Olfactory neurons expressing transient receptor potential channel M5 (TRPM5) are involved in sensing semiochemicals

Abstract

Olfactory sensory neurons (OSNs) in the main olfactory epithelium respond to environmental odorants. Recent studies reveal that these OSNs also respond to semiochemicals such as pheromones and that main olfactory input modulates animal reproduction, but the transduction mechanism for these chemosignals is not fully understood. Previously, we determined that responses to putative pheromones in the main olfactory system were reduced but not eliminated in mice defective for the canonical cAMP transduction pathway, and we suggested, on the basis of pharmacology, an involvement of phospholipase C. In the present study, we find that a downstream signaling component of the phospholipase C pathway, the transient receptor potential channel M5 (TRPM5), is coexpressed with the cyclic nucleotide-gated channel subunit A2 in a subset of mature OSNs. These neurons project axons primarily to the ventral olfactory bulb, where information from urine and other socially relevant signals is processed. We find that these chemosignals activate a subset of glomeruli targeted by TRPM5-expressing OSNs. Our data indicate that TRPM5-expressing OSNs that project axons to glomeruli in the ventral area of the main olfactory bulb are involved in processing of information from semiochemicals.

Fig. 1.

TRPM5 promoter-driven GFP and TRPM5-like antigenicity in the MOE. (A) TRPM5 promoter drove GFP expression in two populations of cells with distinct morphology: sparsely distributed short cells <15 μm in length (arrowheads and Inset a) and densely packed OSNs (representative OSNs are shown in Inset b). (B) A representative confocal image showing immunolabel for TRPM5 at a higher magnification. (C) Negative control, image taken from an olfactory epithelium section from a TRPM5 knockout mouse labeled with the anti-TRPM5 antibody. (D) OSNs expressing GFP in TRPM5-GFP mice (green) also displayed immunoreactivity for TRPM5 (red). (Scale bars: A, B, and C, 20 μm; Inset a, 5 μm; Inset b, 10 μm; D, 50 μm.)

Fig. 2.

Analysis of spot size for STED images of TRPM5 immunoreactivity in the cilia layer. (A and B) Confocal (A) and STED (B) images of TRPM5 immunofluorescence in the cilia layer of the olfactory epithelium. (A Inset) Confocal image at a lower magnification taken with a confocal microscope. (B Inset) Smallest spot gained from the faintest antibody cluster observed in the sample is indicative of the maximum size of the effective point-spread function in the confocal (189-nm) and STED (35-nm) imaging modes. (C and D) Higher-magnification images of the areas enclosed by the dashed boxes in A and B, respectively. The image in D is the STED image after a linear deconvolution (LD). (E) Histogram showing the distribution of full width at half maxima (FWHM) for the clusters in three separate images (130 individual clusters from three separate images). To estimate the FWHM, background was subtracted from the STED images, and each cluster was fit with Lorentz-shaped profiles.

Fig. 3.

Colocalization of TRPM5 or GFP immunostaining with OMP or CNGA2 immunoreactivity. (A) OMP immunoreactivity (green) was seen in mature OSNs. A subset also labeled with TRPM5 antibody (red). (B) A confocal image from a section cut perpendicular to the dendrites at the level of the dendritic knobs and cilia displays immunolabel for OMP and TRPM5. (C) Sagittal image of the ventral portion of the olfactory bulb of a TRPM5-GFP mouse. Axons from TRPM5-expressing OSNs (green) project to a subset of glomeruli that were immunopositive for OMP (red, overlap appears as yellow). (D) An antibody against CNGA2 (red) labeled the apical layer and soma of the majority of OSNs in a TRPM5-GFP mouse. GFP (green) was present in many such OSNs, an indication of coexpression of these ion channels. (E) Immunoreactivity for CNGA2 (red) and GFP (green) in the ventral olfactory bulb (saggital section) in a TRPM5-GFP mouse. Most of the GFP-positive glomeruli also were immunoreactive for CNGA2. A relatively small number of glomeruli displayed GFP expression but stained weakly for CNGA2. (Scale bars: A and D, 20 μm; B, 5 μm; C and E, 50 μm.)

Fig. 4.

Glomeruli targeted by TRPM5-expressing OSNs were detected as GFP-positive glomeruli in TRPM5-GFP mice. (A and B) Whole-mount images showing GFP-positive axons and targeted glomeruli in medial (A) and lateral (B) surfaces of the olfactory bulb. (C) A transverse section showing GFP-positive glomeruli. A line drawn through the subventricular zone is the axis used to map glomeruli in D and E. m, medial; l, lateral. (D) A representative 2D map showing location of GFP-positive glomeruli along the rostrocaudal distance and angle around the transverse section. (E) A pseudocolor rendering of the average number of glomeruli in bins of 10° and 72 μm (average from six bulbs), showing that the highest density of GFP-positive glomeruli was located in the ventral region of the bulb. (F–H) Representative glomeruli that were activated by the pheromone DMP (F), the sex-specific odorant MTMT (G), and soiled bedding from a mating pair (H). Activated glomeruli were identified by numerous Fos-expressing periglomerular neurons (red). GFP antibody immunohistochemistry was used in C and F–H. (Scale bars: A and B, 1 mm; C, 100 μm; F–H, 20 μm.)

Fig. 5.

EOG recordings from wild-type (TRPM5 +/+) and TRPM5 knockout (TRPM5 −/−) mice. (A) Representative EOG traces. (B) Average peak EOG responses to different odors (n = 4–12). There was no significant difference between knockout and control (P < 0.7 for ANOVA). All odorants were presented at 100 μM in Ringer's, and forskolin was presented at a concentration significantly below saturation (20 μM). A 1 mM concentration of 3-isobutyl-1-methylxanthine (IBMX) elicited a saturating response (≈4 mV; data not shown). (C) In wild-type mice, the ACIII inhibitor SQ33256 (300 μm) suppressed EOG responses induced by forskolin (an ACIII activator) and lilial to a larger extent than responses induced by the pheromones 2-heptanone and DMP (n = 4). This differential inhibition disappeared in the knockout animals. A mixed ANOVA with mouse strain and odor as fixed factors and mouse as a random effect yielded a P value of 0.036 for a strain–odor interaction. Post hoc Tukey–Kramer tests showed that the difference in inhibition between forskolin or lilial and the pheromones DMP and 2-heptanone is not significant in the knockouts (post hoc P values >0.05 in the knockout and <0.05 in the controls). Error bars are SEM.