Odorant Receptor Proteins in the Mouse Main Olfactory Epithelium and Olfactory Bulb

NSC 17525 No. of Pages 11 7 January 2017

Please cite this article in press as: Low VF, Mombaerts P. Odorant receptor proteins in the mouse main olfactory epithelium and olfactory bulb. Neu- roscience (2017), http://dx.doi.org/10.1016/j.neuroscience.2016.12.044

1 Neuroscience xxx (2017) xxx–xxx

2 ODORANT RECEPTOR PROTEINS IN THE MOUSE MAIN OLFACTORY

3 EPITHELIUM AND OLFACTORY BULB

4 VICTORIA F. LOW AND PETER MOMBAERTS * INTRODUCTION 9 5 Max Planck Research Unit for Neurogenetics, In mammals such as mouse and rat, odorants are 10 6 Max-von-Laue-Strasse 4, D-60438 Frankfurt, Germany detected in the main olfactory epithelium (MOE) by 11 olfactory sensory neurons (OSNs) via odorant receptors 12 7 Abstract—In the mouse, odorant receptor proteins (ORs) are (ORs), which are G-protein-coupled seven- 13 G-protein-coupled receptors expressed in mature olfactory transmembrane proteins (Buck and Axel, 1991). In the 14 sensory neurons (OSNs) of the main olfactory epithelium mouse, there are 1099 OR genes with an intact open 15 (MOE). ORs mediate odorant reception at the level of the reading frame, of which 1,087 are expressed at the 16 OSN cilia. Most 1100 OR genes in the mouse genome are RNA level in OSNs (Saraiva et al., 2015). 17 expressed, at the RNA level, in mature OSNs. The literature Due to the difficulty in raising antibodies against 18 on antibodies against ORs is limited, and most reports are with antibodies that are not commercially available. Here G-protein-coupled receptors, there have been relatively 19 we have screened 40 commercial antibodies against human few reports about OR proteins in the literature. Initial 20 and mouse ORs by immunofluorescence staining of coronal studies with antibodies against rat ORs suggested that 21 cryosections of the MOE of 21-day-old C57BL/6J mice. the antigen is localized to the cilia of OSNs in the rat 22 Various methods of antigen retrieval were tested. Of the 19 MOE (Koshimoto et al., 1992; Krieger et al., 1994). In situ 23 antibodies raised against human ORs, three yielded a con- hybridization against OR RNAs in mouse revealed that 24 sistent immunoreactive signal in the mouse MOE; of these OR gene expression is restricted to OSNs within discrete 25 three, two appeared to cross react against one or more, regions of the MOE called zones (Ressler et al., 1993; 26 unknown, mouse ORs. Of the 21 antibodies raised against Miyamichi et al., 2005). Gene-targeted mice enabled the 27 mouse ORs, six yielded a consistent immunoreactive signal visualization of entire OSNs expressing a given OR by 28 in the mouse MOE; of these six, two also stained specific glomeruli in the olfactory bulb. Antibody specificity could targeted integration of a cassette that affords cotransla- 29 be validated with gene-targeted mouse strains in the case tion of the OR with a marker such as taulacZ via an inter- 30 of three ORs. The number of OSNs immunoreactive for the nal ribosome entry site (IRES) (Mombaerts et al., 1996). 31 MOR28/Olfr1507 antibody is greater in C57BL/6J than in Axons of OSNs expressing the same OR coalesce into 32 129S6/SvEvTac wild-type mice. Taken together, our results two or a few glomeruli per mouse olfactory bulb, which 33 are encouraging: 20–30% of these commercially available reside in highly restricted regions (Ressler et al., 1994; 34 antibodies are informative in immunohistochemical analy- Mombaerts et al., 1996; Zapiec and Mombaerts, 2015). 35 ses of the mouse MOE. The commercial availability of these Replacement of the OR coding region showed that the 36 antibodies should facilitate the study of OR proteins in the OR does not only mediate odorant detection (Bozza 37 MOE and the olfactory bulb, and the replicability of results et al., 2002), but is also a determinant of where in the 38 in the literature. Ó 2016 The Author(s). Published by Elsevier Ltd on behalf of IBRO. This is an open access article under olfactory bulb the axons coalesce and form glomeruli 39 the CC BY-NC-ND license (http://creativecommons.org/ (Mombaerts et al., 1996; Feinstein et al., 2004; Feinstein 40 licenses/by-nc-nd/4.0/). and Mombaerts, 2004). 41 The widespread success of the gene-targeted 42 approach of the OR-IRES-marker design may have 43 Key words: olfactory receptor, olfactory sensory neuron, overshadowed or diminished efforts to raise and 44 glomerulus, antibody, immunofluorescence, antigen retrieval. characterize OR antibodies. But these gene-targeted 45 8 mutations are designed to visualize the OSNs (including 46 their cilia, dendrite, cell body, axon, and axon terminals) 47 that express a given OR, and do not enable the 48 visualization of the subcellular localization of the OR 49 protein within the OSNs. An attempt at creating an OR: 50 *Corresponding author. GFP C-terminal fusion protein by gene targeting 51 E-mail addresses: [email protected] (V. F. Low), peter.mom- resulted, for reasons that remain unclear, in altered 52 [email protected] (P. Mombaerts). Abbreviations: aa, amino acid; AR, antigen retrieval; EDTA, positions of the corresponding glomeruli in the olfactory 53 ethylenediaminetetraacetic acid; GFP, green fluorescent protein; bulb (Feinstein et al., 2004). 54 IRES, internal ribosome entry site; MOE, main olfactory epithelium; Antibodies against mouse ORs revealed that OR 55 OMP, olfactory marker protein; OR, odorant receptor; OSN, olfactory 56 sensory neuron; PBS, phosphate-buffered saline; PFA, proteins are present not only in OSN cilia, dendrite, and paraformaldehyde; SDS, sodium dodecyl sulfate. cell body but also within the axon and axon terminals 57

http://dx.doi.org/10.1016/j.neuroscience.2016.12.044 0306-4522/Ó 2016 The Author(s). Published by Elsevier Ltd on behalf of IBRO. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 1 NSC 17525 No. of Pages 11 7 January 2017

2 V. F. Low, P. Mombaerts / Neuroscience xxx (2017) xxx–xxx

58 (Barnea et al., 2004; Strotmann et al., 2004), thus sup- 5 mg/kg body weight, respectively) and perfused with 115 59 porting the notion that the OR protein is involved in axonal 5 ml ice-cold phosphate-buffered saline (PBS), followed 116 60 wiring of OSNs. by 15 ml 2% paraformaldehyde in PBS (PFA). The 117 61 A quarter of a century after the discovery of mouse heads were dissected, postfixed in 2% PFA for 118 62 mammalian OR genes (Buck and Axel, 1991), there are 2 h at 4 °C, and decalcified in 0.45 M EDTA in PBS 119 63 still few papers on immunofluorescence staining of the overnight at 4 °C. Samples were cryoprotected in, 120 64 mouse MOE and olfactory bulb with OR antibodies that successively, 15% and 30% sucrose in PBS at 4 °C, 121 65 are commercially available. There are numerous antibod- frozen in O.C.T. Compound (Tissue-Tek), sectioned at 122 66 ies against ORs on the market, but they remain largely 12 lm with a Leica CM3500 cryostat, and collected onto 123 67 uncharacterized. The signal yielded by such antibodies glass slides. Alternative fixation methods were also 124 68 can be affected by the high sequence homology between tested; perfusion with 4% PFA resulted in poorer 125 69 ORs, by post-translational modifications, by protein fold- antigenicity, and immersion fixation in 2% PFA for 2 h 126 70 ing, by protein–protein interactions, and by fixation without perfusion did not yield improved antigenicity. 127 71 crosslinking. To increase the accessibility of the antigen 72 to the antibody, antigen retrieval can be used to break Antigen retrieval. Various antigen retrieval steps were 128 73 some of the crosslinks formed during fixation and to alter tested to optimize visualization of the target protein. 129 74 the protein re-folding following heat denaturation (Shi These included a 10-min incubation in 2 M HCl at room 130 75 et al., 2001; Emoto, 2005). Epitope unmasking through temperature, or 10 min in ice-cold 100% methanol 131 76 antigen retrieval can allow for a more accurate detection, followed by 0.5% sodium dodecyl sulfate (SDS) in PBS 132 77 and therefore representation, of the antigen. at room temperature. Antigen retrieval with a pressure 133 78 Here, we describe our screening of 40 commercially cooker (2100 Retriever; BioVendor) was carried out with 134 79 available antibodies against human and mouse ORs in 10 mM citric acid pH 6.0 and 10 mM Tris pH 9.0 with or 135 80 the mouse MOE using various antigen retrieval without 0.05% Tween 20 and 2 mM EDTA. After heating 136 81 techniques, in coronal 12-lm sections of the MOE of in a pressure cooker to 121 °C, slides were left to cool 137 82 21- day-old C57BL/6J mice. We obtained reliable for 2 h in the retrieval buffer. Additionally, PBS washes 138 83 immunoreactive signals with two antibodies against were tested with 0.05% Tween 20 or 0.2% Triton X-100, 139 84 human ORs (due to presumed crossreactivity with and incubation with antibody was tested for various 140 85 mouse ORs) and six antibodies against mouse ORs in times during a period of 1–4 days, with and without 141 86 the mouse MOE. Of these six, two antibodies also 0.1% Triton X-100. Best results were obtained with PBS 142 87 stained specific glomeruli in the olfactory bulb. washes and incubation with antibody diluted with 0.1% 143 Triton X-100. There was no obvious difference in 144 88 EXPERIMENTAL PROCEDURES antigen detection with different incubation times. 145 89 Mice 146 90 Mice were 21 days old. Wild-type mice were C57BL/6J or Immunohistochemistry. Slides were washed 147 91 129S6/SvEvTac. Our gene-targeted strains (Feinstein 3 10 min with PBS, and, if necessary, antigen retrieval 148 92 et al., 2004) are publicly available from The Jackson Labo- was carried out and the slides were again washed. 149 93 ratory (Bar Harbor, ME, USA): mI7-IRES-tauGFP (https:// Slides were then blocked with 10% normal goat serum 150 94 www.jax.org/strain/006664), and M50-IRES-GFP-IRES- (Jackson ImmunoResearch Laboratories) or normal 151 95 taulacZ (https://www.jax.org/strain/006686). The gene- donkey serum (Millipore) in PBS for 1 h at room 152 96 targeted strain MOR28-IRES-gap-GFP (Serizawa et al., temperature. After blocking, slides were incubated with 153 97 2000) is publicly available from the RIKEN BioResource the primary antibodies in 1% normal goat serum or 154 98 Center (Tsukuba, Japan) (https://www2.brc.riken.jp/lab/an- normal donkey serum with 0.1% Triton X-100 in PBS ° 155 99 imal/detail.php?brc_no=RBRC02928). Mice were main- overnight at 4 C. Commercial rabbit antibodies against 156 100 tained in specified pathogen-free conditions in individually ORs were tested at dilutions of 1:200, 1:500, and 157 101 ventilated cages of the Tecniplast green line. Mice received 1:2000. The following antibodies against human ORs 158 102 ad libitum gamma-irradiated ssniff V1124-727 (ssniff, were obtained from Osenses (Keswick, Australia): 159 103 Soest, Germany). Nesting, bedding, and enrichment were OR11H4, OR1B8, OR1D4, OR1E1, OR1L8, OR1M1, 160 104 provided as nestpak, Datesand Grade 6 (Datesand, OR1N2, OR2A4, OR2D3, OR2F1, OR2H1, OR2K2, 161 105 Manchester, United Kingdom). Mouse experiments were OR2T1, OR2V1, OR2W1, OR4C11, OR51E1, OR8A1. 162 106 performed in accordance with the German Animal Welfare The following antibodies against mouse ORs were 163 107 Act, European Communities Council Directive 2010/63/EU, obtained from Osenses: Olfr2, Olfr6 (x2; different 164 108 and institutional ethical and animal welfare guidelines of the peptide sequences), Olfr15, Olfr16, Olfr24, Olfr56, 165 109 Max Planck Research Unit for Neurogenetics. All efforts Olfr151, Olfr156, Olfr362, Olfr412, Olfr521, Olfr545, 166 110 were made to minimize the number of animals used and Olfr552, Olfr554, Olfr685, Olfr749, Olfr831, Olfr1507. 167 111 their suffering. We also tested sheep OR51E1 and chicken Olfr73 antibodies (Osenses), and rabbit Olfr1303 antibody (Life 168 Technologies). Mature OSNs were visualized with goat 169 112 Immunohistochemical staining OMP antibodies (olfactory marker protein; 1:1000; Wako 170 Chemicals), and green fluorescent protein (GFP) was 171 113 Tissue preparation. Mice were anesthetized by detected with chicken GFP antibodies (1:1000; Aves 172 114 injection of ketamine HCl and xylazine (120 mg/kg and Labs). After incubation with primary antibodies, slides 173

V. F. Low, P. Mombaerts / Neuroscience xxx (2017) xxx–xxx 3

174 were washed with PBS, then incubated with secondary blocked with TNB (0.5% NEN Blocking reagent in TN; 215 175 antibodies: donkey anti-goat Alexa488, donkey anti- 100 mM Tris–Cl pH 7.5, 150 mM NaCl) for 1 h at room 216 176 sheep Alexa488, goat anti-chicken Alexa488, donkey temperature, and incubated with streptavidin–HRP 217 177 anti-rabbit Alexa555, goat anti-rabbit Alexa594 (1:1000; (1:1500; PerkinElmer) and an OR antibody: rabbit Olfr156 218 178 Invitrogen), and DAPI (1:10,000; Molecular Probes) for antibody, rabbit Olfr412 antibody, and rabbit Olfr552 anti- 219 179 1 h at room temperature. Slides were washed with PBS body (1:2000; Osenses) diluted in TNB with 0.1% Triton 220 180 and coverslipped with Mowiol (Calbiochem). X-100, for 3 nights at 4 °C. Slides were then washed 221 thoroughly with TN-T (TN with 0.05% Tween 20). 222 181 Combined immunohistochemistry and in situ Streptavidin-labeled biotinylated probes were visualized 223 182 hybridization by incubation with tyramide-fluorescein (1:50; PerkinEl- 224 mer) diluted with 1 amplification diluent (PerkinElmer) 225 for 10 min. Finally, in order to visualize the antibody, 226 183 Tissue preparation. Mice were anesthetized by slides were washed with TN-T and incubated with goat 227 184 injection of ketamine HCl and xylazine (120 mg/kg and anti-rabbit Alexa594 (1:1000; Invitrogen) and DAPI 228 185 5 mg/kg body weight, respectively) and perfused with (1:10,000) for 2 h at room temperature. Slides were 229 186 15 ml ice-cold 4% PFA. The mouse heads were washed with TN before coverslipping with Mowiol 230 187 dissected, postfixed in 4% PFA, and decalcified in (Calbiochem). 231 188 0.45 M EDTA in PBS overnight at 4 °C. Samples were 189 cryoprotected in 15% and 30% sucrose in PBS at 4 °C, 190 frozen in O.C.T. Compound (Tissue-Tek), sectioned at Image analysis 232 191 12 lm with a Leica CM3500 cryostat, and collected onto 192 glass slides. Microscopy. Following labeling with immunohisto- 233 193 In situ hybridization and immunohistochemistry. OR chemistry, and immunohistochemistry combined with 234 194 RNAs were detected with biotinylated riboprobes in situ hybridization, slides were examined under a 235 195 against Olfr156 (Strotmann et al., 1999); Olfr412 (forward Zeiss LSM 710 confocal microscope. 236 0 0 0 196 5 -ATGGACGGAGGCAACCAGA-3 , reverse 5 -GGCTT 0 197 CCTTAGCAGTCTTCCC-3 ), and Olfr552 (Addgene Cell counting. Every fifth section from the mouse MOE 237 198 plasmid #15843; Hirota et al., 2007). Riboprobes were was labeled with rabbit Olfr1507 antibody (visualized with 238 199 generated as described in Ishii et al. (2004). Combined Alexa 594) and chicken GFP antibody (visualized with 239 200 immunohistochemistry and in situ hybridization was car- Alexa 488). Slides were scanned with a Pannoramic 240 201 ried out using a modified protocol from Ishii et al. Midi Scanner (3D Histech). The diameter of cell nuclei 241 202 (2004). Briefly, slides were washed with PBS and antigen was measured with Panoramic Viewer Software (3D 242 203 retrieval was carried out with 10 mM citric acid pH 6.0. Histech). Each section was exported as a tiff file using 243 204 Next the slides were fixed with 4% PFA for 15 min, the Panoramic Viewer Software, and cells were counted 244 205 washed 5 min with PBS, followed by blocking endoge- using the cell counter plugin in Fiji ImageJ version 1.50 g. 245 206 nous peroxidases with 0.1% H2O2 in PBS for 30 min. 207 Slides were washed in PBS, incubated in 0.2 M HCl for 208 10 min, followed by another PBS wash. Slides were then Statistics. Statistical analysis was performed using 246 209 incubated for 10 min in 0.1 M Triethanol amine-HCl, GraphPad Prism 5. Values are shown as the mean ± 247 210 0.25% acetic anhydrate; pH 8.0. After being washed in standard error of the mean. A t-test was used to 248 211 PBS, slides were dehydrated in graduated ethanol and compare between two means. For multiple 249 212 left to air-dry. Probes were diluted in hybridization solution comparisons, the one-way ANOVA Newman–Keuls 250 213 and the slides were incubated overnight at 65 °C. Slides multiple comparison test was used to test for 251 214 were then washed as described in Ishii et al. (2004), significance. Significance was set at P 6 0.05. 252

Table 1. List of antibodies against ORs

Antigen Host Antigen peptide region Catalogue Source Optimal antigen retrieval Optimal antibody dilution

OR1D4 Rabbit Not speciﬁed OSR00252W Osenses Citric acid AR 1:2000 OR1L8 Rabbit aa 200–250 OSR00258W Osenses Citric acid AR 1:2000 OR51E1 Rabbit aa 270–317 OSR00201W Osenses Citric acid AR 1:2000

Olfr2 Rabbit aa 270–325 OSR00025G Osenses Citric acid AR 1:2000 Olfr6 Rabbit aa 275–316 OSO00046W Osenses Citric acid AR, Tris AR 1:2000 Olfr156 Rabbit Not speciﬁed OSR00044W Osenses Citric acid AR, methanol/SDS 1:2000 Olfr412 Rabbit aa 260–312 OSR00211W Osenses Citric acid AR 1:2000 Olfr552 Rabbit aa 270–317 OSR00204W Osenses No AR, Citric acid AR 1:2000 Olfr1507 Rabbit aa 265–313 OSR00212W Osenses No AR, Citric acid AR 1:5000

List of commercial antibodies raised against human and mouse ORs that yielded immunostaining in cryosections of the mouse MOE. As the peptide sequence of the antigens is not available, the region is given from which the antigenic peptide was designed. Antigen retrieval conditions, and optimal dilutions of the antibodies are indicated. aa; amino acid, Citric acid AR; Antigen retrieval with citric acid buﬀer pH 6.0 in a pressure cooker, No AR; No antigen retrieval, Tris AR; Antigen retrieval with Tris buﬀer pH 9.0 in a pressure cooker.

4 V. F. Low, P. Mombaerts / Neuroscience xxx (2017) xxx–xxx

253 RESULTS Olfr2, one of the Olfr6 antibodies (OSO00046W), 310 Olfr156, Olfr412, Olfr552, and Olfr1507. 311 254 We screened 40 commercially available antibodies that The Olfr2 and Olfr6 antibody labeling was observed in 312 255 were raised against human or mouse ORs in the mouse the ventral-lateral aspect of the MOE, and appeared weak 313 256 MOE, testing various antigen retrieval methods to (Fig. 2). Olfr2 labeling was best observed after antigen 314 257 optimize visualization of the antigen. Of these 40 retrieval in the pressure cooker with citric acid, and 315 258 antibodies, 19 were raised against human ORs with Olfr6 labeling was observed with both citric acid and 316 259 80% amino acid identity to mouse ORs, and 21 were Tris antigen retrieval buﬀers in the pressure cooker. Low 317 260 against mouse ORs. Here we present the results from 9 signal could be due to low protein expression of the 318 261 of the antibodies that yielded reproducible immuno- Olfr2 and Olfr6 ORs, low sensitivity of the antibodies, or 319 262 staining in the mouse MOE (Table 1). For the sake of suboptimal ﬁxation and antigen retrieval procedures. 320 263 brevity, we refer to the antibody with the name of the OR Fortunately, there are gene-targeted mouse strains 321 264 against which it was raised: ORxxx for human ORs, and available for these two ORs, respectively mI7-IRES- 322 265 Olfrxxx for mouse ORs. tauGFP (Bozza et al., 2002) and M50-IRES-GFP-IRES- 323

266 Screening of antibodies against 267 human ORs in the mouse MOE

268 Of the 19 antibodies against human 269 ORs, three yielded consistent 270 immunoreactivity when antigen 271 retrieval was carried out in a 272 pressure cooker with citric acid: 273 OR1D4, OR1L8, and OR51E1. Both 0 274 the OR1D4 antibody (Fig. 1A, B )and 0 275 the OR1L8 antibody (Fig. 1C, D )co- 276 labeled with the OMP antibody, 277 suggesting that they cross react with 278 one or more mouse OR proteins in 279 the MOE. The mouse ORs with the 280 closest homology to these human 281 ORs are Olfr412 (81% identity with 282 OR1D4 across the entire amino acid 283 sequence of the OR) and Olfr355 284 (78% identity with OR1L8 in the 285 amino acid region 200–250 from 286 which the antigenic peptide was 287 designed). For the OR1D4 and 288 OR1L8 antibodies, immunoreactivity 289 was observed midway along the 290 dorsal-medial and ventral-lateral 291 aspect of the MOE. OR51E1 292 immunoreactivity was observed more 293 ventral-laterally. The rabbit OR51E1 294 antibody reacted with an antigen in 295 the mouse MOE (OR51E1 has 87% 296 identity with mouse Olfr558 in the 297 amino acid region 270–317 from 298 which the antigenic peptide was 299 designed): signal was detected in cell 300 bodies, but there was no evidence of 301 immunoreactivity in cilia or dendrites. 302 Furthermore OR51E1 immuno- Fig. 1. Labeling of the mouse MOE by antibodies against human ORs. (A) OR1D4-immunoreactive 303 reactivity did not colocalize with OMP cells in the MOE (red). Mature OSNs were detected with OMP antibody (green). Merge with DAPI 0 304 immunoreactivity (Fig. 1E, F ). (blue). (B) High magnification of insert in A showing weak OR1D4 antibody labeling in the dendrite and cell body of OMP-immunoreactive cells in the MOE (merged red and green, B0). (C) OR1L8- immunoreactive cells labeled in the MOE (red). Mature OSNs were detected with OMP antibody (green). Merge with DAPI (blue). (D) High magnification of insert in C. Antibody labeling is observed 305 Screening of antibodies against in the cilia and cell body, and colocalized with OMP-immunoreactive cells (merged red and green, 306 mouse ORs in the mouse MOE D0). (E) OR51E1 labeling was detected in a population of cells in the MOE (red). Mature OSNs were detected with OMP antibody (green). Merge with DAPI (blue).(F) High magnification of insert in E 307 Of the 21 antibodies against mouse demonstrating strong OR51E1 immunoreactivity in the cell bodies of a population of cells that were 308 ORs, we observed consistent not OMP-immunoreactive (merged red and green, F0). (For interpretation of the references to color 309 immunoreactivity with six antibodies: in this figure legend, the reader is referred to the web version of this article.)

V. F. Low, P. Mombaerts / Neuroscience xxx (2017) xxx–xxx 5

Fig. 2. Labeling of the mouse MOE by antibodies against mouse ORs Olfr2 and Olfr6. (A) Olfr2-immunoreactive cells (red) in the MOE of an mI7- IRES-tauGFP/ mouse. These cells were also detected with GFP antibody (green). Merge with DAPI (blue). (B) High magnification of insert in A showing Olfr2 immunoreactivity in the cilia, dendrite, and cell body, and colocalization with GFP antibody signal (merged red and green, B0). (C) Olfr2 immunoreactivity was not detected in a GFP-immunoreactive (green) glomerulus of an mI7-IRES-tauGFP/ mouse. Merge with DAPI (blue) (C0). (D) Olfr6 immunoreactivity (red) was weak, and only occasionally detected in association with GFP immunoreactivity (green) in an M50-IRES- GFP-IRES-taulacZ/ mouse. Merge with DAPI (blue). (E) High magnification of insert in D. Weak Olfr6 immunoreactivity is observed in the cell body, dendrite and cilia of cells that express M50-IRES-GFP-IRES-taulacZ (merged red and green, E0). (F) Olfr6 immunoreactivity was not detected in GFP-immunoreactive (green) glomeruli of an M50-IRES-GFP-IRES-taulacZ/ mouse. Merge with DAPI (F0). Scale bar = 20 lm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

324 taulacZ (Feinstein et al., 2004). These strains enabled us dorsal-medial and ventral-lateral aspect of the MOE. 351 0 0 325 to validate that Olfr2 (Fig. 2A, B ) and Olfr6 (Fig. 2D, E ) Thus, although Olfr412 is the mouse OR with the 352 326 antibody labeling was limited to cells that were also highest homology (80%) to human OR1D4, it is unlikely 353 327 GFP positive, indicating that the antibody is specific to to be the antigen that the OR1D4 antibody detects. 354 328 the respective OR. Neither antibody against Olfr2 The Olfr552 antibody also showed strong 355 0 0 0 329 (Fig. 2C, C ) nor Olfr6 (Fig. 2D, D ) detected protein immunoreactivity in the mouse MOE (Fig. 3E, F ). This 356 330 expression in the GFP-positive glomeruli in the olfactory antibody did not require antigen retrieval, but the signal 357 331 bulb. was greatly improved following antigen retrieval with 358 332 The Olfr156 antibody yielded signal midway along the citric acid. Signal was detected through the dorsal- 359 333 dorsal-medial and ventral-lateral aspect of the MOE, and medial aspect of the mouse MOE. Two glomeruli were 360 0 334 appeared to have low immunoreactivity (Fig. 3A, B ). detected in the olfactory bulb. Consistent with Olfr552 361 335 SDS/methanol antigen retrieval sporadically reveled belonging to class I ORs, the labeled glomeruli resided 362 336 positive labeling, and citric acid antigen retrieval yielded within the class I dorsal domain of the olfactory bulb. 363 337 more reproducible signal. In both cases only cilia were The medial glomerulus was located midway along the 364 338 labeled consistently, and the cell body, and dots within anterior-posterior extent, and the lateral glomerulus 365 339 the dendrites, were visible on rare occasions. It is (Fig. 3G) was located more anteriorly. 366 340 possible that the antigen is more accessible in certain For Olfr156, Olfr412, and Olfr552, there are no gene- 367 341 subcellular localizations; there may be modifications targeted mice available to validate that the antibody is 368 342 occurring to the OR protein that allow detection under specific for the respective OR. Instead we used in situ 369 343 certain conditions; or an increased density of OR protein hybridization combined with immunohistochemistry to 370 344 in certain subcellular localizations, such as in the cilia, verify that these antibodies were labeling the intended 371 345 may make it more easy to visualize. OR. Combining these methods compromises the quality 372 346 The Olfr412 antibody showed strong immunoreactivity of both techniques, and we were unable to detect 373 347 following citric acid antigen retrieval, and signal was antibody labeling with Olfr156 in combination with in situ 374 348 detected in the dorsal-medial aspect of the mouse hybridization. We were able to detect both protein and 375 0 0 349 MOE (Fig. 3C, D ). This pattern differs from the RNA in the mouse MOE for Olfr412 (Fig. 4A, B ) and 376 0 350 OR1D4 antibody, which yielded signal midway along the Olfr552 (Fig. 4C, D ). We could thus verify that Olfr412 377

6 V. F. Low, P. Mombaerts / Neuroscience xxx (2017) xxx–xxx

strong immunoreactivity in GFP- 392 positive glomeruli in the olfactory bulb 393 0 (Fig. 5C, C ). Having a reliable anti- 394 body that detects a mouse OR for 395 which we have a gene-targeted 396 mouse strain available gave us the 397 opportunity to investigate monoallelic 398 expression of Olfr1507 in mice 399 heterozygous for MOR28-IRES-gap- 400 GFP, with the other, wild-type 401 MOR28/Olfr1507 allele either of 402 C57BL/6J origin (abbreviated B6) or 403 129S6/SvEvTac origin (abbreviated 404 0 129) (Fig. 5D, E ). The MOR28/ 405 Olfr1507 gene targeting had been 406 carried out in an embryonic stem cell 407 line of 129 origin (Serizawa et al., 408 2000). We observed an even distribu- 409 tion of cells that are doubly GFP 410 positive and antibody positive 411 throughout the MOE, versus cells that 412 are antibody-only positive. In the 413 olfactory bulb we found that the 414 GFP-positive glomeruli were labeled 415 with the Olfr1507 antibody in 416 MOR28-IRES-gap-GFP+/ mice 417 0 (Fig. 5F, F ). The glomeruli did not 418 appear to be homogeneously labeled 419 with Olfr1507 and GFP antibodies in 420 MOR28-IRES-gap-GFP+/ mice. 421 Instead, there is a segregation of 422 axon terminals expressing the gene- 423 targeted and wild-type Olfr1507 alle- 424 les. This segregation was observed 425 in MOR28-IRES-gap-GFP+/ (129) 426 mice (Fig. 5F’) and in MOR28-IRES- 427 Fig. 3. Labeling of the mouse MOE by antibodies against mouse ORs Olfr156, Olfr412, and gap-GFP+/ (B6) mice (data not 428 Olfr552. (A) Olfr156-immunoreactive cells (red). Mature OSNs were detected with the OMP shown). 429 antibody (green). Merge with DAPI (blue). (B) High magnification of insert in A showing Olfr156 immunoreactivity in the cilia and dendrite. Merge with OMP immunoreactivity (B0). (C) Olfr412- Counting Olfr1507-immuno- 430 immunoreactive cells (red). Mature OSNs were detected with OMP (green). Merge with DAPI (blue). (D) High magnification of insert in C showing Olfr412 immunoreactivity in the cilia, dendrite reactive cells in the MOE 431 and cell body, colocalizing with OMP immunoreactivity (merged, D0). (E) Olfr552-immunoreactive cells (red). Mature OSNs were detected with OMP (green), and merged with DAPI (blue). (F) High We counted Olfr1507-immuno- 432 magnification of insert in E showing strong Olfr552 immunoreactivity in the cilia, dendrite and cell reactive cells in every fifth coronal 433 0 body of cells that are labeled with OMP antibody (F ). (G) Olfr552 immunoreactivity was also section of the MOE from five types 434 l detected in a glomerulus. Merge with DAPI. Scale bar = 20 m. (For interpretation of the of mice, 20 in total: MOR28-IRES-g 435 references to color in this figure legend, the reader is referred to the web version of this article.) ap-GFP/ (n = 5), MOR28-IRES- 436 gap-GFP+/ (B6) (n = 4), MOR28- 437 378 and Olfr552 antibody labeling was closely associated with IRES-gap-GFP+/ (129) (n = 3), 438 379 the signal from respective riboprobes in the mouse MOE, wild-type C57BL/6J (n = 5), and wild-type 129S6/ 439 380 indicating the antibodies are labeling the cells that SvEvTac (n = 3) mice. To correct for overcounting in 440 381 express the respective OR gene. The sensitivity and sections, we performed an Abercrombie correction 441 382 specificity of these antibodies could not be verified using (Abercrombie, 1946; Bressel et al., 2016)(Table 2). We 442 383 this method of combined labeling. observed no significant difference in the nuclear diameter 443 384 Finally, the Olfr1507 antibody showed strong between strains, and used a correction factor of 444 385 immunoreactivity in the ventral-lateral aspect of the 0.68–0.69. 445 386 mouse MOE (Fig. 5), and did not require antigen In all five types of mice, we observed a distinct 446 387 retrieval to be visualized. Signal could be detected in the ‘M’-pattern distribution of Olfr1507-immunoreactive 447 388 cilia, dendrite, and cell body of OSNs. Immunostaining OSNs in the MOE, with the peaks corresponding to the 448 389 colocalized with GFP expression in mice homozygous ectoturbinates of the MOE (Fig. 6A). The first Olfr1507- 449 390 for the MOR28-IRES-gap-GFP gene-targeted mutation immunoreactive cells were observed midway along the 450 0 391 (Serizawa et al., 2000)(Fig. 5A, B ). We also detected anterior–posterior dimension, between 3.5 and 4 mm 451

V. F. Low, P. Mombaerts / Neuroscience xxx (2017) xxx–xxx 7

(B6) and MOR28-IRES-gap-GFP 480 +/ (129) mice, we observed a 481 homogeneous distribution of the two 482 populations throughout the MOE 483 (Fig. 6A). However, there was a 484 slight preference for expression of 485 the wild-type, B6-derived allele in 486 MOR28-IRES-gap-GFP+/ (B6) 487 mice, with 56.38 ± 1.9% of cells 488 counted expressing the wild-type 489 Olfr1507 allele and 43.01 ± 1.8% of 490 cells expressing the mutant 491 Olfr1507 allele (P < 0.002; Fig. 6B). 492 (An additional 0.62 ± 0.4% of the 493 cells counted reacted with the GFP 494 antibody only.) In MOR28-IRES- 495 gap-GFP (129) mice, we found that 496 51.12 ± 0.8% of cells expressed 497 the wild-type, 129-derived allele and 498 47.21 ± 0.1% of cells expressed 499 the gene-targeted allele 500 (P < 0.008). (An additional 1.67 ± 501 0.8% of cells counted reacted with 502 the GFP antibody only.) Although 503 the expression of the 129 wild-type 504 Fig. 4. In situ hybridization combined with immunohistochemistry in the mouse MOE. (A) The same sells were detected with the Olfr412 antibody (red) and the Olfr412 riboprobe (green). Merge vs gene-targeted allele is still 505 with DAPI (blue). (B) High magnification of insert in A showing Olfr412 immunoreactivity in the cilia, statistically different, it is closer to 506 0 dendrite, and cell body, and colocalizing with Olfr412 riboprobe hybridization in the cell body (B ). the expected 50/50 distribution than 507 (C) Cells were detected with Olfr552 antibody (red) and Olfr552 riboprobe (green) Merge with DAPI when the wild-type allele is of B6 508 (blue). (D) High magnification of insert in C showing Olfr552 immunoreactivity in the cilia, dendrite, and cell body, and colocalizing with Olfr552 riboprobe hybridization detected in the cell body (D0). origin. Note that expression of the 509 Scale bar = 20 lm. (For interpretation of the references to color in this figure legend, the reader is gap-GFP reporter is driven by a 510 referred to the web version of this article.) 129-derived promoter sequence in 511 the gene-targeted mutation 512 (Serizawa et al., 2000). 513 452 from the start of the MOE. The expression peaked after Further support for the idea that 514 453 approximately 1 mm, then began to drop off, but the probability of expression of Olfr1507 is higher for the 515 454 expression quickly increased again, and continued to B6 allele than for the 129 allele, came from comparing 516 455 the most posterior portion of the MOE. the numbers of immunoreactive cells in the various 517 456 In MOR28-IRES-gap-GFP/ mice, we found that strains. In MOR28-IRES-gap-GFP/ mice we counted 518 457 0.94 ± 0.4% of the counted cells were labeled with only 31,657 ± 2305 Olfr1507-immunoreactive cells, and 519 458 the Olfr1507 antibody but not with the GFP antibody, these numbers were not statistically different from 520 459 suggesting that the antibody is highly specific for MOR28-IRES-gap-GFP+/ (129) mice (36,897 ± 521 460 Olfr1507-expressing cells in the mouse MOE. 1547), or 129S6/SvEvTac wild-type mice (24,985 ± 522 461 Conversely, in these mice only 0.40 ± 0.2% of the 1448) (Fig. 6C). Likewise, the numbers of cells counted 523 462 counted cells were labeled with the GFP antibody but in MOR28-IRES-gap-GFP+/ (B6) mice (41,988 ± 524 463 not with the Olfr1507 antibody, indicating that the 3142) and in C57BL/6 wild-type mice (43,500 ± 3606) 525 464 antibody is also highly sensitive. Therefore, we conclude were not significantly different from each other, but were 526 465 that the antibody against an Olfr1507 peptide that is significantly higher than in MOR28-IRES-gap-GFP/ 527 466 commercially available from Osenses (OSR00212W) is mice and in 129S6/SvEvTac wild-type mice (one-way 528 467 a reliable, specific, and sensitive detector of Olfr1507- ANOVA P < 0.003; Fig. 6C). Incidentally, we did not 529 468 expressing OSNs in the mouse MOE, with 98.66 ± identify a significant correlation between the number of 530 469 0.4% of counted cells labeled with both the Olfr1507 Olfr1507-immunoreactive OSNs counted in the MOE 531 470 antibody and the GFP antibody in MOR28-IRES-gap-GF and the weight of the mice, indicating that the 532 471 P/ mice (Fig. 6B). differences in cell numbers were not due to trivial 533 472 Next, we examined the two populations of Olfr1507- variations in mouse weight (Fig. 6D). 534 473 expressing OSNs for each allele in MOR28-IRES-gap- 474 GFP+/ mice. OSNs that were labeled with both the DISCUSSION 535 475 Olfr1507 and GFP antibodies were considered to 476 express the MOR28-IRES-gap-GFP gene-targeted Here, we have demonstrated that some commercially 536 477 allele, and OSNs that were labeled with only the available antibodies against ORs can be a reliable tool 537 478 Olfr1507 antibody were considered to express the to examine cells expressing a given OR in the mouse 538 479 wild-type allele. In both MOR28-IRES-gap-GFP+/ MOE. 539

8 V. F. Low, P. Mombaerts / Neuroscience xxx (2017) xxx–xxx

Fig. 5. Labeling of the MOE of MOR28-IRES-gap-GFP/ and +/ mice by antibodies against Olfr1507/MOR28. (A) Olfr1507-immunoreactive cells (red) in a MOR28-IRES-gap-GFP/ mouse are colabeled with GFP antibody (green). Merge with DAPI (blue). (B) High magnification of insert in A. Olfr1507 immunoreactivity is detected in the cell body, dendrite, and cilia, and colocalizes with GFP immunoreactivity (B0). (C) Olfr1507 immunoreactivity colocalizes uniformly with GFP immunoreactivity within a glomerulus of a MOR28-IRES-gap-GFP/ mouse. Merge with DAPI (C’). (D) Olfr1507-immunoreactive cells (red) in a MOR28-IRES-gap-GFP+/ (129) mouse are colabeled with GFP antibody (green). Merged with DAPI (blue). (E) High magnification of insert in D. Olfr1507 immunoreactivity is detected in the cell body, dendrite, and cilia, and colocalizes with GFP immunoreactivity (E’). (F) Olfr1507 immunoreactivity (red) segregates within a GFP-immunoreactive (green) glomerulus of an MOR28-IRES- gap-GFP+/ (129) mouse. Merge with DAPI (F0). Scale bar = 20 lm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

was required, the best results were often obtained with 550 Table 2. Abercrombie correction; Nucleus diameter of OSNs antigen retrieval in a pressure cooker with a citric acid 551 Strain Nucleus Abercrombie buffer, pH 6.0. While alternative antigen retrieval 552 diameter correction methods also worked, we found that if signal could still 553 l ( m) factor not be detected after antigen retrieval with citric acid, 554 MOR28-IRES-gap-GFP/ 5.60 ± 0.13 0.68 alternative methods would also be without success. 555 MOR28-IRES-gap-GFP+/ (B6) 5.46 ± 0.11 0.69 Some antibodies yielded strong signals, and for others 556 MOR28-IRES-gap-GFP+/ 5.43 ± 0.10 0.69 the signal was barely or inconsistently detectable. It 557 (129) would be interesting to determine if the variation in 558 C57BL/6J 5.40 ± 0.12 0.69 signal is a result of variable antigenicity of the 559 129S6/SvEvTac 5.30 ± 0.09 0.69 antibodies, or reflects widely differing levels of protein 560 The average nucleus diameter and the standard error of the mean and the expression within OSNs expressing these particular ORs. 561 calculated Abercrombie factors are given for the five types of mice. Scale We confirm and extend the observations of OR 562 bars = 20 lm. immunoreactivity in the cell body and dendrite as well 563 as in the cilia of OSNs (Barnea et al., 2004; Strotmann 564 et al., 2004). In the dendrite, and sometimes in the cell 565 body of OSNs, a distinct dot-like pattern could be 566 540 Antibodies against ORs detected. This pattern of signal could be an artifact of anti- 567 541 We have demonstrated that a substantial fraction (at least body staining. Nevertheless, it is a repeatable observation 568 542 20%) of commercially available antibodies are a viable with various antibodies against ORs, and we reported it 569 543 way for assessing OR expression in the mouse MOE. earlier in the M71::GFP fusion mouse strain (Feinstein 570 544 We found that tissue fixation had an effect on the quality et al., 2004), suggesting that this dot-like pattern of OR 571 545 of antigen detection, with perfusion with 2% PFA being protein signal reflects a biological process. It could be 572 546 optimal. For the antibodies that we were unable to get related to the visualization of specific forms of protein 573 547 to work, other fixation methods may yield better and folding, it may be due to increased density of protein as 574 548 consistent signals. We also tested various antigen a result of pulsed transcription, or it may be related to pro- 575 549 retrieval methods, and found that when antigen retrieval tein trafficking. This issue can be examined further 576

V. F. Low, P. Mombaerts / Neuroscience xxx (2017) xxx–xxx 9

Fig. 6. Numbers of Olfr1507-immunoreactive cells in the MOE of wild-type and gene-targeted mice. (A) Average number of cells counted per coronal section through the anterior–posterior dimension of the MOE in five types of mice: MOR28-IRES-gap-GFP/, MOR28-IRES-gap-GFP+/ (B6), MOR28-IRES-gap-GFP+/ (129), C57BL/6J, and 129S6/SvEvTac. Cells labeled with the Olfr1507 antibody (red), the GFP antibody (green), and both the Olfr1507 and GFP antibodies (orange). (B) Percentages of cells labeled with the Olfr1507 antibody (red), the GFP antibody (green), and both the Olfr1507 and GFP antibodies (orange) in three types of mice: MOR28-IRES-gap-GFP/, MOR28-IRES-gap-GFP+/ (B6), and MOR28-IRES-gap-GFP+/ (129) mice. (C) Comparison of the number of cells counted in each strain. Cell counts in every fifth section were multiplied by five, and subjected to Abercrombie correction. (D) There is no correlation between the numbers of Olfr1507-immunoreactive cells and mouse weight. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

577 through the use of antibodies against ORs in combination In both MOR28-IRES-gap-GFP+/ (B6) and 606 578 with membrane markers, and with super-resolution MOR28-IRES-gap-GFP+/ (129) mice, we observed a 607 579 microscopy. compartmentalization of immunoreactive signal within 608 the glomeruli: they were labeled uniformly with the 609 580 In situ hybridization combined with Olfr1507 antibody, but visualized only partially with the 610 581 immunohistochemistry GFP antibody. The segregation in glomeruli in MOR28- 611 IRES-gap-GFP+/ (B6) mice has been reported 612 582 Due to the high homology of mouse ORs, a gene-targeted previously (Ishii et al., 2001), and there is a single amino 613 583 mouse strain is ideal to validate specificity and to assess acid polymorphism between B6 (isoleucine) and 129 614 584 sensitivity of an OR antibody. We have demonstrated that (valine) at residue 50. Interestingly, we also observed a 615 585 in some cases immunohistochemistry can be combined compartmentalization of the two populations of OSNs in 616 586 with in situ hybridization as a form of validation, but this glomeruli of MOR28-IRES-gap-GFP+/ (129) mice. This 617 587 technique has its limitations. It can indicate that an compartmentalization may reflect a subtle difference in 618 588 antibody is targeting the population of ORs that is RNA or protein expression level between the wild-type 619 589 intended, but until this method is further improved, it and 129 alleles, or an undocumented effect of the gap- 620 590 cannot reveal how specific or sensitive an antibody is. GFP axonal marker. In this marker, the 20 N-terminal 621 591 Correlating OR RNA expression with the corresponding amino acid residues of GAP43 are fused to the 622 592 OR protein expression could reveal interesting information N-terminus of the GFP to target it to the plasma mem- 623 593 about OR transcription and translation within OSNs. brane (Moriyoshi et al., 1996). 624

594 Glomerular labeling Probability of OR gene expression 625 595 In most cases we were unable to detect a reliable signal in 596 glomeruli. But with 3,600 glomeruli in the mouse olfactory We used MOR28-IRES-gap-GFP/ mice (Serizawa 626 597 bulb (Richard et al., 2010), it remains possible that we et al., 2000) as validation that the Olfr1507 antibody is 627 598 overlooked the glomerular signal for some antibodies. highly speciﬁc and sensitive for Olfr1507-expressing 628 599 For two antibodies (Olfr552 and Olfr1507/MOR28) we OSNs. We also determined that Olfr1507-expressing 629 600 were able to visualize speciﬁc glomeruli. Interestingly, OSNs are distributed in an ‘M’-like pattern in the MOE in 630 601 both of these antibodies did not require antigen retrieval a series of coronal sections. We have described a similar 631 602 to be detected. Even when antigen retrieval was pattern of distribution of OSNs in the MOE for OSNs 632 603 performed, the glomeruli were still detected, suggesting expressing mI7/Olfr2, M50/Olfr6, or SR1/Olfr124 633 604 that antigen retrieval itself was not inhibiting the (Bressel et al., 2016). The number of Olfr1507- 634 605 detection of glomeruli with the other antibodies. expressing OSNs in a 21-day-old C57BL/6J mouse 635

10 V. F. Low, P. Mombaerts / Neuroscience xxx (2017) xxx–xxx

636 (43,500 ± 3606) is slightly higher than the number of Bozza T, Feinstein P, Zheng C, Mombaerts P (2002) Odorant 692 637 MOR256-17-expressing OSNs (37,023 ± 6318) receptor expression. J Neurosci 22:3033–3043. 693 Bressel OC, Khan M, Mombaerts P (2016) Linear correlation between 694 638 (Bressel et al., 2016). the number of olfactory sensory neurons expressing a given 695 639 When we assessed Olfr1507 expression in MOR28- mouse odorant receptor gene and the total volume of the 696 640 IRES-gap-GFP+/ (B6) mice, we found that there was corresponding glomeruli in the olfactory bulb. J Comp Neurol 697 641 a preferential expression of the wild-type, B6-derived 524:199–209. 698 642 allele. We observed a reduced preference of expression Buck L, Axel R (1991) A novel multigene family may encode odorant 699 643 of the wild-type, 129-derived allele in MOR28-IRES-gap- receptors: a molecular basis for odor recognition. Cell 700 65:175–187. 701 644 GFP+/ (129) mice. When we compared the number Emoto K (2005) Mechanisms of heat-induced antigen retrieval: does 702 645 of Olfr1507-expressing OSNs in wild-type mice, we pH or ionic strength of the solution play a role for refolding 703 646 found that there were more cells in B6 mice than in 129 antigens? J Histochem Cytochem 53:1311–1321. 704 647 mice, suggesting that the probability of Olfr1507 gene Feinstein P, Bozza T, Rodriguez I, Vassalli A, Mombaerts P (2004) 705 648 choice is higher when driven by the B6 promoter than Axon guidance of mouse olfactory sensory neurons by odorant 706 649 the 129 promoter. There does appear to be an receptors and the b2 adrenergic receptor. Cell 117:833–846. 707 708 650 additional effect of the targeted mutations in terms of a Fuss SH, Omura M, Mombaerts P (2007) Local and cis effects of the H element on expression of odorant receptor genes in mouse. 709 651 slight reduction in the probability of gene choice. Cell 130:373–384. 710 652 Interestingly, Olfr1507 is the most proximal OR gene Hirota J, Omura M, Mombaerts P (2007) Differential impact of Lhx2 711 653 among a cluster of seven OR genes that are regulated deficiency on expression of class I and class II odorant receptor 712 654 by the H element (Tsuboi et al., 1999; Serizawa et al., genes in mouse. Mol Cell Neurosci 34:679–688. 713 655 2000, 2003). When the H element is deleted by gene tar- Ishii T, Serizawa S, Kohda A, Nakatani H, Shiroishi T, Okumura K, 714 656 geting, the expression of Olfr1507 is abolished entirely, Iwakura Y, Nagawa F, Tsuboi A, Sakano H (2001) Monoallelic 715 expression of the odourant receptor gene and axonal projection of 716 657 and the expression of the more distal OR genes is olfactory sensory neurones. Genes Cells 6:71–78. 717 658 reduced in a distance-correlated fashion (Fuss et al., Ishii T, Omura M, Mombaerts P (2004) Protocols for two- and three- 718 659 2007; Nishizumi et al., 2007; Khan et al., 2011). More- color fluorescent RNA in situ hybridization of the main and 719 660 over, the position of the Olfr1507 coding region in accessory olfactory epithelia in mouse. J Neurocytol 33:657–669. 720 661 C57BL/6 mice is 11 kb closer to the H element compared Khan M, Vaes E, Mombaerts P (2011) Regulation of the probability of 721 662 to mice with a 129 background (Fuss et al., 2007). We mouse odorant receptor gene choice. Cell 147:907–921. 722 Koshimoto H, Katoh K, Yoshihara Y, Mori K (1992) Distribution of 723 663 speculate that the difference in distance of the H element putative odour receptor proteins in olfactory epithelium. 724 664 to the Olfr1507 gene is causally related to the difference in NeuroReport 3:521–523. 725 665 probability of gene choice, measured operationally as the Krieger J, Schleicher S, Strotmann J, Wanner I, Boekhoff I, Raming 726 666 number of Olfr1507-expressing OSNs. K, Degeus P, Breer H (1994) Probing olfactory receptors with 727 sequence-specific antibodies. Eur J Biochem 219:829–835. 728 Miyamichi K, Serizawa S, Kimura HM, Sakano H (2005) Continuous 729 667 CONCLUSIONS and overlapping expression domains of odorant receptor genes in 730 the olfactory epithelium determine the dorsal/ventral positioning of 731 668 Antibodies against mouse ORs are a valuable tool for glomeruli in the olfactory bulb. J Neurosci 25:3586–3592. 732 669 understanding the biological distribution of OR proteins, Mombaerts P, Wang F, Dulac C, Chao SK, Nemes A, Mendelsohn M, 733 670 and to characterize OR-specific glomeruli in the olfactory Edmondson J, Axel R (1996) Visualizing an olfactory sensory 734 671 bulb when no gene-targeted strain is available. Their value map. Cell 87:675–686. 735 672 to olfactory neuroscience may have been overshadowed Moriyoshi K, Richards LJ, Akazawa C, O’Leary DD, Nakanishi S 736 737 673 by the success of the gene-targeted approach. The (1996) Labeling neural cells using adenoviral gene transfer of membrane-targeted GFP. Neuron 16:255–260. 738 674 significance of reports based on custom-designed or Nishizumi H, Kumasaka K, Inoue N, Nakashima A, Sakano H (2007) 739 675 home-made polyclonal antibodies is constrained by the Deletion of the core-H region in mice abolishes the expression of 740 676 lack of the commercial or public availability of these three proximal odorant receptor genes in cis. Proc Natl Acad Sci 741 677 reagents, hampering replicability studies. Here we have USA 50:20067–20072. 742 678 characterized 40 commercially available antibodies Ressler KJ, Sullivan SL, Buck LB (1993) A zonal organization of 743 679 against ORs. We have demonstrated that they have a odorant receptor gene expression in the olfactory epithelium. Cell 744 73:597–609. 745 680 reasonable probability of success for immunofluorescence Ressler KJ, Sullivan SL, Buck LB (1994) Information coding in the 746 681 in the MOE (20–30%), and can be highly specific and olfactory system: evidence for a stereotyped and highly organized 747 682 sensitive. The production of monoclonal antibodies against epitope map in the olfactory bulb. Cell 79:1245–1255. 748 683 ORs will be an important next step. Saraiva LR, Ibarra-Soria X, Khan M, Omura M, Scialdone A, 749 Mombaerts P, Marioni JC, Logan DW (2015) Hierarchical 750 deconstruction of mouse olfactory sensory neurons: from whole 751 684 Acknowledgment—P.M. acknowledges generous support from mucosa to single-cell RNA-seq. Sci Rep 5:18178. 752 685 the Max Planck Society. Serizawa S, Ishii T, Nakatani H, Tsuboi A, Nagawa F, Asano M, Sudo 753 K, Sakagami J, Sakano H, Ijiri T, Matsuda Y, Suzuki M, Yamamori 754 T, Iwakura Y, Sakano H (2000) Mutually exclusive expression of 755 odorant receptor transgenes. Nat Neurosci 3:687–693. 756 686 REFERENCES Serizawa S, Miyamichi K, Nakatani H, Suzuki M, Saito M, Yoshihara 757 Y, Sakano H (2003) Negative feedback regulation ensures the 758 687 Abercrombie M (1946) Estimation of nuclear population from one receptor-one olfactory neuron rule in mouse. Science 759 688 microtome sections. Anat Rec 94:239–247. 302:2088–2094. 760 689 Barnea G, O’Donnell S, Mancia F, Sun X, Nemes A, Mendelsohn M, Shi SR, Cote RJ, Taylor CR (2001) Antigen retrieval techniques: 761 690 Axel R (2004) Odorant receptors on axon termini in the brain. current perspectives. J Histochem Cytochem 49:931–937. 762 691 Science 304:1468.

V. F. Low, P. Mombaerts / Neuroscience xxx (2017) xxx–xxx 11

763 Strotmann J, Hoppe R, Conzelmann S, Feinstein P, Mombaerts P, (1999) Olfactory neurons expressing closely linked and 772 764 Breer H (1999) Small subfamily of olfactory receptor genes: homologous odorant receptor genes tend to project their axons 773 765 structural features, expression pattern and genomic organization. to neighboring glomeruli on the olfactory bulb. J Neurosci 774 766 Gene 236:281–291. 19:8409–8418. 775 767 Strotmann J, Levai O, Fleischer J, Schwarzenbacher K, Breer H Zapiec B, Mombaerts P (2015) Multiplex assessment of the positions 776 768 (2004) Olfactory receptor proteins in axonal processes of of odorant receptor-speciﬁc glomeruli in the mouse olfactory bulb 777 769 chemosensory neurons. J Neurosci 24:7754–7761. by serial two-photon tomography. Proc Natl Acad Sci USA 112: 778 770 Tsuboi A, Yoshihara S, Yamazaki N, Kasai H, Asai-Tsuboi H, E5873–E5882. 779 771 Komatsu M, Serizawa S, Ishii T, Matsuda Y, Nagawa F, Sakano H 780 781 (Received 27 October 2016, Accepted 23 December 2016) 782 (Available online xxxx)