A neural circuit for male sexual behavior and reward

SUMMARY

Male sexual behavior is innate and rewarding. Despite its centrality to reproduction, a molecularly-specified neural circuit governing innate male sexual behavior and reward remains to be characterized. We have discovered a developmentally-wired neural circuit necessary and sufficient for male mating. This circuit connects chemosensory input to BNSTpr^Tac1 neurons, which innervate POA^Tacr1 neurons that project to centers regulating motor output and reward. Epistasis studies demonstrate that BNSTpr^Tac1 neurons are upstream of POA^Tacr1 neurons, and BNSTpr^Tac1-released Substance P following mate recognition potentiates activation of POA^Tacr1 neurons through Tacr1 to initiate mating. Experimental activation of POA^Tacr1 neurons triggers mating, including in sexually-satiated males and is rewarding such that it elicits dopamine release and self-stimulation of these cells. Together, we have uncovered a neural circuit that governs the key aspects of innate male sexual behavior: motor displays, drive, and reward.

In brief statement

Molecular genetics, neuronal activity monitoring, and circuit epistasis reveal a developmentally-wired neural circuit for male sexual behavior that senses pheromones, recognizes mates, triggers mating, and governs sexual drive and reward.

Graphical Abstract

INTRODUCTION

Mating, an essential social behavior, is developmentally-wired into the brain and can be performed without prior experience. This innate program enables animals to procreate upon sexual maturation and is especially advantageous if encounters with potential mates are infrequent. Indeed, feral mice are solitary, individual males can occupy large territories^1,2, and sexually naive males mate with females³. Despite the experimental tractability of developmentally-patterned processes in mice, a molecularly-specified neural circuit that governs male mating remains to be defined.

Chemosensory cues are critical for detecting potential mates and eliciting sexual behavior^4–8. Male BNSTpr^Tac1 neurons utilize chemosensory cues to distinguish between sexes and guide mating toward females and aggression toward males^9,10. Other neuronal populations are also critical for male mating, including most prominently the preoptic area of the hypothalamus (POA)^11–18, but these do not appear to encode sex of conspecifics in sexually naive males, their molecular identity remains to be defined, and they regulate both mating and aggression^{9–12,15–27}. Indeed, it has been challenging to pinpoint neurons that regulate male mating but not aggression. These findings have led to a model in which mating, aggression, and other social behaviors are regulated by a common set of brain regions, including the principal nucleus of the bed nucleus of the stria terminalis (BNSTpr) and POA, such that these behaviors arise as an emergent output of the network^24,28–33. Finally, most studies on mating have been done in sexually experienced males. Recent work on these neuronal populations reveals significant plasticity in response to internal states and social experience, thereby rendering it difficult to discern their contribution to innate displays of mating^12,14,34,35.

Given the primacy of mating to reproduction, neural circuits must also generate a desire to mate. Such sexual motivation promotes mate-seeking, and male rodents will endure foot-shocks to gain access to females^36,37. Dopaminergic pathways modulate sexual desire or drive, but how these influence circuits that promote mating is poorly understood^38–40.

Sexual behavior is rewarding, presumably as a means to enhance reproduction, but how neural circuits implement this is elusive. Moreover, it is unclear whether such neural pathways are specific to mating-related reward. Classic neuronal self-stimulation studies have suggested the presence of a “pleasure center” for mating in the brain⁴¹. Both the identity of this pleasure center and how it relates to the mating circuit remain to be determined.

Here we use a combination of molecular genetics, neuronal activity monitoring, and functional studies to address these long-standing issues. Our studies reveal a molecularly-specified, developmentally-wired neural circuit that distinguishes between sexes, is innervated by chemosensory pathways, implements motor programs for mating but not aggression, and governs sexual drive and reward.

RESULTS

Male BNSTpr^Tac1 neurons are preferentially activated by females and sufficient for mating

Estrogen receptor alpha (Esr1 or ERα)-expressing BNSTpr (BNSTpr^Esr1) neurons consist of molecularly distinct cells as defined by snRNAseq⁹ (Fig. 1A). Male aromatase-expressing BNSTpr^Esr1 (BNSTpr^Aro) neurons are differentially activated by females and required to distinguish between sexes¹⁰. Of all BNSTpr^Esr1 neurons, only the transcriptomically-defined BNSTpr^Tac1 cell type (a subset of BNSTpr^Aro neurons) is essential to distinguish between sexes⁹. Furthermore, inhibition of male BNSTpr^Tac1 but not other BNSTpr^Esr1 neurons essentially abrogates mating and aggression. We tested whether BNSTpr^Tac1 neurons would also be differentially activated by females, using fiber photometry in sexually naive Tac1^Cre males⁴² (Fig. 1B-E, S1A-M; Table S1). Sexually naive males distinguish between female and male cues^9,10 even though they do not reliably emit courtship ultrasonic vocalizations (USVs) (Table S2). Accordingly, BNSTpr^Tac1 neurons were more responsive to female urine (Fig. 1B-E). Similarly, they were more responsive to a female (Fig. S1A-E), activated during mating, and quiescent during aggression (Fig. S1G-M). Thus, the activity of BNSTpr^Tac1 neurons of sexually naive males distinguishes between sexes and reflects mating.

Figure 1: Activity of male BNSTpr^Tac1 neurons identifies sex of conspecifics and promotes mating.

A. Schematic of section through the adult mouse brain showing that BNSTpr^Tac1 neurons are a subset of BNSTpr^Aro neurons, which are a subset of BNSTpr^Esr1 neurons.

B-E. Schematic of fiber photometry of BNSTpr^Tac1 neurons in males investigating swabs wetted with female urine, male urine, or saline (B). Peri-event time plot (PETP) of normalized GCaMP6s fluorescence (ΔF/F; dark line, mean, and shaded area, SEM for all Figure panels with fiber photometry; dashed vertical line marks insertion of swab into the cage) (C). BNSTpr^Tac1 neurons are activated by urine (D), with larger response to female than male urine (E).

F-J. GRIN lens imaging of BSNTpr^Tac1 GCaMP6s fluorescence with segmented neurons (F). Schematic of miniscope imaging of BNSTpr^Tac1 neurons in males investigating swabs as in panel B (G). Heatmaps of activation of individual neurons during presentation of urine or saline; row numbers correspond to the same neuron across panels, with rows sorted by activation to female urine (H-J).

K-N. Traces of GCaMP6s fluorescence (ΔF/F) of individual neurons in response to stimulus presentation (dashed lines) (K). Percent neurons activated by female or male urine or both (L). Percent co-activated neurons with differential response to female or male urine (M). Among co-activated neurons, the response to female urine was greater than that to male urine (N). Neurons were classified as activated if the peak z-score during the 10s period following swab insertion was >2σ of the peak z-score during the 10s period preceding swab insertion.

O-R. Schematic of optogenetic activation of BNSTpr^Tac1 neurons of a resident male interacting with an intruder male (O). Activation during the first 90s eliminates attacks and promotes mating (P). Raster plot of a male showing aggression toward the intruder without optogenetic activation (Q). Raster plot of behavior of male showing mating with the intruder following optogenetic activation (R).

S-V. Schematic of optogenetic activation of BNSTpr^Tac1 neurons of a resident male interacting with intruder female (S). Optogenetic activation of these cells during the first 90s does not alter the probability (T), number (U), or latency (V) of sniffing or mounting.

Mean ± SEM. n = 6 mice (B-E), 105 neurons from 3 mice (F-N), and 5 mice (O-V). * p < 0.05, ** p < 0.01. Scale bar = 100 μm (F).

See also Fig. S1; Tables S1-3.

To understand activation of BNSTpr^Tac1 neurons at single cell resolution, we imaged these cells from sexually naive males with a miniscope (Inscopix) while presenting pheromonal cues (Fig. 1F-N, S1N-O; Table S1). Many cells (~42%) were activated by female or both female and male cues, whereas few (~3%) were activated solely by male cues (Fig. 1H-L). Of cells co-activated by female and male urine, most showed larger activation with female cues (Fig. 1M-N). BNSTpr^Tac1 neurons therefore distinguish between sexes in two ways: female cues activate more of these cells and at higher levels (Fig. 1L-N, S1O). A recent study reported that sex recognition by BNSTpr^Esr1 cells resulted solely from more cells being activated by female cues¹⁹. This difference likely reflects the fact that we were imaging the transcriptomically-defined cell type (BNSTpr^Tac1) essential for male-typical social behaviors rather than the larger BNSTpr^Esr1 population whose function(s) remain unknown.

The activity of male BNSTpr^Aro and BNSTpr^Tac1 neurons distinguishes between sexes early during the encounter, with higher and longer activation by females¹⁰ (Fig. 1B-N, S1A-E) (BNSTpr^Tac1 neurons, time for GCaMP6s signal to return to baseline: female 90 ± 7.3 s; male 65.8 ± 8.5 s; n = 6; p = 0.009). Optogenetic activation of male BNSTpr^Aro neurons for the first 90 s of a social encounter inhibits aggression and promotes mating with a male without altering mating with a female¹⁰, and these phenotypes were also observed by similar activation of male BNSTpr^Tac1 neurons (Fig. 1O-V, S1P-S; Tables S1, S3). As with BNSTpr^Aro neurons¹⁰, activation of BNSTpr^Tac1 neurons did not immediately trigger mating with males; rather, mating probability, mating latency, and the number of mounts with the male were indistinguishable from those observed in mating with females without optogenetic activation (Mating probability: with female, 0.6; with male, 0.8; p = 0.999; Mating latency: with female, 600 ± 72.5 s; with male, 496.8 ± 79.1 s; p = 0.461; Mount #: with female, 8.5 ± 2.5; with male, 5.8 ± 1.1; p = 0.291) (Fig. 1P,T-V, S1Q,R). As with BNSTpr^Aro neurons¹⁰, intermittent activation of BNSTpr^Tac1 neurons did not elicit hypersexual behavior with females (Fig. S1T-V) (Mating latency: laser on, 724.8 ± 279.1 s; laser off, 859.9 ± 263.2 s; n = 5; p = 0.758; Mount #: laser on, 19.8 ± 2.1; laser off, 16 ± 5; n = 5; p = 0.476).

Together, activation of BNSTpr^Tac1 neurons induces males to treat males as females, and it does not promote hypersexual behaviors with females. These findings support the notion^9,10 that activity of BNSTpr^Tac1 neurons does not enhance sexual motivation but rather reports the presence of a female.

The projection of BNSTpr^Tac1 to POA^Tacr1 neurons governs mating but not aggression

To identify postsynaptic targets of BNSTpr^Tac1 neurons, we expressed synaptophysin-mRuby (Syp:mRuby) in these cells. The largest zone of mRuby+ innervation was the POA (Fig. 2A-B), which regulates male mating^17,18,43,44. Expressing a reporter in the POA of Tacr1^Cre males⁴⁵ revealed putative postsynaptic partners of BNSTpr^Tac1 cells (Fig. 2C-D). Indeed, using an EGFP-encoding monosynaptic rabies vector^46,47 we observed starter POA^Tacr1 neurons (TVA+, EGFP+) in the POA and presynaptic neurons (TVA–, EGFP+) within the BNSTpr (Fig. 2E-F). Most (~93%) BNSTpr^EGFP cells expressed Tac1, and ~45% of BNSTpr^Tac1 neurons were EGFP+ (Fig. 2G-I). Taken together (Fig. 1A, 2H-I), BNSTpr^Tac1 cells represent the major presynaptic input from BNSTpr to POA^Tacr1 neurons.

Figure 2: Innervation of POA^Tacr1 neurons by BNSTpr^Tac1 neurons is essential for male mating.

A-B. Syp:mRuby expression in BNSTpr^Tac1 neurons (A). Syp:mRuby+ termini of BNSTpr^Tac1 neurons in the POA; inset shows area boxed in gray (B).

C-D. mCherry+ BNSTpr^Tac1 neurons (C) and EGFP+ POA^Tacr1 neurons (D) visualized in coronal sections.

E-I. TVA (mCherry) and Rabies (EGFP) expression in POA^Tacr1 starter neurons (E). EGFP+ and TVA–BNSTpr neurons innervating POA^Tacr1 neurons (F). Co-labeling for Tac1 mRNA and EGFP in BNSTpr neurons innervating POA^Tacr1 neurons; arrows show Tac1+, EGFP+ cells (G). Most BNSTpr neurons presynaptic to POA^Tacr1 neurons are BNSTpr^Tac1 neurons (H), and nearly half of BNSTpr^Tac1 neurons innervate POA^Tacr1 neurons (I).

J-L. Strategy to activate BNSTpr^Tac1→POA projections (J). Schematic of optogenetic activation of BNSTpr^Tac1→POA projections of a resident male interacting with intruder male (K). Activation during the first 90s eliminates attacks and promotes mating (L).

M-Q. Strategy to inhibit BNSTpr^Tac1→POA projections (M). Schematic of optogenetic inhibition of BNSTpr^Tac1→POA projections of a resident male interacting with intruder female (N) or male (P). Inhibition suppresses mating with females (O) but does not alter aggression toward males (Q).

Mean ± SEM. n = 4 (A-I), 7 (J-L), and 9 (M-Q) mice. * p < 0.05. Scale bars = 1 mm (C,D), 200 μm (A,B,E,F), 20 μm (B inset, G).

See also Fig. S2; Tables S1,3.

We optogenetically tested whether BNSTpr^Tac1→POA^Tacr1 projections regulated mating or aggression (Fig. 2J; Table S1). Activation of these projections at the start of a social encounter suppressed aggression and promoted mating with males without altering mating with females (Fig. 2K-L, S2A-H; Table S3). This recapitulates the phenotype obtained by activation of BNSTpr^Tac1 soma, indicating that BNSTpr^Tac1→POA^Tacr1 projections convey the presence of a mate. Conversely, inhibiting these projections with the optogenetic actuator parapinopsin (PPO)⁴⁸ suppressed mating, reducing mating probability ~6-fold and mount number >8-fold (Fig. 2M-O, S2I-N; Tables S1, S3). By contrast, such inhibition did not discernibly alter aggressive displays toward a male (Fig. 2P-Q, S2O-S). Together, BNSTpr^Tac1→POA^Tacr1 projections are necessary and sufficient for male mating but not aggression.

POA^Tacr1 neurons are activated during and critical for mating but not aggression

The foregoing findings suggest that POA^Tacr1 neurons regulate mating specifically. Indeed, virtually all POA^Tacr1 neurons express Esr1 (Fig. S3A-D), and POA^Esr1 neurons are active during and critical for male social behaviors, including mating^17,18. Fiber photometry imaging of sexually naive males interacting with a female showed that POA^Tacr1 neurons were activated during anogenital chemoinvestigation (sniffing), mounting, and intromission (penetration) (Fig. 3A-C, S3E-I; Table S1), whereas they were not activated when sniffing or attacking a male (Fig. 3D-F). We next tested whether experimentally activating male POA^Tacr1 neurons regulated mating. Optogenetic activation of POA^Tacr1 cells reduced mating latency and elicited time-locked mating (Fig. 3G-J; Tables S1, S3), with >6-fold more mating bouts during light epochs (15 s on and 45 s off) (Fig. S3J-L). Activation of POA^Tacr1 neurons suppressed aggression and elicited time-locked mating attempts toward males (Fig. 3K-N, S3M; Movie S1). Together, activating male POA^Tacr1 neurons drives mating at the expense of aggression.

Figure 3: POA^Tacr1 neurons are active during male mating and drive male sexual behavior.

A-C. Fiber photometry of POA^Tacr1 neurons in males interacting with intruder female. PETP of GCaMP6s fluorescence (ΔF/F) during sniffing (A) and mounting (B). BNSTpr^Tac1 neurons are activated during sniffs and mounts (C).

D-F. Fiber photometry of POA^Tacr1 neurons in males interacting with intruder male. PETP of GCaMP6s fluorescence (ΔF/F) during sniffing (D) and attacks (E). No discernible activation of BNSTpr^Tac1 neurons during sniffs or attacks (F).

G-J. Schematic of optogenetic activation of POA^Tacr1 neurons of males interacting with females (G). Activation (15/45 s on/off) does not alter percent males mounting females (H) but reduces mating latency ~100-fold (I). Raster plot shows male mounting time-locked to light.

K-N. Schematic of optogenetic activation of POA^Tacr1 neurons of males interacting with intruder males (K). Activation (15/45 s on/off) abrogates aggression and elicits mating (L). Fight latency is prolonged whereas mating latency is reduced by ≥100-fold (M). Raster plot shows male mounting time-locked to light (N).

O-V. Schematic of optogenetic activation (30/30 s on/off; 2 min encounter) of POA^Tacr1 neurons of males interacting with inanimate objects (O). Activation progressively increased investigation of a beaker (P), wood block (Q), and tube (R) and elicited mounting toward the tube outfitted with a toy mouse tail (S), toy mouse (T), male (U), and an unreceptive female (V).

Mean (dark trace) ± SEM (lighter shading) of GCaMP6s activity shown in PETPs of all Figures. Mean ± SEM. n = 8 (A-F), 6 (G-N), and 4 (O-V) mice. * p < 0.05, *** p < 0.001.

See also Fig. S3; Tables S1,3,4; Movie S1.

Activating POA neurons elicits USVs, but the molecular identity of these cells remains to be fully defined^49,50. Optogenetic activation of male POA^Tacr1 cells induced time-locked USVs, but no mounting-type displays (0/4 males), in sexually naive males in the absence of a conspecific (Fig. S3N-Q; Tables S1, S3). Together, POA^Tacr1 neurons trigger USVs but not mating routines in naive, isolated males.

The alacrity with which males mated upon activation of POA^Tacr1 neurons prompted us to test whether they would mate with inanimate objects (Movie S1). Optogenetic activation of male POA^Tacr1 neurons increased investigation of a beaker, wooden block, and a 5 mL Eppendorf tube without inducing mounting (Fig. 3O-R; Tables S1, S4). Strikingly, outfitting the tube with the tail-end of a toy mouse induced mounting during light epochs (Fig. 3S; Table S4). Activation of male POA^Tacr1 neurons progressively increased mounting towards a toy mouse, another male, and a sexually unreceptive female. (Fig. 3T-V; Table S4; Movie S1). Such mounts were largely directed toward the tail-end of the object or mouse, indicating that, in conjunction with activation of POA^Tacr1 neurons, this feature was a releaser stimulus for mating. More broadly, activation of these cells induces time-locked mating with females and atypical targets, including ersatz mice.

We tested whether activity of POA^Tacr1 neurons was essential for mating. Optogenetic inhibition of POA^Tacr1 neurons abrogated mating with females (0/11 males mounted), without altering sniffing (Fig. 4A-C, S3R-T; Tables S1, S3). POA^Tacr1 neurons are also activated upon transitioning from mount to intromission, suggesting that this activity regulates intromission. Indeed, we observed that mount-triggered inhibition reduced intromission >10-fold, and across all mounts, intromission probability dropped from 0.82 to 0.07 (Fig. 4D-F). Male POA^Tacr1 neurons show no discernible activation during aggression (Fig. 3D-F), and correspondingly, their inhibition did not alter fighting (Fig. 4G-K, S3U). Taken together, POA^Tacr1 neurons are active during mating, and their activity is necessary and sufficient for mating but not aggression.

Figure 4: POA^Tacr1 neurons are functionally downstream of BNSTpr^Esr1 neurons for male mating.

A-C. Schematic of optogenetic inhibition of POA^Tacr1 neurons in males interacting with intruder females (A). Silencing eliminates mounting without altering sniffing (B,C).

D-F. Schematic of closed-loop optogenetic inhibition of POA^Tacr1 neurons during mounting (D). Silencing reduced transitions from mount to intromission (E,F). n = 44 mounts (F, laser off), 29 mounts (F, laser on).

G-K. Schematic of optogenetic inhibition of POA^Tacr1 neurons in males interacting with intruder males (G). Silencing did not alter the probability (H), number (I), latency (J), or duration (K) of attacks.

L-P. Strategy to activate BNSTpr^Esr1 neurons while inhibiting POA^Tacr1 neurons (L).

Schematic of chemogenetic inhibition of POA^Tacr1 neurons in males (red cross over POA^Tacr1 neurons in circuit models above M-P) interacting with females (M).

Inhibition of POA^Tacr1 neurons eliminates mating (N).

Schematic of optogenetic activation of BNSTpr^Esr1 neurons (shaded blue in circuit model above O-P) in males given CNO and interacting with females (O).

Inhibition of POA^Tacr1 neurons eliminated mating even when BNSTpr^Esr1 neurons were activated (P).

Q-U. Strategy to activate POA^Tacr1 neurons while inhibiting BNSTpr^Esr1 neurons (Q).

Schematic of chemogenetic inhibition of BNSTpr^Esr1 neurons (red cross over BNSTpr neurons in circuit models above R-U) in males interacting with females (R).

Inhibition of BNSTpr^Esr1 neurons suppressed mating (S).

Schematic of optogenetic activation of POA^Tacr1 neurons (shaded blue in circuit model above TU) in males given CNO and interacting with females (T).

Activation of POA^Tacr1 neurons induced mating even when BNSTpr^Esr1 neurons were inhibited (U).

Mean ± SEM. n = 11 (A-C), 3 (D-F), 11 (G-K), 6 (L-P) and 7 (Q-U) mice. * p < 0.05, ** p < 0.01.

See also Fig. S3,4; Tables S1,3.

POA^Tacr1 neurons function downstream of BNSTpr^Tac1 neurons in promoting male mating

Our findings suggest that male POA^Tacr1 neurons function downstream of BNSTpr^Tac1 neurons in the mating circuit. We tested this with neural circuit epistasis studies using intersectional genetics. In the absence of a Tac1^Flpo line, we employed the Esr1^Flpo strain⁹ to access the BNSTpr^Tac1 population, the only BNSTpr^Esr1 cell type essential for male mating^9,10. We expressed the inhibitory chemogenetic actuator DREADDi⁵¹ in POA^Tacr1 neurons and ChR2 in BNSTpr^Esr1 neurons of Esr1^Flpo;Tacr1^Cre males (Fig. 4L, S4A-C; Tables S1, S3). As expected, males did not mate with females when given the DREADD ligand CNO but territorial aggression or other behaviors were unaffected (Fig. 4M-N, S4D-E, H-L); in separate control experiments, optogenetic activation of BNSTpr^Esr1 neurons, similar to activating BNSTpr^Tac1 neurons, induced mating toward males (Fig. 1O-R, S4F-G). However, activation of BNSTpr^Esr1 neurons while inhibiting POA^Tacr1 neurons did not elicit male sexual behavior (Fig. 4O-P, S4M-P). Together, the mating-promoting effect of BNSTpr^Esr1 neurons is upstream of POA^Tacr1 neurons.

We tested whether POA^Tacr1 neurons were functionally downstream of BNSTpr^Esr1 neurons. Accordingly, we expressed DREADDi and ChR2 in male BNSTpr^Esr1 and POA^Tacr1 neurons, respectively (Fig. 4Q, S4Q-S; Tables S1, S3). As expected, light activated POA^Tacr1 neurons and, in separate tests, CNO inhibited males from mating with females (Fig. 4R-S, S4S-V). Activating POA^Tacr1 neurons while inhibiting BNSTpr^Esr1 neurons elicited mating with females, with a >9-fold increase in mating probability (Fig. 4T-U, S4W-Y). Altogether, POA^Tacr1 neurons are anatomically and functionally downstream of BNSTpr^Tac1 neurons, which are a subset of BNSTpr^Esr1 neurons.

Substance P released by male BNSTpr^Tac1 neurons acts on POA^Tacr1 neurons to promote mating

Substance P, encoded by Tac1, is expressed in the BNSTpr and, among tachykinin neuropeptides, it is the cognate ligand for the receptor Tacr1^52–54, which is important for male mating^55,56. Accordingly, most Tac1^−/− males did not even initiate mating (Tac1^−/−, 1/7 mated; WT, 5/6 mated; p = 0.029). Substance P and Tacr1 are expressed widely and regulate diverse behaviors beyond social interactions^52–54, and it is unclear where they act to regulate mating. We hypothesized that BNSTpr^Tac1 and POA^Tacr1 neurons use this neuropeptide-receptor pair to regulate mating.

We first recorded from POA^Tacr1 neurons in slices to test this hypothesis. Although Substance P can increase excitability of other neurons^57,58, it did not change the firing rate of POA^Tacr1 neurons (S5A-D). We next tested whether Substance P enhanced excitatory input onto POA^Tacr1 neurons. Similar to its effects in other neurons^59–61, Substance P (50 nM) induced long-term potentiation (LTP) of evoked excitatory postsynaptic potentials (EPSPs) in POA^Tacr1 neurons, and this LTP was abolished by L-703,606, a Tacr1 antagonist^56,62 (Fig. 5A-D). Furthermore, optogenetic activation of BNSTpr projections also induced LTP in POA^Tacr1 neurons in a Tacr1-dependent manner (Fig. 5E-H). Thus, Substance P released by BNSTpr^Tac1 neurons signals through Tacr1 to induce excitatory LTP in POA^Tacr1 neurons.

Figure 5: Substance P-Tacr1 signaling potentiates activation of POA^Tacr1 neurons and promotes male mating.

A-D. Substance P induces LTP in POA^Tacr1 neurons.

Strategy to record POA^Tacr1 neurons with local stimulation of glutamatergic inputs in the presence of picrotoxin (100 μM) (A).

tdTomato+ POA^Tacr1 neurons (upper) and recording pipette and theta stimulation electrode (lower) (B).

Summary plot of normalized EPSP amplitudes (Amp) and input resistance (R_in) before and after 5 min (green line) bath application of Substance P (50nM). Inset: Representative EPSP traces of baseline (1) and the last 5 min (2) (C-D).

Substance P induces LTP in POA^Tacr1 neurons (C) and the Tacr1 antagonist L-703,606 (10 μM for slice recordings) blocks it (D). Black line (D) represents average trace with vehicle.

E-H. Activation of BNSTpr projections induces LTP onto POA^Tacr1 neurons.

Strategy to record POA^Tacr1 neurons while optogenetically activating BNSTpr→POA projections in the presence of picrotoxin (E).

Trace of light-evoked firing of BNSTpr neurons (F, upper). Light elicits reliable firing early as well as later (F, lower).

Summary plot of normalized EPSP Amp and R_in before and after transient activation (90 s, blue arrow) of BNSTpr→POA projections (G-H). Inset: Representative EPSP traces of baseline (1) and the last 5 min (2) (G-H).

Light induces LTP in POA^Tacr1 neurons (G) and L-703,606 blocks it (H). Gray line (H) represents the average trace with vehicle.

I-M. Female encounter strengthens excitatory synapses onto POA^Tacr1 neurons in a Substance P-dependent manner.

Schematic of assay in which singly-housed males were administered vehicle or L-703,606 (10 mg/kg for in vivo studies) 30 min prior to initiating testing (I-K, upper panels), and strategy to record POA^Tacr1 neurons with local stimulation of glutamatergic inputs in the presence of picrotoxin (L).

Darker and lighter traces show example excitatory postsynaptic currents recorded at −70 mV and +40 mV, respectively (I-K, lower panels).

MPA/NMDA ratio is increased only in males given vehicle prior to encountering a female (M).

N-P. Schematic of paradigm to infuse Substance P into the POA of WT males interacting with females. Substance P (10 ng in 200 nL) or vehicle was infused 15 min prior to female entry. Substance P did not alter probability of (O) but reduced latency to initiate (P) mating.

Q-S. Schematic of infusion of L-703,606 (500 pmol in 200 nL) into the POA of WT males interacting with a receptive female (Q). L-703,606 reduced the probability of (R) and increased latency to initiate (S) mating.

T-X. Strategy to activate BNSTpr^Tac1→POA projections (T). Schematic of optogenetic activation of BNSTpr^Tac1→POA projections in males given L-703,606 and interacting with females (U). L-703,606 suppresses mating even upon activation of BNSTpr^Tac1→POA projections (V). Schematic of optogenetic activation of BNSTpr^Tac1→POA projections in males given L-703,606 and interacting with males (W). L-703,606 eliminates the male-male mating promoted by activation of BNSTpr^Tac1→POA projections (X).

Mean ± SEM. n = 7 neurons from 4 mice (C), 5 neurons from 3 mice (D), 8 neurons from 4 mice (G), 7 neurons from 3 mice (H), 30 neurons from 4 mice (Control), 33 neurons from 5 mice (Encounter + Vehicle), 31 neurons from 4 mice (Encounter + L-703,606) (M), 8 mice (N-S), 7 mice (T-X). * p < 0.05, ** p < 0.01, **** p < 0.0001.

See also Fig. S5,6; Tables S1,3.

Given that BNSTpr^Tac1 neurons are GABAergic⁹, we wondered how Substance P induced LTP in POA^Tacr1 neurons. Optogenetic activation of BNSTpr projections in the POA inhibited firing of POA^Tacr1 neurons acutely without perdurant effects (Fig. S5E-H). As standard in the field^63–65, our LTP recordings included the GABA-A blocker picrotoxin (Fig. 5E-H). However, we also observed robust Tacr1-dependent LTP without picrotoxin (Fig. S5I-K). Thus, activation of GABAergic BNSTpr^Tac1 neurons initially inhibits but ultimately elicits Tacr1-dependent excitatory LTP in POA^Tacr1 neurons.

Male BNSTpr^Tac1 neurons are activated for ~90 s when first encountering a female whereas mating ensues 10–15 min later (Fig. S1A-E). Sustained activity of neuromodulatory neurons is thought to elicit presynaptic release of neuropeptides⁶⁶. We hypothesized that this ~90 s activation of BNSTpr^Tac1 neurons would release Substance P and potentiate excitatory transmission onto POA^Tacr1 neurons in vivo. We observed a Tacr1-dependent, ~1.6-fold increase in the ratio of evoked AMPA and NMDA receptor-mediated currents (AMPA/NMDA ratio) in POA^Tacr1 neurons of virgin males interacting with females prior to mounting (Fig. 5I-M). The paired-pulse ratio in POA^Tacr1 neurons was not discernibly altered (Fig. S5L-P). Together, encountering a female elicits a Substance P-induced enhancement of excitatory transmission onto POA^Tacr1 neurons that precedes mating.

We wondered if the time-course of this Substance P-induced plasticity of POA^Tacr1 neurons in vivo provided a timer mechanism for male mating initiation. If so, Substance P should bidirectionally regulate mating onset. Substance P infusion into the POA reduced male mating latency by 2.4-fold (Fig. 5N-P, S6A-C). Conversely, L-703,606 increased mating latency such that most males did not mate (Fig. 5Q-S). Importantly, aggressive and locomotor behaviors appeared unchanged (Fig. S6D-J). Our findings therefore indicate that Substance P-induced LTP of POA^Tacr1 neurons provides a molecular timer for male mating onset.

We tested whether male mating induced by POA^Tacr1 neurons is downstream of Substance P signaling from BNSTpr^Tac1 projections. We optogenetically activated these projections or POA^Tacr1 neurons in separate experimental cohorts ± L-703,606 (Fig. 5T,U,W, S6O; Tables S1, S3). We delivered L-703,606 systemically because the optogenetic setup precluded delivery to the POA. L-703,606 inhibited mating despite activation of BNSTpr^Tac1→POA projections, whereas activation of POA^Tacr1 neurons bypassed this pharmacologically-enforced mating block (Fig. 5V,X, S6K-N, P-U). Taken together, Substance P released by BNSTpr^Tac1 neurons potentiates activation of POA^Tacr1 neurons in a Tacr1-dependent manner to initiate male mating.

Male POA^Tacr1 neurons regulate sexual drive and reward

Males lose mating drive following ejaculation and enter a refractory period that lasts ~5 days in C57BL/6J mice^67,68 (Data S1). Activation of hypothalamic dopaminergic AVPV/PVpo^TH neurons restores mating, albeit with variable delays, during the refractory period³⁹. This effect depends on dopamine receptor DRD1 that is thought to be widely expressed in POA^Esr1 neurons³⁹. Non-targeted activation of rat POA neurons also shortens the refractory period⁶⁹. RNAseq studies show that DRD1 is expressed only in a small subset of Esr1+, Tacr1–, POA neurons^9,70. Given that Tacr1 activation, like DRD1 activation, elevates intracellular cAMP⁵² and optogenetic activation of POA^Tacr1 neurons drives time-locked mating, we tested whether activation of these cells restored mating drive. Optogenetic activation of POA^Tacr1 neurons immediately following ejaculation triggered time-locked mating and many males ejaculated again (Fig. 6A-D, S7A-C; Tables S1, S3; Movie S2). Thus, activation of POA^Tacr1 neurons restores mating drive and shortens the ~5 day refractory period to ~1 s.

Figure 6: Forced activation of POA^Tacr1 neuron overrides the post-ejaculatory refractory period and is self-reinforcing.

A-D. Schematic of optogenetic activation of POA^Tacr1 neurons in post-ejaculatory males interacting with females (A). Activation (30/30 s on/off) of these cells re-ignites mating drive, increasing the probability (B) and number (C) of mating routines. Raster plot of sexually satiated male subsequent to activating POA^Tacr1 neurons (D).

E-I. Schematic of optogenetic activation of POA^Tacr1 neurons in the SPP test (E). Activation governs behavioral preference of males (F). Raster plots (each row is a male) showing that optogenetic activation governs investigation by males (H-I).

J-L. Schematic of optogenetic self-stimulation (0.5 s, 40 Hz) of POA^Tacr1 neurons in virgin and sexually experienced males (J). Both virgin and experienced males show more nose pokes to the active port (K,L), with sexual experience increasing self-stimulation (L).

M-N. Schematic of optogenetic inhibition of POA^Tacr1 neurons during the SPP test (M). Inhibition does not alter male preference for side containing the female (N).

O-P. Schematic of optogenetic inhibition of POA^Tacr1 neurons during sucrose consumption test (O). Inhibition does not alter sucrose consumption (P).

Mean ± SEM. n = 10 mice (A-D), n = 10 mice (E-I), n = 7 virgin mice and 10 experienced mice (J-L), n = 10 mice (M-N), n = 6 mice (O-P). * p < 0.05, ** p < 0.01, *** p < 0.001.

See also Fig. S7; Tables S1,3; Data S1; Movie S2.

The hypothalamus contains sites whose activation elicits neuronal self-stimulation^41,71–75. Given that mating is rewarding, we wondered whether POA^Tacr1 neurons encoded mating-related reward. We developed a social place preference (SPP) paradigm in which males displayed preference for an area housing a mating-inaccessible female (Fig. 6E-G; Tables S1, S3). Area-specific optogenetic activation of POA^Tacr1 neurons strengthened preference for that area regardless of the presence of a female (Fig. 6F,H-I, S7D-E). Together, activation of male POA^Tacr1 neurons is strongly reinforcing and even overrides the attraction to females.

We next tested whether males would optogenetically self-stimulate POA^Tacr1 neurons by nose-poking to a light-delivering (active) port in a 2-port chamber (Fig. 6J; Tables S1, S3). Males self-stimulated POA^Tacr1 neurons, with sexually experienced males nose-poking more than naive males (Fig. 6J-L, S7F-G; Movie S2). In summary, POA^Tacr1 neurons constitute a molecularly-defined mating center that encodes reward that is enhanced by sexual experience.

Activation of POA^Tacr1 neurons is rewarding, and we wondered whether inhibition of these cells would be aversive. Optogenetic inhibition of these cells however did not alter males’ preference for the female-occupied area in the SPP paradigm, showing that inhibiting POA^Tacr1 cells suppresses mating but does not lead to avoidance of females (Fig. 6M-N, S7H-J; Tables S1, S3). Similarly, inhibition of POA^Tacr1 neurons did not discernibly alter consumption of sucrose-sweetened water, a fluid innately attractive to mice⁷⁶ (Fig. 6O-P). Thus, inhibiting these neurons does not induce general anhedonia. Altogether, male POA^Tacr1 neurons govern sexual drive, reward, and displays.

POA^Tacr1 neurons act independent of dopaminergic AVPV/PVpo^TH neurons to modulate sexual drive and reward

In contrast to other neuronal populations^39,77, activation of POA^Tacr1 neurons immediately restores mating during the refractory period. We tested whether POA^Tacr1 cells are functionally downstream of one such population, the AVPV/PVpo^TH neurons³⁹, in regulating mating drive. Local delivery of 6-hydroxy DOPA (6OHDA), a dopaminergic neuron-specific toxin, reduced AVPV/PVpo^TH neuron number ~3-fold, with a smaller diminution in TH neurons in the ventral tegmental area (VTA), presumably via diffusion of 6OHDA (Fig. 7A, S8A-B; Tables S1, S3). This ablation did not discernibly alter male mating, and furthermore optogenetic activation of POA^Tacr1 neurons immediately restored mating following ejaculation in all of these males (Fig. 7C-D, S8C-D). Although dopamine signaling is considered central to behavioral reinforcement⁷⁸, optogenetic activation of POA^Tacr1 cells guided place preference and induced self-stimulation in these males (Fig. 7E-F). Our findings indicate that POA^Tacr1 neurons are functionally downstream of AVPV/PVpo^TH neurons in regulating sexual drive. Further, POA^Tacr1 neurons reinforce behavior independent of any dopamine released by AVPV/PVpo^TH neurons.

Figure 7: The BNSTpr^Tac1→POA^Tacr1 pathway is embedded in a neural circuit linking pheromone sensing to motor output and reward.

A-D. Strategy to activate POA^Tacr1 neurons following bilateral ablation of AVPV/PVpo^TH neurons with 6OHDA (10 μg in 1 μL) (A). Schematic of optogenetic activation of POA^Tacr1 neurons in post-ejaculatory males interacting with females (B). Activation (30/30 s on/off) re-ignites mating drive, increasing the probability (C) and number (D) of mating routines.

E. Schematic of optogenetic activation of POA^Tacr1 neurons in males lacking AVPV/PVpo^TH neurons in the SPP test (left). Activation governs behavioral preference of males (right).

F. Schematic of optogenetic self-stimulation (0.5 s, 40 Hz) of POA^Tacr1 neurons in males lacking AVPV/PVpo^TH neurons (left). Males show more nose pokes to the active port (right).

G-K. Strategy for optogenetic activation of VTA termini of male POA^Tacr1 neurons (G).

Males show more nose pokes to the active port (H,I).

More mounts (J) but not intromissions (K) in light epochs in encounters with females.

L-P. Strategy for ptogenetic activation of PAG termini of POA^Tacr1 neurons (L). Males show more nose pokes to the active port (M,N).

More mounts (O) and intromissions (P) in light epochs in encounters with females.

Q-T. Optogenetic activation of POA^Tacr1→VTA projections induces DA release in NAc. Strategy for fiber photometry imaging of DA release in NAc during activation (Q).

PETP (R) and change in fluorescence (S) of NAc neurons during activation.

GRAB_DA signal trace (T) during light epochs (laser 10 s on, 30 s off) shows time-locked DA release.

U. Schematic of the male mating circuit outlined in this study. BNSTpr^Tac1 neurons receive input from the AOB and release Substance P to potentiate activation of POA^Tacr1 neurons. Activation of POA^Tacr1 neurons is necessary and sufficient to elicit mating, and it re-motivates sexually satiated mice to mate and is rewarding. Projections of these cells to the VTA or PAG also drive mating, mating in satiated males, and reinforcement, and activation of VTA projections elicits DA release in NAc. Arrows indicate pathways established by prior work⁷⁹ (black) or from our study (blue). Dashed arrows (purple) indicate pathways emanating from VTA and PAG that remain to be identified and connect with centers that drive mating-related motor programs, establish a motivational state for sexual behavior, and make sexual behavior rewarding.

Mean ± SEM. n = 6 mice (A-F), n = 7 mice (H-I), n = 11 mice (J-K), n = 7 mice (M-N), n = 14 mice (O-P), n = 8 mice (Q-T). * p < 0.05, ** p < 0.01, *** p < 0.001.

See also Fig. S8,9; Tables S1,3; Data S1.

The BNSTpr^Tac1→POA^Tacr1 circuit directly links to centers for sensory input, motor output, and reward

We sought to place the BNSTpr^Tac1→POA^Tacr1 pathway in context of an anatomical framework that endows male mating with its defining features: pheromonal control, reward, and motor displays. We used monosynaptic rabies to determine if BNSTpr^Tac1 or POA^Tacr1 neurons were innervated by main or accessory olfactory bulbs (MOB and AOB), which receive pheromonal input from the nose⁷⁹. We observed presynaptic neurons exclusively in the AOB and only when BNSTpr^Tac1 neurons comprised the starter population (Fig. S8E-N). This is in agreement with work showing that only the BNST receives olfactory bulb input^79–81, and we further show that AOB innervates BNSTpr^Tac1 neurons. Together, pheromonal input converges onto BNSTpr^Tac1 neurons, whose activity distinguishes between sexes in a pheromone sensing-dependent manner¹⁰.

To understand how POA^Tacr1 neurons drive mating and reinforcement, we traced their projections using Syp:mRuby. In addition to other targets, these neurons projected to the periaqueductal gray (PAG) and VTA, centers that gate motor output and reward, respectively (Fig. S8O-Q). We optogenetically activated PAG and VTA projections of POA^Tacr1 neurons in separate experimental cohorts (Fig. 7G,L; Tables S1, S3). Surprisingly, activation of either VTA or PAG projections elicited nose-pokes to the active port and mounting toward both sexes (Fig. 7H-J, M-O, S9A-D). Activation of these projections elicited comparable nose-pokes to the active port (PAG projections, 238 ± 25, and VTA, 284 ± 35.3; n = 7 each; p > 0.99). By contrast, mounting induced by activation of VTA projections did not progress to intromissions whereas activation of PAG projections led to mounting and intromission (Fig. 7K,P). Together, activation of VTA and PAG projections of POA^Tacr1 neurons elicits self-stimulation, and VTA projections can drive mounting whereas PAG projections drive both mounting and intromission.

The foregoing findings could reflect antidromic activation of POA^Tacr1 soma upon activating their projections to the VTA or PAG. We delivered differently-labeled cholera toxin B⁸² (CTB), a retrograde tracer, to the VTA and PAG to test whether POA^Tacr1 neurons collateralize to both targets. We found that ~50% of VTA-projecting POA^Tacr1 cells also projected to the PAG and vice versa (Data S1). We have likely underestimated the extent of collateralization because the rostrocaudal extent (>1 mm) of the VTA and PAG precluded comprehensive labeling of POA^Tacr1 projections. Given such collateralization of projections, we eliminated antidromic activation of POA^Tacr1 soma with locally-delivered tetrodotoxin (TTX⁸³) and optogenetically activated these projections. Activation of projections to either the VTA or PAG induced self-stimulation and mating displays, in a manner indistinguishable from males administered vehicle instead of TTX (Fig. S9E-J). We conclude that projections of POA^Tacr1 neurons to the VTA or PAG regulate mating behaviors and reward-related displays.

Mesolimbic dopamine signaling is considered a sine qua non for reward-related responses^38,40,78,84. We therefore tested whether activation of POA^Tacr1 neurons would elicit dopamine release in the nucleus accumbens (NAc), a reward hub⁷⁸. The anatomical proximity of the POA and NAc precluded concurrent POA^Tacr1 soma activation and dopamine imaging. We therefore optogenetically activated POA^Tacr1 projections to the VTA, which sends massive projections to the NAc^78,85, and imaged dopamine release in the NAc with GRAB_DA⁸⁶. This experiment revealed a time-locked increase in GRAB_DA signal (Fig. 7Q-T, S9K-P; Tables S1, S3). This signal was neither perdurant nor rapidly diminishing across activation epochs, thereby suggesting a cellular mechanism whereby mice repeatedly self-stimulate these cells to obtain reward (Fig. 6J-L). More generally, our findings indicate that the rewarding aspect of POA^Tacr1 activation involves mesolimbic dopamine signaling.

DISCUSSION

We have identified a neural circuit for innate male sexual behavior (Fig. 7U). This circuit enables recognition of potential mates and motor displays of mating, rejuvenates sexual drive, and makes mating rewarding. Moreover, our findings show that the brain employs distinct neural circuits to generate male mating and aggression.

A male mating circuit

Male sexual behavior is regulated by a motivational state or drive – libido in humans – that enables even sexually naive animals to mate. It is also rewarding in the sense that humans self-report it to be pleasurable and animals seek to mate repeatedly. The neural circuit we have uncovered incorporates these and other essential features of male sexual behavior. It receives sensory input, recognizes potential mates, elicits mating even in sexually naive males, governs sexual drive, and its activation is rewarding. BNSTpr^Aro neurons, of which BNSTpr^Tac1 neurons are a subset, exhibit extreme sexual dimorphism in that they elicit male sexual behavior only in males^9,10. Together, this suggests a mechanism whereby the neural circuit we have described preferentially elicits male-typical mating displays in males.

BNSTpr^Tac1 neurons of sexually naive males provide a neuronal activity signature of sex recognition, a signature predictive of subsequent mating with females and fighting with males. USVs have been used as a surrogate for sex recognition and to categorize male mating as either sexual or aggressive in nature^18,19. However, only sexually experienced males emit USVs differentially and reliably to females, and these USVs are governed by context, genetic background, and prior experience, including with males^87–91. Thus, there are confounds to using USV to categorize male mating as sexual or aggressive in intent or as a surrogate for sex recognition. It may be more informative to label male mating agnostically with motor routines shared across species and to use the activity signature of BNSTpr^Tac1 neurons as a measure of sex recognition.

It has been challenging to pinpoint neurons that regulate male mating but not aggression. This has led to a model wherein all innate social behaviors are co-regulated by several brain regions that comprise a “social behavior network”²⁴. This network is proposed to be a fully connected, single layer system, with individual behaviors arising as an emergent output. However, we find that there are developmentally-patterned asymmetries in the neuronal encoding of male mating. BNSTpr^Tac1, but not POA^Tacr1, neurons receive pheromonal input and disabling this input disables BNSTpr neurons from distinguishing between sexes¹⁰. Our epistasis studies further establish an upstream-downstream relationship between BNSTpr^Tac1 and POA^Tacr1 neurons. Recent studies have identified additional centers that regulate male mating and aggression and could be considered to comprise an extended social behavior network^16,21,92,93. Nevertheless, information flow even within this extended network cannot compensate for disabling POA^Tacr1 neurons. Thus, in contrast to a social behavior network, we find directional transfer of information in a neural circuit that drives male mating but not another innate social behavior, aggression.

POA^Tacr1 neurons do not merely relay information from BNSTpr^Tac1 neurons. First, they regulate mating and not aggression. They also encode a timer mechanism that delays mating onset following sex recognition. This lag in mating onset lasts several minutes whereas male BNSTpr^Tac1 neurons recognize females within seconds. The Substance P-potentiated excitatory transmission onto POA^Tacr1 neurons, which occurs over several minutes, provides a timer mechanism for the delay in mating onset. In agreement with this notion, experimental activation of POA^Tacr1 but not BNSTpr^Tac1 neurons bypasses this delay to trigger mating. More broadly, a biological timer may have evolved to enable animals to assess other contingencies such as proximity of predators prior to mating.

POA^Tacr1 projections to the VTA and PAG drive male mating, reinforcement, and immediate recovery from the refractory period (PAG, 7/7 males; VTA, 1/1 male). There may exist POA^Tacr1 neurons projecting uniquely to these targets and regulating subsets of these behavioral features. Alternatively, simultaneous broadcasting of these features may be an integral property of the mating circuit that links behavior and internal states. Such a circuit model differs from the one for parenting⁹⁴, in which POA^Gal neurons broadcast behavioral features to different brain regions to govern parenting. Together, different innate social behaviors governed by the POA may utilize distinct circuit architectures, presumably reflecting unique evolutionary histories or needs.

Sexual drive and reward

Animals desire to mate and to do so repeatedly. Male mating is appetitive or rewarding and associated with a drive that promotes mating even in sexually naive animals^3,36,37,95. So-called pleasure centers in the brain also elicit self-stimulation^41,71,72,96. However, neurons whose activity is self-reinforcing and promotes male mating were unknown. In principle, sexual reward and displays could be encoded by distinct centers. However, the activity of POA^Tacr1 neurons triggers time-locked mating and drives self-stimulation of these cells, which likely entails mesolimbic dopamine signaling. Thus, the same neuronal population drives mating and reward, providing an intimate cellular link between mating and its hedonic value. These neurons are not essential for rewarding prosocial behaviors such as approach^97–99, aggression, or for reward associated with non-social behaviors. Thus, POA^Tacr1 neurons may be exclusively purposed for male mating and its associated reward.

Studies of drive mechanisms for drinking and feeding have provided two modes of operation that enable animals to satisfy needs^100–106. Aversive drives motivate animals to perform behaviors that enable return to a homeostatic setpoint. Other drives regulate behavior bidirectionally such that activation of key neurons is rewarding whereas their inhibition is aversive. By contrast, activation of POA^Tacr1 neurons is rewarding whereas their inhibition is not aversive, thereby identifying another drive mechanism to satisfy innate needs. More broadly, POA^Tacr1 cells represent a molecularly-identified neuronal subset that is active during behavior, is necessary and sufficient for the behavior, drives it in a time-locked manner, regulates behavioral drive, and is rewarding.

POA^Tacr1 neurons provide an entry point to understand sexual drive. The male primate POA is active during mating, and its activation can promote mating^107–109. This close concordance between male primate and rodent sexual behavior suggests that this region also regulates libido in men. Drugs that help with erectile dysfunction¹¹⁰ are ineffective for loss of libido. POA^Tacr1 neurons may offer a therapeutic target to modulate libido in humans.

Limitations of the study

Response profiles of BNSTpr^Tac1 neurons were obtained using undiluted urine, and these could change when different amounts or dilutions of urine are used. Rabies vectors exhibit tropism, and we cannot exclude the possibility that some presynaptic neuronal populations were not labeled. Substance P signaling modulates mating latency and induces LTP onto POA^Tacr1 neurons. Further studies are needed to directly link LTP to mating latency and determine whether Substance P utilizes other signaling pathways to regulate latency. We used systemic antagonism of Tacr1 for some studies, and in future studies it will be important to test whether Tacr1 functions in POA^Tacr1 neurons to regulate mating. The circuit or synaptic mechanisms whereby GABAergic POA^Tacr1 neurons govern behavioral and physiological outcomes remain to be determined. We imagine that projections of BNSTpr^Tac1 and POA^Tacr1 neurons that we have not examined in this study may also modulate the social interactions that we have assayed as well as other behaviors. GCaMP6s fluorescence is a surrogate for neuronal firing, and in vivo voltage recordings during social interactions will be important to reveal the firing patterns within the mating circuit we have identified.

STAR METHODS

Resource Availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled upon reasonable request by the lead contact, Dr. Nirao M. Shah (nirao@stanford.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• Data reported in this paper will be shared by the lead contact upon request.

• All scripts used in this manuscript are available from the lead contact upon reasonable request.

• Any additional information required to reanalyze data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animals

Adult mice 10–24 weeks of age were used for all studies. All mice were bred in our colony (Tac1^Cre, Tacr1^Cre, Tac1 null, Esr1^Flpo)^9,42,45,112 or purchased from Jax (C57BL/6J, used as WT resident males and stimulus females in mating assays) and Charles River (BALB/c, used as WT intruder males). Mice were housed under a reverse 12:12 hour light:dark cycle (lights off at 1pm) with controlled air, temperature, and humidity, and food and water were provided ad libitum unless otherwise mentioned. Mice were group housed by sex after weaning at 3 weeks of age and were therefore sexually naive prior to initiation of behavioral testing. All animal studies were done in compliance with Institutional Animal Care and Use Committee guidelines and protocols approved by Stanford University’s Administrative Panel on Laboratory Animal Care and Administrative Panel of Biosafety.

Viruses

AAV-hSyn-DIO-GCaMP6s (serotype 1), AAV-EF1α-DIO-PPO:Venus (serotype 9), AAV-EF1α-DIO-hChR2(H134R)-EYFP (serotype 1), AAV-hSyn-hChR2(H134R)-EYFP (serotype 1), AAV-CAG-DIO-EGFP (serotype 1), and AAV-CAG-DIO-tdTomato (serotype 1) were purchased from Addgene. AAV-EF1α-DIO-hChR2(H134R):mCherry (serotype 2), AAV-EF1α-DIO-eNpHR3.0:mCherry (serotype 2), AAV-EF1α-DIO-mCherry (serotype 2), AAV-EF1α-DIO-hM4Di:mCherry (serotype DJ, encoding DREADDi), AAV-EF1α-fDIO-hM4Di:mCherry (serotype DJ, encoding DREADDi), and AAV-EF1a-fDIO-hChR2(H134R):EYFP (serotype DJ) were purchased from UNC Vector Core. AAV-hSyn-DIO-mGFP-2A-Synaptophysin:mRuby (serotype DJ), and AAV-DIO-TVA:mCherry-2A-oG (serotype 8.2) were custom packaged by Virovek (Hayward, CA). EnvA G-deleted Rabies-EGFP was purchased from Salk Institute. Cholera Toxin Subunit B (CTB), Alexa Fluor 555 and 647 conjugates were purchased from ThermoFisher Scientific. AAV-hSyn-GRAB_DA2m (serotype 9) was purchased from Addgene. All virus titers were > 10¹² genomic copies/mL.

METHOD DETAILS

Stereotaxic surgeries

Viruses were delivered into brains of male mice at 10–16 weeks of age, using a Kopf stereotaxic alignment system (model 1900), as described previously¹⁰. Viruses were injected into empirically determined coordinates for the BNSTpr and POA (BNSTpr: ±0.85 mm mediolateral (ML), −0.20 mm anteroposterior (AP), and −4.30 mm dorsoventral (DV) relative to bregma; POA: ±0.60 mm ML, +0.05 mm AP, and −5.20 mm DV). For fiber photometry and miniscope studies, 0.5 μl of virus was injected unilaterally (counterbalanced for injections into left or right hemisphere across animals), and for all circuit mapping and functional manipulation studies, 0.5 μl of virus was injected bilaterally. Viruses were infused at a rate of 100 nl/min using a syringe pump (Harvard Apparatus), and the needle was left for an additional 5 min and withdrawn at 1 min/mm.

For fiber photometry and optogenetic manipulation studies, following viral injection during the same surgery, mice were implanted with an optic fiber (0.5 NA, 400 μm diameter and 0.39 NA, 200 μm diameter, respectively; RWD Life Sciences) placed 0.5 mm above the viral injection site of either the BNSTpr, POA, or NAc, depending upon the functional manipulation. For the POA^Tacr1 projection activation experiments, optic fibers were placed at empirically determined coordinates for the VTA and PAG (VTA: ±0.50 mm ML, −2.92 mm AP, and −4.50 mm DV; PAG: ±0.50 mm ML, −4.60 mm AP, and −2.50 mm DV). For POA^Tacr1 projection activation with concurrent GRAB_DA photometry experiments, an activating optic fiber was placed over the VTA and the recording optic fiber was placed over the ipsilateral NAc (±1.20 mm ML, +1.2 mm AP, and −4.20 mm DV). An additional activating optic fiber was placed over the contralateral POA for USV experiments. For local POA infusion of Tacr1 antagonist studies, mice were implanted with dual cannulas (RWD Life Sciences) placed above the POA (±0.60 mm ML, +0.05 mm AP, and −4.70 mm DV). For miniscope imaging studies, a 0.6 mm x 7.3 mm (diameter x length) GRIN lens (Inscopix) was implanted 3 weeks after viral injection. The GRIN lens was connected to a miniscope imaging system (nVista, Inscopix), lowered at 0.5 mm/min while monitoring fluorescence, and placed between 150 and 250 μm dorsal to the coordinates used for viral delivery. The GRIN lens was capped with a small piece of parafilm and silicon adhesive (Kwik-Sil, WPI) prior to closing the skin incision. The silicon cover was removed 14 days after GRIN lens implantation and a baseplate (Inscopix) was installed above the GRIN lens. The baseplate was connected to the miniscope, lowered until clear cellular morphology was detected across the imaging plane, anchored to the skull with adhesive, and covered with a baseplate cover (Inscopix). Cannulas, GRIN lens, and baseplates were secured to the skull using adhesive dental cement (C&B Metabond, Parkell). Following surgery, mice were allowed to recover individually over a heat pad and then returned to their home cages.

For monosynaptic rabies vector-based trans-synaptic labeling, we first delivered an AAV (0.4 μL) encoding Cre-dependent TVA and rabies G glycoprotein to the POA (Tacr1^Cre males) or BNST (Tac1^Cre males). Two weeks following AAV delivery, we delivered an EGFP-encoding rabies vector (0.2 μL) to the same regions. We euthanized males 1 week following delivery of rabies vector and performed immunolabeling for TVA and EGFP to visualize presynaptic cells. For CTB labeling of EGFP+ POA^Tacr1 projections to the VTA and PAG, 0.2 μL of CTB labeled with different fluorophores was delivered to each of these projection fields. Males were euthanized one week following CTB delivery and processed for histology.

Histology

We confirmed expression of all virally delivered genes in all experimental animals using previously published procedures^9,10,12,113. Mice were anesthetized with 2.5% avertin and perfused with HBS followed by 4% paraformaldehyde (PFA). Brains were dissected, post-fixed in 4% PFA overnight, sectioned at 65μm thickness with a vibratome (Leica VT1000S), and immunolabeled and counter-stained with DAPI (0.2 μg/mL). Primary antisera were sheep anti-GFP (BioRad; 1:2000), rat anti-RFP (Chromotek; 1:2000), rat anti-mCherry (ThermoFisher; 1:2000), rabbit anti-Fos (Synaptic Systems; 1:1000), rabbit anti-Esr1 (Millipore; 1:10000), and rabbit anti-TH (Aves; 1:500). Secondary antisera were Alex Fluor 488 donkey anti-sheep (Jackson ImmunoResearch; 1:300), Alexa Fluor 488 donkey anti-rabbit (Jackson ImmunoResearch; 1:300), Cy3 donkey anti-rat (Jackson ImmunoResearch; 1:800), Cy3 donkey anti-rabbit (Jackson ImmunoResearch; 1:800), Cy5 donkey anti-chicken (Jackson ImmunoResearch; 1:500), and Alexa Fluor 647 donkey anti-rabbit (Jackson ImmunoResearch; 1:500). Optic fiber and GRIN lens placement above target locations was verified for all mice, and Fos induction by ChR2 was confirmed by immunostaining for Fos in mice perfused 1 hr after 5 min of laser illumination. To determine the co-expression of Tac1 and Rabies-EGFP, we performed fluorescent hybridization chain reaction based in situ hybridization for Tac1 in combination with the natural fluorescence emitted by Rabies-EGFP. We recently described this HCR ISH protocol in detail⁹. The Tac1 HCR ISH probes and fluorescent amplifiers conjugated to AlexaFluor-647 were purchased from Molecular Instruments (Pasadena, CA). Sections were imaged using confocal microscopy (LSM800, Zeiss) and quantified using ImageJ software (NIH) as described previously^9,10,12.

Drugs

Females used as stimulus animals in mating assays were hormonally primed as described before¹⁰. In brief, females were injected subcutaneously with 10 μg of 17-β-estradiol benzoate (Sigma, cat# E8515) in 100 μL sesame oil at 48 hours before the assay, 5 μg in 50 μL sesame oil at 24 hours before the assay, and 50 μg of progesterone (Sigma, cat# P0130) in 50 μL sesame oil at 4–6 hours before the assay on the day of the mating test.

CNO solution was prepared as previously described¹⁰. In brief, CNO (Enzo) was dissolved in sterile saline at 5 mg/mL, aliquots were frozen, and each aliquot was freshly diluted with sterile saline prior to intra-peritoneal (i.p.) administration. The final dose of CNO for chemogenetic studies was 1 mg/kg.

Substance P (Sigma, Cat# S6883) was dissolved in sterile saline. For local infusion into POA, 200 nL of Substance P (10 ng) was infused per side, a dose based on a previous study⁵⁵. For slice recording, Substance P was dissolved in ddH₂O to a stock concentration of 1 mM and diluted in ACSF to 50 nM for bath application. This concentration was chosen because it is ~2-fold > its EC₅₀ (28 nM)⁵⁸.

L-703,606, (Sigma, # L119) was chosen as the Tacr1 antagonist for use in our studies as i.p. injection of 10 mg/kg of L-703,606 into male mice has previously been shown to disrupt recognition of female urine⁵⁶. Therefore, we used the same 10 mg/kg dose for our Tacr1 antagonist studies with i.p. injections. As in published studies, L-703,606 was dissolved in 45% 2-hydroxypropyl-β-cyclodextrin to a stock concentration of 10 mg/mL (Sigma, cat# 332607). For local infusion into the POA, 1 μL of L-703,606 (500 pmol) was infused per side. This dose was chosen based on published doses of similar Tacr1 antagonists infused into other brain regions^114,115. For slice recording, L-703,606 was dissolved in dimethyl sulfoxide (DMSO, Sigma) to a stock concentration of 10 mM and diluted to 10 μM in ACSF during recording.

Picrotoxin (PTX), was dissolved in DMSO to a stock concentration of 100 mM and diluted to 100 μM during recording.

6OHDA was used to ablate TH⁺ neurons in the AVPV/PVpo as previously described¹¹⁶. On the day of the injection, the buffer solution was made, and 6OHDA (Sigma #H4381) was dissolved to 10 mg/ml in 0.1% ascorbic acid saline (0.9 % NaCl and 0.1 % Ascorbic acid; Sigma # PHR1008) and filter-sterilized. Bilateral stereotaxic injection of 1 μL of 6OHDA solution was delivered to empirically determined coordinates for the AVPV/PVpo (+/− 0.30 mm ML, +0.25 mm AP, and −5.50 mm DV).

The sodium channel blocker, tetrodotoxin (TTX), was used to inhibit POA activity during optogenetic activation of POA^Tacr1 projections. TTX was diluted to a concentration of 1 μM in sterile saline and bilaterally infused (0.4 μL/side) via cannula into the POA.

Fiber photometry

Fiber photometry was conducted as described previously^10,113. Briefly, the implanted optic fiber cannula on the mouse was connected via a patch cable (RWD Life Science) to a previously described custom-built fiber photometry setup (Bayless et al., 2019). An excitation light emitted from a 473 nm diode laser (Omicron LuxX) passed through an optic chopper (Thorlabs, MC2000) running at 400 Hz, neutral density filters (Thorlabs NE10B-A, NE30B-A, NE50B-A), a GFP excitation filter (Thorlabs, MF369–35), a dichroic mirror (Semrock, FF495-Di03–25×36), and a fiber collimator (Thorlabs, F240FC) before being directed to the patch cable connected to the experimental mouse. Fluorescence emitted by GCaMP6s during the behavioral assay was then passed through a fiber collimator, a GFP emission filter (Thorlabs, MF525–39), and a dichroic mirror, and focused by a plano-convex lens (Thorlabs, LA1255-A) onto a femtowatt photoreceiver (Newport, 2151). The signal from the photoreceiver was relayed to a lock-in amplifier (Stanford Research System, SR810), which also received a phase lock-in signal from the optic chopper. The output signal from the amplifier was recorded on a computer via a data acquisition device (LabJack, U6-Pro) at a 250 Hz sampling rate.

Prior to any behavioral testing, mice were habituated to the weight and feel of the optic fiber cable. The cable was attached to the optic fiber implants on the mice, and mice were allowed to move freely in their home cages during 3 separate 15 min habituation sessions. Behavioral video files and fluorescence data were time-locked via a light flash present in both datasets that was initiated by a pulse generator (Doric OTPG-4). The raw fluorescence data was normalized to the median fluorescence of the 5 min baseline period before the entrance of any animal or object into the cage. For PETP, time zero was set to the start of a behavior or event of interest, and the average fluorescence during the time window 10 s prior to a behavioral event was used as the normalization factor to calculate change in fluorescence from baseline (ΔF/F). For peri-event analysis of amplitude changes, the 95 percent peak fluorescence (peak ΔF/F) was calculated and compared for the 10 s time window before and after a behavioral event.

Miniscope calcium imaging

We used a miniaturized fluorescence microscopy setup (nVista, Inscopix) to perform miniscope calcium imaging. During imaging sessions, the baseplate cover was removed and a miniscope was mounted and secured with a side screw. To synchronize fluorescence signals with annotated behaviors, we aligned both the imaging data and behavior videos to the event of a LED flash captured by the camera and recorded in the imaging data by a LED-triggered TTL signal generated by a data acquisition device. We used identical LED power, lens focus, digital gain, exposure time, and recording frame rate for all sessions for the same animal.

Imaging data were loaded on Inscopix data processing software (IDPS, Inscopix) and the size of the image was cropped to the area of the GRIN lens. The cropped data was processed to rectify defective pixels, spatially down-sampled by a factor of two to reduce data size, filtered with a spatial bandpass to remove low and high spatial frequency content, and corrected for motion so that each pixel corresponded to the same location across all frames. Imaging data were then converted into ΔF/F values. To identify the spatial locations of neurons {spatial masks of identified regions of interest (ROIs)} and its associated fluorescence signal from the processed imaging data, a constrained nonnegative matrix factorization-extended (CNMF-E) algorithm was applied in IDPS¹¹⁷. Identified ROIs were further screened based on all pixels being singly connected, morphology, location in imaging field, size, dynamics of associated raw calcium signal, and signal:noise of calcium signal. Calcium signals associated with each identified ROI (neuron) were synchronized with annotated behaviors and z-scored based on the mean and standard deviation of the entire imaging session. To determine whether an identified ROI generated a significant response during a behavioral event, the mean z-score between −10 s and 0 s to the onset of the behavioral event was compared with that between 0 s and 10 s following the onset of the behavioral event. A ROI (neuron) was considered significantly activated if the mean z-score following the behavioral event (0–10s) was >2 σ above the mean z-score prior to the behavioral event (−10–0s).

Optogenetic manipulations

Optogenetic manipulations were conducted as described previously¹⁰. Briefly, the implanted optic fiber cannula on the mouse was connected via a patch cable (RWD Life Science) to a diode laser (Opto Engine). All optogenetic stimuli were produced by a pulse generator (Doric OTPG-4) that triggered a blue light (473 nm) laser for ChR2 and PPO studies and a yellow light (593.5 nm) laser for eNpHR3.0 studies. Laser illumination commenced as soon as an intruder was placed into the resident’s cage. Descriptions of the laser parameters are detailed below for each experiment. As with fiber photometry, prior to any behavioral testing, mice were habituated to the weight and feel of the optic fiber cable. Mice were given 3 separate 15 min habituation sessions. Mice were tested on each behavioral assay once each with laser illumination and no laser illumination, with the order of “Laser on” and “Laser off” assays counterbalanced across mice.

Chemogenetic inhibition

Chemogenetic inhibition studies were performed as described previously¹⁰. Briefly, experimental mice were i.p. injected with CNO at 1 mg/kg or sterile saline 30 min prior to behavioral assays. Mice were tested on each behavioral assay once each with CNO and saline, with the order of CNO and saline administration counterbalanced across animals.

Tacr1 antagonist manipulations

Fifteen minutes prior to behavioral testing, the bilateral cannula (RWD Life Science) implanted above the POA of experimental males was connected to two Hamilton syringes via polyethylene tubing, loaded with 1 μL of L-703,606 (500 pmol) or control vehicle per side and infused slowly over a period of 2 min using a syringe pump (Harvard Apparatus). Needles were left connected for an additional 2 min and then withdrawn. Prior to any behavioral testing, mice were habituated to the handling and the weight and feel of being connected to the Hamilton syringes via polyethylene tubing. Mice were given 3 separate “mock infusion” habituation sessions on separate days.

Electrophysiology

Slice preparation

Brain slices (300 μm) were obtained using standard techniques. Briefly, animals were anesthetized with isoflurane and decapitated. The brain was exposed and chilled with ice-cold artificial cerebrospinal fluid (ACSF) containing 125 mM NaCl, 2.5 mM KCl, 2 mM CaCl₂, 1.25 mM NaH₂PO₄, 1 mM MgCl₂, 25 mM NaHCO₃, and 15 mM D-glucose. ACSF was saturated with 95% O₂ and 5% CO₂. Osmolarity was adjusted to 300–305 mOsm. Coronal brain slices containing BNST or POA were prepared with a vibrating microtome (Leica VT1200 S, Germany) and left to recover in ACSF for 30 min at 34°C and then at room temperature for an additional 30 min. Slices were then moved to a submerged recording chamber perfused with ACSF at a rate of 2–3 ml/min at 30–31°C, and brain slices were recorded within 5 hours after recovery.

Whole-cell recordings

To identify POA^Tacr1 neurons, we injected AAV-DIO-tdTomato into the POA of Tacr1^Cre males and prepared brain slices 4 weeks after viral delivery. POA neurons were visualized under infrared illumination using an Olympus BX51WI microscope equipped with Differential Interference Contrast (DIC), a water-immersion objective (40× NA 0.8), and a CMOS camera (Hamamatsu Photonics). TdTomato+ neurons were identified using fluorescence illumination (Lambda XL, Sutter instrument) coupled with a Texas Red filter set (Chroma Technology Corp).

To record spontaneous activity, whole-cell current-clamp with zero holding current (Ih = 0) was performed with borosilicate glass microelectrode (3–3.5 MΩ) filled with a K⁺-based internal solution (135 mM KMeSO₃, 8.1 mM KCl, 10 mM HEPES, 8 mM Na²-Phosphocreatine, 0.3 mM GTP-Na, 4 mM ATP-Mg, 0.1 mM CaCl₂, 1 mM EGTA; pH 7.2–7.3; osmolarity 285–290 mOsm). Baseline firing rate was recorded for at least 5 min, before Substance P perfusion or BNSTpr terminal stimulation. Substance P (1 μM) was applied in bath perfusion for 5 min while the membrane potential was continuously recorded for 10 min. To stimulate BNSTpr terminals, we injected AAV-ChR2:EYFP to the BNSTpr as well as AAV-DIO-tdTomato to the POA of Tacr1^Cre males, and slices were prepared 8–10 weeks after viral delivery. BNSTpr terminals were stimulated with 90 s 473 nm laser (OptoEngine) illumination (0.5 ms pulse, 5 Hz, 90 s) through the objective while the membrane potential responses of POA^Tacr1 neurons were continuously recorded. To record optically evoked action potentials in ChR2+ BNSTpr neurons, whole-cell current-clamp (Ih = 0) was performed, and the same laser illumination as above was delivered.

To record AMPA/NMDA ratios and paired-pulse ratios, picrotoxin (100 μM) was included in the ACSF perfusion and whole-cell voltage-clamp was performed with a Cs⁺-based internal solution (126 mM CsMeSO₃, 8 mM NaCl, 10 mM HEPES, 2.9 mM QX-314, 8 mM Na²-Phosphocreatine, 0.3 mM GTP-Na, 4 mM ATP-Mg, 0.1 mM CaCl₂, 1 mM EGTA; pH 7.2–7.3; osmolarity 285–290 mOsm). Membrane potential was held at −70 mV to record AMPA receptor mediated current or +40 mV to record NMDA receptor-mediated current. Glutamatergic terminals were stimulated locally with a concentric bipolar microelectrode (FHC) placed 100–200 μm from the recorded POA^Tacr1 neuron. Stimulation intensity (0.2 ms, 100–200 μA) was adjusted to evoke EPSCs with an amplitude of around 50–100 pA while recording at +40 mV. AMPA/NMDA ratios were calculated by dividing the amplitude of AMPA current, measured as the peak current at −70 mV, by the amplitude of NMDA current, measured at 50 ms after stimulus at +40 mV when the contribution of the AMPA component was minimal. In the same neurons, two EPSCs were evoked 50 ms apart while being recorded at −70 mV, and the paired-pulse ratio was calculated by dividing the peak amplitude of the second EPSC by the first.

Recordings were obtained with a Multiclamp 700B amplifier (Molecular Devices) using the WinWCP software (University of Strathclyde, UK). Liquid junction potential was not corrected. Signals were filtered at 2 kHz, digitized at 10 kHz (NI PCIe-6259, National Instruments), and analyzed offline using Matlab (Mathworks).

Perforated patch and LTP recording

Electrical access to the POA^Tacr1 neurons was achieved through the perforated-patch method using Gramicidin A (Sigma). Perforated patch was performed with a borosilicate glass microelectrode (3–3.5 MΩ), front-filled with 1 μl K⁺-based internal solution and back-filled with 10 μl Gramicidin A-containing internal solution. The Gramicidin A-containing internal solution was made fresh before use: Gramicidin A (Sigma) was dissolved in DMSO to 20 mg/mL and then diluted in the K⁺-based internal solution yielding a final concentration of 200 μg/mL. The solution was thoroughly mixed by vortexing and then sonicated for 5 min and filtered with a centrifuge tube filter (0.22 μm, Spin-X, Costar). After the microelectrode formed a giga seal with the cell membrane, access resistance was continuously monitored during perforation by applying a −5 mV pulse from a holding potential of −70 mV, under the voltage-clamp mode. A stable perforated patch normally formed within 30–60 minutes, and the access resistance stayed around 30–50 MΩ without further decreasing. Then the recording was switched to current-clamp mode. The membrane potential was adjusted to ~−70 mV by injecting a negative holding current (Ih = − 50–200 pA) since POA^Tacr1 neurons were intrinsically firing. The serial resistance was compensated with amplifier bridge balance. The data were excluded if the input resistance (R_in) changed more than 25% over the course of the experiment.

Picrotoxin (100 μM) was included in ACSF perfusion throughout the recording, EPSPs were evoked by focal extracellular stimulation at 0.05Hz with a small theta glass electrode positioned 50–100 μm from the recorded cell body. Stimulation intensity (0.2 ms, 5–30 μA) was adjusted to evoke stable EPSPs with an amplitude of around 2–5 mV. Basal EPSPs were recorded for at least 5 min, before LTP was induced by Substance P perfusion or BNSTpr terminal stimulation. Substance P (50 nM) was applied in bath perfusion for 5 min, and EPSPs were continuously recorded for another 25 min to monitor the change of amplitudes. BNSTpr terminals were stimulated with 90 s 473 nm laser illumination as above, and EPSPs were recorded afterwards for 30 min. For Tacr1 antagonist experiments, L-703,606 (10 μM) was added in ACSF perfusion throughout the recordings.

Behavioral testing

All behavioral testing was initiated ≥ 1 hr after onset of the dark cycle and recorded using camcorders (Sony) under infrared illumination as described previously¹⁰. Videos were played at 30 frames per second and manually annotated using custom software described previously^11,118,119. This permitted analysis of multiple parameters (including number, duration, latency, probability, and inter-event interval) of different behavioral routines. In particular, anogenital investigation (sniff), mounting, repeated pelvic thrust (intromission), and ejaculation were scored for sexual behaviors. Aggression was scored as occurring when physical attacks (episodes of biting, wrestling, tumbling, chasing) were observed.

Both sexual and aggressive displays were annotated during female-male and male-male interactions. Male-female interactions lasted for 30 min to enable display of different mating routines. Male-male interactions were restricted to 15 min to allow fighting to occur but were short enough to preclude excessive attacks toward intruders. When female and male mating were directly compared, the behavioral statistics for interactions with female intruders were derived from only including behaviors that occurred in the first 15 min of the assay to make the analysis window the same as the 15 min assay with the male intruder. All behavioral studies were conducted using previously published procedures^{11,20,118–120}. The various behavioral assays used in the study are described briefly below.

Behaviors for Fig. 1, S1

To examine the response of sexually naïve male BNSTpr^Tac1 neurons to pheromones (Fig. 1B-E), sexually naïve males with GCaMP6s expressed in and optic fibers above BNSTpr^Tac1 neurons were single housed for ≥ 7 days prior to fiber photometry studies. Males were exposed for 3 min on separate days in their home cage to a 1”x1” cotton swab wetted with 80 μl of undiluted urine from urine from WT group housed primed C57BL/6J females, WT group housed C57BL/6J males, or saline. The order of the swab assays was counterbalanced across mice. Urine was collected 3–6 hours prior to use and kept on ice until pipetted onto the swab.

To examine the response of sexually naïve male BNSTpr^Tac1 neurons during initial encounters with females and males (Fig. S1A-E), after completion of urine swab assays, males were exposed for 3 min on separate days in their home cage to a WT group housed hormonally primed female, a WT group housed male, or an inanimate toy mouse, with the order of the intruder assays counterbalanced across mice. The short 3 min assays precluded the occurrence of any mating or aggression displays during the assays.

To examine the response of male BNSTpr^Tac1 neurons during mating and aggression (Fig. S1F-M), after completion of the brief encounter assays, males were assayed on separate days in their home cage during a 30 min interaction with a WT group housed hormonally primed female and a 15 min interaction with a WT group housed male. For both assays, males had no prior experience mating with a female intruder or attacking a male intruder. We imaged GCaMP6s fluorescence in multiple assays until the experimental male exhibited typical social displays, mating toward female intruder and aggression toward male intruder. Most males showed social behaviors even in the first assay (3/6 mated with females and 4/6 attacked males), and all males exhibited mating with females and aggression with males within 3 rounds of testing. In our imaging studies using sexually naïve males, we did not observe any significant increase in the Ca²⁺ activity of BNSTpr^Tac1 neurons during attack bouts. However, in a recent study¹²¹, we did observe an increase in Ca²⁺ activity of BNSTpr^Tac1 neurons during attack bouts of males that had extensive mating and aggression experience. Together, these findings indicate that mating and/or aggression experience increases the responsivity of BNSTpr^Tac1 neurons during attack bouts.

To examine the response of individual BNSTpr^Tac1 neurons to pheromones (Fig. 1F-N, S1N-O), sexually naïve males with GCaMP6s expressed in and GRIN lenses above BNSTpr^Tac1 neurons were single housed for ≥ 7 days prior to miniscope imaging studies. Males were exposed for 3 min each in their home cage to a 1”x1” cotton swab wetted with 80 μl of undiluted urine from WT group housed primed C57BL/6J females, WT group housed C57BL/6J males, or saline. To ensure reliable imaging from the same set of BNSTpr^Tac1 neurons in response to each swab, the swabs were presented during the same imaging session with the order of the presentation of the swabs counterbalanced across mice and a 7 min interval between swab presentations. Urine was collected 3–6 hours prior to use and kept on ice until pipetted onto the swab.

To examine the effect of optogenetic activation of BNSTpr^Tac1 neurons during social interactions with males and females (Fig. 1O-V, S1P-V), sexually naïve males with ChR2 expressed in and optic fibers above BNSTpr^Tac1 neurons were single housed for ≥ 7 days prior to behavioral testing. To assay male-female interactions, a WT group-housed hormonally primed female was inserted into the experimental male’s home cage for 30 min. To assay male-male interactions, a WT group-housed male was inserted into the experimental male’s home cage for 15 min. Laser illumination parameters were based on published parameters for BNSTpr^Aro neurons that promote male-male mating¹⁰. For transient photostimulation, laser illumination (473 nm, 5 Hz, 5 ms pulse width, 8 mW from the fiber tip) was provided for the first 90s of the assay, and then illumination was turned off for the rest of the assay. For intermittent photostimulation, laser illumination (473 nm, 5 Hz, 5 ms pulse width, 8 mW from the fiber tip) was provided in cycles of 30 s laser on and 30 s laser off throughout the entire assay.

Behaviors for Fig. 2, S2

To examine the effect of optogenetic activation of BNSTpr^Tac1→POA projections during social interactions with males and females (Fig. 2J-L, S2A-H), sexually naïve males with ChR2 expressed in BNSTpr^Tac1 neurons and optic fibers above the POA were single housed for ≥ 7 days prior to behavioral testing. Behavioral assays and transient photostimulation were performed as described above for Fig. 1O-V.

To examine the effect of optogenetic inhibition of BNSTpr^Tac1→POA projections during social interactions with males and females (Fig. 2M-Q, S2I-R), sexually naïve males with PPO expressed in BNSTpr^Tac1 neurons and optic fibers above the POA were single housed for ≥ 7 days prior to behavioral testing. Behavioral assays were performed as described above for Fig. 1O-V. Laser illumination parameters were based on published parameters that inhibit synaptic transmission at axon terminals⁴⁸. Laser illumination (473 nm, 10 Hz, 10 ms pulse width, 8 mW from the fiber tip) was delivered constantly for the entire assay.

Behaviors for Fig. 3, S3

To examine the response of male POA^Tacr1 neurons during mating and aggression (Fig. 3A-F, S3E-I), sexually naïve males with GCaMP6s expressed in and optic fibers above POA^Tacr1 neurons were single housed for ≥ 7 days prior to fiber photometry studies. Behavioral assays were performed as described above for Fig. S1F-M.

To examine the effect of optogenetic activation of POA^Tacr1 neurons during social interactions with males and females (Fig. 3G-N, S3J-N), sexually naïve males with ChR2 expressed in and optic fibers above POA^Tacr1 neurons were single housed for ≥ 7 days prior to behavioral testing. Behavioral assays were performed as described above for Fig. 1O-V. Laser illumination parameters were based on published parameters for POA^Esr1 neurons that elicit mating behavior¹⁷. Laser illumination (473 nm, 40 Hz, 10 ms pulse width, 8 mW from the fiber tip) was provided in cycles of 15 s laser on and 45 s laser off throughout the entire assay.

To examine the effect of optogenetic activation of POA^Tacr1 neurons during interactions with inanimate objects (Fig. 3O-V), after completion of social interaction assays, males were given separate 2 min interaction assays with a range of inanimate object and mouse stimuli. During each assay, laser illumination (473 nm, 40 Hz, 10 ms pulse width, 8 mW from the fiber tip) was provided for two cycles of 30 s laser off and 30 s laser on. The order of the object interaction assays was counterbalanced across mice.

To examine the effect of optogenetic activation of POA^Tacr1 neurons on USV production (Fig. 3N-Q), sexually naïve group housed males were tested in a clean cage alone with no social stimulus. An ultrasonic microphone (Noldus, Pettersson M500–384 ultrasonic microphone) was placed above the testing cage. During the assay, laser illumination (473 nm, 40 Hz, 10 ms pulse width, 8 mW from the fiber tip) was provided for three cycles of 30 s laser off and 30 s laser on. The USV frequency analysis window was restricted to 50k Hz – 80,000k Hz and analyzed using Noldus UltraVox XT software.

Behaviors for Fig. 4, S4

To examine the effect of optogenetic inhibition of POA^Tacr1 neurons during social interactions with males and females (Fig. 4A-K, S3R-U), sexually naïve males with eNpHR3.0 expressed in and optic fibers above POA^Tacr1 neurons were single housed for ≥ 7 days prior to behavioral testing. Behavioral assays were performed as described above for Fig. 1O-V. Laser illumination parameters were based on published parameters for BNSTpr^Aro neurons that inhibit male mating behaviors¹⁰. For constant inhibition studies, laser illumination (593.5 nm, 6 mW from the fiber tip) was on throughout the entire assay. For mount-triggered inhibition studies, laser illumination (593.5 nm, 6 mW from the fiber tip) was delivered immediately upon mounting of a female. Laser illumination was turned off as soon as the mounting behavior ended. This paradigm was repeated for each subsequent mounting attempt.

To examine the effect of optogenetic activation of BNSTpr^Esr1 neurons in combination with chemogenetic inhibition of POA^Tacr1 neurons during social interactions with males and females (Fig. 4L-P, S4A-P), sexually naïve males with ChR2 expressed in and optic fibers above BNSTpr^Esr1 neurons and DREADDi expressed in POA^Tacr1 neurons were single housed for ≥ 7 days prior to behavioral testing. Behavioral assays and laser illumination parameters were performed as described above for Fig. 1O-V. CNO at 1 mg/kg or control saline was i.p. injected into males 30 min prior to behavioral assays with the order of the CNO and control injection assays counterbalanced across mice.

To examine the effect of chemogenetic inhibition of POA^Tacr1 neurons on non-social behaviors, experimental males were also assayed for performance in locomotor activity (Fig. S4J), latency to find hidden food (Fig. S4K), and elevated plus maze (Fig. S4L), as described previously^11,20,119. Briefly, males were tested for locomotor activity by quantifying the number of midline crosses that each male made during a social interaction. Then, males were tested for olfactory motivated behavior with a food finding assay. After being food deprived for 6hrs, males were inserted into a clean cage with a cracker (Cheez-It) buried under the bedding. Males were given 5 min to find the cracker at which time the assay ended. Latency to find the food was scored. Then, males were tested for anxiety-like behavior with the elevated plus maze. Males were placed in the center of an elevated plus maze facing the open arm at the start of the assay. Time spent in the closed or open arms during the 5 min assay was scored. Mice were given ≥1 days between each behavioral test. Chemogenetic inhibition was performed as described above for Fig. 4L-P.

To examine the effect of optogenetic activation of POA^Tacr1 neurons in combination with chemogenetic inhibition of BNSTpr^Esr1 neurons during social interactions with males and females (Fig. 4Q-U, S4Q-Y), sexually naïve males with ChR2 expressed in and optic fibers above POA^Tacr1 neurons and DREADDi expressed in BNSTpr^Esr1 neurons were single housed for ≥ 7 days prior to behavioral testing. Behavioral assays were performed as described above for Fig. 1O-V. Laser illumination parameters were performed as described above for Fig. 3G-N. Chemogenetic inhibition was performed as described above for Fig. 4L-P.

To test for effects of laser illumination absent expression of optogenetic actuators, we expressed mCherry in male POA^Tacr1 neurons instead of ChR2 (Fig. 3G-V, S3J-M) or eNpHR3.0 (Fig. 4A-K, S3R-U). We assayed these control mice using the behavioral tests and laser illumination paradigms described for experimental mice expressing ChR2 or eNpHR3.0. The same cohort of control mice was tested without light or with 473 nm or 593.5 nm laser illumination. Pairwise comparisons between no-light and laser illumination conditions revealed no effect of light and are reported in Table S3.

Behaviors for Fig. 5, S5, S6

To examine the effect of encountering a female on the AMPA/NMDA ratio and paired-pulse ratio of POA^Tacr1 neurons (Fig. 5I-M, S5L-O), sexually naïve males were single-housed for ≥ 7 days prior to behavioral testing. L-703,606 at 10 mg/kg or control vehicle was injected i.p. into experimental males. After 30 min, the male continued to be singly-housed for 15 min or a hormonally primed sexually receptive WT female was inserted into the male’s cage for 15 min. No mounting of the female was observed in these 15 min for assays in which the male was used for electrophysiology studies shown in this study. At the end of behavioral assays, male mice were anesthetized and brains were collected for slice preparation and recording.

To examine the effect of local POA infusion of Substance P during social interactions with females (Fig. 5N-P, S6B-C), sexually naïve males with bilateral cannulas placed above the POA were single-housed for ≥ 7 days prior to behavioral testing. Behavioral assays were performed as described above (Fig. 1O-V). Fifteen min prior to behavioral assays, 200 nL of Substance P (10 ng) or control vehicle was infused per side into the POA. The order of Substance P and vehicle infusion was counterbalanced across mice and performed ≥2 days apart.

To examine the effect of local POA infusion of the Tacr1 antagonist, L-703,606, during social interactions with males and females (Fig. 5Q-S, S6D-J), sexually naïve males with bilateral cannulas placed above the POA were single housed for ≥ 7 days prior to behavioral testing. Behavioral assays were performed as described above for Fig. 1O-V. Fifteen min prior to behavioral assays, 1 μL of L-703,606 (500 pmol) or control vehicle was infused per side into the POA. The order of L-703,606 and control infusion assays were counterbalanced across mice and performed ≥2 days apart.

To examine the effect of optogenetic activation of BNSTpr^Tac1→POA projections in combination with systemic administration of the Tacr1 antagonist, L-703,606, during social interactions with males and females (Fig. 5T-X, S6K-N), following experiments of Fig. 2J-L, S2A-H, males with ChR2 expressed in BNSTpr^Tac1 neurons and optic fibers above the POA were used for behavioral testing. Behavioral assays and laser illumination parameters were performed as described above for Fig. 1O-V. L-703,606 at 10 mg/kg or control vehicle was i.p. injected into males 30 min prior to behavioral assays with the order of the L-703,606 and control injection assays counterbalanced across mice.

To examine the effect of optogenetic activation of POA^Tacr1 neurons in combination with systemic administration of the Tacr1 antagonist, L-703,606, during social interactions with males and females (Fig. S6O-U), following experiments of Fig. 3G-N, S3J-M, males with ChR2 expressed in and optic fibers above POA^Tacr1 neurons were used for behavioral testing. Behavioral assays and laser illumination parameters were performed as described above for Fig. 3G-N. L-703,606 administration was performed as described above for Fig. 5T-X.

Behaviors for Fig. 6, S7

To examine the effect of optogenetic activation of POA^Tacr1 neurons during the post-ejaculatory refractory period (Fig. 6A-D, S7A-C), males with ChR2 expressed in and optic fibers above POA^Tacr1 neurons were single housed for ≥ 7 days prior to behavioral testing. Prior to measuring refractory period behavior, the experimental males were allowed to interact with a female until ejaculation. In a subset of the mating assays, POA^Tacr1 neurons were activated to ensure that mating behavior and ejaculation occurred. Refractory period mating experiments began when the male became ambulatory after the initial ejaculation. The refractory period mating assay consisted of a 30 min assay with a WT female intruder. Laser illumination (473 nm, 40 Hz, 10 ms pulse width, 5 mW from the fiber tip) was provided in cycles of 30 s laser on and 30 s laser off throughout the entire refractory period mating assay. Laser illumination and no laser illumination assays were counterbalanced and performed at least one week apart. In a subset of the assays, the female that received the initial ejaculation was replaced by a new WT female. In these cases, the male interacted with a new female during both the laser illumination and no laser illumination assays.

To examine the effect of optogenetic activation of POA^Tacr1 neurons during the real time social place preference assay (Fig. 6E-I, S7D-E), males with ChR2 expressed in and optic fibers above POA^Tacr1 neurons were single housed for ≥ 7 days prior to behavioral testing. In the real time social place preference assay, experimental males were permitted to freely explore a rectangular (19.5 cm x 13.0 cm) behavioral arena. At the ends of the arena were two containers, one which was empty and the other which contained a stimulus WT female. The containers were constructed by securing together two disposable feeders (Innovive), and both experimental and stimulus animals were acclimated to the behavior arena for 10 minutes 1 day prior to the experiment. Each experimental mouse underwent three 10 minute trials: “Laser off” in which there was no laser illumination, “Laser on, female side” in which laser illumination (473 nm, 40 Hz, 10 ms pulse width, 5 mW from the fiber tip) was provided only when the experimental mouse was located in the half of the behavior arena containing the stimulus female, and “Laser on, empty side” in which laser illumination (same as above) was provided only when the experimental mouse was located in the half of the behavior arena with the empty container. The location of the experimental mouse was tracked using custom code (MATLAB). The 3 trials were conducted consecutively, with the “Laser off” trial always occurring first and the “Laser on, female side” and “Laser on, empty side” being counterbalanced across mice. Female preference index was calculated as (investigate female container – investigate empty container) / (investigate female container + investigate empty container). Sniffing or placement of forepaws onto the female or empty container was manually classified as investigation behavior, and duration of investigation behavior towards the containers was quantified to construct the female preference index.

To test if males perform self-stimulation for optogenetic activation of POA^Tacr1 neurons (Fig. 6J-L, S7F-G), males with ChR2 expressed in and optic fibers above POA^Tacr1 neurons were single housed for ≥ 7 days prior to behavioral testing. In the optogenetic self-stimulation assay, experimental males were permitted to freely explore a behavior chamber containing two ports. Each mouse was acclimated to the behavior chamber for 10 minutes 1 day prior to the experiment. When the mouse made a nose poke to the port designated as “active,” the mouse received a short period of laser illumination (0.5 s, 473 nm, 40 Hz, 10 ms pulse width, 5 mW from the fiber tip). When the mouse made a nose poke to the other port, designated “inactive”, nothing happened. Each experimental mouse underwent two 10 min trials with the identity of the “active” port alternating between the two port locations in each trial. The two trials were performed consecutively, and the order of the active ports were counterbalanced. Each trial began when the mouse made a nose poke to both ports at least once. For the experiments of Fig. 6J-L, S7F-G, initiation of laser illumination and nose-poke quantification were performed manually from an overhead camera. For the experiments of Fig. 7F,H,I,M,N, S9E-F) custom code and Arduino were used for initiation of laser illumination and nose-poke quantification.

To examine the effect of optogenetic inhibition of POA^Tacr1 neurons during the real time social place preference assay (Fig. 6M-N), males with eNPHR3.0 expressed in and optic fibers above POA^Tacr1 neurons were single housed for ≥ 7 days prior to behavioral testing. The real time social place preference assay was performed as described above for Fig. 6E-I. “Laser off”, “Laser on, female side”, and “Laser on, empty side” trials were performed as described above for Fig. 6E-I, except with 593.5 nm laser illumination constant at 5 mW from the fiber tip.

To examine the effect of optogenetic inhibition of POA^Tacr1 neurons during sucrose consumption (Fig. 6O-P), males with eNPHR3.0 expressed in and optic fibers above POA^Tacr1 neurons were tested on a sucrose consumption assay. In the sucrose consumption assay, an experimental mouse was permitted to freely explore a rectangular behavior arena (19.0 cm x 28.5 cm) containing a spout on one end of the arena that dispensed water or sucrose water. When the mouse drank from the spout, the spout dispensed additional liquid once the mouse crossed the midpoint of the arena. A repeating pipette was manually used to dispense 4 uL liquid. Mice were acclimated to the arena 1 day before the first trial and gained experience drinking from a spout that dispensed 30% sucrose water. In a 10 min trial, the number of times the mouse drank the liquid and crossed the midpoint of the arena was quantified. During “Laser on” trials, the mouse received constant 5 mW 593.5 nm laser illumination, and during “Laser off” trials, no laser illumination was provided. The trials were conducted in a block of two consecutive trials of “Laser on” and “Laser off” for each concentration of sucrose (0%, 15%, and 30%), and each block was performed twice, with one block performing the “Laser on” trial first and the other block performing the “Laser off” trial first. Each block was performed at least two days apart.

Behaviors for Fig. 7, S8, S9

To examine the effect of optogenetic activation of POA^Tacr1 neurons during the post-ejaculatory refractory period in males with ablated AVPV/PVpo^TH neurons (Fig. 7A-D, S8A-D), males with ChR2 expressed in and optic fibers above POA^Tacr1 neurons and injected with 6OHDA or control vehicle into the AVPV/PVpo were single housed for ≥ 7 days prior to behavioral testing. Refractory period mating experiments and laser illumination conditions were performed as described above for Fig. 6A-D.

To examine the effect of optogenetic activation of POA^Tacr1 neurons during the real time social place preference assay in males with ablated AVPV/PVpo^TH neurons (Fig. 7E), males with ChR2 expressed in and optic fibers above POA^Tacr1 neurons and injected with 6OHDA or control vehicle into the AVPV/PVpo were assayed. Real time social place preference experiments and laser illumination conditions were performed as described above for Fig. 6E-I.

To test if males with ablated AVPV/PVpo^TH neurons perform self-stimulation for optogenetic activation of POA^Tacr1 neurons (Fig. 7F), males with ChR2 expressed in and optic fibers above POA^Tacr1 neurons and injected with 6OHDA or control vehicle into the AVPV/PVpo were assayed. Optogenetic self-stimulation experiments and laser illumination conditions were performed as described above for Fig. 6J-L.

To test if males perform self-stimulation for optogenetic activation of POA^Tacr1→VTA projections (Fig. 7G-I), males with ChR2 expressed in POA^Tacr1 neurons and optic fibers above the VTA were single housed for ≥ 7 days prior to behavioral testing. Optogenetic self-stimulation experiments and laser illumination conditions were performed as described above for Fig. 6J-L.

To examine the effect of optogenetic activation of POA^Tacr1→VTA projections during social interactions with males and females (Fig. 7J-K, S9A-B), males with ChR2 expressed in POA^Tacr1 neurons and optic fibers above the VTA were assayed. Behavioral assays were performed as described above for Fig. 1O-V, and laser illumination conditions were performed as described above for Fig. 3G-N.

To test if males perform self-stimulation for optogenetic activation of POA^Tacr1→PAG projections (Fig. 7L-N), males with ChR2 expressed in POA^Tacr1 neurons and optic fibers above the PAG were single housed for ≥ 7 days prior to behavioral testing. Optogenetic self-stimulation experiments and laser illumination conditions were performed as described above for Fig. 6J-L.

To examine the effect of optogenetic activation of POA^Tacr1→PAG projections during social interactions with males and females (Fig. 7O-P, S9C-D), males with ChR2 expressed in POA^Tacr1 neurons and optic fibers above the PAG were assayed. Behavioral assays were performed as described above for Fig. 1O-V, and laser illumination conditions were performed as described above for Fig. 3G-N.

To examine if silencing of POA activity alters the behavioral effects of optogenetic activation of POA^Tacr1→VTA or PAG projections during the self-stimulation assay (Fig. S9E-F) or during social interactions with males and females (Fig. S9G-J), males with ChR2 expressed in and bilateral cannula placed above POA^Tacr1 neurons and optic fibers above the VTA or PAG were assayed. For optogenetic self-stimulation experiments, behavioral assays and laser illumination conditions were performed as described above for Fig. 6J-L. For social interaction assays, behavioral assays were performed as described above for Fig. 1O-V, and laser illumination conditions were performed as described above for Fig. 3G-N. Twenty min prior to behavioral assays, 0.4 μL of 1 μM TTX or control vehicle was manually infused per side into the POA over a period of 2 min. The order of TTX and control infusion assays were counterbalanced across mice and performed ≥2 days apart.

To test if optogenetic activation of POA^Tacr1→VTA projections elicited DA release in NAc, ChR2 (Fig. 7Q-T) or control AAV (Fig. S9N-P) was expressed in POA^Tacr1 neurons and GRAB_DA was expressed constitutively in NAc. Optic cables were placed dorsal to the VTA and NAc for activation and imaging, respectively. Pilot experiments were attempted to place optic cables over the POA and the NAc, but the two regions’ proximity to each other precluded this configuration. While recording GRAB_DA signal from NAc, laser illumination (473 nm, 40 Hz, 10 ms pulse width, 5 mW from the fiber tip) was provided to the VTA for 10 seconds. For GRAB_DA photometry during repeated activation of VTA (Fig. 7T), the same laser 10 s laser illumination was applied 6 times with 30 s intervals.

QUANTIFICATION AND STATISTICAL ANALYSIS

Calcium imaging data, behavioral video recordings, electrophysiological studies, and histological samples were conducted or analyzed blind to relevant variables, including identity of administered solutions, genotypes, light illumination conditions, surgical procedures, and viruses injected. Statistical analysis was performed using GraphPad PRISM (GraphPad Software). To compare categorical data including percentage of mice displaying a behavior, Fisher’s exact test was performed from a 2×2 contingency table. To compare non-categorical data, we first determined if the data values were from a normal distribution using the D’Agostino-Pearson omnibus normality test. In experiments with paired samples, we used a paired t test or repeated measures ANOVA for parametric data and a Wilcoxon matched-pairs signed rank test or Friedman test for non-parametric data. In all other experiments, we used a t test or ANOVA for parametric data and a Mann-Whitney or Kruskal-Wallis test for non-parametric data. All multiple comparisons were corrected for using Bonferroni corrections.

Supplementary Material

1. Figure S1: Male BNSTpr^Tac1 neurons are more active during interactions with females than males and forced activation of these cells promotes male-male mating. Related to Figure 1.

A-E. Schematic of fiber photometry of BNSTpr^Tac1 neurons in males following insertion of a receptive female, male, or toy mouse (A). PETP of normalized GCaMP6s fluorescence (ΔF/F) during insertion of mice or a toy mouse (B). BNSTpr^Tac1 neurons are activated by mice (C), with greater activation by females (D). Activation perdures longer following encounter with females than males (E).

F. GCaMP6s expression in BNSTpr^Tac1 neurons. FT, Fiber optic tract.

G-J. Schematic of fiber photometry of BNSTpr^Tac1 neurons during male sexual behavior (G). PETP of normalized GCaMP6s fluorescence during mounts (H) and ejaculation (I) shows significant increase in activity above baseline (J)

K-M. Schematic of fiber photometry of BNSTpr^Tac1 neurons in males attacking an intruder male (K). PETP of normalized GCaMP6s fluorescence during attacks (L) reveals no discernible activation of BNSTpr^Tac1 neurons (M).

N-O. GCaMP6s expression in BNSTpr^Tac1 neurons (N). GRIN, GRIN lens tract. More neurons are activated by female compared to male urine (O).

P. Laser illumination of ChR2+ BNSTpr^Tac1 neurons induces Fos expression.

Q-S. Optogenetic activation of ChR2+ BNSTpr^Tac1 neurons in males for the first 90s of an encounter with an intruder male eliminates attacks and induces mounts (Q,R) without altering locomotor activity (S) of the experimental male.

T-V. Schematic of optogenetic activation of BNSTpr^Tac1 neurons in males interacting with a receptive female (T). Optogenetic activation (30/30 s on/off ) does not alter mounting behavior (U). Raster plot of behavior of male mounting under this optogenetic activation paradigm (V).

Mean ± SEM. n = 6 mice (A-M), 105 neurons from 3 mice (N,O), and 5 mice (P-V). * p < 0.05, ** p < 0.01. Scale bar = 200 μm (F,N,P).

2. Figure S2: Activity of BNSTpr^Tac1→POA projections is necessary and sufficient for male sexual behavior. Related to Figure 2.

A. Laser illumination of ChR2+ BNSTpr^Tac1 neurons induces Fos expression.

B-C. Optogenetic activation of BNSTpr^Tac1→POA projections during first 90s of encounter with an intruder male eliminates attacks and induces mounting during the rest of the assay

D-H. Schematic of optogenetic activation of BNSTpr^Tac1→POA projections in males interacting with a receptive female (D). Activation of these projections during first 90s of encounter does not alter the probability (E), number (F), or latency (G) of mounting behaviors, nor does it alter locomotor activity of the male (H).

I-J. PPO expression in soma of BNSTpr^Tac1 neurons (I) and their projections to the POA (J).

K-N. Schematic for optogenetic inhibition of BNSTpr^Tac1→POA projections in males interacting with a receptive female (K). Inhibition reduces the number (L) and increases the latency (M) of mounting behavior without altering locomotor activity of the male (N).

O-S. Schematic for optogenetic inhibition of BNSTpr^Tac1→POA projections in males interacting with an intruder male (O). Inhibition does not alter the number (P), latency (Q), or mean duration (R) of attacks nor does it alter locomotor activity of the experimental male (S).

Mean ± SEM. n = 7 (A-H) and 9 (I-S) mice. * p < 0.05, ** p < 0.01. Scale bars = 200 μm (A,I,J).

3. Figure S3: Activity of male POA^Tacr1 neurons during mating is necessary and sufficient for the behavior. Related to Figures 3 and 4.

A-D. Schematic of coronal section through the adult mouse brain showing that POA^Tacr1 neurons are a subset of POA^ESR1 neurons. Co-expression of Tacr1 (visualized with AAV-DIO-GCaMP6s delivered to POA of Tacr1^Cre males) and Esr1 in the POA (B). Vast majority ofe POA^Tacr1 neurons express Esr1 (C) whereas a small subset of POA^Esr1 neurons express Tacr1 (D).

E. GCaMP6s expression in POA^Tacr1 neurons.

F-I. Schematic of fiber photometry of POA^Tacr1 neurons in males mating with a receptive female (F). PETP of normalized GCaMP6s fluorescence during intromission (G) and ejaculation (H) shows that these cells are active during intromission (I).

J. Laser illumination of ChR2+ POA^Tacr1 neurons induces Fos.

K-L. Optogenetic activation of POA^Tacr1 neurons in males mating with a receptive female induces males to mount significantly more during epochs with laser illumination (15/45 s on/off) (K) without altering their locomotor activity (L).

M. Optogenetic activation of POA^Tacr1 neurons in males interacting with an intruder male eliminates attacks and induces mounting behavior toward the intruder (laser illumination, 15/45 s on/off).

N-Q. Schematic of paradigm to test for ultrasonic vocalizations (USVs) upon laser illumination of POA^Tacr1 neurons expressing ChR2 (N) or mCherry (P). Laser illumination elicits USVs from mice expressing ChR2 (O) but not mCherry (Q) in their POA^Tacr1 neurons.

R. eNpHR3.0 expression in POA^Tacr1 neurons.

S-T. Optogenetic inhibition of POA^Tacr1 neurons in males interacting with a receptive female increases latency to initiate mating (S) without altering locomotor activity of the male (T).

U. Optogenetic inhibition of POA^Tacr1 neurons in males interacting with an intruder male does not alter locomotor activity of the experimental male.

Mean ± SEM. n = 5 (A-D), 8 (E-I), 6 (J-M), 4 (N-O), 3 (P-Q), 11 (S), and 5(T-U) mice. * p < 0.05, *** p < 0.001. Scale bars = 20 μm (B), 200 μm (E,J,O).

4. Figure S4: POA^Tacr1 neurons are functionally downstream of BNSTpr^Esr1 neurons. Related to Figure 4.

A-C. Strategy to activate BNSTpr^Esr1 neurons while inhibiting POA^Tacr1 neurons (A). DREADDi expression in POA^Tacr1 neurons (B). ChR2 expression in BNSTpr^Esr1 neurons (B). Laser illumination of BNSTpr^Esr1 neurons induces Fos expression.

D-E. CNO eliminates mounting behavior.

F-G. Optogenetic activation of ChR2+ BNSTpr^Esr1 neurons in males given vehicle (rather than CNO) interacting with an intruder male increases mounts (F) and latency to fight (G), and reduces latency to mate (G). BNSTpr^Esr1 neurons were optogenetically activated only for the first 90 s of the assay.

H-L. Chemogenetic inhibition of male POA^Tacr1 neurons does not alter number of (H), or latency to (I), attack, nor does it alter locomotor activity (J), motivation to find hidden food (K), or performance on an elevated plus maze (L).

M-P. Optogenetic activation of ChR2+ BNSTpr^Esr1 neurons in males given CNO and interacting with a receptive female (M-N) and or male intruder (O-P). CNO eliminates mounting behavior even when BNSTpr^Esr1 neurons have been activated (M-N). Optogenetic activation of BNSTpr^Esr1 neurons tends to reduce aggression but it cannot induce mating in the presence of CNO (O-P).

Q-S. Strategy to activate POA^Tacr1 neurons while inhibiting BNSTpr^Esr1 neurons (Q). DREADDi expression in BNSTpr^Esr1 neurons (R). Laser illumination of ChR2+ POA^Tacr1 neurons induces Fos expression (S).

T-V. Schematic of chemogenetic inhibition of BNSTpr^Esr1 neurons in males interacting with a receptive female (T). CNO reduces mounts (U) and increases latency to mount (V).

W-Y. Schematic of optogenetic activation of POA^Tacr1 neurons in males given CNO and interacting with a receptive female (W). Activation of POA^Tacr1 neurons overrides suppression of mating by inhibition of BNSTpr^Esr1 neurons such that it increases mounts number (X) and reduces latency to mount (Y).

Mean ± SEM. n = 6 (A-P) and 7 (Q-Y) mice. * p < 0.05, ** p < 0.01. Scale bar = 200 μm (B,C,R,S).

5. Figure S5: Substance P released from BNSTpr^Tac1 neurons potentiates excitatory LTP onto POA^Tacr1 neurons. Related to Figure 5.

A-D. Substance P does not alter activity of POA^Tacr1 neurons.

Strategy for recording from POA^Tacr1 neurons (A).

Representative traces from a POA^Tacr1 neuron recorded before (B) and after bath application of 1 μm Substance P (C).

Spontaneous firing rate of POA^Tacr1 neurons is not altered by Substance P (D).

E-H. Inhibition of POA^Tacr1 neurons by BNSTpr neurons is transient.

Strategy for recording from POA^Tacr1 neurons while activating ChR2+ BNSTpr projections in the POA in the absence of picrotoxin (E).

Representative trace from POA^Tacr1 neurons showing that their firing is inhibited by laser illumination (F).

Trace from outlined box in (F) expanded to show time-locked optically-evoked inhibitory postsynaptic potentials (G).

Spontaneous firing rate is reduced during laser illumination (laser-on) and recovers to baseline state thereafter (H).

I-K. BNSTpr neurons induce LTP onto POA^Tacr1 neurons in a Tacr1-contingent manner.

Strategy for recording POA^Tacr1 neurons while activating ChR2+ BNSTpr projections in the POA in the absence of picrotoxin (I).

Summary plot of normalized EPSP Amp and R_in in POA^Tacr1 neurons before and after 90 s laser illumination (blue arrow) of BNSTpr→POA projections (J-K). Insets: Representative average EPSP traces of the baseline (1) and the last 5 min (2). Optogenetic activation of BNSTpr projections in the POA induces LTP in POA^Tacr1 neurons (J) and L-703,606 blocks this LTP (K). Gray line (K) represents the average trace during experiments in which vehicle was used instead of L-703,606.

L-P. Encounter with a female does not alter excitatory presynaptic release probability onto POA^Tacr1 neurons.

Schematic of behavioral assay in which singly-housed males were administered vehicle or L-703,606 30 min prior to start of the behavioral assay (L-N, upper panels), and recording strategy for POA^Tacr1 neurons in which glutamatergic inputs are locally stimulated with a bipolar microelectrode in the presence of picrotoxin (O).

Representative traces of excitatory postsynaptic currents, recorded at −70 mV and evoked 50 ms apart, in POA^Tacr1 neurons from males from each of the three experimental conditions (L-N, lower panels).

No difference in paired-pulse ratio between groups (P).

Mean ± SEM. n = 13 neurons from 4 mice (A-D), n = 10 neurons from 3 mice (E-H), n = 8 neurons from 4 mice (J), 6 neurons from 3 mice (K), n = 27 neurons from 4 mice (Control), 30 neurons (5 mice in Encounter + Vehicle), 27 neurons from 4 mice (Encounter + L-703,606) (P). * p < 0.05, ** p < 0.01.

6. Fig. S6: Substance P-Tacr1 signaling in the BNSTpr^Tac1→POA^Tacr1 pathway regulates male sexual behavior. Related to Figure 5.

A. Dual cannula placement above the POA.

B-C. Raster plots of male mating displays 15 following vehicle (B) or Substance P (C) infusion into WT male POA shows that Substance P accelerates onset of mating.

D-E. Infusion of L-703,606 into WT male POA increases latency to start mating with a female (B) without altering sniffing (D) or locomotor activity (E).

F-J. Infusion of L-703,606 into WT male POA does not alter probability (F), number (G), latency (H), mean duration (I) of sniffing or attacks nor does it alter locomotor activity (J).

K-N. Tacr1 antagonist dominantly suppresses facilitation of male sexual behavior induced by optogenetic activation of BNSTpr→POA projections. Optogenetic activation of BNSTpr^Tac1 neurons cannot overcome suppression of male sexual behavior induced by inhibition of Tacr1 signaling in the POA with L-703,606 (K-L). Optogenetic activation of BNSTpr^Tac1 neurons cannot overcome suppression of male mating induced by inhibition of Tacr1 signaling in the POA with L-703,606 (M,N).

O-U. Optogenetic activation of POA^Tacr1 neurons dominantly drives male sexual behavior even upon supplementing males with L-703,606.

Strategy to optogenetically activate male POA^Tacr1 neurons with systemic provision of L-703,606 (O).

Activation of POA^Tacr1 neurons overrides L-703,606-induced inhibition of male sexual behavior, increasing probability of mating (P) and mount number (Q) while reducing latency to mate (R). Activation of POA^Tacr1 neurons elicits mating toward the intruder male (S) despite provision of L-703,606, with the experimental male showing more mounts (T) and reduced latency to mate (U).

Mean ± SEM. n = 7 (B-C), n = 8 (D-E), 7 (K-N), and 6 (O-U) mice. * p < 0.05, ** p < 0.01, *** p < 0.001. Scale bar = 200 μm (A).

7. Figure S7: Forced activation of POA^Tacr1 neurons overcomes the post-ejaculatory refractory period and is rewarding. Related to Figure 6.

A. Raster plot of mating behavior in all males tested during the post-ejaculatory refractory period. Optogenetic activation following ejaculation rejuvenates mating drive such that all males display mounting, intromission, and some also ejaculate again.

B-C. Schematic of laser illumination of mCherry+ POA^Tacr1 neurons in post-ejaculatory males interacting with a receptive female (B). Without ChR2, laser illumination does not drive mating in the post-ejaculatory period (C).

D-E. Schematic of laser illumination of mCherry+ POA^Tacr1 neurons in males during the SPP test (D). Without ChR2, laser illumination does not alter female preference index (E).

F-G. Schematic of self-stimulation (0.5 s, 40 Hz) of mCherry+ POA^Tacr1 neurons in males (F). Without ChR2, males nose poke equivalently infrequently between the active (delivering laser illumination) and inactive port (G).

H-J. Schematic of optogenetic inhibition of POA^Tacr1 neurons in males who performed the SPP and sucrose consumption tests (Fig. 6M-P) (H). Inhibition decreases the precent of males who perform mating routines (I) and the number of mounts and intromissions performed (J).

Mean ± SEM. n = 10 mice (A), n = 6 mice (B-C), n = 6 mice (D-E), n = 6 mice (F-G), n = 6 mice (H-J). * p < 0.05

8. Figure S8: Relation of the BNSTpr^Tac1-POA^Tacr1 pathway to sensory input, motivational centers, motor output, and reward. Related to Figure 7.

A-D. TH expression in the AVPV and VTA of 6OHDA or vehicle injected males (same mice as shown in Fig. 7A-F) (A).

Fewer TH neurons in AVPV and VTA in mice given 6OHDA (B).

6OHDA injected males mated and the probability of mating routines (C) or number of mating events were not altered (D) even upon optogenetic activation of ChR2+ POA^Tacr1 neurons.

E-I. BNSTpr^Tac1 neurons receive mono-synaptic input from AOB but not MOB.

TVA (mCherry) and rabies vector (EGFP) expression in BNSTpr^Tac1 starter neurons (E).

mCherry– and EGFP+ presynaptic neurons in the AOB (F,H).

No mCherry or EGFP-expressing neurons in the MOB (G,I).

J-N. POA^Tacr1 neurons do not receive mono-synaptic input from AOB or MOB. TVA (mCherry) and rabies vector (EGFP) expression in POA^Tacr1 starter neurons (J). No mCherry or EGFP-expressing neurons in AOB or MOB (K-N).

O-Q. Syp:mRuby expression in soma of POA^Tacr1 neurons (O). Syp:mRuby+ termini of POA^Tacr1 neurons in the VTA (P) and PAG (Q), with insets showing area outlined in gray at higher magnification.

Mean ± SEM. n = 9 control mice and 6 experimental mice (A-B), n = 6 mice (C-D), 6 mice (E-I), 4 mice (J-N), 3 mice (O-Q). Scale bars = 100 μm (A, E-G, J-L, O-Q) and 20 μm (insets in P, Q).

9. Figure S9: Projections of POA^Tacr1 neurons drive mating, self-reinforcement, and DA release. Related to Figure 7.

A-B. Schematic of optogenetic activation of VTA termini of POA^Tacr1 neurons in males interacting with an intruder male (A). Activation (15/45 s on/off) elicits mounting toward the intruder in a laser illumination-epoch linked manner (B).

C-D. Schematic of optogenetic activation of PAG termini of POA^Tacr1 neurons in males interacting with an intruder male (C). Activation (15/45 s on/off) elicits mounting toward the intruder in a laser illumination-epoch linked manner (D).

E-F. Schematic of optogenetic self-stimulation (0.5 s, 40 Hz) of POA^Tacr1 termini in PAG or VTA following inactivation of POA with TTX (E). Silencing the POA does not affect the number of nose pokes made to the active ports with activation of POA^Tacr1 neuronal termini in PAG or VTA (F).

G-J. Schematic of optogenetic activation of VTA or PAG termini of POA^Tacr1 neurons subsequent to inactivating the POA with TTX, in males interacting with a receptive female (G). Silencing the POA reduces percent males mating, but the activation of PAG or VTA termini of POA^Tacr1 neurons (15/45 s on/off) overrides POA silencing and elicits mating (H). Silencing the POA does not reduce laser illumination-epochs linked mounts elicited by activation of VTA (I) or PAG (J) termini of POA^Tacr1 neurons.

K-M. To test for DA release in NAc, GRAB_DA was constitutively expressed in this region (K), and ChR2 was expressed in a Cre-dependent manner in POA^Tacr1 neurons enabling visualization of their soma in the POA (L) and projections in the VTA (M).

N-P. DA release in NAc upon optogenetic activation of POA^Tacr1 neuronal projections in VTA is contingent on ChR2 expression in POA^Tacr1 neurons.

Strategy for fiber photometry imaging of DA release in NAc during laser stimulation of VTA projections of ChR2– and mCherry+ POA^Tacr1 neurons (N).

Laser illumination does not induce DA release in NAc (O, P).

Mean ± SEM. n = 7 mice (A-B), 11 mice (C-D), 7 mice each for PAG and VTA studies (E-F), 7 mice (G-J), 4 mice (K-P). Scale bar = 100 μm.

10. Table S1: Histological analyses for experimental mice. Related to Figures 1–7.

Viral expression cell counts for each experimental mouse listed per hemisphere with approximate stereotaxic locations of the tip of each optic fiber/cannula labeled on a corresponding coronal brain section¹¹¹. Optic fiber/cannula placements are color-coded for each mouse.

11. Table S3: Specificity of activating optogenetic or chemogenetic actuators and behavioral phenotypes. Related to Figures 1–7.

Laser illumination or CNO administration does not alter social behaviors in males expressing mCherry in BNSTpr^Tac1 or POA^Tacr1 neurons. Forced activation or inhibition of BNSTpr^Tac1 or POA^Tacr1 neurons has no effect on grooming behavior.

13

Movie S1: Optogenetic activation of POA^Tacr1 neurons elicits mating toward females, males, and inanimate objects that provide a releaser stimulus (mouse tail-like structure). Related to Figure 3.

14

Movie S2: Optogenetic activation of POA^Tacr1 neurons enables males to resume mating immediately following ejaculation, and mice will nose poke to receive optogenetic activation of their POA^Tacr1 neurons. Related to Figure 6.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Sheep anti-GFP	Biorad	Cat # 4745–1051; RRID: AB_619712
Rat anti-RFP	Chromotek	Cat # 5f8–100; RRID: AB_2336064
Rat anti-mCherry	ThermoFisher	Cat # M11217; RRID: AB_2536611
Rabbit anti-Fos	Synaptic Systems	Cat # 226 008; RRID: AB_2891278
Rabbit anti-Esr1	Millipore	Cat # 06–935; RRID: AB_310305
Rabbit anti-TH	Aves	Cat # TYH; RRID: AB_10013440
Donkey anti-sheep, Alexa 488 conjugate	Jackson ImmunoResearch	Cat # 713–545-147; RRID: AB_2340745
Donkey anti-rabbit, Alexa 488 conjugate	Jackson ImmunoResearch	Cat # 711–545-152; RRID: AB_2313584
Donkey anti-rat, Cy3 conjugate	Jackson ImmunoResearch	Cat # 712–165-150; RRID: AB_2340666
Donkey anti-rabbit, Cy3 conjugate	Jackson ImmunoResearch	Cat # 711–165-152; RRID: AB_2307443
Donkey anti-chicken, Cy5 conjugate	Jackson ImmunoResearch	Cat # 703–175-155; RRID: AB_2340365
Donkey anti-rabbit, Alexa 647 conjugate	Jackson ImmunoResearch	Cat # 711–605-152; RRID: AB_2492288
Bacterial and virus strains
Recombinant adeno-associated virus: AAV1-Syn-DIO-GCaMP6s	Addgene	Addgene number: 100845
Recombinant adeno-associated virus: AAV9-EF1α-DIO-PPO:Venus	Addgene	Addgene number: 139505
Recombinant adeno-associated virus: AAV1-EF1α-DIO-hChR2(H134R):EYFP	Addgene	Addgene number: 20298
Recombinant adeno-associated virus: AAV1-hSyn-hChR2(H134R):EYFP	Addgene	Addgene number: 26973
Recombinant adeno-associated virus: AAV1-CAG-DIO-EGFP	Addgene	Addgene number: 59331
Recombinant adeno-associated virus: AAV1-CAG-DIO-tdTomato	Addgene	Addgene number: 28306
Recombinant adeno-associated virus: AAV9-hSyn-GRAB_DA2m	Addgene	Addgene number: 140553
Recombinant adeno-associated virus: AAV2-EF1α-DIO-hChR2(H134R):mCherry	UNC Vector Core	Addgene number: 20297
Recombinant adeno-associated virus: AAV2-EF1α-DIO-eNpHR3.0:mCherry	UNC Vector Core	Addgene number: 26966
Recombinant adeno-associated virus: AAV2-EF1α-DIO-mCherry	UNC Vector Core	Addgene number: 50462
Recombinant adeno-associated virus: AAVDJ-EF1α-DIO-hM4DGi:mCherry	UNC Vector Core	Addgene number: 50461
Recombinant adeno-associated virus: AAVDJ-EF1α-fDIO-hM4DGi:mCherry	UNC Vector Core	Cat # AV7932
Recombinant adeno-associated virus: AAVDJ-EF1α-fDIO-hChR2(H134R):EYFP	UNC Vector Core	Addgene number: 55639
Recombinant adeno-associated virus: AAVDJ-hSyn-DIO-mGFP-2A-Synaptophysin:mRuby	Virovek	Addgene number: 71760
Recombinant adeno-associated virus: AAV8.2-TRE-DIO-TVA:mCherry-2A-oG	Virovek	Addgene number: 166602
EnvA G-deleted Rabies-EGFP	Salk Institute	Lot # 02282016
Cholera Toxin Subunit B, Alexa Fluor 555	ThermoFisher Scientific	Cat # C34776
Cholera Toxin Subunit B, Alexa Fluor 647	ThermoFisher Scientific	Cat # C34778
Chemicals, Peptides, and Recombinant Proteins
DAPI	Sigma-Aldrich	Cat # D9542; CAS # 28718-90-3
Clozapine-N-oxide (CNO)	Enzo Life Sciences	Cat # BML-NS105-0005
Estradiol	Sigma-Aldrich	Cat # E8875; CAS # 50-28-2
Progesterone	Sigma-Aldrich	Cat # P0130; CAS # 57-83-0
Substance P	Sigma-Aldrich	Cat # S6883; CAS # 137348-11-9
L-703,606	Sigma-Aldrich	Cat # L119; CAS # 351351-06-9
Picrotoxin	Sigma-Aldrich	Cat # P1675; CAS # 124-87-8
6OHDA	Sigma-Aldrich	Cat # H4381; CAS # 28094-15-7
Tetrodotoxin	Tocris	Cat # 1069; CAS # 18660-81-6
Experimental models: Organisms/Strains
Mouse: C57BL/6J	The Jackson Laboratory	Stock # 000664; RRID: IMSR_JAX:000664
Mouse: BALB/c	Charles River	Stock # 028
Mouse: Tac1Cre	The Jackson Laboratory	Stock # 021877; RRID: IMSR_JAX:021877
Mouse: Tacr1Cre	Diagle et al., 2018	N/A
Mouse: Tac1 null	The Jackson Laboratory	Stock # 004103; RRID: IMSR_JAX:004103
Mouse: Esr1Flpo	The Jackson Laboratory	Stock # 037009; RRID: IMSR_JAX:037009
Software and Algorithms
ImageJ	NIH	https://imagej.nih.gov/ij/index.html; RRID: SCR_003070
MATLAB	MathWorks	https://www.mathworks.com/products.html; RRID: SCR_001622
GraphPad Prism 6	GraphPad Software	https://www.graphpad.com/scientificsoftware/prism/; RRID: SCR_002798
Inscopix Data Processing Software	Inscopix Inc.	https://www.inscopix.com/software-analysisminiscope-imaging#software_idps

Highlights.

• A multisynaptic neural circuit that converts sensory input to male sexual behavior.

• An embedded molecular timer delays mating onset following mate-recognition.

• This neural circuit elicits dopamine release and governs male sexual drive and reward.

• This neural circuit governs libido and motor displays of male sexual behavior.

INCLUSION AND DIVERSITY

One or more of the authors of this paper self-identifies as an underrepresented ethnic minority in their field of research or within their geographical location, or living with a disability, or has received support from a program designed to increase minority representation in their field of research. While citing references scientifically relevant for this work, we also actively worked to promote gender balance in our reference list.