Opposing Roles of Dopamine Receptor D1- and D2-Expressing Neurons in the Anteromedial Olfactory Tubercle in Acquisition of Place Preference in Mice

Olfaction induces adaptive motivated behaviors. Odors associated with food induce attractive behavior, whereas those associated with dangers induce aversive behavior. We previously reported that learned odor-induced attractive and aversive behaviors accompany activation of the olfactory tubercle (OT) in a domain- and cell type-specific manner. Odor cues associated with a sugar reward induced attractive behavior and c-fos expression in the dopamine receptor D1-expressing neurons (D1 neurons) in the anteromedial OT. In contrast, odor cues associated with electrical shock induced aversive behavior and c-fos expression in the pamine receptor D2-expressing neurons (D2 neurons) in the anteromedial OT, as well as the D1 neurons in the lateral OT. Here, we investigated whether the D1 and D2 neurons in the anteromedial OT play distinct roles in attractive or aversive behaviors, using optogenetic stimulation and real-time place preference (RTPP) tests. Mice expressing ChETA (ChR2/E123T)-enhanced yellow fluorescent protein (EYFP) in the D1 neurons in the anteromedial OT spent a longer time in the photo-stimulation side of the place preference chamber than the control mice expressing EYFP. On the other hand, upon optogenetic stimulation of the D2 neurons in the anteromedial OT, the mice spent a shorter time in the photo-stimulation side than the control mice. Local neural activation in the anteromedial OT during the RTPP tests was confirmed by c-fos mRNA expression. These results suggest that the D1 and D2 neurons in the anteromedial OT play opposing roles in attractive and aversive behaviors, respectively.

Olfaction induces adaptive motivated behaviors. Odors associated with food induce attractive behavior, whereas those associated with dangers induce aversive behavior. We previously reported that learned odor-induced attractive and aversive behaviors accompany activation of the olfactory tubercle (OT) in a domain-and cell type-specific manner. Odor cues associated with a sugar reward induced attractive behavior and c-fos expression in the dopamine receptor D1-expressing neurons (D1 neurons) in the anteromedial OT. In contrast, odor cues associated with electrical shock induced aversive behavior and c-fos expression in the pamine receptor D2-expressing neurons (D2 neurons) in the anteromedial OT, as well as the D1 neurons in the lateral OT. Here, we investigated whether the D1 and D2 neurons in the anteromedial OT play distinct roles in attractive or aversive behaviors, using optogenetic stimulation and real-time place preference (RTPP) tests. Mice expressing ChETA (ChR2/E123T)-enhanced yellow fluorescent protein (EYFP) in the D1 neurons in the anteromedial OT spent a longer time in the photo-stimulation side of the place preference chamber than the control mice expressing EYFP. On the other hand, upon optogenetic stimulation of the D2 neurons in the anteromedial OT, the mice spent a shorter time in the photo-stimulation side than the control mice. Local neural activation in the anteromedial OT during the RTPP tests was confirmed by c-fos mRNA expression. These results suggest that the D1 and D2 neurons in the anteromedial OT play opposing roles in attractive and aversive behaviors, respectively.

INTRODUCTION
Odor sensation elicits various motivations, which enable adaptive behavioral responses such as obtaining food rewards or avoiding potential dangers (Doty, 1986). Although some odorants elicit innate motivated behaviors in mice, such as fear responses to predator odors (Kobayakawa et al., 2007;Saito et al., 2017) or attractive responses to social odors (Inokuchi et al., 2017), animals can acquire appropriate behaviors to odor cues according to their experience, through odor-reward or odor-danger associative learning. However, the neural circuit mechanisms engaged in these odor-induced adaptive behaviors are still unclear.
Recent studies have revealed the importance of the olfactory tubercle (OT) in the odor-induced motivated behaviors (DiBenedictis et al., 2015;Gadziola et al., 2015;Yamaguchi, 2017;Zhang et al., 2017a;Murofushi et al., 2018). The OT is a part of the olfactory cortex that receives olfactory inputs directly from the olfactory bulb as well as indirectly from other parts of the olfactory cortex and the orbitofrontal cortex (Shepherd, 2004;Zhang et al., 2017b). The OT is also a part of the ventral striatum, in addition to the nucleus accumbens (NAc), which receives massive dopaminergic inputs from the ventral tegmental area (Ikemoto, 2007;Park et al., 2017;Zhang et al., 2017b;Poulin et al., 2018). The OT is composed of three major types of neurons: medium spiny neurons, dwarf cells, and granule cells (Millhouse and Heimer, 1984;Xiong and Wesson, 2016). The medium spiny neurons are distributed in the whole OT, forming the layer II (dense cell layer) of the cortex-like region (Millhouse and Heimer, 1984). A majority of the medium spiny neurons in the OT as well as the NAc and dorsal striatum express either dopamine receptor D1 or D2 (Yung et al., 1995;Murata et al., 2015). Dwarf cells are clustered in the lateral part of the OT, forming the cap region, which is interspersed throughout the antero-posterior axis (Hosoya and Hirata, 1974;Murata et al., 2015). The dwarf cells are considered a smaller type of the medium spiny neurons, and express D1 but not D2 (Murata et al., 2015). Granule cells are clustered through the anteromedial surface to the central deep part of the OT, forming the Islands of Calleja, which is presumably a continuous structure (Fallon et al., 1978;de Vente et al., 2001). The granule cells weakly express D1, and do not express D2 (Murata et al., 2015). In addition to these three types of neurons in the striatal component, the OT contains the ventral pallidal component and axon bundles that project from the striato-pallidal structure to other brain areas, forming the medial forebrain bundle (Heimer, 1978).
In our previous study, we divided the OT into domains, using the cap and Islands of Calleja as a landmark, and mapped c-fos expression when mice showed learned odor-induced attractive or aversive behaviors (Murata et al., 2015). Odor cues associated with a sugar reward induced attractive behavior and c-fos expression in the D1-expressing neurons (D1 neurons) in the cortex-like region of the anteromedial domain, which is covered by the superficially located Islands of Calleja. In contrast, odor cues associated with electrical shock induced aversive behavior and c-fos expression in the D2-expressing neurons (D2 neurons) in the cortex-like region of the anteromedial domain, as well as D1 neurons in the cap and cortex-like regions of the lateral domain, which is surrounded by the cap region. These results raise the possibility that the D1 and D2 neurons in the anteromedial OT play opposing roles in odor-guided motivated behaviors. Consistent with this idea, the D1 and D2 neurons in the NAc have distinct roles in attractive and aversive learning (Hikida et al., 2010). Here, we investigated whether activation of the D1 and D2 neurons induces attractive and aversive behaviors, respectively, by combining optogenetic stimulation and real-time place preference (RTPP) tests (Zhang et al., 2017a).

Animals
All experiments were conducted in accordance with the Guidelines for Animal Experimentation in Neuroscience of the Japan Neuroscience Society, and were approved by the Experimental Animal Research Committee of University of Fukui. The D1-Cre and D2-Cre mice used were heterozygotes and bred from D1-Cre (the Mutant Mouse Resource and Research Centers, STOCK Tg(Drd1a-cre)FK150Gsat/Mmucd, stock number: 029178-UCD; Gong et al., 2003Gong et al., , 2007 and D2-Cre mice (the Mutant Mouse Resource and Research Centers, B6.FVB(Cg)-Tg(Drd2-cre)ER44Gsat/Mmucd, stock number: 032108-UCD; Gong et al., 2003Gong et al., , 2007, by mating the heterozygote transgenic mice with wild-type C57BL/6J mice (Japan SLC, Inc., Shizuoka, Japan). Experiments were performed using male mice. After weaning, male mice were housed with their male littermates (2-6 mice per cage, wild-type and heterozygous mice were housed together) until the surgery, and then individually housed with a 12/12-h light/dark cycle. Food and water were freely available.

Stereotaxic Surgery
Stereotaxic surgeries were performed on mice aged 10-16 weeks. Mice were anesthetized with a mixture of three anesthetics (medetomidine, midazolam, and butorphanol;Nakamura et al., 2017), and then placed in a stereotaxic apparatus (SR-5M; Narishige Co., Ltd., Tokyo, Japan). The skull above the targeted areas was thinned using a dental drill and removed carefully. The AAVs were injected using a syringe pump (UltraMicroPump III; World Precision Instruments, LLC, Sarasota, FL, USA) connected to a Hamilton syringe (RN-1701; Hamilton, Reno, NV, USA), and a mounted glass micropipette with a tip diameter of 50 µm connected by an adaptor (55750-01, Hamilton, Reno, NV, USA).
We unilaterally injected 300 nL AAV5-EF1a-DIO-EYFP for confirmation of cell type-specific expression ( Figure 1A) and as a control for optogenetic stimulation, or AAV2-EF1a-DIO-ChETA-EYFP into the left hemisphere of the anteromedial OT of D1-Cre or D2-Cre mice using the following coordinates: anterior-posterior, +1.5 mm; medial-lateral, 0.7 mm from the bregma; and dorso-ventral, 4.35 mm from the brain surface. Two to three weeks later, the mice were ipsilaterally implanted with a chronic optical fiber (numerical aperture = 0.39, 200-µm diameter; CFMC12U; Thorlabs, Inc., Newton, NJ, USA) targeted to the anteromedial OT with the same coordinates described above. One to two weeks after fiber implantation, the following behavioral tests were conducted.

Optogenetic Stimulation and RTPP Tests
For optogenetic stimulation, the implanted optic fiber was connected to a blue light laser via patch cords with a fiberoptic rotary joint (RJPSF2; Thorlabs, Inc., Newton, NJ, USA). All photo-stimulation experiments used 5-ms, 5-7-mW, 473-nm light pulses at 20 Hz via a solid-state laser for light delivery (CST-L-473-50-OEM; Ultralasers, Inc., Newmarket, ON, Canada) triggered by a stimulator (STO2; Bio Research Center Co., Ltd., Nagoya, Japan). The intensity of the light-stimulation was verified by c-fos expression (Figure 4).
After being connected to the blue light laser, the mice were placed in a place preference chamber [30 (width) × 30 (depth) × 25 (height) cm 3 ] equipped with vertical or horizontal striped wall, as shown in Figure 1B, for 20 min. The non-stimulation control side was as assigned at the start of the experiment. Laser stimulation at 20 Hz was initiated when the mice entered the stimulation side, constantly delivered when the mice were in the stimulation side, and stopped when they returned back to the initial non-stimulation side. All behavioral tests were recorded using a USB digital video camera (Logicool c920r; Logitech, Lausanne, Switzerland). Offline analyzes of the time spent in each chamber, tracking data, and total traveled distance were performed using a video-tracking software (SMART 3.0; Panlab, Barcelona, Spain). Thirty minutes after the end of the RTPP tests, mice were deeply anesthetized by intraperitoneal injection of sodium pentobarbital, and then fixed for histochemical analysis.

Histochemistry
Mice were transcardially perfused with phosphate-buffered saline (PBS), followed by 4% paraformaldehyde (PFA). After cryoprotection with sucrose solution, the brain was frozen and sliced into coronal sections with a thickness of 20 µm.
The sections were rinsed in PBS and 0.1 M phosphate buffer, mounted on glass slides using a paint brush, dried overnight in a vacuum desiccator, and then stored at 4 • C until histochemistry.
To confirm cell type-specific EYFP expression, we performed double fluorescence labeling for EYFP and mRNAs of D1 or D2 as follows. Digoxigenin (DIG)-labeled RNA probes were prepared using an in vitro transcription kit (Roche, Basel, Switzerland) according to the manufacturer's protocol with a plasmid kindly provided by Dr. Kazuto Kobayashi (Sano et al., 2003). The dried sections were fixed in 4% PFA, digested using proteinase K (10 µg/mL) for 30 min, and post-fixed in 4% PFA. After prehybridization, the sections were incubated overnight at 65 • C with DIG-labeled RNA probes. After stringent washing, the sections were incubated in 1% blocking buffer (11096176001, Roche, Basel, Switzerland) for 1 h. Primary antibody against EYFP (1:1,000; Medical and Biological Laboratories Co., Ltd., Nagoya, Japan) and an anti-DIG antibody conjugated with alkaline phosphatase (1: For c-fos mRNA detection, we performed in situ hybridization using DIG-labeled antisense RNA probes. The RNA probe was prepared using an in vitro transcription kit (Roche, Basel, Switzerland) according to the manufacturer's protocol with a plasmid kindly provided by Dr. Hirohide Takebayashi (Bepari et al., 2012). Hybridization and washing were performed as described above. Subsequently, the sections were blocked with 10% normal sheep serum, 1% bovine serum albumin, and 0.1% Triton X-100 in PBS. The sections were then incubated overnight at 4 • C with alkaline phosphatase-conjugated anti-DIG antibody (1:1,000, Roche, Basel, Switzerland). The sections were washed in TNT, followed by alkaline phosphatase buffer [100 mM NaCl, 100 mM Tris-HCl (pH 9.5), 50 mM MgCl 2 , 0.1% Tween 20, 5 mM levamisole]. The sections were treated overnight with nitro-blue tetrazolium/5-bromo-4-chloro-3 -indolylphosphate (Roche, Basel, Switzerland) mixture at room temperature in a dark room for color development. Subsequently, they were rinsed in PBS and mounted in PermaFluor (Thermo Fisher Scientific, Waltham, MA, USA).

Microscopy and Image Analysis
Sections were examined using a confocal laser microscope (FV1200, Olympus, Tokyo, Japan) and a bright-field virtual slide system (NanoZoomer; Hamamatsu Photonics, Shizuoka, Japan). To quantify the number of c-fos mRNA-expressing cells, three coronal sections of the anteromedial OT with 20 µmthickness were selected (centering [antero-posterior axis] section and 100 µm-anterior and posterior sections). The number of cells in the anteromedial domain of the OT and NAc medial shell was counted using ImageJ (National Institutes of Health, Bethesda, MD, USA).

Criteria for Data Analysis
We confirmed that all the mice analyzed in the EYFP-expressing control group (shown in Figures 3, 4) expressed EYFP in the anteromedial OT and that the tip of the optic fiber was placed above the anteromedial OT. In the ChETA-EYFP group (shown in Figures 3, 4), the mice in which the center of the c-fos activation was located outside of the anteromedial OT (including the NAc) were excluded from data analysis [a total of 12 D1-Cre and D2-Cre mice (6 each) were excluded].  Franklin and Paxinos (2008). We injected Cre-dependent AAVs encoding EYFP or ChETA-EYFP and implanted optic fiber canula into the anteromedial OT of D1-Cre or D2-Cre mice. (B) Place preference chamber. In the RTPP tests, mice were randomly placed on either side of the chamber, which was assigned as the control side (no photo-stimulation). Blue light was delivered when mice were on the opposite side of the initial position. OT, olfactory tubercle; AAV, adeno-associated virus; EYFP, enhanced yellow fluorescent protein; RTPP, real-time place preference; D1, dopamine receptor D1; D2, dopamine receptor D2.

Statistics
Normality of the quantitative data presented in Figures 3, 4 was confirmed with the Kolmogorov-Smirnov test, using the R software (version 3.5.2; The R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was tested with parametric tests [Figure 3; two-way analysis of variance (ANOVA) with post hoc Tukey's test, Figure 4; three-way ANOVA with post hoc Tukey's test], using the Prism 8 software (GraphPad Software, San Diego, CA, USA). Quantitative data are shown as mean ± standard deviation or mean with individual data plots. Differences were considered statistically significant at p < 0.05.

DISCUSSION
In this study, we demonstrate that cell type-specific activation of the D1 and D2 neurons in the anteromedial OT elicits attractive and aversive behaviors, respectively. To achieve selective manipulation of the D1 or D2 neurons, which are intermingled in the cortex-like region of the OT, we used D1-Cre and D2-Cre transgenic mouse lines and Cre-dependent AAV vectors (Figures 1A, 2). This combination enabled us to deliver AAV-derived genes preferentially to the D1 or D2 neurons in the anteromedial OT. Optogenetic activation of the D1 and D2 neurons in the anteromedial OT induced attraction to and aversion from the photo-stimulation chamber, respectively (Figures 1B, 3). The photo-stimulation did not induce significant changes in traveled distance during the RTPP tests, suggesting that locomotor activity was not significantly influenced by the optogenetic activation ( Figure 3C). After the RTPP tests, activation of the anteromedial OT was confirmed by local increase in the c-fos expression (Figure 4). These results suggest that the D1 and D2 neurons in the anteromedial OT are involved in eliciting attractive and aversive behaviors, respectively (Figure 5).
Previous studies have shown that the anteromedial domain of the OT plays an important role in the reward system. Local self-administration of cocaine into the OT and NAc revealed that the anteromedial domain of the OT more robustly mediates the rewarding action of cocaine than other domains FIGURE 5 | Summary diagram. Both D1 and D2 neurons are intermingled in the anteromedial OT. Both cell types project their axons mainly to the pallidum, hypothalamus, amygdala, and midbrain. Cell type-specific activation using optogenetics revealed the opposing roles of the D1 and D2 neurons in the anteromedial OT in eliciting attractive and aversive behaviors, respectively. OT, olfactory tubercle; D1 neuron, dopamine receptor D1-expressing neuron; D2 neuron, dopamine receptor D2-expressing neuron. Stereotaxic atlas from Franklin and Paxinos (2008).
of the OT and NAc (Ikemoto, 2003). Optogenetic stimulation of the dopaminergic fiber from the ventral tegmental area to the medial OT elicits rewarding effects which generate place and odor preference (Zhang et al., 2017a). These local manipulations should exert excitatory effect on the D1 neurons via increased dopamine level in the anteromedial OT because D1 is coupled with Gs (Stoof and Kebabian, 1981). In line with these previous reports, our result directly demonstrates the role of the D1 neurons in the anteromedial OT in eliciting attractive behavior. In contrast, D2 is coupled with Gi (Stoof and Kebabian, 1981). Aversive stimuli reduce tonic firing of dopaminergic neurons, resulting in decreased ambient dopamine level at the target structure (Bromberg-Martin et al., 2010;Cohen et al., 2012;McCutcheon et al., 2012), which should exert excitatory effect on the D2 neurons. As blunting the tonic dopamine release in the ventromedial striatum leads to conditioned place aversion (Liu et al., 2008), our results support the idea that the D2 neurons in the anteromedial OT detect aversive stimuli via decreased dopamine release and elicit aversive behavior. Both D1 and D2 neurons in the medial OT project their GABAergic axons to mainly the reward-related brain regions, including the pallidum, hypothalamus, amygdala, and midbrain (Zhang et al., 2017b). Future neuroanatomical studies should address how the downstream neural circuits elicit attractive and aversive behaviors ( Figure 5). As we previously reported, neurons in the OT are activated by odor cues that induce motivated behaviors. The understanding of these cell type-specific roles of the D1 and D2 neurons in the anteromedial OT will provide a neural basis for odor-guided adaptive motivated behaviors.