Panx1^+/+ mice (Panx1fl/fl) with three LoxP consensus sequences integrated into the Panx1 gene and knockout mice with global loss of Panx1 (Panx1^−/−, CMV-Cre/Panx1) were described previously (Gründken et al., 2011; Dvoriantchikova et al., 2012; Prochnow et al., 2012). Handling and housing of animals used in this study was performed in compliance with the German Animal Rights law and approved by the Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen, Germany (Permission No. 50.8735.1 Nr. 100/4) and formal approval by the Animal Care Committee (York University, Canada). Adult male mice (4–8 months of age) were housed individually 1 week prior to, and during, behavior testing.

Digoxigenin (dig)-labeled sense and antisense cRNA probes were prepared from a full length Panx1 cloned into the pcDNA3 plasmid as described previously (Ray et al., 2006). After linearization of the plasmid, sense and antisense cRNA probes were transcribed using T7 and SP6 RNA polymerase with dig-RNA labeling mix (Roche, Germany). The ISH was performed as described (Larsson et al., 2004) with minor modifications. OE from P7 mice were dissected and immediately embedded in tissue freezing medium (Leica, Germany) at −30°C and cryostat sections (12 μm) were cut immediately. Slides were subsequently fixed in 4% paraformaldehyde in PBS at 4°C for 20 min, washed in PBS and acetylated by a 15 min treatment in 0.1 M triethanolaminhydrochloride solution with 0.25% acetic anhydride on a stir plate. Sections were rinsed in 2× SSC (30 mM NaCl and 3 mM sodium citrate) and prehybridized in hybridization buffer (50% formamide, 5× SSC, 5× Denhardts' solution, 2.5 mM EDTA, 50 μg/ml heparin, 250 μg/ml tRNA, 500 μg/ml salmon sperm DNA, and 0.1% Tween-20) for 1 h at 55°C. Riboprobes were added to the hybridization buffer (50 ng in 200 μl hybridization buffer), denaturized at 80°C for 2 min and applied to sections. Sections were incubated over night at 55°C for hybridization. Post-hybridization, slides were washed with 0.2× SSC for 1 h and then with 0.1× SSC for 15 min, to remove non-specific binding. Sections were subsequently equilibrated for 10 min in PBS containing 0.1 % TritonX-100 (PBST), blocked with 10% goat serum in PBST buffer for 1 h, and then incubated with 1:1000 alkaline phosphatase (AP) conjugated anti-dig Fab fragment (Roche, Germany) in blocking solution overnight (ON) at 4°C. Finally, slides were washed in PBST, equilibrated in B3-Buffer (0.1 Tris-HCl, 0.1 M NaCl, 50 mM MgCl₂, 0.1% Tween-20), followed by treatment with NBT/BCIP (Roche, Germany) (20 μl/ml B3) to visualize the hybridized probes.

After the fur and palate were removed, heads from adult male mice were fixed in 4% PFA at 4°C ON, then immersed in 30% sucrose at 4°C ON. 12 μm cryosections were prepared, blocked with 5% cold-water fish skin gelatine for 1 h at RT, and primary antibodies (1:250, Santa Cruz, CA, G_α olf sc-383; CNG sc-13700, ACIII sc-588, acetylated tubulin sc-23950) were applied in 1% cold-water fish skin gelatin in PBS containing 0.1% Triton X-100, at 4°C ON. After 30 min washing in PBS, secondary goat anti-rabbit antibodies Alexa Fluor 568 (Invitrogen, Germany) were applied for 30 min at RT in PBS. After 30 min washing in PBS, sections were embedded in ProlongGold Antifade (Invitrogen, Germany).

The Laird laboratory generously provided an antibody for Panx1 IHC (Penuela et al., 2007). For Panx1 detection the following modifications were introduced. For antigen retrieval, fixed cryostat sections were incubated for 5 min with 1% SDS, followed by three washes for 5 min with PBS. After blocking for 1 h at RT with 5% normal goat serum (NGS), 1% bovine serum albumin (BSA), and 0.1% Triton X100 in PBS, the primary antibody was diluted (1:100) in 1% BSA, 0.1% Triton X100 in PBS. After ON incubation, specimen were washed three times with PBS for 10 min each. The secondary goat anti-rabbit Alexa Fluor 488 antibody was diluted (1:1000) in PBS and applied for 30 min at RT, followed by three washes for 10 min with PBS.

Confocal microscopy was performed using a ZEISS LSM700 microscope. ZEISS ZEN software was used to control all parameters during imaging. Identical settings were used to allow a direct comparison of IHCs of Panx1^+/+ and Panx1^−/− mice. LSM images were exported into tiff format and assembled using Photoshop CS.

RNA was isolated from adult male mice using the RNAeasy Fibrous Tissue Mini Kit (Invitrogen, Canada) and cDNA was synthesized from 1 μg total RNA with the ReadyScript cDNA Synthesis Kit (Sigma-Aldrich, Canada), according to the manufacturer's instructions. qPCR was performed using the SsoFast EvaGreen Supremix (Bio-Rad, Canada) and the following oligonucleotide pairs: Panx1fw: CAGGCTGCCTTTGTGGATTC Panx1rev: CGGGCAGGTACAGGAGTATG Panx2fw: GGTACCAAGAAGGCCAAGACT Panx2rev: GGGGTACGGGATTTCCTTCTC Panx3fw: CTTACAACCGTTCCATCCGC Panx3rev: CAGGTACCGCTCTAGCAAGG 18Sfw: TGACTCTTTCGAGGCCCTGTA 18Srev: TGGAATTACCGCGGCTGCTG. fw, forward; rev, reverse. Experiments were performed in triplicates, using three biological replicates. Relative gene expression was calculated using the REST software (2009) (Pfaffl et al., 2002). 18S served as the reference gene.

Skulls from adult male mice were cut parasagittal to the septum to expose the nasal cavity. The turbinates were removed and EOGs were recorded from the OE on the septum. A constant, deodorized, humidified air stream was delivered to the OE and adjusted to 2.4 l/min. Henkel 100 (H100, Henkel, Germany), a mixture of 100 different odors (Wetzel et al., 1999), was used as a stimulus. Dilutions were made in distilled water and soaked into a piece of felt, which was placed into a custom-made injection device. Recording electrodes were made from pulled glass capillaries containing chloride silver wires and filled with Ringer's solution (140 mM NaCl, 5 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, pH 7.4). During recording, heads were placed on 2% agarose dissolved in Ringer's solution containing the reference electrode. Student's t-test was used for statistical analysis.

The intact OE was exposed after skulls from adult male mice were cut parasagittal to the septum to expose the nasal cavity. Once the OE was entirely exposed, the tissue was positioned in upside down orientation. Since the OE rapidly degenerates once axons get transsected, this ex vivo preparation was used instead of more invasive dissection procedures to extract extracellular ATP from the cilia surface. To avoid mechanical stimulation small droplets (25 μl) of Ringers solution, H100 [Henkel 100, 1:10000 in Ringers solution, Henkel, Germany (Wetzel et al., 1999)], and potassium gluconate (25 mM, Sigma-Aldrich, diluted in Ringers solution) were subsequently and gently pipetted onto the cilia surface at the center of the OE. The tissue was incubated for the indicated timespans. Between steps droplets were fully removed. Solutions contained 100 μM ARL 67156 trisodium salt hydrate (Sigma-Aldrich) to inhibit ATPases. Samples were heated at 95°C for 1 min after extraction, flash frozen, and stored at −80°C until used. ATP assays were performed in 96 well format; using the Molecular Probes® ATP Determination Kit (Life Technologies, USA) and the Synergy H4 hybrid multiwell plate reader (Biotek, USA). ATP concentrations were determined from ATP standard curves included in each assay. To test for statistical significance the student's T-test was used.

Pannexin expression was determined in the mouse OE and whole brain lysate using qPCR and primer pairs specific for Panx1 (GI:86262133). Panx1 mRNA expression was found in the brain and OE adult mice (Figure 1), with mean cycle threshold (C_t) values of 29.8 ± 0.5 (brain) and 31.3 ± 0.9 (OE). Panx2 was expressed in the OE at significantly lower expression levels compared to the brain (p < 0.001). Mean cycle threshold (C_t) values were 27.2 ± 0.7 (brain) and 33.9 ± 0.7 (OE). No Panx3 expression was detected (C_t < 35). Neither in the brain, nor in the OE a compensatory upregulation of Panx2 or Panx3 mRNA was observed in Panx^−/− mice. Consistent with the genetic phenotype, no Panx1 mRNA expression (C_t < 35) was found in Panx1^−/− animals lacking exons 3 and 4 (Dvoriantchikova et al., 2012).

Localization of Panx1 expression sites in the OE were determined using in-situ hybridization technology and cRNA probes specific for mouse Panx1 (Ray et al., 2006). Figure 2 shows a representative distribution of Panx1 mRNAs in the OE. The sense cRNA probes generated only a very weak background staining, demonstrating the antisense probes' specificity. In the OE, a strong staining of the OSN layer and a less intense staining of the sustentacular (SC, arrows) and basal cell (BC) layers were observed.

To validate mRNA localization data, immunohistochemistry (IHC) was performed using an antibody specific for mouse Panx1 (Penuela et al., 2007). In Panx1^+/+ animals (Figure 3), the IHC of the OE revealed a significant labeling of the OSN axon bundles, projecting to the olfactory bulb. Virtually no staining was found in OSN axon bundles of Panx1^−/− mice. However, a diffuse signal was observed in the basal-, neuronal-, and SC layers. No pronounced staining was found in OSN cilia, which inherit olfactory receptor proteins and signaling proteins involved in OSN signaling/depolarization after odorant activation.

When an olfactory receptor (OR) is activated, the signal transduction cascade elicited through the OR typically evokes a massive influx of calcium mediated by cyclic nucleotide gated (CNG) channels. Since Panx1 channels can be activated by raised intracellular calcium levels (Locovei et al., 2006), we investigated whether the lack of Panx1 alters the response of the OE upon odor application by performing electroolfactogram (EOG) measurements in adult male mice. A mixture of 100 different odorants [Henkel 100 (H100) (Wetzel et al., 1999)] was applied via a constant, humidified airstream to activate a broad range of ORs. Figure 4A shows a typical response amplitude after odorant application for 100 ms. Results were quantified and summarized in Figure 4B. No significant difference in the response amplitudes for Panx1^+/+ (N = 7) and Panx1^−/− (N = 8) mice were detected (minimum 10 measurements on different locations per mouse, Panx1^+/+ = 4.2 ± 0.2 mV, Panx1^−/− = 4.7 ± 0.3 mV, p = 0.17). Further, the response kinetics, namely the rise time (Panx1^+/+ = 86 ± 3 ms, Panx1^−/− = 94 ± 4 ms, p = 0.1) and decay time (Panx1^+/+ = 989 ± 191 ms, Panx1^−/− = 984 ± 128 ms, p = 0.98), did not differ between both animal groups.

To test the ability of the OE to adapt to odorants, we measured odor responses after a repetitive stimulation paradigm (Brunert et al., 2009) (N_Panx1^+/+ = 6, N_{Panx1^−/−} = 6, 1 s stimulus duration, 4 s interstimulus interval, 1 measurement per mouse), as depicted in Figures 4C,D. Using this paradigm, we achieved a robust adaptation to less than 30 % of the control level, but adaptation kinetics observed in Panx1^−/− animals were statistically insignificant to those in wild type mice. Since Panx1^+/+ and Panx1^−/− responded equally to both experimental conditions, we concluded that loss of Panx1 did not alter physiological odorant responses. Further confirming this finding, ablation of Panx1 did not cause a significant change in the localization and distribution of the four major olfactory signal transduction proteins detecting adenylyl cyclase 3 (ADCYIII), cyclic nucleotide gated channel alpha 2 (CNGA2), olfactory neuron specific-G protein (G_olf), and acetylated tubulin (AcTub) (Supplementary Figure 1). Similarly, quantification of the mRNA expression of adenylyl cyclase 3 (ADCYIII), cyclic nucleotide gated channel alpha 2 (CNGA2), and olfactory neuron specific-G protein (G_olf)using qPCR revealed no difference (Supplementary Figure 2).

Panx1 is considered to be a major ATP release channel in various tissues. We quantified extracellular ATP using ex vivo preparations of the OE (Figure 5). During this procedure the OE is placed upside down for extracellular ATP extraction. Due to the small size of the OE, ATP was extracted in a 25 μl droplet of Ringers solution gently placed at the center of the OE. After equilibration for 3 min, droplets recovered from the Panx1^+/+ and Panx1^−/−OEs had an ATP concentration in the pM range. Extracellular ATP concentrations in the 25 μl droplet increased more than 100fold when this extraction step was repeated with Ringers solution containing the ATPase inhibitor (ARL 67156). No significant difference was detected (p = 0.66), suggesting that ablation of Panx1 did not compromise efflux of extracellular ATP. After Ringers solution with ATPase inhibitor was completely removed, the odorant mixture Henkel 100 was applied for 90 s in a small droplet (25 μl) placed again at the center of the OE. No significant differences in extracellular ATP release in both mouse strains were detectable. The concentration of extractable ATP was slightly reduced to the previous step reflecting the reduced extraction time. It was concluded that alternative ATP release pathways primarily promoted the observed ATP efflux in response to the odorant. Subsequently 25 mM potassium gluconate in Ringers solution (25 μl) was applied for 90 s. Significant differences in extracellular ATP levels were detected (Panx1^+/+, 3.4 ± 0.4 pM; Panx1^−/− 2.2 ± 0.3 pM; p = 0.022), suggesting that Panx1 channels contributed significantly to ATP release in this experimental condition.

The results shown suggested that loss of Panx1 does not impair detection of odorants in the OE. However, Panx1 is expressed in other cells of the olfactory system, including subsets of neurons in the olfactory bulb and the piriform cortex (Bruzzone et al., 2003; Ray et al., 2005). To test for odorant perception, a cookie-finding test was performed. In this behavioral test, an odorous cookie was hidden beneath the bedding in the home cage of each mouse and the latency until the mice found the treat was determined. Tests were conducted on four subsequent days (Figure 6A). On the training day (T), a large cookie (1 g) was hidden. The size (and odor) allowed the mice to familiarize with the test. Panx1^+/+ (N = 15) and Panx1^−/− (N = 14) mice performed equally on test day 1, when test conditions were more difficult for the mice (50 mg cookie; Panx1^+/+ = 627 ± 83 s, Panx1^−/− = 630 ± 82 s). We performed two more tests using a large [150 mg (day 2)] and a small [50 mg (day 3)] cookie. On day 2, Panx1^+/+ animals showed a tendency to find the cookie faster than Panx1^−/− mice, however, this difference was not significant (Panx1^+/+ = 212 ± 83 s, Panx1^−/− = 428 ± 82 s, p = 0.1). When the difficulty of the test was increased on day 3 (50 mg cookie), Panx1^−/− mice took significantly longer to find the cookie (Panx1^+/+ = 149 ± 58 s, Panx1^−/− = 439 ± 107 s, p = 0.02). Further, they did not improve compared to the previous days (day 2) and training day (T) when conditions were easier (Figure 6A). This suggests that processing of olfactory information was not affected, as both animal cohorts did find the 50 mg cookie on test day one at the same time, but the ability of the animals to learn the test was significantly altered. We also tracked the velocity of the animals and found that Panx1^−/− mice showed a significantly higher mobility (Panx1^+/+ = 4.7 ± 0.3 cm/s, Panx1^−/− = 6.2 ± 0.5 cm/s, p = 0.02) compared to Panx1^+/+ mice (Figure 6B).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.