEMBright random primer reverse-transcription labeling kit (Incyte Genomics). Labeled cDNA was prepared from each poly(A)+ RNA sample using nucleotides labeled with the fluorescent dye Cy5. Labeled cDNAs were then subjected to competitive hybridization to mouse GEM1 microarrays, composed of spotted cDNAs corresponding to 8,638 sequence-verified IMAGE expressed sequence tag (EST) clones (500–5,000 nucleotides in length) with a low redundancy rate with respect to their corresponding gene products (a complete list of the genes represented on the GEM1 microarray can be found at http://microarray.uc.edu/DataBases/Incyte_Mouse_GEM1.xls). For all samples in the database, hybridizations were performed in competition with Cy3-labeled mRNA from whole postnatal day 1 (P1) mouse. This reference sample was chosen based on independent tests of its ability to generate reproducible competitive hybridizations from three different P1 mouse isolates (data not shown). For each mRNA sample, duplicate arrays were hybridized. Additional dye-flip control experiments indicated a low dye effect, particularly compared with other Cy3/Cy5 labeling methods (data not shown). Dye effect was, however, subtracted from relative sample-specific gene expression based on per gene median normalization across a large number of samples (see below). Fluorescence intensity analyses and background subtraction were performed using an Axon Instruments scanner and GenePix software.

Data analyses.
Primary quantitative data, spot geometry, and background fluorescence were examined using Incyte GEMTools software. Defective cDNA spots (irregular geometry, scratched, or less than 40% area compared with average) were eliminated from the data set. All data sets were normalized first using a balancing coefficient of the median of all Cy5 channel measurements divided by the median of all Cy3 channel measurements. Each microarray contained 192 control genes present as either nonmammalian single gene "spikes" or complex targets consisting of pools of genes expressed in most cell types, and each experimental mRNA sample was augmented with incremental amounts of nonmammalian gene RNA (2-fold, 4-fold, 16-fold, etc.) to permit assessment of the dynamic range attained within each microarray. Less than twofold variation was observed across the microarray series with respect to the 192 control genes (data not shown), providing additional support for the feasibility of interarray comparisons to detect genes regulated in a tissue-specific manner. Second-stage data analyses were carried out using GeneSpring software (Silicon Genetics, Redwood City, CA). A total of 269 cDNAs corresponding to genes highly expressed in the adult mouse OM were identified based on expression that was >=1.7-fold their median expression in all other tissues in the database. We found that this cutoff was capable of identifying many known important genes and was more efficient than ANOVA-based approaches (59), principally because of the number of genes expressed in OM that are also expressed, albeit relatively selectively, in other compartments such as liver and central nervous system.

The relative expression patterning of the 269 OM-overexpressed cDNAs was determined using the hierarchical tree clustering algorithm as implemented in the GeneSpring program using Pearson correlation applied to the log ratio of gene expression values. Relative gene expression ratio values were obtained using a three-step normalization strategy consisting of a per-spot normalization in which sample channel was divided by the whole mouse reference channel followed by a LOWESS ("locally weighted scatter plot smoothing") intensity-dependent normalization, and then per gene across the series of probe arrays using the median expression ratio observed for the gene across the entire University of Cincinnati-Children’s Hospital Medical Center (UC-CHMCC) Mouse Tissue Specific Gene Expression Database.

The measured intensity of each gene was divided by its control channel value in each sample. When the control channel value was below 10.0, the data point was considered invalid. Intensity-dependent normalization was also applied, where the ratio was reduced to the residual of the LOWESS fit of the intensity vs. ratio curve as implemented in GeneSpring 4.2.1. The 50th percentile of all measurements was used as a positive control for each sample; each measurement for each gene was divided by this synthetic positive control, assuming that this was at least 0.01. Each gene was normalized to itself by making a synthetic positive control for that gene and dividing all measurements for that gene by this positive control, assuming it was at least 0.01. This synthetic control was the median of the gene’s expression values over all the samples.

Hierarchical tree clustering allowed grouping of genes based on relative expression across the sample series; we used this approach to partition genes highly expressed in the OM into those that were also highly expressed in other tissues. For example, this allowed for clear identification of 105 of the 269 genes with strong expression in both OM and nervous system regions. A similar result could be obtained by using a Venn diagram function within GeneSpring to obtain the conjunction (Venn intersection) of the 269 OM genes and those identified by Student’s t-test ANOVA as being nervous system-expressed vs. non-nervous system-expressed.

Sequence analysis, identification, and functional classification of "olfactorome" genes.
The genes represented among the 269 olfactorome gene cDNAs were identified and categorized using a variety of resources. The Incyte GEM1 microarray contains a set of sequenced cDNA clones that have been mapped to representative UniGene sequences (http://www.ncbi.nlm.nih.gov/UniGene/Mm.Home.html) and BLASTN searches (http://www.ncbi.nlm.nih.gov/BLAST/) vs. the nonredundant nucleotide database over a period from July to November, 2002. This procedure allowed 1) matching of the ESTs to full-length gene products (principally using UniGene), 2) their assignment to corresponding representative mRNA from RefSeq (http://www.ncbi.nlm.nih.gov/RefSeq), and then 3) classification of the encoded protein function using SwissProt (http://us.expasy.org/sprot/), Gene Ontology (http://www.geneontology.org), BLINK (http://www.ncbi.nlm.nih.gov/sutils/static/blinkhelp.html), HomoloGene (http://www.ncbi.nlm.nih.gov/HomoloGene), and additional Medline-based literature analyses (http://www.ncbi.nlm.nih.gov/PubMed). For unannotated EST sequences, additional BLASTN and BLAT (http://genome.ucsc.edu) searches were performed to identify genomic region hits, and potential gene hits were based on the use of University of California, Santa Cruz (UCSC) Genome Browser (http://genome.ucsc.edu) and ENSEMBL (http://www.ensembl.org/) resources (December, 2002).

Gene annotations (functional classification and groupings) were based on high scoring homologies of the corresponding encoded proteins to known proteins using biochemical function and molecular process concepts represented in SwissProt and Gene Ontology into the following categories: cytoskeletal, metabolic enzymes, xenobiotic metabolizing enzymes, intracellular protein trafficking, apoptosis related, cell cycle control related, immune system/inflammatory response related, signal transduction, extracellular matrix/cell adhesion, proteinase inhibitors, DNA/RNA binding/transcription factors, calcium homeostasis, membrane proteins, and others (Table 1). From the 269 gene cDNAs that we designated as top members of the "olfactorome," we identified 41 cDNAs/ESTs that lacked useful genomic annotations or designation as a gene. We subcategorized these 41 cDNAs/ESTs into 6 major groups, principally based on their expression pattern: highly expressed in 1) OM alone; 2) olfactory bulb and OM; 3) OM and respiratory mucosa; 4) whole brain and olfactory bulb and OM; 5) whole brain and olfactory bulb; and 6) a mixed patterns. Each of these cDNAs/ESTs was then subjected to a BLAT search against the UCSC Genome Browser (on Mouse Feb. 2003 Freeze) and BLASTN searches against Celera (http://www.celera.com/) and ENSEMBL databases for full-length known and putative mouse mRNAs.


View this table:
[in this window]
[in a new window]
  		Table 1. Characterization of highly expressed olfactory mucosal genes (with GenBank/RefSeq accession numbers)

 
Identification of cis-elements in coordinately expressed genes.
To identify putative cis-acting regulatory regions and elements in genes coordinately expressed in the olfactory system, we selected three genes from each of the first three expression subcategories above and subjected each gene to mouse-human ortholog pair analysis for conserved cis-regulatory elements (40). To do this, the complete genomic sequences of the selected mouse genes and the corresponding human orthologs were extracted from the Celera/ENSEMBL/UCSC mouse and human databases, respectively. Advanced PipMaker (http://bio.cse.psu.edu) was used to compute alignments of similar regions in the mouse-human orthologous DNA sequences using the chaining option (67). The TRANSFAC database, a database of eukaryotic transcription factors and DNA-binding site profiles (http://transfac.gbf.de) (88), was utilized in the program MatInspector (Professional ver. 4.3, 2000; http://www.genomatix.de/) to locate putative transcription factor binding sites in each of these genomic sequences. We then compared genomic regions within and across the three groups to identify transcription factor binding sites in common.

In situ hybridization analysis of highly expressed OM genes.
In situ hybridization was performed for localization of selected highly expressed genes using 35S-labeled probes which were generated from the Incyte clones (Incyte Pharmaceuticals; St. Louis, MO) available from the University of Cincinnati Microarray Core (http://microarray.uc.edu). The probes were as follows: W08172 for sphingosine phosphate lyase (Sgpl1); AA028678 for Plunc; W17967 for Pon1; AA220582 for cytochrome P-450 2f2 (Cyp2f2); AA498773 for crp ductin (Dmbt1); AA474964 for lactotransferrin (Ltf); and W11846, which was originally an EST, but which has been associated with the recently characterized Alms1 locus (16, 17). We sequenced each clone from the T3 and T7 primer sites and found two chimeric clones: W08172 contained an unidentified sequence at the T3 (antisense) end, and AA220582 had high identity with CYP2f2 from the T3 (antisense) end and with Dmbt1 from the T7 (sense) end. To generate a vector containing only the Sgpl1 sequence for use as a probe, a restriction map of the Sgpl1 sequence was generated (based on Ref. 65), and based on known restriction sites, the NotI linearized plasmid was cut with BbsI, blunt-end subcloned, and validated for contaminating fragment removal using EcoRI/BamHI and HindIII/DraI. This modified clone was used as a template for 35S-labeled sense and antisense probe synthesis.

Sphingosine phosphate lyase activity measurements.
Sphingosine phosphate lyase (Sgpl1) activity was measured in tissues isolated from male Long-Evans rats because of the relative abundance of several tissues, including OM, in rat vs. mouse, and for ease of comparison with previously published results (79). Tissues were homogenized in 0.25 M sucrose, 5 mM MOPS-NaOH buffer, pH 7.2, 1 mM DTT, 1 mM EDTA, either with a Dounce homogenizer or, for more refractive tissues, a Polytron PT7 device (Kinematica Benelux). Sgpl1 activity was measured as described before using [4,5-3H]sphinganine-1-phosphate (80).

Sgpl1 antiserum preparation.
Antiserum against (human) Sgpl1 was obtained as follows. A His-tagged human Sgpl1 (amino acids 59–568) fusion protein, expressed in DH5{alpha} Escherichia coli cells transformed with plasmid pVB001 (81) and induced with 0.2% (wt/vol) D-arabinose, was purified by means of Ni-NTA agarose (Qiagen), followed by SDS-PAGE and blotting to nitrocellulose. Portions of the blot containing ~100 µg poly-His-fusion protein were dissolved in DMSO (1), mixed with an equal volume of Freund’s complete adjuvant and injected subcutaneously in a New Zealand White rabbit after procurement of an aliquot of preimmune serum. Booster injections were given every 4 wk with 50 µg of protein mixed with incomplete adjuvant. Ten days after the third booster, the animal was bled. The final preparation recognized a single band of 65 kDa in blots prepared from human, rat, and mouse liver (P. P. Van Veldhoven, unpublished data).

Immunohistochemistry.
Sections (5 µm in thickness) were cut from archived paraffin-embedded nasal cavity sections from untreated male Long-Evans rats and male C57BL/6J mice. Tissues had been prepared for embedding by fixation in 10% neutral buffered formalin and decalcified in either 10% formic acid or 0.3 M EDTA (pH 7.4). Anti-Cyp2f2 antiserum (previously described, see Ref. 63) was provided by Dr. Gerold Yost (University of Utah). Immunohistochemistry was performed with 1:100 dilution of preimmune serum or 1:100 dilution of Sgpl1 antiserum or 1:2,500 dilution of anti-Cyp2f2 antiserum. Immunoreactivity was detected using anti-rabbit HRP-conjugated secondary antibody (Dako, Carpenteria, CA) and aminoethylcarbazole as chromogen as previously described (29).

Data archive.
Gene identities, expression data, cluster groups, and the functional categories of the dynamically regulated genes are available on our microarray database web server [http://genet.cchmc.org; log in (as guest or with username: olfepi; password: olfepi), select "IncyteMouseGEM1 (127_CloneID)," then select the "Olfactorome" subdirectory from the "Gene Lists" folder; select the list of genes of interest, and click on the paperclip icon to the right of the list].


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Identification of genes highly expressed in adult OM.
To define genes important for neurogenesis and OM homeostasis, we identified arrayed cDNAs that hybridized to adult mouse OM cDNA at a level at least 1.7-fold higher than their median expression in all other tissues. We found 269 cDNAs and ESTs in OM with increased expression in replicate hybridizations compared with reference. Of the known genes, the highest expression in OM was noted for Plunc (86). Spgl1, Pon1, Cyp2f2, and Ltf were among other highly expressed OM genes. We subjected the log2-transformed ratios of the overexpressed genes to hierarchical tree cluster analysis and clustered the genes into those that were highly expressed only in OM vs. those also highly expressed in other tissues (Fig. 1).



View larger version (108K):
[in this window]
[in a new window]
  		Fig. 1. Hierarchical tree clustering of the 269 highly expressed olfactorome genes and expressed sequence tags (ESTs). Intensity in the red and blue color range indicates upregulated and downregulated RNAs, respectively. Each row represents a gene. Each column represents an individual tissue. Each tissue sample RNA was hybridized in duplicate with the mean value shown. The position of the olfactory marker protein gene (Omp), a marker of mature olfactory neurons, is indicated. Note that there is a large cluster of olfactory mucosal (OM) genes that are also highly expressed in the brain (larger box). Brain regions included in this cluster are embryonic day 18 brain, postnatal day 2 brain, and the following regions of the adult mouse brain: hippocampus, hypothalamus, nucleus accumbens, olfactory bulb, striatum, cerebellum, spinal cord, and dorsal root ganglion. Genes included in this cluster include synuclein-{alpha} (Snca; NM_009221); calcium channel ß3-subunit (Cacnb3; NM_007581); leucine-rich repeat protein 1, neuronal (Lrrn1; NM_008516); Tau4 (NM_010838); and adducin 2ß (Add2; NM_013458). There are also separate clusters of genes with high coexpression in OM and liver [smaller box; e.g., paraoxonase 1 (Pon1; NM_011134, shown); aldehyde dehydrogenase 1 (Adh1; NM_007407); and glycine N-methyltransferase (Gnmt; NM_010321)]; gastrointestinal tract, skin, and lung (e.g., cytochrome P-450 2f2; Cyp2f2; NM_007817, shown). Both genes and samples are clustered using Pearson correlation as implemented in GeneSpring 6.

 
The 269 highly expressed OM genes showed both temporally and spatially restricted expression in other regions of the central nervous system. Seventy-seven genes were highly expressed in OM and in at least one brain region. Some of these genes exhibited very different behavior from one brain region to another. For example, lumican (Lum) and procollagen, type I, {alpha}1 (Colla1) were highly expressed only in OM, postnatal day 2 brain and dorsal root ganglion, and significantly underexpressed in many other brain regions represented in the analysis. Genes that were highly expressed in both olfactory bulb and OM encoded many cytoskeletal proteins and proteins associated with vesicular trafficking [e.g., RAB3B (Rab3b); vesicle-associated membrane protein 2 (Vamp2); pantophysin (Pphn); and amyloid-ß (A4) precursor protein binding (Aplp2) (35, 47, 53, 75)], axonal guidance [e.g., plexin A3 (Plxna3) and dihydropyrimidinase-like 3 (Dpysl3); NM_009468; orthologous to unc-33, which is associated with axonal guidance in Caenorhabditis elegans (12, 52, 68)], and signal transduction [e.g., the neuronal early immediate gene homer (Homer2-pending) and calcium channel ß3-subunit (Cacnb3)]. Fifteen genes and ESTs were highly expressed in both OM and in E18.5 mouse brain. Of the known genes highly coexpressed in OM and in developing brain, eight are clearly implicated in nervous system function, neuronal development, and/or epithelial differentiation. These are tau (Mapt), stathmin (Stmn4; 50, 51, 57, 60), fatty acid binding protein 7 (Fabp7; 23), Meg3 (66), pleiotrophin (Ptn; 74), Mad4 (39), Dpysl3, and Plxa3.

Sgpl1 activity measurements.
Several tissues in the mouse gene expression database displayed relatively high Sgpl1 gene expression, including OM, thymus, jejunum, ileum, and skin (Table 2). In a screening of the main organs of the rat, highest activities were previously detected in intestinal mucosa, Harderian gland, and liver (79). Measurement of the specific activity of Sgpl1 in rat OM in the present study confirmed that it was indeed high compared with other rat tissues (Table 2).


View this table:
[in this window]
[in a new window]
  		Table 2. Sphingosine phosphate lyase (Sgpl1) gene expression and activity in selected tissues

 
In situ hybridization and immunohistochemistry studies.
We performed in situ hybridization to independently confirm results of the gene expression analysis and to localize several previously undescribed OM genes, Sgpl1; W11846, originally an EST that has recently been associated with the Alms1 locus (16, 17); Dmbt1, Cyp2f2, and Pon1. The expression of Pon1 and Cyp2f2 was highly localized to the OM airway surface and in acinar cells and ducts of Bowman’s glands (Fig. 2). This distribution was further confirmed by immunohistochemistry for Cyp2f2. This is a classic distribution for many xenobiotic metabolizing enzymes in OM (10). In contrast, Sgpl1 and Alms1 were distributed in the olfactory epithelium (i.e., above the epithelial basement membrane) but not in the subepithelial compartment (Fig. 2). Immunohistochemistry suggested Sgpl1 protein localization to the perinuclear region of mature olfactory neurons (Fig. 2); colocalization studies with olfactory marker protein (OMP) to confirm localization to mature neurons were deemed unnecessary based on the localization of the histochemical reaction products in tissue sections. Dmbt1 expression was localized in the lateral nasal gland (Fig. 2). Plunc expression was included as a positive control for the in situ methodology, as its high and punctate expression has been previously described in both developing and adult OM (86) (Fig. 2).



View larger version (134K):
[in this window]
[in a new window]
  		Fig. 2. Localization of novel genes and ESTs in the OM using in situ hybridization and immunohistochemistry. A: cytochrome P-450 2f2 (Cyp2f2) is shown by in situ hybridization in this bright-field photomicrograph to localize in the Bowman’s glands (black arrow) and along the epithelial surface (white arrow) of the dorsal septum and the turbinates lining the dorsal medial airways of the mouse. B: immunohistochemistry using an anti-Cyp2f2 antibody confirms localization in apical cytoplasm of sustentacular cells (white arrow) and duct and acinar cells of Bowman’s glands (black arrow). C: in situ hybridization with antisense Sgpl1 reveals expression in the olfactory epithelial layer (*), but not in the subepithelial compartment (**), suggesting that Sgpl1 is present in either olfactory sensory neurons or in sustentacular (supporting) cells. D: in situ hybridization with Sgpl1 sense probe does not result in hybridization in the olfactory epithelial layer (*) or the subepithelial compartment (**). E: immunohistochemistry with anti-Sgpl1 antibody reveals localization in the perinuclear layer of mature olfactory sensory neurons (example indicated with arrow). F: Alms1 antisense probe similarly localizes to the olfactory epithelial compartment (*), but not subepithelial structures (**). G: Dmbt1 localizes to the lateral nasal gland, but not in OM (white arrow). H: Plunc expression (arrow) was the highest of any gene in the OM compared with other genes in the mouse gene expression database. Its expression is punctate and extremely high in olfactory epithelium (often displaying high expression at the junctions of OM and respiratory mucosa), consistent with previous localization studies (86).

 
Functional classification of genes highly expressed in the adult mouse olfactory system.
We classified each of the highly expressed genes based on biochemical function of the encoded protein (Table 1). Overall, the function of ~60% of the genes/ESTs could be identified based on public databases. The largest fraction of identified genes is related to metabolism, with the encoded proteins participating in metabolism of both endogenous and exogenous substrates (although we separated metabolic enzymes and xenobiotic metabolizing enzymes into separate categories, we realize that these distinctions are not absolute, i.e., a given enzyme might well metabolize endogenous and exogenous substrates). The predicted protein products of these genes bear homology to members of protein families that participate in DNA-dependent regulation of transcription, cell differentiation, positive regulation of cell proliferation, carbohydrate transport, cell adhesion, metabolism, or signal transduction.

Gene discovery in the olfactory system.
The use of EST and cDNA microarrays corresponding to unknown genes provides the opportunity for gene discovery relevant to the multiple cellular components of OM. As shown in Fig. 3, a total of 137 cDNAs and ESTs out of the 269 could be associated with a well-annotated mouse gene (having a gene symbol and an mRNA). Of these 137 genes, 103 had a corresponding NCBI RefSeq protein entry for which an ortholog could be identified. Of the 34 mouse genes that did not show a human entry in the initial search, 27 of them could be mapped to the human genome to a likely human ortholog. However, there were 7 mouse genes for which no human ortholog could be found after BLASTN and BLAT searches against human genome. Of the remaining 120 cDNA/ESTs, of 269 that were poorly annotated, 47 could be assigned to a gene with a valid symbol and an mRNA entry in the RefSeq database of GenBank (largely representing Riken long clones). To find the human orthologs for these 47 mRNAs, BLAT and BLASTN searches were performed against the human genome using corresponding full-length Riken clone sequence or mouse genomic sequence. Taken together, of the total 269 olfactorome cDNA/ESTs, there were 31 for which no human ortholog is available in the present databases.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 3. Flow chart for the annotation of 269 gene elements of the olfactorome. A total of 137 cDNA/ESTs out of the 269 could be associated with a well-annotated mouse gene, and of these 137 genes, for 103 the human ortholog could be identified. Of the 34 mouse genes which did not show a human entry in the initial search, 27 of them could be mapped to the human genome. However, there were 7 mouse genes for which no human ortholog was found even after a BLASTN and BLAT search against human genome. Of the remaining 120 cDNA/ESTs out of 269, and that were poorly annotated, there were 47 that had an approved gene symbol and an mRNA entry in the RefSeq database of GenBank. To find the human orthologs for these 47 mRNAs, a BLAT and BLASTN search was performed against the human genome. There were 46 human hits, and for one there was no human ortholog identified. The 73 cDNA/ESTs out of 120 with no approved gene symbol association were also subjected to a BLAT and BLASTN search against human genome. For 49, we could identify an associated human counterpart. However, 23 were unmatched. One (CA490663) was withdrawn from the GenBank database. Thus, of the total 269 olfactorome cDNA/ESTs, there were 31 for which no human ortholog is available in the present databases. The complete table for these annotations can be obtained at the http://genet.cchmc.org data archive ("269.xls") as an attachment to the gene list file titled "269_treeOrder".

 
To further investigate the 120 poorly annotated genes, we selected a representative group of 41 clones likely to represent real genes but could not be otherwise assigned using established annotations (summarized in Fig. 4; complete table is posted on the microarray database web server at http://genet.cchmc.org; for login details, see Data archive, in MATERIALS AND METHODS, above). Many of these could be credibly annotated using mouse-human homology and EST-to-genome mapping. For example, sequence BF232834 could first be mapped onto aldose-1-epimerase-like gene in mouse, and the novel human ortholog for this gene was then identified. Sequence BC006717 could be mapped to CPR8 (cell cycle progression 8 protein) in human, and we could then identify the novel mouse ortholog for this gene by human-mouse homology. A search for functional domains in the proteins encoded by these genes revealed that 23 of the 41 did not have any known functional protein domain. Of the 18 that mapped onto a known, functional domain, there were 3 (2 with high expression in OM) WD-repeat protein-encoding genes. Overall, based on public databases [LocusLink (http://www.ncbi.nlm.nih.gov/LocusLink/) and SwissProt], functional annotations (Gene Ontology Attributes) could be found for only 9 of 41 clones. Interestingly, many of the unknown ESTs represented on the Mouse GEM1 cDNA microarray appear to encode entirely novel proteins of potential importance for OM. In addition, there were nine mouse genes for which no human ortholog was found. A BLASTN and BLAT search of these nine mouse mRNAs against human genome also did not yield any significant hits. For instance, there was no human ortholog reported for Expi (extracellular proteinase inhibitor, with high expression in OM and respiratory mucosa). Likewise, the EST sequence AA060184 mapped onto the intergenic region of a putative gene (NM_025582) and Mell1 in mouse. However, the intergenic space in human is smaller, and no human ortholog was found.



View larger version (66K):
[in this window]
[in a new window]
  		Fig. 4. Gene discovery from olfactorome-implicated ESTs. Forty-one genes expressed across brain olfactory bulb, lung, skin, OM, and respiratory mucosa. Genes within the table are organized by clone ID and expression pattern and are tabulated for BLAST-implicated hits to the mouse genome. Sequence-predicted features of the implicated proteins are also listed. A more complete version of this annotation table is downloadable from the Expression Data Server located at http://genet.cchmc.org (see Excel worksheet attached to the gene list file titled "41_olfactorome_gene-discovery"; username: olfepi; password: olfepi).

 
Common cis-elements in coordinately expressed genes.
Based on in situ hybridization analysis, several of the genes that are coordinately expressed across multiple mouse tissues appeared to be localized within the same OM cell types. We therefore sought to test the hypothesis that cell-type-specific gene expression within the OM could be due to shared cis-acting regulatory elements. To do this, we selected three different gene group clusters reflecting expression in 1) OM only; 2) OM plus respiratory mucosa; and 3) whole brain, olfactory bulb, and OM. The complete genomic sequences of the selected mouse genes and their corresponding human orthologs were extracted from the Celera/ENSEMBL/UCSC mouse and human databases, respectively. We next identified clusters of regulatory elements conserved between each pair of mouse and human orthologs within their conserved noncoding region using the TraFaC server (http://trafac.chmcc.org; Ref. 40). Using an extension of this system, we next identified cis-element modules consisting of compositionally similar cis-element clusters that occurred between three different coordinately regulated ortholog pairs.

Analysis of the upstream 3-kb region for the gene group showing high levels of expression in OM revealed cis-elements CLOX, OCT1, BRNF, FKHD, CREB, SORY, MEIS, and ETSF, occurring in clusters of four elements each with in a maximum span of 200 bp. Likewise, common cis-elements were found among each group of genes examined: OM-specific and respiratory mucosa-specific genes showed cis-elements GATA, FKHD, CREB, ETSF, GFI1, SMAD, and MEF2. Genes of the group OM and brain shared cis-elements CREB, NKXH, CLOX, MEF2, BRNF, HNF1, FKHD, LEFF, MYT1, SORY, and MEIS (Fig. 5).



View larger version (36K):
[in this window]
[in a new window]
  		Fig. 5. Shared cis-clusters in OM-specific genes. Compositionally similar cis-regulatory element clusters that occur in groups of coregulated genes within each of their ortholog-pair evolutionarily conserved cis-regulatory regions were identified by comparative genomics-based cis-element analyses (http://trafac.cchmc.org). The composition of each cluster (a set of two or more cis-elements occurring in conserved sequence region between human and mouse) is displayed at bottom. The cis-clusters are represented as different color squares along the genomic sequence. The vertical yellow bars are the exons of the human gene. The numbers indicate nucleotide coordinates of the human genomic sequence. Two of the genes (XM_134943 and XM_143313) encode hypothetical proteins, while the third is TPD52 ("tumor protein D52"). The mouse clone ID 443638 (XM_134943) is orthologous to human hCG17229 Celera gene (unknown WD domain protein XP_028413.4), a conserved ortholog to "Abnormal chemotaxis protein 2" [Caenorhabditis elegans] AAK68383.1. The Riken mouse gene 1700017B05Rik is the ortholog of the hCG2007005 Celera gene, which corresponds to unknown human proline-rich protein DKFZP434H132. TraFaC images display the composition of the cis-clusters conserved in human and mouse. The two gray vertical bars are the two genes that are compared. The TF-binding sites occurring in both the genes are highlighted as various colored bars drawn across the two genes. The numbers indicate the nucleotide coordinates of the genomic sequences. Two of the genes (Tpd52, first intron; and 1700017B05Rik, second intron) with high expression in OM shared cis-elements NOLF, ETSF, and HNF1.

 
Since the potential regulatory switches that direct tissue-specific expression are not always limited to promoter region alone, but are also known to occur in the intronic, downstream, and upstream regions, we extended the query to identify compositionally similar cis-regulatory element clusters across these regions. The window size ranged from 200 to 300 bp. Two of the genes (Tpd52 and 1700017B05Rik) with high expression in OM shared cis-elements NOLF, ETSF, and HNF1, whereas the genes with high expression in both OM and respiratory mucosa shared cis-elements HNF4, NOLF, and MYT1. MYT1, a Xenopus C2HC-type zinc finger protein, has been associated with a regulatory function in neuronal differentiation (7). NOLF ("neuron-specific olfactory factor") has been previously reported in olfactory-neuron-specific genes (85).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
The results presented in this study have identified a relatively large number of genes strongly expressed in the OM, and we have classified them by function and by their relative expression patterning across a large number of other tissues. However, it should be noted that there are several limitations at the present stage of analyses. First, as for any microarray study, the findings are necessarily a reflection of the content of the array, as well as the precise samples that have been profiled. The Incyte Mouse GEM1 array is relatively old and incomplete. For example, candidate odorant receptor genes, which are undoubtedly among the highly expressed OM genes and comprise ~1% of the mammalian genome (46), are not present on the GEM1 microarray. This result illustrates the limitation of using less-than-whole genome representation on a particular microarray platform. A recent publication details changes in gene expression in the senescence-accelerated mouse, using Affymetrix U74Av2 GeneChip oligonucleotide arrays (30). These investigators noted significant expression and aging-related changes in a number of chemosensory receptor genes. However, many genes on the GEM1 microarray are not present on the Affymetrix U74Av2 array. Of more interest, many genes and gene groups that are differentially regulated in senescence-accelerated mice correspond well to the genes that we found to be highly expressed in the mucosa; e.g., a large number of immune system-related genes and metabolic enzyme genes. Thus the results of our study serve more as model of the power that can be obtained in the definition of tissue-specific genomic anatomy if a very large number of different biological samples of diverse differentiation could be carefully intercompared. In addition, our assay of mucosal gene expression is relatively coarse-grained because of the use of tissue samples with mixed cell types. Very interesting results could also be obtained from isolated individual cell types of the mucosa as a number of technical hurdles are progressively overcome.

Gene expression patterns in the OM strongly reflect its unique anatomy and physiology. The cellular anatomy of the OM includes basal cells, supporting (sustentacular) cells, both immature and mature olfactory neurons, plus subepithelial Bowman’s glands and nerve bundles. Olfactory neurons interact directly with the external environment by means of their surface dendrites. Therefore, olfactory neurons are in contact with, and potentially susceptible to damage by, environmental agents. This provides a strong rationale for the high-level coexpression of a series of xenobiotic metabolizing enzyme genes in OM that are also strongly expressed in liver and/or lung. Since the OM continuously generates new olfactory neurons from neuronal progenitor cells, accounting for its regenerative capacity following traumatic or toxicant-induced damage, we hypothesized that a significant component of OM gene expression might overlap with genes that participate in neurodevelopment. Consistent with this hypothesis, we observed a significant overlap between the OM and the embryonic and early postnatal brains with respect to genes involved in neurogenesis, neuronal differentiation, and migration. Finally, the OM contains a family of odorant receptors, as well as the signal transduction machinery necessary to conduct odorant-related information to the rest of the nervous system.

A major goal of the present study was to identify genes involved in neurogenesis and neuronal development, using the OM as a model system, given that the OM supports sustained generation of mature neurons from neuronal progenitor cells. A number of genes that had not previously been localized to the OM were identified as a result of these efforts, although their roles, if any, in neurogenesis are still to be determined. Among these are Sep3 (NM_011889; Refs. 49 and 83), Mad4, and Pdgfec ("platelet-derived endothelial growth factor/thymidine phosphorylase/gliostatin"; NM_138302). Pdgfec had previously been localized to the peripheral nervous system (20), identified as a protein that has a potential role in development and regeneration of the central nervous system (4), and determined to be highly expressed in invasive colorectal tumors (44); therefore, this gene/protein may regulate the switch between proliferation and differentiation in the process of neurogenesis in the OM. Mad4 is highly induced in differentiated cells (39, 43), so we would expect to find this protein localized to one or more nonproliferating cell populations in the OM. In contrast to a previous report, in which brain fatty acid binding protein 7 (Fabp7; NM_021272) was found to be absent from OM (but present in olfactory bulb; Ref. 45), we found expression in OM to be well above our cutoff criterion, with an expression level of approximately one-half that of olfactory bulb and one-tenth that of whole embryonic and postnatal day 2 brain. This protein has been characterized as a potential brain morphogen during development (45).

The olfactory system requires a high degree of synaptic plasticity, as even in the absence of toxicant exposure mature neurons die and new olfactory neurons are continuously generated. As a result, synapses between olfactory sensory cell axons and mitral cells in the olfactory bulb are continuously lost and reformed. Consequently, there is the constant need for axonal guidance to direct the axons of newly generated neurons to the olfactory bulb. Cell adhesion molecules are important in this process, and we observed high expression of NCAM in OM, olfactory bulb, cerebellum, and postnatal day 2 brain. Dpysl3 (NM_009468), another highly expressed OM gene, is orthologous to the C. elegans axonal guidance phosphoprotein unc-33 (12). Similarly, Plxna3, together with neuropilins, acts as a receptor for class 1 semiphorins; these aid in axonal guidance, likely through growth avoidance cues (12, 58, 71).

Three different amyloid precursor protein (APP)-related transcripts were highly expressed in OM. While multiple functions for amyloid-related proteins have been proposed in nervous system homeostasis and pathology, of particular interest, APP is highly expressed in the developing brain, and specifically in the olfactory system (15, 38, 48, 78). Aplp2 has previously been reported to be abundant in olfactory sensory axons and in postsynaptic connections in the olfactory bulb and is proposed to play a role in axonal pathfinding, axogenesis, and/or synaptogenesis (15, 76). APP-like proteins were also highly expressed in the regenerating OM following administration of the reversible OM toxicant diethyldithiocarbamate (70). However, amyloid ß-peptide was found to inhibit neurogenesis in the subventricular zone of mice and in human cortical neuronal precursor cells in vitro (36). Furthermore, OM and olfactory bulb contain high levels of ß-amyloid and amyloid precursor proteins in patients with Alzheimer’s disease, Parkinson’s disease, and Down syndrome (18, 84). Therefore, ß-amyloid-related proteins are clearly pleiotropic, given their expression during brain development, in multiple regions of the adult brain, and in degenerative nervous system conditions.

Many of the highly expressed OM genes are associated with signal transduction. Examples of these genes include those related to Ca2+ ion channels and Ca2+ homeostasis; G-protein-coupled receptor-related genes; and homer, the neuronal early immediate gene associated with glutamate signaling (77). Following interaction of odorants with cell surface receptors, two signal transduction pathways are activated. One class of odorants activate G-protein-mediated formation of adenosine 3',5'-cyclic monophosphate (cAMP), causing opening of cAMP-gated channels and influx of calcium. The increase in intracellular Ca2+, in turn, causes opening of Ca2+-activated chloride channels, further increasing membrane depolarization (41). Other odorants stimulate G-protein-mediated formation of inositol-1,4,5-triphosphate (IP3). An increase in intracellular IP3 causes the opening of a Ca2+ conductance and a nonspecific cation conductance and may also activate Ca2+-activated K+ channels (41). The existence of an IP3-mediated pathway in mammals has been questioned, whereas it has been well accepted in nonmammalian species (e.g., Refs. 11, 22, 32). The rich representation of genes within these categories provides strong support for the role of their corresponding pathways in OM physiology.

The neurotransmitter(s) used by olfactory sensory neurons has also been a source of controversy over the years. Initially, the tripeptide carnosine was regarded as the primary neurotransmitter released following activation of odorant receptors (64). It now appears to be more widely accepted that glutamate is the excitatory olfactory neurotransmitter (8, 24, 34). L-Glutamate acts at both ligand-gated (ionotropic) and G-protein-coupled (metabotropic) glutamate receptors (5). Homer2 encodes a protein that functions directly at the synapse, linking group 1 metabotropic glutamate receptors with IP3 receptors (77).

Among the genes most highly coexpressed in OM, nasal respiratory mucosa, and lung are those encoding proteins with epithelial defense and immune function, including multiple immunoglobulin-related genes, Ltf, and a cathelin-like protein (a member of a family of antimicrobial peptides; Ref. 25). Ltf is an iron-binding multifunctional protein, produced primarily by neutrophils, and is believed to be important in defense against bacteria, fungi, and viruses (3). Plunc proteins have recently been found to be associated with respiratory tract irritation and may also function in mucosal defenses (9, 31).

Genes encoding xenobiotic-metabolizing enzymes were highly expressed in OM, including multiple cytochromes P-450 (CYP), sulfotransferases, and paraoxonase 1. These observations corroborate previous reports documenting sulfotransferases in OM, with some specifically expressed in OM (72, 73). Although the presence of Cyp2f2 in OM had been inferred based on enzymatic activity (69), we demonstrate herein that Cyp2f2 is localized by in situ hybridization and immunohistochemistry to sustentacular cells and Bowman’s glands. Cyp2f2 has been demonstrated to bioactivate naphthalene in the lung; its presence in OM is not surprising, given that naphthalene induces lung and OM tumors in rodents (54, 55, 69). Pon1, which detoxifies organophosphorous pesticides and nerve gases (2), was similarly highly expressed in the Bowman’s glands, a distribution shared among many metabolic enzymes (10, 28). The {alpha}/ß-hydrolase catalytic domain is found in a wide range of enzymes, including esterases and epoxide hydrolases (37). The esterase 1 and 22 transcripts were highly expressed in OM in the present study, and both esterases and microsomal epoxide hydrolase have previously been localized to OM (10, 28).

Our results strongly demonstrate the power of large-scale microarray profiling approaches for compartment-specific gene discovery. One gene exhibiting a particularly interesting expression profile encodes sphingosine phosphate lyase (Sgpl1). Sgpl1 expression was high in thymus, multiple regions of the gastrointestinal tract, skin, and OM. Direct Sgpl1 activity measurements confirmed the high expression in OM, relative to other tissues (Table 2). Although the aforementioned tissues are sites of relatively high cell proliferation, Sgpl1 expression was not exclusively associated with proliferative states, as Sgpl1 expression was higher in unmanipulated mouse liver than in regenerating liver in our database (confirmed by unpublished enzymatic activity measurements from the laboratory of P. P. Van Veldhoven). The role of Sgpl1 in OM is currently unclear, but it may play a role in the kinetics of neuronal proliferation and apoptosis; a similar role for Sgpl1 has been proposed in skin (26). A function of Sgpl1 is the degradation of sphingosine-1-phosphate (S-1-P), which is not only a sphingolipid catabolite, but also a potent second messenger with roles in proliferation and differentiation in various model systems (56, 62, 82). In many cell types, S-1-P is associated with inhibition of apoptosis (21); the high Sgpl1 activity we have observed in OM may be important in modulating cell kinetics in OM, perhaps regulating apoptosis in the oldest population of olfactory neurons that are targeted for destruction. Our studies show that Fas apoptotic inhibitory molecule (Faim; NM_011810) and Sgpl1 share a common pattern of high expression in OM, suggesting that mitochondria-dependent Fas-induced apoptosis may be a major signaling pathway in olfactory neuronal homeostasis (19).

The recent completion and large-scale annotation of the mouse genome has greatly improved our ability to characterize EST gene sequences. We originally selected an EST (W11846) for localization within OM because of its relatively high expression and because it appeared to segregate with Sgpl1 by hierarchical tree analysis. This gene has recently been associated with the Alms1 locus (2p13); mutations of this locus cause Alstrom syndrome, with a wide range of symptoms, including neurosensory degeneration (16, 17). The sodium bicarbonate transporter NBC4 also maps to chromosome 2p13 in humans and is regarded as a new candidate gene for Alstrom syndrome (61). The function of this protein in OM is currently unknown.

The importance of independent verification of genes that appear to be highly expressed in microarray experiments cannot be overemphasized. We chose to do so by in situ hybridization, enzymatic activity measurements, and immunohistochemistry. In our experiments, crp ductin (Dmbt1) appeared to be among the highest expressed genes in OM. However, in situ hybridization studies revealed negligible Dmbt1 expression in OM but very high expression in the nearby lateral nasal gland (Fig. 2). Furthermore, the rat homolog of crp ductin, ebnerin, has recently been immunolocalized to nasal respiratory mucosa and alachlor-induced OM tumors but was not detected in control OM (29). Therefore, we suspect that the speed required for isolation of tissue for preparation of high-quality RNA resulted in accidental collection of the lateral nasal gland from one or more of the mice. This tissue localization issue would not have been resolved by use of reverse transcription, followed by polymerase chain reaction, as the same RNA would have been used for the confirmatory studies. Although the function of crp ductin in the lateral nasal gland remains unknown, we do now know that we should eliminate this gene/protein from those candidates for serving a role in neurogenesis.

Finally, we have also sought to exploit the identification of OM-specific genes to aid in the identification of potential regulatory regions responsible for cell- and region-specific gene expression. Switches that direct tissue-specific expression are not always limited to promoter region alone, but are also known to occur in the intronic, downstream, and upstream regions also. Hence, we extended the query to identify compositionally similar cis-regulatory element clusters within each of their ortholog-pair evolutionarily conserved cis-regulatory regions. The window size ranged from 200 to 300 bp. Genes with high expression in OM shared cis-elements NOLF, ETSF, and HNF1, while the genes with high expression in both olfactory and respiratory mucosa shared cis-elements HNF4, NOLF, and MYT1. MYT1, a Xenopus C2HC-type zinc finger protein, has been associated with a regulatory function in neuronal differentiation (7). NOLF ("neuron-specific olfactory factor") has been reported in olfactory-neuron-specific genes (85). These results support the hypothesis that the approach of tissue-specific expression mining coupled to detailed genomic analyses will be very helpful in the identification of compartment-specific gene regulatory regions.


View this table:
[in this window]
[in a new window]
  		Table 3. Shared clusters of cis-elements in the conserved (human-mouse) noncoding regions of coexpressed genes in different segments of the olfactory system

 

   ACKNOWLEDGMENTS
 
We thank the many contributors to the UC-CHMCC Mouse Tissue Specific Gene Expression Database.

This work was supported by the Howard Hughes Medical Institute; the Children’s Hospital Research Foundation; the University of Cincinnati Center for Environmental Genetics (ES-06096); and by National Institutes of Health Grants U01-ES-11038 (to B. J. Aronow) and R01-ES-08799 (to M. B. Genter) and by Grant G.0405.02 from the Fonds Wetenschappelijk Onderzoek-Vlaanderen (to P. P. Van Veldhoven).


   FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: M. B. Genter, Dept. of Environmental Health, Univ. of Cincinnati, Cincinnati, OH 45267-0056 (E-mail: MaryBeth.Genter{at}UC.edu).

10.1152/physiolgenomics.00117.2003.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Abou-Zeid C, Filley E, Steele J, and Rook GA. A simple new method for using antigens separated by polyacrylamide gel electrophoresis to stimulate lymphocytes in vitro after converting bands cut from Western blots into antigen-bearing particles. J Immunol Methods 98: 5–10, 1987.[CrossRef][ISI][Medline]
  2. Akgur SA, Ozturk P, Sozmen EY, Delen Y, Tanyalcin T, and Ege B. Paraoxonase and acetylcholinesterase activities in humans exposed to organophosphorous compounds. J Toxicol Environ Health A 58: 469–474, 1999.[CrossRef][ISI][Medline]
  3. Andersen JH, Osbakk SA, Vorland LH, Traavik T, and Gutteberg TJ. Lactoferrin and cyclic lactoferricin inhibit the entry of human cytomegalovirus into human fibroblasts. Antiviral Res 51: 141–149, 2001.[CrossRef][ISI][Medline]
  4. Asai K, Nakanishi K, Isobe I, Eksioglu YZ, Hirano A, Hama K, Miyamoto T, and Kato T. Neurotrophic action of gliostatin on cortical neurons. Identity of gliostatin and platelet-derived endothelial cell growth factor. J Biol Chem 267: 20311–20316, 1992.[Abstract/Free Full Text]
  5. Bailey A, Kelland EE, Thomas A, Biggs J, Crawford D, Kitchen I, and Toms NJ. Regional mapping of low-affinity kainate receptors in mouse brain using [3H](2S,4R)-4-methylglutamate autoradiography. Eur J Pharmacol 431: 305–310, 2001.[Medline]
  6. Bates MD, Erwin CR, Sanford LP, Wiginton D, Bezerra JA, Schatzman LC, Jegga AG, Ley-Ebert C, Williams SS, Steinbrecher KA, Warner BW, Cohen MB, and Aronow BJ. Novel genes and functional relationships in the adult mouse gastrointestinal tract identified by microarray analysis. Gastroenterology 122: 1467–1482, 2002.[CrossRef][ISI][Medline]
  7. Bellefroid EJ, Bourguignon C, Hollemann T, Ma Q, Anderson DJ, Kintner C, and Pieler T. X-MyT1, a Xenopus C2HC-type zinc finger protein with a regulatory function in neuronal differentiation. Cell 87: 1191–1202, 1996.[CrossRef][Medline]
  8. Berkowitz DA, Trombley PQ, and Shepherd GM. Evidence for glutamate as the olfactory receptor cell neurotransmitter. J Neurophysiol 71: 2557–2561, 1994.[Abstract/Free Full Text]
  9. Bingle CD and Craven CJ. PLUNC: a novel family of candidate host defence proteins expressed in the upper airways and nasopharynx. Hum Mol Genet 11: 937–943, 2002.[Abstract/Free Full Text]
  10. Bogdanffy MS. Biotransformation enzymes in the rodent nasal mucosa: the value of a histochemical approach. Environ Health Perspect 85: 177–186, 1990.[Medline]
  11. Brunet LJ, Gold GH, and Ngai J. General anosmia caused by a targeted disruption of the mouse olfactory cyclic nucleotide-gated cation channel. Neuron 17: 681–593, 1996.[CrossRef][ISI][Medline]
  12. Byk T, Dobransky T, Cifuentes-Diaz C, and Sobel A. Identification and molecular characterization of Unc-33-like phosphoprotein (Ulip), a putative mammalian homolog of the axonal guidance-associated unc-33 gene product. J Neurosci 16: 688–701, 1996.[Abstract/Free Full Text]
  13. Camara CG and Harding JW. Thymidine incorporation in the olfactory epithelium of mice: early exponential response induced by olfactory neurectomy. Brain Res 308: 63–68, 1984.[CrossRef][Medline]
  14. Cancalon P. Degeneration and regeneration of olfactory cells induced by ZnSO4 and other chemicals. Tissue Cell 14: 717–733, 1982.[CrossRef][ISI][Medline]
  15. Clarris HJ, Key B, Beyreuther K, Masters CL, and Small DH. Expression of the amyloid protein precursor of Alzheimer’s disease in the developing rat olfactory system. Brain Res Dev Brain Res 88: 87–95, 1995.[CrossRef][Medline]
  16. Collin GB, Marshall JD, Ikeda A, So WV, Russell-Eggitt I, Maffei P, Beck S, Boerkoel CF, Sicolo N, Martin M, Nishina PM, and Naggert JK. Mutations in ALMS1 cause obesity, type 2 diabetes and neurosensory degeneration in Alstrom syndrome. Nat Genet 31: 74–78, 2002.[CrossRef][ISI][Medline]
  17. Collin GB, Marshall JD, Boerkoel CF, Levin AV, Weksberg R, Greenberg J, Michaud JL, Naggert JK, and Nishina PM. Alstrom syndrome: further evidence for linkage to human chromosome 2p13. Hum Genet 105: 474–479, 1999.[CrossRef][Medline]
  18. Crino PB, Martin JA, Hill WD, Greenberg B, Lee VM, and Trojanowski JQ. Beta-amyloid peptide and amyloid precursor proteins in olfactory mucosa of patients with Alzheimer’s disease, Parkinson’s disease, and Down syndrome. Ann Otol Rhinol Laryngol 104: 655–661, 1995.[ISI][Medline]
  19. Cuvillier O, Edsall L, and Spiegel S. Involvement of sphingosine in mitochondria-dependent Fas-induced apoptosis of type II Jurkat T cells. J Biol Chem 275: 15691–15700, 2000.[Abstract/Free Full Text]
  20. Eccleston PA, Funa K, Heldin CH. Neurons of the peripheral nervous system express thymidine phosphorylase. Neurosci Lett 192: 137–141, 1995.[CrossRef][ISI][Medline]
  21. Edsall LC, Cuvillier O, Twitty S, Spiegel S, and Milstien S. Sphingosine kinase expression regulates apoptosis and caspase activation in PC12 cells. J Neurochem 76: 1573–1584, 2001.[CrossRef][Medline]
  22. Fadool DA and Ache BW. Plasma membrane inositol 1,4,5-phosphate-activated channels mediate signal transduction in lobster olfactory receptor neurons. Neuron 9: 907–918, 1992.[CrossRef][ISI][Medline]
  23. Feng L, Hatten ME, and Heintz N. Brain lipid-binding protein (BLBP): a novel signaling system in the developing mammalian CNS. Neuron 12: 895–908, 1994.[CrossRef][ISI][Medline]
  24. Ferraris N, Perroteau I, De Marchis S, Fasolo A, and Bovolin P. Glutamatergic deafferentation of olfactory bulb modulates the expression of mGluR1a mRNA. Neuroreport 8: 1949–1953, 1997.[ISI][Medline]
  25. Gallo RL, Kim KJ, Bernfield M, Kozak CA, Zanetti M, Merluzzi L, and Gennaro R. Identification of CRAMP, a cathelin-related antimicrobial peptide expressed in the embryonic and adult mouse. J Biol Chem 16: 13088–13093, 1997.
  26. Geilen CC, Barz S, and Bektas M. Sphingolipid signaling in epidermal homeostasis. Current knowledge and new therapeutic approaches in dermatology. Skin Pharmacol Appl Skin Physiol 14: 261–71, 2001.[CrossRef][Medline]
  27. Genter MB, Llorens J, O’Callaghan JP, Peele DB, Morgan KT, and Crofton K. Olfactory toxicity of ß,ß'-iminodipropionitrile (IDPN) in the rat. J Pharmacol Exp Ther 263: 1432–1439, 1992.[Abstract/Free Full Text]
  28. Genter MB, Owens DM, and Deamer NJ. Distribution of microsomal epoxide hydrolase and glutathione S-transferase in the rat olfactory mucosa: relevance to distribution of lesions caused by systemically-administered olfactory toxicants. Chem Senses 20: 385–392, 1995.[Abstract/Free Full Text]
  29. Genter MB, Burman DM, Vijayakumar S, Ebert CL, and Aronow BJ. Genomic analysis of alachlor-induced oncogenesis in rat olfactory mucosa. Physiol Genomics 12: 35–45, 2002. First published November 5, 2002; 10.1152/physiolgenomics.00120.2002.
  30. Getchell TV, Peng X, Stromberg AJ, Chen KC, Paul Green C, Subhedar NK, Shah DS, Mattson MP, and Getchell ML. Age-related trends in gene expression in the chemosensory-nasal mucosae of senescence-accelerated mice. Ageing Res Rev 2: 211–243, 2003.[CrossRef][Medline]
  31. Ghafouri B, Stahlbom B, Tagesson C, and Lindahl M. Newly identified proteins in human nasal lavage fluid from non-smokers and smokers using two-dimensional gel electrophoresis and peptide mass fingerprinting. Proteomics 2: 112–120, 2002.[CrossRef][ISI][Medline]
  32. Gold G. Controversial issues in vertebrate olfactory transduction. Annu Rev Physiol 61: 857–871, 1999.[CrossRef][ISI][Medline]
  33. Graziadei PP and Graziadei GA. Neurogenesis and neuron regeneration in the olfactory system of mammals. I. Morphological aspects of differentiation and structural organization of the olfactory sensory neurons. J Neurocytol 8: 1–18, 1979.[Medline]
  34. Griff ER, Greer CA, Margolis F, Ennis M, and Shipley MT. Ultrastructural characteristics and conduction velocity of olfactory receptor neuron axons in the olfactory marker protein-null mouse. Brain Res 866: 227–236, 2000.[Medline]
  35. Haass NK, Kartenbeck AM, and Leube RE. Pantophysin is a ubiquitously expressed synaptophysin homologue and defines constitutive transport vesicles. J Cell Biol 134: 731–746, 1996.[Abstract/Free Full Text]
  36. Haughey NJ, Liu D, Nath A, Borchard AC, and Mattson MP. Disruption of neurogenesis in the subventricular zone of adult mice, and in human cortical neuronal precursor cells in culture, by amyloid beta-peptide: implications for the pathogenesis of Alzheimer’s disease. Neuromolecular Med 1: 125–135, 2002.[Medline]
  37. Heikinheimo P, Goldman A, Jeffries C, and Ollis DL. Of barn owls and bankers: a lush variety of {alpha}/ß hydrolases. Structure 7: R141–146, 1999.[Medline]
  38. Hsu LJ, Mallory M, Xia Y, Veinbergsm I, Hashimoto M, Yoshimoto M, Thal LJ, Saitoh T, and Masliah E. Expression pattern of synucleins (non-Aß component of Alzheimer’s disease amyloid precursor protein/{alpha}-synuclein) during murine brain development. J Neurochem 71: 338–344, 1998.[ISI][Medline]
  39. Hurlin PJ, Queva C, Koskinen PJ, Steingrimsson E, Ayer DE, Copeland NG, Jenkins NA, and Eisenman RN. Mad3 and Mad4: novel Max-interacting transcriptional repressors that suppress c-myc dependent transformation and are expressed during neural and epidermal differentiation. EMBO J 15: 2030, 1996.[Medline]
  40. Jegga AG, Sherwood SP, Carman JW, Pinski AT, Phillips JL, Pestian JP, and Aronow BJ. Detection and visualization of compositionally similar cis-regulatory element clusters in orthologous and coordinately controlled genes. Genome Res 12: 1408–1417, 2002.[Abstract/Free Full Text]
  41. Kaur R, Zhu XO, Moorhouse AJ, and Barry PH. IP3-gated channels and their occurrence relative to CNG channels in the soma and dendritic knob of rat olfactory receptor neurons. J Membr Biol 181: 91–105, 2001.[ISI][Medline]
  42. Kelley-Loughnane N, Sabla GE, Ley-Ebert C, Aronow BJ, and Bezerra JA. Independent and overlapping transcriptional activation during liver development and regeneration in mice. Hepatology 35: 525–534, 2002.[CrossRef][ISI][Medline]
  43. Kime L and Wright SC. Mad4 is regulated by a transcriptional repressor complex that contains Miz-1 and c-Myc. Biochem J 370: 291–298, 2003.[CrossRef][ISI][Medline]
  44. Kitabatake T, Kojima K, Fukasawa M, Beppu T, and Futagawa S. Correlation of thymidine phosphorylase staining and the Ki-67 labeling index to clinicopathologic factors and hepatic metastasis in patients with colorectal cancer. Surg Today 32: 322–328, 2002.