Introduction

Sexual behavior is a fundamental and universal component of the observable behavioral spectrum in all mammalian species. This behavior results in reproduction, the vital process for species propagation. However, reproduction carries significant burdens, particularly for females, spanning pregnancy, birth, nursing, offspring rearing, and heightened vulnerability to environmental threats. Consequently, the female ability to strictly control sexual drive is adaptive and confers individual survivability. This can involve both increasing and decreasing motivation to engage in sexual behavior, dependent on anticipated beneficial outcome. Rodents provide a powerful resource for studying female sexual behavior and the underlying neural basis (Agrati, 2022; Micevych and Meisel, 2017). Typically, rodent sexual behavior is divided into three periods: an appetitive, consummatory, and post-copulatory phase (Hashikawa et al., 2016; Ishii and Touhara, 2018; Lenschow and Lima, 2020). The transition from the consummatory to the post-copulatory phase is triggered by the completion of mating, which is often associated with ejaculation from the male animal. Importantly, in males, sexual motivation acutely reduces after the completion of mating (Hull et al., 2005). However, less is known about how mating completion impacts female sexual behavior and motivation. Furthermore, how the transient experience of mating completion is represented in the brain to change the behavioral outcome is not well understood.

The medial preoptic area (MPOA), a molecularly and functionally heterogenous brain region, has classically been known to regulate wide variety of social behaviors (Fang et al., 2018; Hashikawa et al., 2021; Karigo et al., 2021; Kohl et al., 2018; Moffitt et al., 2018; Tsuneoka and Funato, 2021; Wei et al., 2018). Indeed, the MPOA has also been described to be an important brain area for the regulation of, including female sexual behavior. Studies in female rats using electrophysiological recording or histological approaches have reported that the MPOA contains neurons with various response specificity to sexual appetitive behaviors, consummatory behaviors and completion of mating (Kato and Sakuma, 2000; Pfaus et al., 1993). Similarly, MPOA lesions increased sexual receptivity, suggesting that the MPOA negatively regulates female sexual behavior when intact (Powers and Valenstein, 1972). Interestingly, the MPOA has also been found to positively regulate female sexual appetitive behavior through their projection to the mesolimbic reward system (McHenry et al., 2017). How the MPOA both positively and negatively regulate female sexual behavior, furthermore, sexual motivation, has been a long-lasting question in the field.

Here, we first assess sexual motivation in female mice by quantifying measurable approach or avoidance of a male conspecific. Next, by combining cellular-resolution whole-brain mapping of activity-dependent labeling, in vivo calcium imaging, and chemogenetic neural manipulation, we find that a subset of neurons in the MPOA, which are largely inhibitory, respond to mating completion but not to sexual appetitive behavior. This population is sufficient to suppress female sexual motivation. Our findings illuminate the importance of the MPOA as a brain region that encodes a negative signal that sustains low sexual motivation following the completion of mating in female mice.

Results

Female sexual motivation is suppressed after mating completion

Experimentally, sexual motivation can be quantified by the frequency of a subject displaying appetitive behavior, such as approach and sniffing behavior, toward an opposite sex conspecific. However, in a traditional rodent mating assay, where the female and male both freely interact with each other, it is difficult to isolate female-initiated appetitive behavior, and thus their sexual motivation. To directly address aspects of female sexual motivation independent from the male’s behavior, we used a female self-paced mating assay (Erskine, 1989; Peirce and Nuttall, 1961). In this procedure, a female subject and a sexually experienced male are placed in a behavior apparatus divided by a wall with a hole small enough for a female to move through, but not the male. The female subject may move into the room where the male animal was placed (“Interaction zone”) or move out to the other room (“Isolation zone”) (Figure 1A). Thus, in this assay the female has direct control of when interaction happens and provides a more ethologically valid measure of female appetitive sexual behavior. Specifically, the amount of time the female subject spends in the interaction zone and the number of transitions they make serve as metrics to quantify female appetitive behavior. On the experiment day, the female subject was placed in the apparatus with the male intruder (“Sexual interaction” trial) or alone (“Control” trial). Each animal went through both trials on different experiment days. The position of the animal was tracked to quantify the time spent in each zone. First, we found that the virgin female mice stayed in the isolation zone significantly longer than sexually experienced females, suggesting that previous sexual experience increased sexual motivation in the female mice (Figure S1). Next, in the sexually experienced animals, we observed robust behavioral changes after the completion of mating (Figure 1B). First, we observed that female subjects tended to spend more time in the isolation zone after mating completion compared to post-sniffing, mounting and intromission (Figure 1C–F). To quantify the changes in animal behavior, we divided the sexual interaction trial into pre- and post-mating completion epochs (Figure 1G–I). As result, we found that there was a significant decrease in the number of zone transitions as well as a significant increase in the time spent in the isolation zone after the completion of mating. In addition, while there was no change in the amount of female-to-male sniffing, there was a significant decrease in the number of mounting and intromission episodes (Figure 1J–L). There was also a significant increase in the number of self-grooming behavioral episodes (Figure 1M). Collectively, these results suggest that the completion of mating decreases sexual motivation.

Figure 1:

Brain-wide mapping of neural response to mating completion using activity-dependent labeling in the female brain

We next questioned how mating completion suppresses female sexual motivation. To investigate this question, we first determined which brain regions were associated with neural activity following the completion of mating in an unbiased manner. We conducted a cellular-resolution brain wide activity mapping analysis using targeted recombination in active populations method (TRAP), which allows genetic labeling and access to neurons that are active during a specific time window determined by tamoxifen administration (DeNardo et al., 2019). To capture the activity pattern associated with mating completion, Fos-2A-iCreER^T2::RCL-tdTomato (TRAP2::Ai14) female mice were administered 4-hydroxytamoxifen (4-OHT) immediately after the completion of mating (“Completion” group) (Figure 2A). In this method, cells that show behaviorally induced Fos activity were labeled with tdTomato. In addition to these animals, we also prepared a group of animals that were administrated with 4-OHT while the animals were in the appetitive phase. Female subjects interacted with male mice but did not experience any mount or intromission behavior, nor did it complete mating (“Appetitive” group). After 2 weeks, the tissue was collected, cleared, imaged, and processed for registration on a template brain atlas (Park et al., 2019; Renier et al., 2016) (Figure 2A). The tdTomato+ cells were segmented using a trained classifier and normalized by the area of each brain region (Figure 2B). As result, we found mating completion increased tdTomato expression a brain wide manner (Figure 2C–D, Table 2). We further noticed that regions well known to be related to social behavior had significantly higher tdTomato density in the Completion group compared to the Appetitive group (Figure 2E–I, Figure S2, Table 3). This includes paraventricular hypothalamic areas (PVHd, PVH, PaV) which are important for endocrinological regulations, medial amygdala (MeA) which is critical for olfactory sensory information processing, and the BNST (Flanigan and Kash, 2022; Marsh et al., 2021; Raam and Hong, 2021). The medial preoptic area nucleus (MPN), a subregion of the MPOA, had significantly higher tdTomato density in the Completion group compared to the Appetitive group. The MPN also had the 4^th largest difference of density between Completion group and Appetitive group among all the brain regions. From this result, we decided to further investigate the role of MPOA and how it is related to the regulation of female sexual motivation after the completion of mating.

Figure 2:

Table 1:

Table 2:

Table 3:

Mating completion is represented by both inhibitory and excitatory cell types in the MPOA

MPOA is a molecularly heterogenous brain region. The area is a mixture of 80% inhibitory neurons and 20% of excitatory neurons which can be further classified into neuronal subtypes by marker gene expressions (Hashikawa et al., 2021; Moffitt et al., 2018). These subtypes are often associated with a biological function. One example are the Nts expressing cells, which also tend to co-express Esr1 and Vgat (McHenry et al., 2017). In female mice, these neurons were shown to respond to male odor when the subject was sexually receptive, and the activation of these neurons were sufficient to increase sexual appetitive behaviors. Therefore, identifying the specific cell types which respond to mating completion but not to appetitive behavior is critical to understand how the MPOA complex regulates female sexual motivation. Here, we utilized multiplexed in situ RNA hybridization chain reaction (HCR) to identify previously characterized cell type markers of MPOA and an immediate early gene Fos (Figure S3A) (Choi et al., 2014). Brain samples were collected after the completion of mating (“Completion” group) or during the appetitive phase; while the subject was interacting with the male mice without any consummatory behavior (“Appetitive” group) (Figure S3B). First, to understand which subregion of the preoptic area was activated, we quantified the number of Fos+ cells. As result, we found that the number of Fos+ cells in the Completion group were significantly higher than the Appetitive group in the MPN but not in the nearby regions (Figure S3C–D). In the MPN, the amount of Vglut2+ and Vgat+ cells expressing Fos were significantly higher in the Completion group than in the Appetitive group, suggesting that the completion of mating activates both excitatory and inhibitory neurons (Figure S3E–I). The proportion of Fos+ cells was higher in Vgat+ population. To further investigate the cell types in the MPN that were activated by mating completion, we analyzed the co-expression of Fos and the MPOA cell type markers (Figure S3J). As result, in the Vglut2+ expressing cells, we found that Fos induction was only observed in the cells that co-expressed Calcr. In contrast, in the Vgat+ expressing cells, Fos induction was observed in Calcr, Gal and Prlr co-expressing cells. Next, we conducted clustering on the gene expression matrix to classify cells and analyzed the distribution of Fos among the clusters (Figure S3K–M). As result, we were able to identify 6 potential cell types which we labeled as type 1: Vglut2+, type 2: Vgat+, type 3: Gal+, type 4: Nts+, type 5: Calcr+ and type 6: NA, or a cell group that could not be categorized by a specific gene expression pattern. Type 1: Vglut2+, type 2: Vgat+ and type 3: Gal+ had significantly more Fos+ cells from the Completion group compared to the Appetitive group (Figure S3M). In contrast, Fos+ cells were not induced in type 4: Nts+ indicating that MPOA neurons which regulate appetitive sexual behavior was not activated by mating completion. Collectively, these results indicate that mating completion induces more neural activity in the MPN when compared to appetitive behavior, both in excitatory and inhibitory cell types.

A subpopulation of MPOA Vgat cells respond to mating completion in the female brain

While mapping neural activity based on Fos activity provided critical insight in how the brain responds to mating completion, the low temporal resolution of the method makes it difficult to identify cells that specifically respond to the onset to each behavior. For this reason, we next conducted single-cell resolution in vivo imaging in a freely moving animal (Figure 3A–B). We virally labeled MPOA inhibitory and excitatory cells with a calcium indicator GCaMP6s by using a Cre-dependent virus and female Vgat-Cre or Vglut2-Cre respectively (Figure 3C, Figure S4) (Vgat-Cre: n = 6, 87 ± 27 cells. Vglut2-Cre: n = 4, 68 ± 42 cells). We recorded the calcium activity of MPOA Vgat and Vglut2 cells in female mice during mating in a home cage mating assay (Figure 3E). First, we analyzed the magnitude of calcium activity for 15-seconds after behavior onset. From peri-event analysis, we found that as a population, Vgat cells showed significant changes in calcium activity after the onset of all behaviors observed, with the highest increase after mating completion (Figure 3F–G). Vglut2 cells showed increased activity after sniffing and mounting, intromission, and mating completion. Furthermore, when we compared the response of Vgat and Vglut2 cells, Vgat cells showed significantly higher average response magnitude after sniffing, self-grooming, anogenital sniffing and mating completion than the Vglut2 cells (Figure 3H). Vgat cells also contained a significantly higher proportion of cells that positively respond (> 2δ) to sniff and mating completion onset (Figure 3I). These results suggest that mating completion signal is more dominantly represented in the MPOA inhibitory cells than in the excitatory cells. On the other hand, the response magnitude and the proportion of cells that respond after mounting and intromission had no significant difference between the two cell types. This indicates that the excitatory/inhibitory balance in the MPOA fluctuates during mating. The MPOA has high inhibitory activity during appetitive phase, with more Vgat cells responding to sniffing, similar activity during the consummatory phase, and higher inhibitory activity after mating completion.

Figure 3

To further investigate the heterogeneity of MPOA during mating, we subset the cells into those which positively responded to any of sniff, mount or mating completion, and then classified the cells based on their specificity (Figure 4A,B and Figure S5). Consistent with Figure 3I, the number of cells that specifically responded was largest to mating completion with 177 cells, next to mount with 46 cells and to sniff with 12 cells. When compared to other cell populations, sniff cells and completion cells only showed significantly stronger responses than the other cells to sniff and mating completion respectively (Figure 4C). Mount responsive cells showed significantly stronger response than the other cells to mount, intromission and anogenital sniffing, but not to sniffing and mating completion (Figure 4C). We also found that the cells that respond to these 3 behaviors partially overlap but were mostly distinct (Figure 4D). The proportion of cells that respond specifically to a behavior was highest for mating completion (Vgat: 80%, Vglut2: 82%), and lower in mounting (Vgat: 55%, Vglut2: 68%) and sniffing (Vgat: 36%, Vglut2: 0%). There was no notable bias in the anatomical distribution of these cells within the field of view (Figure 4E). Overall, these results suggest that the MPOA contains highly heterogenous functional population which contains a large proportion of cells specifically responding to mating completion in female mice.

Figure 4:

MPOA Vgat cells display prolonged activity after the onset of mating completion

Female mice avoid sexual interaction with a male animal after mating completion for hours and even days, suggesting a persistent changes in the animal’s behavior (McGill and Coughlin, 1970; Zhou et al., 2023). How the MPOA regulate this persistent change in female sexual motivation is unclear. To analyze the response of MPOA cells on a longer time scale, we focused on the cells that positively responded to mating completion and characterized the calcium response pattern for 120-seconds after the onset (Figure 5). First, we observed that Vgat cells showed more diverse response patterns during this time-window (Figure 5A,B). To quantify this difference, we calculated the decay length of calcium signal, the time for the increased calcium signal to return to baseline (Figure 5C). As result, we found that Vgat cells had significantly longer decay length compared to the Vglut2 cells, suggesting increased activity late after the onset of the completion of mating. To further classify the temporal dynamics of responses to mating completion, we first concatenated calcium activity that occurred within this time-window from both Vgat and Vglut2 cells and conducted spectral clustering in an unbiased fashion. As result, we were able to classify the responses into 5 types that could be characterized as: 1) fast response, 2) slow response, 3) late response, 4) weak response and 5) negative response (Figure 5E). Type 3 contained cells that show increased calcium dynamics late after the onset of mating completion and had significantly larger decay length than the other cell types (Figure 5F). Then we separated the Vgat and Vglut2 cells to compare their proportions within each cluster. Interestingly, all the cells in type 3 were Vgat cells, suggesting that the increased activity in the MPOA late after the onset of mating completion is mediated by inhibitory neurons (Figure 5H–K). In contrast, Vglut2 cells were more likely to show fast responses than Vgat cells (Figure 5K). Collectively, these results suggest that the MPOA contains a subpopulation of inhibitory neurons which increase their activity late after the onset of mating completion, which may contribute to persistent suppression of the female animals’ sexual motivation.

Figure 5

An activity defined subset of neurons in the MPOA is sufficient to suppress female sexual behavior

Next, we examined whether the increased activity of neurons in the MPOA is sufficient to suppress female sexual motivation after mating completion. We injected an AAV to express the chemogenetic neural activator hM3Dq or eYFP as control into the MPOA of TRAP2 female mice (Figure 6A). To capture the neural activity, each mouse was administrated 4-OHT immediately after the after the completion of mating (“Completion-hM3Dq” or “Completion-eYFP” group). In addition to these animals, we also prepared a group of animals in which we injected an AAV expressing hM3Dq and then administrated 4-OHT during the appetitive phase (“Appetitive-hM3Dq” group) (Figure 6C). We found that while the overall number of TRAP-labeled cells had no difference between the Completion groups and the Appetitive group, the proportion of cells located in the MPN was significantly higher in the Completion groups (Figure 6D and Figure S7A–D). From in situ RNA hybridization, we found that most TRAP-labeled cells in the MPN were Vgat+ (Vgat+: 0.76 ± 0.16. Vglut2+: 0.15 ± 0.13.). The proportion of Vgat+ cells was significantly higher than the Vglut2+ cells (Figure S7E–H). Together these results suggest that the Completion group labels more inhibitory neurons in the MPN than in the Appetitive group. Furthermore, we analyzed the accuracy and efficiency of TRAP labeling by exposing the animal to mating completion before perfusing the animal. By labeling c-Fos protein signal, we analyzed how well the TRAP ensemble overlaps with the c-Fos positive-ensemble in the MPN. As result, we first found that the number of TRAP-labeled cells as well as c-Fos+ and TRAP-labeled cells in the Completion group was higher than the Appetitive group, consistent with the previous results (Figure S7J, 6K). Next, we found that the accuracy of TRAP labeling (the proportion of TRAP-labeled cells that were also c-Fos positive) was similar between the groups (Figure S7J, Completion group 0.44 ± 0.08, Appetitive group 0.51 ± 0.31). The efficiency of TRAP labeling (the proportion of c-Fos+ cells that were also TRAP labeled) tended to be higher in the Completion group (Figure S7J; 0.12 ± 0.10) than in the Appetitive group (0.03 ± 0.02). This also suggests that the Appetitive group labels a different set of neurons than the Completion group. These results indicate that by using the TRAP2 method, we were able to specifically label a proportion of MPOA neurons, which are primarily inhibitory, that respond specifically to mating completion. To test the causality of these neurons, we first investigated how the activation affects female sexual behavior in a home cage mating assay (Figure S8A–F). The subjects were intraperitoneal injection (i.p.) injected with CNO 30-minutes before the assay. As result, first we found that Completion-hM3Dq groups showed significantly fewer sniffing events compared to Completion-eYFP and Appetitive-hM3Dq groups (Figure S8C). For consummatory behaviors, while Completion-hM3Dq group showed no difference in mounting episodes compared to Completion-eYFP and Appetitive-hM3Dq groups, there were significantly less intromission episodes, resulting in lower receptive scores (Figure S8D–F). Thus, from these results, activation of MPOA neurons that respond to mating completion suppresses both appetitive and consummatory sexual behavior in female mice.

Figure 6:

As a next step, we investigated how the activation of these MPOA neurons affects female sexual motivation in a self-paced mating assay (Figure 6). The subjects were i.p. injected with CNO or saline 30-minutes before the assay (Figure 6B). As result, Completion-hM3Dq group, but not Completion-eYFP and Appetitive-hM3Dq group, showed significantly fewer number of transitions and spent more time in the isolation zone on the CNO trial than on the saline trial (Figure 6E–G). We also found that Completion-hM3Dq group, but not Completion-eYFP and Appetitive-hM3Dq group, displayed fewer numbers of self-grooming, mounting, intromission and lower receptivity score on the CNO trial than on the saline trial (Figure 6H–L). Interestingly, activation of these neurons did not induce conditional place preference or aversion (Figure S8G–J), implying that the suppression of sexual behavior is not associated with negative valence. Collectively, these results suggest that the activation of a subset of inhibitory neurons in the MPOA that respond to mating completion is sufficient for the suppression of female sexual motivation (Figure 6M).

Discussion

How sexual motivation is represented in the brain has been a long-lasting question in the field. Here we found that MPOA plays a critical role in regulating female sexual motivation. We found that MPOA neurons, especially the inhibitory neurons, respond significantly and specifically to the completion of mating. These neurons show persistent activity after the onset of mating completion. Reactivation of these neurons result in suppression of female sexual motivation. We propose that MPOA together with other neural substrates such as the BNST (Zhou et al., 2023), encode a negative feedback signal that sustains low sexual motivation state after the completion of mating in female mice. Here, we will further discuss how sensory inputs contribute to the mating completion response in the MPOA, how MPOA represents appetitive behavior and mating completion signal and how these neurons work in concert with other sexual behavior related hypothalamic regions.

Sensory representation of mating completion in the female brain and genitalia

Suppression of female sexual behavior and motivation after the completion of mating is not a unique phenomenon found in mice and can be observed in various insect species (Gillott, 2003). Studies in Drosophila has demonstrated that the primary trigger of this behavioral switch is the sex peptide (SP), a peptide released in the male seminal fluid (Chapman et al., 2003; Chen et al., 1988; Liu and Kubli, 2003). Studies in flies have further identified the receptor for SP as well as the neural circuitry which conveys the sensory information to the mating regulating neurons in the brain (Feng et al., 2014; Rezával et al., 2012; Yapici et al., 2008). In contrast to the well-studied system in Drosophila, little is known about how the mammalian brain senses the completion of mating. Histological studies in rats have shown that mating completion induces c-Fos expression widely in the brain, including the MPOA (Pfaus et al., 1993). Indeed, anesthesia to the female genitalia has been shown to disrupt pacing of mating, indicating that sensory input from the genitalia can impact female sexual behavior (Parada et al., 2014). Furthermore, artificial stimulation of the vagina or the clitoris of the female rat was further shown to induce c-Fos expression in the MPOA in a similar fashion as mating completion, indicating that at least tactile sensory input can be conveyed to this brain area (Parada et al., 2010; Pfaus et al., 1996). Ejaculation from vasectomized male animals were sufficient to suppress sexual behavior suggesting that at least chemical components from the testis is not necessary for this effect (Zhou et al., 2023). A recent study has shown that mechanical stretch induces serotonergic activity in the dorsal raphe, which resembles changes after mating completion, suggesting that stretch sensation is mediating mating completion signal to the brain (Troconis et al., 2023). However, it is unclear how stretch sensation in the genitalia is conveyed to the brain, and equally unclear how this information is ultimately transmitted to the MPOA. Our findings on how the mating completion signal is represented in the MPOA will be an ideal entry point to study the neural basis of genital-brain communication.

Appetitive behavior and mating completion ensembles in MPOA

The MPOA is a molecularly and functionally heterogenous brain region. One function MPOA has heavily been related to is parenting behavior. MPOA neurons which express Gal and/or Calcr has been demonstrated to regulate parenting behavior (Kohl et al., 2018; Wu et al., 2014; Yoshihara et al., 2021). From our data, neurons that respond to mating completion tended to co-express Gal and Calcr more than neurons that responded to appetitive behavior (Figure S3). However, we also found that chemogenetic activation of mating completion responsive MPOA neurons did not induce parenting behavior (Figure S8). While it is necessary to conduct extensive analysis of how MPOA neurons respond to other behaviors, these results suggest that mating completion responding MPOA neurons may specifically regulate sexual behavior and motivation but not parenting behavior. A previous study also found that MPOA contains a subset of neurons which regulate female appetitive sexual behavior. MPOA neurons which express Vgat, Esr1 and Nts, and projects to the ventral tegmental area (VTA) has been shown to respond to olfactory signals from the male mouse and to enhance female-to-male sniffing (McHenry et al., 2017). There is also a significant release of dopamine after the first female to male sniffing interaction as well as after mating completion (Dai et al., 2022). How the brain, especially the MPOA, increases sexual motivation before and during the appetitive phase, and how it suppresses motivation after the completion of mating, has been a long-lasting question in the field that is addressed throughout this work. The whole brain activity-dependent labeling experiment identified the MPOA as one of the brain regions which strongly increases Fos activity (Figure 2).

However, there were many other brain regions showing increased activity associated with mating completion, including ones that are not canonically involved in the regulation of social behavior. While it is possible that all these brain regions work in concert to suppress sexual motivation after the completion of mating, our results demonstrate that a specific brain region, MPOA, is sufficient to evoke behavioral response. This suggests that all brain regions showing higher Fos activity are not necessarily causal to behavioral changes. It is also possible that these widespread brain regions may be a consequence of noise due to the long time-window in this method, which may be resolved by further development of genetic methods for neural ensemble labeling. In the MPOA, our in vivo calcium imaging data shows nearly half of the inhibitory neurons that positively respond to appetitive behavior also respond to mating completion (Figure 4). Our data also identified that there was no difference between the number of Nts+ cells activated by appetitive behavior and mating completion (Figure S3). These results suggest that there is a small but significant number of neurons which respond both to appetitive behavior and mating completion in the MPOA, and potentially regulate release of dopamine.

Although a large amount of appetitive behavior responding neurons also respond to mating completion, the opposite is not true. Most of the mating completion responding neurons did not respond to female-to-male sniffing (Figure 4). Importantly, the chemogenetic activation of this population suppressed the amount of sexual interaction in the self-paced mating assay, which suggest that it reduces sexual motivation, opposite to what was described when activating Nts+ MPOA neurons (McHenry et al., 2017). This suggests that there are at least two-distinct populations in the MPOA which regulate appetitive sexual behavior in opposing directions.

Neural circuitry for female sexual behavior regulation

The neural circuitry that regulates female sexual behavior has been extensively studied in rodent models. From early lesion and electrical stimulation studies, it has been proposed that the rodent brain has a positive and negative regulatory system for female sexual behavior. The positive system has been proposed in rats, where Pfaff and Sakuma demonstrated that the lesion of the ventromedial hypothalamus (VMH) abolished sexual receptivity while electrical stimulation of the same region induces receptivity (Pfaff and Sakuma, 1979a, 1979b). Mouse genetics further showed that a subset of neurons in the ventrolateral part of the VMH is crucial for the regulation of female sexual behavior (Inoue et al., 2019; Yang et al., 2013). More specifically, Esr1+ Cckar+ (Nts-) cells in the VMHvl were found to be necessary and sufficient for female sexual receptivity (Liu et al., 2022; Yin et al., 2022). These results suggest that the VMHvl is the “command center” which positively regulate female sexual behavior.

The MPOA together with other brain regions has been proposed to form the negative regulating system. Lesion of the MPOA was shown to increase the sexual receptivity of the female rat and electrical stimulation was shown to reduce receptivity, suggesting that MPOA negatively regulates female sexual behavior (Pfaff and Sakuma, 1979b; Powers and Valenstein, 1972). In the current study, we demonstrate that chemogenetic activation of neurons in the MPOA which respond to mating completion is sufficient to suppress female sexual behavior and sexual motivation. This result suggests that these neurons in the MPOA constitute the negative regulating system. Importantly, both our single-cell calcium imaging results and histological analysis of the chemogenetic activation cohort suggest that the majority of these neurons are inhibitory (Figure 3I and Figure S7H). We further demonstrate that the inhibitory population contains neurons that show prolonged responses to mating completion which resemble the persistent change of sexual motivation. Other studies have also supported the idea that the inhibitory cell population is more related to the regulation of social behavior than the excitatory population (Fang et al., 2018; Hashikawa et al., 2021; Moffitt et al., 2018; Tsuneoka and Funato, 2021; G.-W. Zhang et al., 2021). Our results are also in line with this idea and suggest that the inhibitory neuronal population is the main regulator of the change in sexual motivation. The role of excitatory neurons that respond to mating completion will be a future topic to examine.

Recently, BNST was also found to play a role in negatively controlling female sexual behavior (Zhou et al., 2023). There, they use a behavioral assay which introduces 2 different males to the female subject before and after the completion of mating. They demonstrate that after the completion of mating, the female will be less receptive toward the second male. Importantly, this assay clearly demonstrate that the reduction of female sexual behavior is due to the female herself, not due to the change in their male conspecific. This result together with the result from our self-paced mating assay (Figure 1) strongly indicates that the behavioral change after mating completion reflects change in internal motivational states.

The study further used single-cell resolution in vivo calcium imaging techniques to demonstrate that BNST^Esr2 neurons show specific and robust responses after mating completion, like what we observed in the MPOA. While inhibition of BNST^Esr2 neurons rescued female receptivity after mating completion, activation of these neurons did not affect sexual behavior in female mice, suggesting that BNST^Esr2 neurons are necessary but not sufficient to suppress female sexual behavior or potentially sexual motivation. This result may suggest that BNST^Esr2 neurons require activation of other neural populations to affect female sexual behavior, such as the populations we identify in the MPOA. Still, the overall role of BNST and MPOA seems to be similar suggesting that the two regions work in concert to suppress female sexual behavior. This is contrary to the finding from Mei et al. showing that the BNST^Esr1 neurons and MPOA^Esr1 neurons forms a bidirectional inhibitory circuit to antagonistically regulate infanticidal behavior and maternal behavior respectively (Mei et al., 2023). A mechanism which is yet to be revealed that can override the negative feedback loop between MPOA and BNST neurons is necessary for these neurons to regulate the same behavior change. Interestingly, (Yin et al., 2022) showed that VMHvl^Cckar neurons show strong inhibition of spontaneous bulk calcium events after mating completion (Yin et al., 2022). This is opposite to the persistent increase of activity in the MPOA and BNST, suggesting that the activity in these regions is reflected onto the VMHvl. Indeed, MPOA and BNST both send dense projection to the VMHvl, and VMHvl^Esr1 neurons project to the MPOA and to the BNST, making the MPOA-BNST-VMHvl network highly interconnected (Dimén et al., 2021; Lo et al., 2019; Osakada et al., 2018).

Long-lasting change in neural activity within this network, potentially through monoamine or steroid molecules such as dopamine or opioids may be the neural basis of suppressed sexual motivation after mating completion (Micevych and Meisel, 2017; S. X. Zhang et al., 2021). Another candidate is serotonin, which is known to be an important mediator for sexual behavior (Uphouse, 2014). Recent study has revealed release of serotonin in response to mating completion in female mice (Troconis et al., 2023). Our whole brain activity mapping data also indicates increased Fos activity in the dorsal raphe (Figure S2). Whether serotonin mediates the changes in persistent activity across the entire network and whether there is a difference in how it impacts MPOA and BNST will be an important topic of future studies.

Taken together, our study identifies a subset of neurons in the MPOA which is dedicated to negatively sustain female sexual motivation. This finding fills the missing hole in how the brain neurons regulate female sexual motivation and proposes that the basal forebrain complex of MPOA and BNST work in concert as the negative regulating system of female sexual behavior.

Materials and methods

Animals

All experiments were approved by the Institutional Animals Care and Use Committee at the University of Washington (Protocol #4450-01). Animals were group-housed with littermates on a 12-hr light cycle at ~22°C with food and water available ad libitum. Wild-type Swiss-Webster mice were purchased from Taconic Biosciences (NY, U.S.). Wild-type C57BL/6 mice were purchased from the Jackson laboratory and bred in-house. Vgat-Cre (also known as Slc32a1-ires-Cre, Jax#016962), Vglut2-Cre (also known as Slc17a6-ires-Cre, Jax#028863) and Ai14 (Rosa-CAG-LSL-tdTomato-WPRE, Jax#007914) were purchased from Jackson laboratory. TRAP2 mice (also known as Fos-2A-iCreER^T2, (DeNardo et al., 2019)) were provided by Dr. Palmiter (University of Washington) and crossed with Ai14 to generate TRAP2::Ai14 mice. All male mice which were used as intruders were singly housed to avoid male-male aggression. Female mice with GRIN-lens implant were singly housed. All other mice were group housed.

Surgery

For targeting AAV into a certain brain region, stereotactic coordinates were defined for each brain region based on the brain atlas (Franklin and Paxinos, 2007). Mice were anesthetized isoflurane and head-fixed to stereotactic equipment (Kopf). For in vivo calcium imaging experiments, 500 nL of AAVDj EF1a-DIO-GCaMP6s (2.8*e12, UNC viral core, lot#av78310) was injected into two coordinates in the target brain region at a speed of 60 nl/min using Nanoject 3 pump (Drummond Scientific). The following coordinates were used [Distance in millimeters from the Bregma for the anterior (A)-posterior (P) and lateral (L) positions, and from the brain surface for ventral (V) direction]: MPOA, A 0.2, L 0.15, V 5.2. and A 0.4, L 0.35, V 5.2. Immediately following viral injection, 5 screws were placed into the skull and a GRIN lens (Inscopix, 0.6 mm x 7.3 mm) was slowly lowered to position above the injection site: MPOA, A 0.3, L 0.3, V 4.95. GRIN lens was glued to the skull with C&B Metabond Kit (Parkell, # S380) and dental cement (Lang Dental, #1304CL). After placement of GRIN lens, a custom-made head ring was glued to the skull and screws. For bare GRIN lenses, more than 3 weeks after the virus injection and GRIN lens placement surgery, a baseplate (Inscopix) was positioned to while visualizing GCaMP6s expression with a microscope. Once a field of view with GCaMP6s expression was found, the baseplate was glued to the head ring and skull. For pharmacogenetic experiments, 300-400 nL of AAV5 hSyn-DIO-hM3D(Gq)-mCherry (2.0*e13, Addgene, #44361-AAV5, lot#v141469) or AAV5 EF1a-DIO-eYFP (4.4*e12, UNC viral core, lot#av4802B) was bilaterally injected into the target brain region at a speed of 60 nl/min using Nanoject 3 pump. The following coordinates were used: MPOA, A 0.3, L 0.25, V 5.2. All female mice that experienced a mating assay was ovariectomized. Animals that were used in in vivo calcium imaging experiments or pharmacogenetic experiments were ovariectomized during the virus injection or GRIN lens placement surgery. After surgery, the incision was sutured, and the animal was warmed using a heat pad to facilitate recovery from anesthesia.

Behavior experiments and analysis

For all mating assays, female mice older than 6 weeks were ovariectomized (OVX) under isoflurane anesthesia. If necessary, mice underwent surgeries for AAV injections, GRIN lens implantation on the same day. After more than 2 weeks of recovery, the OVX female subjects were subcutaneously (s.c.) injected with 0.1 mL and 0.05 mL Estradiol (E8515, Sigma-Aldrich, 0.1 mg/ml) dissolved in sesame oil (Sigma-Aldrich) at 24 and 48h before assay, and 0.05 mL Progesterone (P0130, Sigma-Aldrich, 10 mg/ml in sesame oil) at 4h before assay. If necessary, mice underwent I.P. injection of CNO/saline. C57BL/6 or Swiss-Webster mice were used as a stud male. A 4 – 6 cm to head bar was placed onto the male mice under isoflurane anesthesia, which was used for the female self-paced mating assay described below. The male animals were kept isolated until the assay and had been trained prior to the assay to show mounting and ejaculation.

The home cage mating assay was conducted as previously described with modifications (Ishii et al., 2017). Each assay was conducted 5-10 hrs after the onset of dark period, under an IR light and recorded (30 Hz) for subsequent 30 to 120 min for analysis of sexual behavior from a dorsal view. For neural activation experiments using hM3Dq (Figure S5, Figure S6), 0.1 mL of 1 mg/kg clozapine-N-oxide dissolved in saline (CNO, RTI International, NIMH Code C-929) or 0.1 mL of saline was I.P. injected into the animal 30 minutes prior to the male entry.

The female self-paced mating assay was conducted using a behavior apparatus modified from an apparatus used in rat (Nyuyki et al., 2011). A 1-cm thick black wall with a 20 mm x 20 mm hole 3D-printed with PFA filament was placed in a 25 cm x 35 cm cage to divide the room in a 1:2 ratio along the longer edge. The smaller room was defined as the isolation zone and the larger room was defined as the interaction zone. Female subjects were habituated to the room for 30 minutes before the assay. The female subjects that were not able to pass through the hole during the habituation was eliminated from the cohort. On the experiment day, the female subjects were placed into the behavior apparatus for 30 minutes for analysis of movement tracking from a dorsal view, which was used as control (Control trial). After the habituation, the male intruder (Swiss-Webster) was placed in the interaction zone. Each assay was conducted 5-10 hrs after the onset of dark period, under an IR light and recorded (15 Hz) for subsequent 30 to 120 min for analysis of movement tracking and sexual behavior from a dorsal view (Sexual interaction trial). For Figure 6, females underwent two mating assays at > 1 week interval, with different treatment (e.g., saline or CNO). Different males were used for each assay. For neural activation experiments using hM3Dq (Figure 6, Figure S5, Figure S6), 0.1 mL of 1 mg/kg clozapine-N-oxide dissolved in saline (CNO, RTI International, NIMH Code C-929) or 0.1 mL of saline was I.P. injected into the animal 30 minutes prior to the male entry.

Sexual behavior was annotated from the recorded video using Behavioral Observation Research Interactive Software (BORIS) in a frame-by-frame fashion (Friard and Gamba, 2016). Male mounts, intromission, male to female anogenital sniffing, female to male sniffing, female self-grooming and mating completion were annotated based on characteristic postures described previously (Ishii et al., 2017). Briefly, mount was defined as a stud male using both forepaws to climb onto a female from behind for copulation. Intromission was defined by stable thrusting from the male animal during mount. The completion of mating was identified by increase and then termination of thrusting and more than 10 seconds of immobility from the male, which are postures frequently associated with male ejaculation. Male to female anogenital sniffing was identified by the male animal following the female subjects anogenital area. Female to male sniffing was identified by female subject nose approaching male subject. When analyzing sniff-evoked GCaMP activity, to restrict the analysis to appetitive sniffing behaviors, sniffing behaviors which happened after the first mounting was excluded from the analysis. Female self-grooming was identified by the female subject moving their head close to their genital region. The receptive ratio was calculated by dividing the total number of intromissions with the total number of mountings. Behavior annotation was conducted by volunteers who were blind for the stimulation or AAV condition. For the female self-paced mating assay, the position of the female subject was tracked from the recorded video using DeepLabCut (Nath et al., 2019). The amount of time spent in the interaction zone was quantified using SimBA region of interest (ROI) interface (Nilson et al., 2020). Behavior data from in vivo calcium imaging experiments were down-sampled to 10-Hz to match the imaging data.

For conditional place preference (Figure S8G–J), a two chambered apparatus was used with visual (horizontal and vertical stripes) cues. The floor was backlit and an overhead video camera recorded position (Basler, 3.75 Hz frame rate with Ethovision software, Noldus). The apparatus was in a sound isolation chamber. Ad libitum fed Fos-CreER mice with hM3Dq and eYFP controls were acclimatized (60 min). The following day, the mouse’s position was recorded (1,800 s) and tracked offline using Ethovision (pre-test). Daily conditioning consisted of two 1,800 s sessions: (1) one side of the chamber paired with saline injection; (2) the other side paired with CNO (1 mg/kg) injection. Saline or CNO was administrated 10 min prior to the entry to the chamber. The side on which saline or CNO was administered was counterbalanced. After three conditioning days, mice were given free access to the entire apparatus and their position was tracked (post-test). The difference of time spent in the CNO zone, the number of entries into the CNO zone and the velocity between the post- and pre-test.

For pup retrieval test (Figure S8K–O), a home cage of the female subject was used. Female subject was i.p. injected with saline as control or CNO in their home cage. 30-minutes after administration, three 7-day-old pups were placed in the corners of the home cage. Animal behavior was recorded from the dorsal view for 10 minutes. Parenting behavior was annotated from the recorded video using Behavioral Observation Research Interactive Software (BORIS) in a frame-by-frame fashion. Pup retrievals, nest building, pup grooming and pup sniffing were annotated based on characteristic postures. Briefly, pup retrieval was defined as a female subject carrying pup. Nest building was defined by female subject digging their bedding or carrying nesting material. Pup grooming was identified by grooming pups without carrying. Female to pup sniffing was identified by female subject nose approaching pup. For neural activation experiments using hM3Dq (Figure S6), 0.1 mL of 1 mg/kg clozapine-N-oxide dissolved in saline (CNO, RTI International, NIMH Code C-929) or 0.1 mL of saline was I.P. injected into the animal 30 minutes prior to the pup introduction.

In vivo calcium imaging experiments and analysis

Miniature microscope nVoke 2.0 or nVue (Inscopix) was used to obtain GCaMP6s signal (20 Hz). The microscope was connected above the behavior apparatus using a commutator (Inscopix) to support the weight of the microscope. On the experiment day, the subject was habituated to the microscope for 10 minutes. After the habituation, the LED power and gain was adjusted to 0.2-0.5 and 6.0-8.0 respectively depending on the intensity of GCaMP6s expression. Data acquisition started after the adjustment of imaging parameters. The animal was kept habituated to the microscope for additional > 5 minutes to obtain background signal, and then the intruder male animal was introduced to the subject’s home cage. Calcium activity as well as animal behavior was recorded until 2-10 minutes after the completion of mating. Imaging data were spatially down-sampled by factor of 4 and temporally down-sampled to 10 Hz using the Inscopix Data Processing software (1.8.0). The data was further filtered by a spatial band-pass filter and motion corrected using the same software. Masks were manually drawn to extract data from single cell. The data frame containing GCaMP6s signal intensity traces were exported from the Inscopix Data Processing software and further processed in Python. Signal intensity array was z-scored using the background signal. For peri event activity analysis in Figure 3 and Figure 4, z-scored signal was collected from −5 to + 15 seconds around the behavior onset. Peri event activity from multiple behavior events were concatenated and averaged. Finally, the mean value of pre-behavior background (−5 to 0 seconds from onset) was subtracted. The magnitude of z-scored signal was quantified by averaging 0 to +5 seconds from onset. Cells that had larger than 2 d z-scored signal were classified as “responding” cells. For persistent activity analysis in Figure 5, z-scored signal was collected from −5 to + 120 seconds around the onset of mating completion. The mean value of pre-behavior background (−5 to 0 seconds from onset) was subtracted. The decay length was quantified by calculating the time the z-scored signal was smaller than 0. The magnitude of z-scored signal was quantified by averaging 0 to +5 seconds from onset. Spectral clustering was conducted using scikit-learn package and previously used custom script (Namboodiri et al., 2019). Briefly, first dimensionality reduction using principal component analysis was conducted for z-scored signal from −5 to +120 seconds around the onset of mating completion. Next, n = 4 principal components were used to project the data. This was further used as the input into the clustering algorithm while systematically testing nearest neighbor and number of cluster variables. The best variable that resulted in maximizing the silhouette score was determined (number of clusters = 5, number of nearest neighbors = 55, average silhouette score = 0.405). All clusters contained cells from multiple animal subjects, indicating the variability of the dataset can be explained by heterogenous cell population, rather than subject variability.

TRAP experiments

TRAP method was conducted as previously described with modifications (DeNardo et al., 2019). TRAP2::Ai14 or TRAP2 mice were OVX during isoflurane anesthesia and AAV injected if necessary. After 2 – 3 weeks from OVX surgery, animals were prepared for the activity labeling assay. Female subjects were subcutaneously (s.c.) injected with 0.1 mL and 0.05 mL Estradiol (E8515, Sigma-Aldrich, 0.1 mg/ml) dissolved in sesame oil (Sigma-Aldrich) at 24 and 48h before assay. On the day for the assay, female subjects were isolated in their home cage and introduced with a male intruder to either experience appetitive behaviors or mating completion. For the appetitive group, the male intruder was removed after 10–20 minutes. To ensure that the female subject did not experience any consummatory behavior from the male, the male intruder was removed whenever an attempt to mount was observed. For the completion group, the male intruder was removed just after the completion of mating. Immediately after the removal of the male animal, the female subject was I.P. injected with 0.125 mL of 50 mg/kg of 4-Hydroxytamoxifen (Sigma-Aldrich, cat#H6278) dissolved in sesame oil. After the injection, the female subject was isolated in their home cage and kept in the behavior room for overnight to avoid background stimulation. On the following day, the mice were returned to the regular housing.

HCR experiments and analysis

Hybridization chain reaction (HCR) was performed based on the protocol provided by Molecular Instruments with some modifications (Figure S3 and Figure S7E–H). For Fos activity analysis (Figure S3), the female subject was first singly housed in the home cage for 1 hour. The male intruder was introduced into the home cage and animal behavior was recorded. For the appetitive group, the male intruder was removed after 10–20 minutes until they first attempt mounting to the female subject. For the completion group, the male intruder was removed just after the completion of mating. Twenty-minutes after the removal of the male intruder, the female subject was immediately sacrificed to collect brain tissue. The brain tissue was then snap-frozen using dry ice. Every fifth 20-micron coronal sections containing the MPOA (3 sections from Bregma +0.10 mm to – 0.10 mm) were collected on a slide glass and fixed with 4% PFA in PBS for 30 min and rinsed with PBS, followed by dehydration with a series of 50%, 70%, 100%, 100% of ethanol. After final dehydration the slides were dried. The tissue was treated with Protease4 (ACDBiosystem) for 5– 7 minutes. The slides were then rinsed with PBS and were applied with hybridization buffer (Molecular Instruments) for 10 minutes at 37°C for pre-hybridization. RNA probes for Fos, Vgat, Vglut2, Calcr, Esr1, Gal, Nts, Prlr (1 mM, Molecular Instruments) were diluted in a ratio of 1:100– 1:250 into the hybridization buffer. After pre-hybridization, hybridization buffer (70 μl) containing the probes was applied to each slide on which the parafilm cover was placed. After 12h–18h of incubation at 37°C in the moisture chamber, the sections were washed, with a series of probe wash buffer and SSC-T buffer (5x SSC, 0.1% Tween 20) mixture. After the final wash, the sections were incubated in amplification buffer (Molecular Instruments) for 30 minutes at room temperature for pre-amplification. Amplification hairpin probes (3 mM, Molecular Instruments) were diluted in a ratio of 1:50 into the hybridization and then snap cooled for 1.5 minutes in 95°C. After pre-amplification, amplification buffer containing hairpin probes (70 μl) was applied to each slide on which the parafilm cover was placed. After 12h–18h of incubation at room temperature in the moisture chamber, the sections were washed in 5xSSC-T buffer for 30 minutes x2. The sections were quenched for autofluorescence using Vector TrueView reagent (Vector, cat# SP-8400). After washed in 2xSSC, the sections were stained with DAPI in PBS (5 μg/ml, Thermo Fisher Scientific, cat#D1306) for 8 minutes. The slide was then mounted using Vibrance Antifade Mounting Medium (VECTASHIELD, cat#H-1800). Images were obtained using a conventional microscope (Zeiss, Apotome Imager.M2) with a 20x objective. After image acquisition, the coverslip was removed and washed in 2x SSC for 10 minutes, then incubated with DNase1 (0.25 U/μl, Roche, cat#4716728001) for 1.5 h at room temperature. The section was then washed in 2x SSC for 5 minutes x6. The sections then proceeded to the pre-hybridization step for the next round of HCR. A total of 3 rounds of HCR were conducted. Images from multiple rounds were aligned using HiPlex image registration software (ACDBiosystem) or manually aligned using ImageJ package TrackEM2 (Cardona et al., 2012).

For anatomical characterization for TRAP2 animals used in Figure S7E–H, subjects were perfused with PBS and 4% PFA-PBS, then brain was harvested and post-fixed with PFA overnight. Every fourth 40-μm coronal sections containing the MPOA were collected on a slide glass. The sections went through 1 round of HCR and were stained for Vgat and Vglut2 following the procedures described above.

The images with each probe signal were aligned to the DAPI channel to identify MPOA subregions and adjacent regions. The image was processed in HALO v3.2 software (Indica Labs) to segment each cell based on the DAPI signal and to quantify RNA transcript signal in each cell. Parameters for detecting puncta and intensity of each gene were manually adjusted for each tissue section of each animal by a volunteer blind to the experimental condition. The cutoff value to determine a gene “expressing cell” was defined as the mean fluorescent intensity of 3 to 5 transcripts depending on the background fluorescent signal. For supervised clustering of the MPN cell gene expression matrix (Figure S3K–M) was conducted using Seurat3.1.1 in R. Cells that had no mRNA signal were excluded from the dataset. This expression matrix contained single cells detected across all 10 animals and the spot counts for every gene; we also included signal intensity measurements as metadata. Once loaded, the HCR digital expression matrix was scaled but not normalized. During scaling, we regressed out gene intensity measurements on the basis that signal intensity may indicate staining quality. Next, cells were clustered using the same pipeline. The cells were UMAP plotted using nearest_neighbors = 100, minimum_distance = 0.05 (Figure S3K). The cluster ID for each cell was exported and used for statistical analysis using custom Python scripts.

Immunohistochemistry experiments

To evaluate the accuracy and efficiency of TRAP labeling, we utilized subjects used in pharmacogenetic experiments (Figure 6, Figure S7) and labeled neurons that respond to mating completion using immunohistochemistry against immediate early gene c-Fos. The TRAP2 female subjects were labeled with eYFP or hM3Dq-mCherry in the appetitive or mating completion responding neurons in the MPOA (Appetitive-hM3Dq group, Completion-hM3Dq or Completion-eYFP group). Ninety-minutes before perfusion, the female subjects experienced mating completion during a home cage mating assay (described in “Behavior experiments and analysis.”) to activate the mating completion responding neurons in the MPOA. The subjects were then perfused with PBS and 4% PFA-PBS, and the brain was harvested and post-fixed with PFA overnight. Every fourth 40-μm coronal sections containing the MPOA were collected. were washed three times with PBS-T (1x PBS, 0.3% Triton-X, Sigma-Aldrich, cat# 11332481001) for 10 min x3, and treated with 5% normal donkey serum (NDS, Sigma-Aldrich, cat# 566460) in PBST for 1 hr at room temperature for blocking. Sections were then incubated with rabbit anti-cFos antibody (Abcam, cat#190289-Rb, 1:500), diluted into 5% BSA in PBST for overnight at 4°C. Antibody signals were detected by goat anti-rabbit Alexa 488 or Alexa 594 (Invitrogen, cat#A21206 or cat#A21209, 1:500) for 2h at room temperature. Sections were washed once with PBST for 10 min, treated with PBS containing DAPI for 20 min, rinsed with PBS, and mounted with cover glass using Vibrance Antifade Mounting Medium (VECTASHIELD, cat#H-1800).

Sections were imaged by FV-3000 confocal microscope (Olympus) with a 20x objective. Images were processed in cellpose to segment DAPI positive cells in the MPOA and their subregions (Stringer et al., 2021). DAPI positive cell masks were used to quantify intensity of c-Fos protein or YFP/mCherry signal. The intensity value was used to determine the negative and positive cell.

Brain-wide analysis of activity-dependent labeling experiments

TRAP2::Ai14 females were used to label neurons responding to appetitive behavior or mating completion. Two-weeks after activity labeling (described in “TRAP experiments.”), the animals were perfused with PBS and 4% PFA-PBS, the brain was harvested and post-fixed with PFA overnight. After post-fixation, the brains were processed for tissue clearing. For tissue clearing, we utilized stabilization under harsh conditions via intramolecular epoxide linkages to prevent degradation method (SHIELD) (Park et al., 2019). The procedures followed the SHIELD tissue clearing protocol (Full Active Pipeline Protocol v5.05, LifeCanvas). Briefly, the brain was incubated in SHIELD OFF solution [2.5 mL DI water, 2.5 mL SHIELD-Buffer Solution, 5 mL SHIELD-Epoxy Solution] at 4oC for 4 days. Then transferred to SHIELD ON buffer and incubated at 37°C for 24 hrs. Following, SHIELD ON buffer incubation, the brain was pre-incubated in Delipidization buffer for overnight at room temperature. The tissue was then placed into SmartClear II Pro (LifeCanvas) to conduct active tissue clearing for 24 hours. To match the refraction index, the tissue was incubated in 50% of EASYINDEX solution (RI = 1.52, LifeCanvas) at 37°C shaking for 24 hrs and then 100% of EASYINDEX solution at 37°C shaking for 24 hrs. The tissue was then embedded in EASYINDEX with 2% agar and mounted onto SmartSPIM light sheet microscope (LifeCanvas). The tissue was mounted ventral side facing toward the objective, imaged from the horizontal view using a 3.6x objective with 4-μm zstep. 594 nm channel was collected as the signal channel, 488 or 647 nm channel was collected as the background channel. Background channel was used for the alignment to the brain atlas. The images were first destriped and then stitched using a custom written script (LifeCanvas). The images were pre-processed and down sampled by 2 times (final resolution (x,y,z) = (3.6 μm, 3.6 μm, 4 μm)). Images were registered to a standard brain atlas Unified Anatomical Atlas (Yongsoo Kim Lab) (resolution (x,y,z) = (20 μm, 20 μm, 50 μm)) using ClearMap pipeline on a cloud-based cluster computer (Chon et al., 2019; Madangopal et al., 2022; Renier et al., 2016). tdTomato expressing cells were segmented through the same pipeline using a classifier created using Ilastik (Berg et al., 2019). To train the classifier, 10 chunks of images from the 4 of the subjects were randomly sampled and used as the training dataset (Image size, (x,y,z) = 1.8 mm x 1.8 mm x 100 μm). One chunk image of the same size from each subject was used as the test dataset. The quality of the classifier was visually examined. The total number of tdTomato positive cells per brain region was quantified and normalized by the size of each brain region (density, mm³). In the standard brain atlas, brain regions were hierarchically sorted from broader regions to finer regions (level 1 to level 8). Level 4 was selected for broader analysis of brain regions (Figure 2C). Level 6 was selected for finer analysis of brain regions (Figure 2I, Figure S2). To compare the means of density of tdTomato cell in each brain area between the Appetitive group and Completion group, we conducted Student’s t-test and further corrected for multiple comparisons with Benjamini/Hochberg correction. False discovery rate (FDR) = 0.05 was used for the broader analysis while a loose discovery rate (FDR = 0.10) was used for finer analysis due to the large number of comparisons.

Quantification, data processing and Statistical Analysis

Data were presented as mean with error bar of 95% confidence interval unless otherwise mentioned. Data processing and plot generation was done in Python and R. Plots were further processed in Adobe Illustrator to create Figures. The sample sizes were chosen based on previous studies using similar behavioral experiments. Codes used for analysis and statistics are provided on Github (https://github.com/stuberlab/Ishii-Mating-completion.git). The statistical details of each experiment including the statistical tests used, exact value of n and what n represents, are shown in each figure legend. Each statistical test result is shown in Table 1. Significance was shown as *p < 0.05, **p < 0.01, ***p < 0.001 and not significant values were not noted or shown as n.s.

Author Contributions

Experiments were conceived and designed by K.K.I. and G.D.S. G.D.S. supervised the project, provided resources, and funding. K.H. provided transcriptional analysis data of the MPOA and assisted in gene selection of HCR experiments. J.C. conducted HCR experiments and histological analysis. S.Y., R.E.F. and S.K. conducted behavioral experiments as well as behavioral annotations. C.Z. provided in vivo imaging data analysis pipelines. J.N., A.D.M, E.R.S., and S.A.G. assisted and supervised tissue clearing experiments and whole-brain data analysis.

Acknowledgements

This work was supported by a JSPS oversea fellowship (K.K.I.), Uehara postdoctoral fellowship (K.K.I.), NIH grants R37 DA032750, R01 DA038168, R01 DA054317, P30 DA048736 and Washington Research Foundation Postdoctoral Fellowship (E.R.S.)

Supplementary Figures

Figure S1:

Figure S2:

Figure S3:

Figure S4:

Figure S5:

Figure S6:

Figure S7:

Figure S8:

Statistical analysis: