The C. elegans Male Exercises Directional Control during Mating through Cholinergic Regulation of Sex-Shared Command Interneurons

Background Mating behaviors in simple invertebrate model organisms represent tractable paradigms for understanding the neural bases of sex-specific behaviors, decision-making and sensorimotor integration. However, there are few examples where such neural circuits have been defined at high resolution or interrogated. Methodology/Principal Findings Here we exploit the simplicity of the nematode Caenorhabditis elegans to define the neural circuits underlying the male’s decision to initiate mating in response to contact with a mate. Mate contact is sensed by male-specific sensilla of the tail, the rays, which subsequently induce and guide a contact-based search of the hermaphrodite’s surface for the vulva (the vulva search). Atypically, search locomotion has a backward directional bias so its implementation requires overcoming an intrinsic bias for forward movement, set by activity of the sex-shared locomotory system. Using optogenetics, cell-specific ablation- and mutant behavioral analyses, we show that the male makes this shift by manipulating the activity of command cells within this sex-shared locomotory system. The rays control the command interneurons through the male-specific, decision-making interneuron PVY and its auxiliary cell PVX. Unlike many sex-shared pathways, PVY/PVX regulate the command cells via cholinergic, rather than glutamatergic transmission, a feature that likely contributes to response specificity and coordinates directional movement with other cholinergic-dependent motor behaviors of the mating sequence. PVY/PVX preferentially activate the backward, and not forward, command cells because of a bias in synaptic inputs and the distribution of key cholinergic receptors (encoded by the genes acr-18, acr-16 and unc-29) in favor of the backward command cells. Conclusion/Significance Our interrogation of male neural circuits reveals that a sex-specific response to the opposite sex is conferred by a male-specific pathway that renders subordinate, sex-shared motor programs responsive to mate cues. Circuit modifications of these types may make prominent contributions to natural variations in behavior that ultimately bring about speciation.


Introduction
Courtship and mating are among the most elaborate of behaviors displayed in the animal kingdom [1]. The ultimate goal of the male in these behaviors is to fertilize oocytes. However, achieving this goal often requires the execution of a behavioral sequence, progression through which depends on detection of appropriate conspecific cues and the expression of appropriate behavioral responses [2]. These sequences serve to ensure that copulation occurs between conspecifics and that unworthy or inappropriate partners are rejected. As such, the study of mating behavior can provide valuable insight into understanding not only the neural bases of sex-specific behaviors, but also mechanisms underlying decision-making, sensorimotor integration and the coordinated timing of sequenced motor behaviors. In many species, delineation of the underlying circuitry and its functional interrogation are hampered, however, by nervous system complexity and relative genetic intractability.
The nematode C. elegans is an attractive model for exploring the circuits controlling sex-specific behaviors. The two sexes, male and hermaphrodite, have comparatively simple nervous systems, consisting of only 302 and 383 neurons, respectively 294 of which are sex-shared (a.k.a. core neurons) [3][4][5]. The connectivity of the hermaphrodite nervous system has been described with single cell resolution [6,7] and, recently, an equally detailed map of the male posterior nervous system has been completed [3]. These wiring diagrams, combined with the amenability of C. elegans to genetic and optogenetic manipulation, enables the functional dissection of circuits supporting specific behaviors, with single cell resolution and in the context of a freely behaving animal (reviewed in Xu and Kim, 2011 [8]).
The C. elegans hermaphrodite is essentially a female that can make a limited supply of sperm for self-fertilization. The hermaphrodite can also fertilize her oocytes using sperm acquired through copulation with a male [9]. The C. elegans male mates the hermaphrodite using a goal-oriented behavioral sequence that is composed of sex-specific motor behaviors [10,11]. In this study, we explore the circuitry controlling the male's decision to initiate mating in response to physical contact with a mate. Prior to a mate encounter, the locomotory patterns of the male are similar to that of the hermaphrodite. He moves through his environment with a predominantly forward directional bias and his movement is powered by propagation of a dorsal-ventral sinusoidal wave along the body (Fig. 1A). If mate contact prompts the mating choice, the male initiates a contact-based search of the hermaphrodite's surface for the vulva region, the vulva search. External sensilla on the male tail, the male sensory rays, are responsible for inducing vulva search behavior and guiding its trajectory in response to hermaphrodite surface cues ( Fig. 1B-1C) [12]. Ray stimulation causes the male to abruptly cease forward movement and reverse onto the hermaphrodite, apposing his tail against her surface (Fig. 1B). This action brings the rays and the vulva-sensing sensilla of the tail (the hook and the post-cloacal sensilla -p.c.s.) into direct contact with the hermaphrodite cuticle. With his tail apposed, the male moves systematically over the hermaphrodite's surface with a backward directional bias so that his sensilla-laden tail leads the way (Fig. 1C). Both this extended bout of backward movement and tail apposition posture are male-specific motor behaviors. Their induction also sex-specifically alters propagation of the sinusoidal wave, such that it does not progress to the tail region and may be variably absent from the posterior half of the body during mating. The vulva search continues until the hook and p.c.s. sense the vulva area and induce the male to pause there and begin to prod for the vulva slit with his copulatory spicules [11,13,14]. The hermaphrodite locomotory system has been extensively studied and homologous cellular components are present in the male nervous system, raising the possibility that male movement during mating may be supported by this sex-shared circuitry [4,5,7]. However, the male nervous system also contains 89 additional neurons that are male-specific, mainly distributed through the tail and the ventral nerve cord [3][4][5]. The extent to which these male-specific cells contribute to locomotory control during mating is the question of current interest. In this study, we show that the rays regulate directional movement during the vulva search using a male-specific pathway that converges on the sexshared locomotory system, rendering it responsive to mate-derived sensory signals. Our data suggest that ray activation triggers the forward-to-backward directional switch by preferentially upregulating the activity of backward command interneurons in this system. The rays act through male-specific decision-making interneuron PVY and its auxiliary cell PVX. These interneurons define the convergence point for multiple sensory pathways, both sex-shared and male-specific, suggesting that PVY and PVX may be a pivotal site for integrating and prioritizing various sensory cues to generate the appropriate behavioral response.

Backward Locomotion during Mating Depends on Sexshared Command Interneurons
In the C. elegans hermaphrodite considerable progress has been made in understanding the cellular and molecular bases for directional control of movement. Laser ablation experiments have uncovered a small group of interneurons that play a prominent role in the choice between backward and forward locomotion [15][16][17][18]. These interneurons, termed command interneurons, can be divided into two functional classes. The backward command interneurons, Figure 1. Male mating behavior is characterized by distinct patterns of locomotion. A cartoon depicting the key changes observed in male movement and body posture that are triggered by mate contact. A. In the absence of mate contact, male locomotion resembles that of the hermaphrodite: the male moves with a forward locomotion bias and the sinusoidal body wave driving movement propagates along the full length of the body. B. Contact with a mate via the male tail induces contact response: the male presses his tail against the hermaphrodite surface and commences backward movement. C. Backward locomotion continues until the vulva is sensed, whereupon the male pauses and prods for the vulva slit opening with his copulatory spicules. The male sensory rays (shown in the inset for B), which sense hermaphrodite contact, are essential for the induction and maintenance of tail apposition and for directional control on the hermaphrodite surface. doi:10.1371/journal.pone.0060597.g001 AVA, AVD and AVE left/right (L/R) bilateral pairs, promote backward locomotion. The forward command interneurons, AVB and PVC L/R bilateral pairs, promote forward movement. The forward and backward command interneurons confer movement directionality by stimulating distinct pools of ventral nerve cord motor neurons ( Fig. 2A), dedicated to forward and backward movement, respectively [15,[19][20][21]. Motor neuron activity, in turn, regulates the sex-shared body wall muscles to produce the sinusoidal wave. The command interneurons are the postsynaptic targets of a number of sensory pathways that promote movement either towards or away from specific stimuli by preferentially stimulating forward or backward command cells, respectively [7,[22][23][24][25]. During spontaneous locomotion (locomotion in the absence of specific stimuli), the worm moves with a forward directional bias but occasionally moves backward (i.e. makes reversals) to change trajectory [26]. It is proposed that this forward locomotion bias is produced by an intrinsic imbalance of ventral nerve cord motor neuron activities favoring the forward motor pathway. This imbalance is established in part by gap junction connections [21] and in part by descending inputs from a second directional control pathway, the disinhibitory pathway [25].
The command interneurons and motor components of the locomotory system are also present in the male nervous system [3][4][5]. In the pre-anal ganglion (PAG) of the male tail, the command cell processes receive direct inputs from a number of male-specific interneurons that are postsynaptic targets of the ray neurons ( Fig. 2A). Although these male inputs target both backward and forward command interneurons, these connections are heavily weighted in favor of the backward command interneurons, in particular AVA. This suggests that the rays may promote reversal by preferentially activating AVA. To test this, we eliminated the AVA interneurons in males using laser-mediated cell-specific ablation then examined its impact on male mating behavior (Materials and Methods). AVA interneurons are born in the embryo, however, presynaptic male-specific neurons (born in the third larval stage (L3)) do not connect with AVA until the stage prior to adulthood and sexual maturation, L4. Therefore, we performed these AVA ablations on L4 males. Mock-ablated (Control) and AVA-ablated L4 males were allowed to mature into adults overnight in the absence of hermaphrodites, so they were virgin, 1-day old adults at the time of assay. All male populations examined in this study were isolated in this way to ensure that they were sexually inexperienced and of the same age as, anecdotally, we have observed that these factors significantly affect the probability of vulva search initiation. Mating behavior assays were performed as follows: a single mock-or cell-ablated virgin adult male was placed on a bacterial mating lawn with five virgin, 1-day old adult hermaphrodites. Genetically wild type hermaphrodites often dart or move during the male's mating attempts [27], making quantification of male locomotory behavior difficult. To reduce hermaphrodite movement during mating, we used unc-64 hermaphrodites as mating partners, which are sluggish due to a mutation in the C. elegans syntaxin gene [28]. Male behavior during mating was digitally recorded for 15 mins or until ejaculation occurred, which ever happened first [12]. We paid particular attention to three aspects of vulva search behavior that would indicate the efficiency of forward to backward switching: the frequency with which backward movement was induced by initial contact (a.k.a. contact response; Fig. 2B) and the strength and duration of the backward movement during the search (scanning speed and sustained tail contact, respectively; Fig. 2C, 2D).
Mock-ablated males typically responded on the first or second contact by placing their tail, ventral side down, against the hermaphrodite and commencing backing (Video S1). Conse-  Fig. 3. Sex-shared cells (pink); male-specific cells (blue). The backward command interneurons (only the left AVA neuron is shown) have cell bodies in the head and send a process along the ventral nerve cord (VNC) where they synapse with motor neurons required for locomotion (the distribution of cell bodies for only a single motor neuron class is shown). PVY and PVX (located in the male pre-anal ganglion) receive inputs from the ray sensory neurons and have outputs onto the command interneuron processes. B-D. The impact of cell-specific ablations on three aspects of vulva search behavior related to locomotory control. B. Contact response efficiency. C. Scanning speed. D. The number of times tail contact was lost per mating. The X-axis indicates the cells ablated in each treatment. A box plot representation of the data is shown, with median and mean values indicated by the line and the black dot within quently these control males had a median contact response success rate of 100% (i.e. 1 out of 1 contacts resulted in a complete contact response, namely both tail apposition and initiation of backward movement; Fig. 2B). Using these motor behaviors, control males then moved along the hermaphrodite's anterior-posterior axis, a behavior referred to as scanning (Fig. 1C). If males reached the hermaphrodite head or tail without sensing the vulva, they made a sharp turn, without losing tail contact, and scanned along the other side of the hermaphrodite, again moving with backward locomotion but in the opposite direction. We calculated the scanning speed of control males (in mm/sec) by dividing the length of the hermaphrodite by the time required for a male to scan the length (see Materials and Methods for details). As encountering the vulva can induce pausing (and consequently affect scanning time), scanning speed measurements were obtained from only the nonvulva side. Control males scanned a hermaphrodite length at a median speed of 180 mm/sec (Fig. 2C). Control males typically maintained tail contact with the hermaphrodite for the entire search ( Fig. 2D; median number of lost contacts during scanning = 1).
Ablation of AVA caused locomotion defects in males that were apparent even when males were not engaged in mating. These defects resembled those reported for AVA-ablated hermaphrodites [15]. Specifically, during spontaneous locomotion, AVA-ablated males had difficulty switching from forward to backward movement and consequently reversed only a short distance before reverting to forward movement. Backward movement defects were also evident in vulva search behavior. AVA-ablated males exhibited highly variable contact response efficiency (-AVA treatment, Fig. 2B). When scanning was initiated successfully, the scanning speed of AVA-ablated males was significantly slower than that of control males (-AVA treatment: median scanning speed = 140 mm/sec cf. control males = 180 mm/sec; Fig. 2C). In addition, AVA-ablated males lost tail contact during scanning at a significantly higher rate than controls, possibly because the uncoordinated phenotype resulting from AVA ablation interferes with tail apposition (-AVA treatment: median number of lost contacts = 5 cf. control males = 1; Fig. 2D). These results suggest that the rays depend significantly on the sex-shared AVA command interneurons to induce backward movement during the vulva search. However, our observation that AVA-ablated males can exhibit limited and stochastic backward movement during mating also reveals that the rays may additionally act through other interneurons involved in directional control. The stochastic nature of ablated-male defects suggests that, in some instances, activity of these alternative pathways is sufficient to drive backward movement, whereas at other moments it is not.

Male-specific Interneuron PVY Promotes Backward Locomotion during Mating
We next investigated the neural pathways by which the rays control AVA. In the male PAG, two major ray neuron targets, the male-specific interneurons PVY and PVX, have significant inputs onto command interneuron processes ( Fig. 2A; [3]). These inputs are biased in favor of the AVA backward command interneurons such that the ratio of synaptic inputs onto AVA, relative to the AVB forward command interneurons is 2:1 [3]. To test whether PVY and/or PVX represent the major route by which the rays control the command cells, we ablated one or both male-specific interneurons and assessed the impact on mating behavior. In males where only PVY was ablated, contact response efficiency and scanning speed were significantly reduced compared to control males (-PVY treatment: median % contact response = 30%, Fig. 2B; median scanning speed 135 mm/sec, Fig. 2C; Video S2). PVY-ablated males also lost tail contact at a significantly higher frequency than control males during the search (median number of contacts lost = 5; Fig. 2D). These various locomotion defects and their stochastic expression resembled the behavioral phenotype of AVA-ablated males, consistent with the model that PVY promotes backwards movement by acting through AVA neurons. However, in contrast to AVA-ablated males, PVY-ablated animals had no obvious defects in reversal when not engaged in mating (data not shown). Thus, PVY is only required for locomotory control in response to ray-mediated mate contact. Although PVX also has inputs onto AVA, its ablation had no obvious affect on vulva search behavior. Males lacking PVX had superficially wild type mating behavior (-PVX treatment, Fig. 2B-2D) and males lacking both PVX and PVY (-PVY-PVX treatment, Fig. 2B-2D) were not significantly different from males lacking PVY only. Together, these ablation experiments reveal that PVY promotes backward movement during the vulva search, while PVX may have a functionally redundant or subtle role in the process. Like the AVA-ablated males, males lacking PVY could still reverse along the hermaphrodite, albeit erratically and slowly. Thus, although PVY and AVA are important for backward moment during mating, other ray neuron targets may be able to partially compensate for their absence.

Optogenetic Manipulation of PVY and PVX Activity Affects Male Movement
To gain further insight into the functional properties of PVY and PVX, we artificially activated these interneurons in solitary males using the heterologous light-inducible cation channel ChannelRhodopsin-2 (ChR2) from Chlamydomonas reinhardtii [29]. To target expression of ChR2 to PVY and PVX, we placed a YFPtagged ChR2 transgene (ChR2-YFP) under the control of the nlp-14 promoter (Pnlp-14; [30]). Previous studies have shown that this particular nlp-14 promoter sequence drives expression of GFP reporters to a subset of nervous system cells common to both sexes: the sensory neurons ASI, ASK, ASE, PHA, two retrovesicular ganglion neurons, ventral nerve cord motor neurons and the interneuron PVT [30]. We find that in the male Pnlp-14::GFP reporters are additionally expressed in male-specific PVY, PVX and two male-specific dorsal rectal ganglion cells. However, we observed that the ChR2-YFP transgene, when placed under the control of this same nlp-14 promoter region, is expressed in only a fraction of these cells, being notably absent from the ventral nerve cord motor neurons, PHA and often the male dorsal rectal ganglion neurons. The limited expression of CHR2-YFP, compared to GFP, may be because the former lacks introns and consequently may be less efficiently processed in some cell types.
Retinal is an essential co-factor for ChR2 and as C. elegans does not endogenously synthesize this compound, it must be provided in their food in the form of all-trans-retinal (ATR) in order for the animal to generate functional ChR2 (Materials and Methods). To perform the artificial activation assays, 1-day old adult, virgin Pnlp-14::ChR2-YFP males were individually exposed to a 500 msec flash of blue light (470/40 nm wavelength) while moving forward. Since blue light is an aversive stimulus for C. elegans and can alone elicit an avoidance response (either forward or backward movement away from the source), these transgenic animals also carried a mutation in the lite-1 gene (lite-1(ce314)), which encodes a putative the box, respectively. Comparisons to the control (mock-ablated) were made using a ranksum test. The number of males assayed for each treatment, n: Control = 63; -PVY-PVX = 15; -PVY = 19; -PVX = 8; -AVA = 22. Significance, *p,0.05; **p,0.005. doi:10.1371/journal.pone.0060597.g002 ultraviolet receptor [31]. Animals carrying lite-1(ce314) are substantially less avoidant of blue light. The behavior of transgenic animals was digitally recorded during the assay and the distance they traveled (either forward or backward) from the time of the flash was determined by analyzing video frames (as described in Materials and Methods). When exposed to the light flash, 90% of Pnlp-14::ChR2-YFP males immediately reversed and traveled a median distance of -5 mm, with the negative value indicating that they reversed away from their position at the time of the flash (M treatment, Fig. 3A, Video S3). Through a series of control experiments we determined that this reversal response was specifically due to activation of male-specific interneurons PVY and PVX. First, this robust reversal behavior was not a nonspecific response to the light flash, as the substantial majority of males expressing non-functional ChR2 did not alter their behavior and continued moving forward (-ATR, M treatment, Fig. 3A, Video S4). Second, Pnlp-14::ChR2-YFP transgenic hermaphrodites, which lack PVY and PVX, did not respond to the light flash (H treatment, Fig. 3A). Moreover, Pnlp-14::ChR2-YFP males lacking PVY and PVX cells, due to their specific ablation with the laser, did not reverse in response to the light flash and continued moving forward (-PVY-PVX treatment, Fig. 3B). These experiments demonstrate that artificial activation of PVY and PVX is sufficient to overcome the forward directional bias and induce backward movement.
We next used the ChR2 artificial activation system to further explore the relative contribution of PVY and PVX to male reversal behavior. In the mating assays discussed above, eliminating PVY alone significantly disrupted backward locomotion during the vulva search ( Fig. 2B-2D). Consistent with this, few Pnlp-14::ChR2-YFP males ablated for PVY reversed when exposed to the light flash (16%, -PVY treatment, Fig. 3B), even though PVX was present and presumably activated. While activation of PVX alone had little impact, PVX co-stimulation with PVY enhanced the robustness of the reversal response, as males with only PVY (-PVX) showed greater variability in their response. Thus, while these experiments confirm that PVY is a crucial effecter of backward locomotion, they also reveal a function for PVX as an auxiliary cell, a role that may be readily compensated for by other ray targets in the context of mating if PVX is absent (Fig. 2).
The ChR2 experiments above showed that PVY+PVX activation is sufficient to induce the switch from forward to backward locomotion. We next asked whether PVY and PVX activity were required continuously during the vulva search to drive backward movement. To test this, we used the heterologous light-inducible hyperpolarizing channel NpHR, from Natronomonas pharaonis [32,33], to inhibit PVY and PVX while the male was performing the search. These experiments were performed on lite-1(ce314) transgenic males in which NpHR expression was targeted to PVY and PVX by placing an NpHR-EYFP transgene under the control of the nlp-14 promoter. Under the control of this promoter, NpHR-YFP is expressed in the same constellation of cells as ChR2-YFP (described above). Like ChR2, NpHR function depends on the presence of ATR so males grown in the absence of ATR provide a convenient negative control for any non-specific effect of light exposure. In each mating assay, a single virgin 1-day old adult transgenic male was placed with 5 virgin adult unc-64; lite-1 hermaphrodites on a mating lawn and allowed to initiate mating. We activated NpHR during scanning by exposing the male to a 500 msec flash of yellow light (540/25 nm wavelength) at 5, evenly spaced, time intervals and counted the number of times pausing coincided with a pulse (Fig. 3C). Pnlp-14::NpHR-EYFP males, grown in the presence of ATR, paused in response to 4 out of 5 flashes, on average (+ ATR+mating treatment: mean pausing frequency = 80%, Video S5). In contrast, Pnlp-14::NpHR-EYFP control males, grown in the absence of ATR, largely ignored the light pulses and continued scanning, pausing at a mean frequency of only 1% (-ATR+mating treatment, Fig. 3C, Video S6). These data therefore argue that PVY and PVX function is required continuously during the vulva search to maintain backward movement. To test whether NpHR activation could induce pausing when males were not engaged in mating, we flashed Pnlp-14::NpHR-EYFP males (grown in the presence of ATR) during spontaneous locomotion while they were reversing. These males rarely paused suggesting that the pausing induced during mating was due to hyperpolarization of cells that are normally active only when the rays sense mate contact (+ATR+no mating treatment: mean pausing frequency = 10% Fig. 3C). The most likely candidates for these cells are PVY and PVX. Taken together, these optogenetic data indicate that PVY promotes and sustains backward movement in the context of mating, and that PVX functions as an auxiliary cell that is largely redundant in the circumscribed mating assays used here with Unc hermaphrodites, but which under certain circumstances might enhance PVYinduced motor output.

PVY+PVX-induced Reversal is Dependent on Activity of AVA
Our cell ablation experiments show that AVA is required for backward movement during the vulva search. If PVY and PVX act through AVA, as the wiring diagram implies [3], then ablation of AVA should block PVY+PVX-induced reversals in our ChR2 assay system. We observed that this was indeed the case. When exposed to a 500 msec flash of blue light, Pnlp-14::ChR2-YFP AVA-ablated males were severely impaired in reversal behavior with a significant proportion either pausing, unable to reverse (25%), or reversing a median distance of 21 mm (56%) (Fig. 3B). Thus, PVY+PVX-induced reversal is dependent on AVA function.
AVA interneurons stimulate backward movement by activating the VA and DA backward motor neurons of the ventral nerve cord [15,20,21]. Consistent with this, we observed that males defective in VA neuron connectivity (unc-4 mutants; [34]) were defective in PVY+PVX-induced reversal response and, like AVA-ablated animals, a significant proportion of unc-4 males either paused (41%) or backed less than a micrometer (27%) (-VA treatment, Fig. 3B). Thus, PVY+PVX-induced backward movement is ultimately dependent on sex-shared backward motor neurons of the ventral cord. Taken together, these experiments delineate a simple neural pathway that consists of the sensory ray neurons, PVY and PVX, AVA and the backward motor neurons.

PVY+PVX-induced Reversal Depends on Cholinergic Neurotransmission
PVY and PVX express cholinergic markers (Fig. S1A), suggesting that these male-specific interneurons stimulate the command cells using cholinergic transmission. Consistent with this possibility, the command interneurons express several genes encoding cationic acetylcholine receptor (acr) subunits. acr-15 and acr-16 encode alpha nicotinic acetylcholine receptor subunits (nAChRs). acr-15 is expressed in both AVA and AVB, while acr-16 is expressed only in AVA [35,36]. We observed that acr-18, which encodes a DEG-3 type acetylcholine receptor, is also expressed in AVA, as well as unc (uncoordinated)-29, which encodes a nonalpha receptor subunit (Fig. S1B, S1C). To test whether any of the cholinergic receptors encoded by these genes are required for PVY+PVX-induced reversal behavior, we artificially activated PVY+PVX in single mutants and in various double and triple mutant combinations of these receptor genes. Mutations in three of these genes (acr-18, unc-29 and acr-16) disrupted reversal behavior and are discussed below (for other mutant combinations tested see Fig. S2).
In acr-18 null mutants, only 28% of males reverse in response to PVY+PVX activation (Fig. 4). The remainder either continued to move forward (56%) or paused (16%). acr-18 reporter expression is not limited to AVA. In both sexes, acr-18 expression is apparent in a number of sex-shared head and ventral nerve cord neurons and, in the male, in a subset of ray and pre-anal ganglion neurons (Fig.  S1B) [37]. To determine whether the absence of acr-18 function specifically in the backward command cells was responsible for the mutant reversal defect, we targeted expression of a wild type acr-18 transgene to AVA in acr-18 mutant males and asked whether this was sufficient to rescue the reversal defect. To this end, we placed a promoterless fragment of the wild type acr-18 gene (a cDNA::genomic hybrid gene and its 39 UTR) under the control of the nmr-1 gene promoter, which is expressed in backward command cells AVA, AVD, AVE and in AVG, RIM and PVC ( [38]; Materials and Methods). The Pnmr-1(AVA)::acr-18(+) transgene was co-injected with a Pflp-18::mCherry construct. flp-18 is expressed in AVA and we used the presence of mCHERRY in AVA as a proxy for the presence of the Pnmr-1(AVA)::acr-18(+) transgene because in C. elegans, transgenic arrays typically contain copies of all injected constructs. We observed that the presence of the acr-18(+) transgene in the nmr-1-expressing cells was sufficient to rescue the acr-18 mutant defect: 76% of acr-18 males carrying the rescuing transgene in AVA reversed in response to PVY+PVX activation, whereas only 28% of acr-18 control males, which lacked this transgene, reversed (Fig. 4). These data suggest that PVY+PVX-induced reversal depends on an ACR-18-containing cholinergic receptor that functions in nmr-1-expressing cells, most likely the backward command cells as these are direct targets of cholinergic PVY and PVX.
While the absence of acr-18 function in AVA significantly impairs reversal behavior in our ChR2 assays, acr-18 male populations can still reverse at an appreciable frequency (28%). This suggests that other receptors must be involved in mediating transmission. Loss of acr-16 function alone, or in combination with acr-18 mutations, caused a modest, but not significant, decrease in reversal frequency, suggesting that acr-16 contributes to reversal response but is not pivotal (Fig. 4). acr-16 has been shown to function redundantly with channels containing the non-alpha subunit encoded by unc-29 in two other tissues in the worm: in the body wall muscles, required for locomotion, and in male-specific muscles that facilitate spicule insertion during copulation [37,39]. During our analyses of unc-29; acr-16 spicule insertion behavior, we noticed that double mutant males exhibit reduced scanning speed during the vulva search, a phenotype also exhibited by PVYablated males (Fig. 2C) [37,39]. This prompted us to test whether unc-29 and acr-16 might also have partially redundant roles in mediating PVY+PVX-induced reversal behavior. The absence of both unc-29 and acr-16 function in body wall muscle renders double mutants severely paralyzed [39][40][41], a defect that precludes mating behavior analyses. However, this locomotion defect can be rescued by expressing an unc-29 wild type cDNA (unc-29(+)) specifically in muscles, using the acr-8 promoter, which is additionally expressed in neurons of the ventral nerve cord [37,41]. We therefore performed all male behavioral assays involving unc-29 mutant alleles with strains carrying an integrated version of the muscle-function rescuing transgene, rgIs1[Pacr-8::unc-29cDNA::SL2::GFP] [37,41]. In this transgenic array, unc-29(+) is expressed from a polycistronic message that additionally encodes GFP. This provides a convenient means of visualizing unc-29(+)-expressing cells as tissues that express GFP should also express unc-29(+). In the PVY+PVX artificial activation assays, the muscle-rescued unc-29; acr-16 males showed impaired reversal response with only 31% of males reversing (Fig. 4). Thus, unc-29 loss of function significantly enhances the severity of reversal defects caused by loss of acr-16 function. As mentioned above, an unc-29 full-length translational reporter is co-expressed with acr-16 in AVA and in RIB (Fig. S1C). To test whether the absence of unc-29 function from AVA might be responsible for the reversal defect of unc-29; acr-16 males, we performed tissue-specific rescue experiments, targeting expression of unc-29(+) to AVA. This was achieved using an unc-29(+)::SL2::GFP transgene placed under the control of the Pnmr-1 promoter. In ChR2-mediated PVY+PVX activation assays, unc-29; acr-16 males expressing unc-29(+) in AVA (indicated by GFP expression in AVA) were rescued, reversing at frequencies comparable to acr-16 single mutants and wild type males (83%) (Fig. 4). These results suggest that the site of action for unc-29 in mediating this behavior may be the backward command cells. Given that unc-29 is partially redundant with acr-16 in cells were the two genes are co-expressed [37,39], AVA may also be the site of action acr-16 in this behavior. As the magnitude of reversals in unc-29(+)-rescued double mutants was consistently less than that of acr-16 single mutant males, unc-29(+) may be additionally required in cells downstream of the command cells to enhance the magnitude of backward movement once initiated.
The data above suggest that unc-29, acr-16 and acr-18 act in AVA to mediate PVY/PVX-to-AVA cholinergic transmission but are partially redundant. If so, then animals carrying mutations in all three receptor genes should be severely defective in reversal when PVY+PVX are artificially activated. We observed that this was in fact the case: only 4% of unc-29; acr-16 acr-18 triple mutant males reversed in response to PVY+PVX activation with ChR2 (Fig. 4, Video S7). This mutant phenotype, together with the tissue-specific rescue data, support a model in which PVY/PVXto-AVA transmission is mediated by three partially redundant AChR subunits.

Cholinergic Receptor Mutants Exhibit Defects in Locomotion during the Vulva Search
We next asked whether disruption of PVY/PVX circuit function in AChR mutants affected search behavior in the context of mating. Surprisingly, we found that although more than 50% of acr-18 single mutants are defective in reversal behavior in the ChR2 assays, acr-18 males were superficially wild type for backward locomotion in the context of mating (Fig. 5). unc-29; acr-16 double mutants also performed this behavior reasonably well although showed reduced scanning speed as previously reported [37] (median scanning speed = 150 mm/sec cf. wild type control male median = 180 mm/sec, Fig. 5). The fact that unc-29; acr-16 mutants are affected in vulva search locomotion speed, and acr-18 mutants are not, suggests that unc-29 and acr-16 encoded channels may be more critical in determining locomotory speed than the acr-18 encoded channel. The fact that neither acr-18 single-nor unc-29; acr-16 double mutants resembled PVY-ablated males (Fig. 2) suggests that the male nervous system can compensate for the partial disruption of PVY-dependent transmission, possibly through upregulation of the remaining functional receptors in these respective mutant backgrounds. Consistent with this possibility, males mutant for all three receptor genes (unc-29; acr-16 acr-18 triple mutants) were significantly affected in all aspects of vulva search locomotion that we measured ( Fig. 5; Video S8). The triple mutant defects were, in fact, similar to those displayed by males lacking PVY and PVX (Fig. 2). unc-29; acr-16 acr-18 mutants had a median contact response efficiency of 50% (cf. 50% for the -PVY-PVX treatment, Fig. 2), a median scanning speed of 140 mm/sec (cf. 110 mm/sec for the -PVY-PVX treatment) and a median loss of contact frequency of 2 (cf. 8 for the -PVY-PVX treatment) (Fig. 2, Fig. 5). The similarity of the triple mutant to cell-ablated males is consistent with the ChR2 assay data, which argues that eliminating all three receptors essentially blocks PVY+PVX-induced backward movement. Like -PVY-PVX males, the triple mutant can still reverse in the context of mating, albeit inefficiently, suggesting that other ray-targeted pathways can partially compensate for the absence of PVY/PVX pathway activity and promote backward movement (see model in Fig. 6).

Discussion
How animals integrate and prioritize external cues and select the most adaptive behavioral response is a broadly relevant question. Sensory control of directional movement in C. elegans provides a tractable system for exploring this issue and has been the subject of intense investigation in the hermaphrodite. Here we examine this issue in the context of a sex-specific behavior and ask: how do mate contact cues override the forward locomotion bias of exploratory behavior and induce the male to pursue sex? Mate contact, sensed by the rays, causes the male to abruptly switch from forward to backward movement in order to conduct a systematic, contact-based search for the vulva. Our data suggest a simple circuit model to explain how ray stimulation by mate contact overrides a bias for forward movement, set by intrinsic activity of the sex-shared locomotory system or induced by external cues that promote exploration. In this model, ray neuron stimulation preferentially upregulates the activity of the backward command cells and consequently resets the activity patterns of locomotory motor neurons in favor of backward movement (Fig. 6). Ray control of the sex-shared command interneurons is primarily mediated via the male-specific PVY interneuron. Both the hard wiring of this ray-to-command cell pathway and the distribution of key molecular components within it would be predicted to induce a high probability of reversal in response to mate contact. First, the wiring diagram argues that the rays, PVY/PVX and the command cells form a feedforward pathway, as there are virtually no reciprocal connections between the rays and PVY/PVX or between PVY/PVX and AVA. Second, although PVY and PVX have inputs into both forward and backward command cells, this connectivity is biased in favor of AVA. Third, the key receptors that mediate cholinergic neurotransmission in this pathway (encoded by acr-18, unc-29 and acr-16) appear to be preferentially expressed on the backward command cells.
Interestingly, we find that removal of PVY, PVX and even AVA does not completely abolish reversal in the context of mating, indicating that although these cells are important for this behavior, other ray neuron targets can partially compensate for their absence. The interneuron ''X'' in Fig. 6 denotes this hypothetical, alternative pathway. The ray neurons make connections with several sex-shared interneurons that have inputs onto the command cells and could potentially fulfill this role (PVN, AVH, AVJ and AVF) [3][4][5]. It is also possible that this putative pathway corresponds to the disinhibitory pathway, a second directional control system that operates in parallel with the command cells in certain behavioral contexts [25]. The use of parallel pathways would be predicted to confer behavioral robustness and, if these pathways have subtly different roles, behavioral acuity. In the mating assays performed here we used Unc mating partners, which are relatively docile. However, wild type hermaphrodites are less compliant and dart or continue moving during the search [27], making it necessary for the male to adjust his speed and direction in order to maintain tail contact. In this context, robust and accurate control of movement is essential for mating success.
By several criteria, PVY corresponds to a decision-making neuron of the male nervous system. Its artificial activation in the absence of upstream ray stimulation is sufficient to induce backward locomotion in solitary males. Conversely, its hyperpolarization or elimination significantly impairs movement on the hermaphrodite in the context of mating. Moreover, PVY and PVX receive inputs from several tail sensilla [3], both male-specific (the rays, the hook and the p.c.s.) and sex-shared (the phasmids), consistent with the possibility that PVY (and PVX) represent a site of sensory stimulus integration and that the outcome of this integration determines whether contact-induced reversal is initiated, maintained or terminated. Recent studies show that phasmid and other sex-shared sensory neurons contribute to matesearching behavior, which is conducted with a forward directional bias [42]. Potentially, the substantial inputs from these sensory cells onto PVX may serve to suppress PVY+PVX-dependent reversal until a mate is located. During mating, detection of the vulva by the male hook and p.c.s. induces the male tail to pause at the vulva and remain there during ensuing copulatory attempts [11]. A simple model would be that the hook neurons induce pausing by inhibiting PVY and PVX activity. This would result in suppression of backward movement without affecting tail apposition. Consistent with this possibility, hook neurons have inputs onto PVY and PVX [3]. However, these inputs are surprisingly sparse compared to those made by hook neurons with their other targets. These latter postsynaptic targets include several neurons with inputs onto the command interneurons or ventral nerve cord motor neurons (e.g., sex-shared AVG and the male-specific EF and PVZ interneurons). Whether the hook induces pausing by inhibiting PVY/PVX directly or indirectly is therefore an open question. Future studies, examining the dynamics of PVY, PVX and command cell activity in these various behavioral contexts should provide insight into how converging sensory pathways alter directional movement through their action on PVY and PVX.
PVY+PVX-induced reversal depends on cholinergic transmission in AVA that is mediated by partially redundant AChRs that contain acr-18-, unc-29and acr-16-encoded subunits. The existence of AChRs in the command cells was initially revealed in studies examining the impact of exogenous nicotine on worm behavior [35]. However, to our knowledge, our study is the first to demonstrate a role for command cell cholinergic transmission in a natural C. elegans behavior. While the command cells have a few inputs from sex-shared cholinergic cells (e.g., the PVC forward command interneurons, the SDQ interneurons of the oxygen sensing pathway and tail interneuron DVC [43,44]), many inputs are from glutamatergic neurons and these regulate the command cells through cationic glutamatergic receptors [25,[45][46][47]. The use of cholinergic transmission in the male may therefore contribute to the specificity of command system control by the male nervous system. It could also serve to coordinate reversal with other motor outputs associated with mating that depend on cholinergic signaling, such as tail apposition and spicule prodding behavior [14,48]. The rays induce tail apposition in part through cholinergic stimulation of tail muscles [48,49] and, during vulva penetration attempts, cholinergic neurons of the p.c.s. may further potentiate ray action to cup the tail over the vulva [37]. Tail apposition posture depends on sex-shared muscles that are normally engaged in wave propagation. Therefore, the absence of wave propagation in the tail during mating may be simply due to the fact these muscles are otherwise engaged in tail apposition, as a consequence of ray cholinergic transmission. The predominance of cholinergic signaling in mating behavior circuits may also relate to the rapid speed with which the male must respond to hermaphrodite surface cues, as cholinergic transmission also features in rapidly executed behaviors in other invertebrate systems, such as escape response in Drosophila, crickets and snails [50][51][52]. In these latter systems, the sensory neurons are cholinergic and act on giant fibers. Giant fibers, in turn, are electrically coupled to their motor neuron targets, a design feature that likely contributes to rapid response. The C. elegans command interneurons could be considered analogous to giant fibers. Like giant fibers, their processes extend long distances and regulate an extensive population of motor neurons, notably through gap junction connections [7].
Although behavioral sexual dimorphism is widely observed in the animal kingdom, identifying their neural and molecular bases in many systems is hindered by genomic intractability and nervous system complexity. Functional dissection of sexual behaviors in simple, invertebrate models has provided some insights into where and how pivotal sex differences in nervous system function are Color intensity indicates a cell activity state (intense color = high; weak color = low). The arrows indicate the positive action of an activated cell on its postsynaptic target. A. When not engaged in mating, the male moves with a forward directional bias due to high levels of activity in the forward pathway of the sex-shared locomotory system: forward command interneurons AVB and their motor neuron targets (FWD mns). This activity bias is conferred by the default state of this system [21], in the absence of specific cues, or by external or internal sensory cues that drive the male to explore his environment. B. Ray neuron stimulation by mate contact activates PVY and PVX, which release acetylcholine (ACh), preferentially stimulating the backward (AVA), and not forward, command interneurons. This is due to a bias in synaptic inputs and in cholinergic receptor expression (ACR-18, ACR-16 and UNC-29) in favor of the backward command cells. AVA activation in turn stimulates the sex-shared backward motor neurons (BK mns). The rays may also promote backward movement through an as yet uncharacterized PVY/PVX-independent pathway (represented as interneuron ''X''). doi:10.1371/journal.pone.0060597.g006 encoded and to what extent cells common to both sexes are employed. In C. elegans, sex-shared cells contribute to several malespecific behaviors. For example, the sex-shared DVA neuron is the source of oxytocin/vasopressin-like peptides that enhance the robustness and coherence of mate-searching and mating subbehaviors by priming activity of the underlying circuitry [53]. Sexshared sensory pathways also provide critical input for matesearching behavior, mate attraction, male-specific odor preference and aspects of vulva search behavior [42,[54][55][56][57][58]. Although sexshared sensory pathways are required for mate attraction and male odorant preference, sexually dimorphic properties of these neurons confer the male-specificity of the response [54,57]. In contrast, we find that a sensory pathway composed of strictly malespecific cell types is primarily responsible for inducing malespecific vulva search behavior. The presence of this ray-PVY/ PVX pathway in the male renders the sex-shared locomotory system responsive to genital contact with a hermaphrodite. It does so by targeting key control centers in the sex-shared locomotory system, namely AVA. In the C. elegans hermaphrodite a similar circuit design underlies the locomotory patterns associated with egg-laying. In this case, the Hermaphrodite-Specific motor Neurons (HSNs) of the egg-laying system cause a burst of forward acceleration to coincide with egg release via direct regulation of sex-shared locomotory system interneurons [59]. Similarly, in Drosophila the reason males produce a courtship song using wing vibrations, and females do not, is primarily due to male-specific modulation of sex-shared central pattern generators for wing vibration [60][61][62]. These various examples, along with our own interrogation of the PVY/PVX pathway, provide insight into how new reproductive behaviors might evolve through relatively simple modifications to shared circuitry. Circuit modifications of these types may make prominent contributions to natural variations in behavior that ultimately bring about speciation.  [37]. All strains used in this study carried the him-5(e1490) mutation which generates a high incidence of males [66]. pha-1(e2123ts) was grown at 15uC. All other strains were maintained at 20uC and cultured as per Brenner (1974) [9]. Transgenic lines were generated using standard microinjection techniques [67]. Constructs were injected at the following concentrations: Pnlp-14::ChR2-YFP (100 ng/mL); Pflp-18::mCherry (40 ng/mL), Pacr-16::mCherry (100 ng/mL); Pnmr-1(AVA)::acr-18(+) (50 ng/mL), Pnmr-1(AVA)::unc-29(+)::SL2::GFP (50 ng/mL). Cotransformation markers, pBX-1(pha-1(+)), Pttx-3(AIY)::mCherry, Pttx-3(AIY)::GFP and Punc-122(coelomocytes)::GFP were each injected at 50 ng/mL.

Laser-mediated Cell Ablations
Laser-mediated cell ablations (for data presented in Fig. 2 and Fig. 3) were performed on L4 transgenic males using standard procedures [69]. Control (mock-ablated) L4 males were mounted with anesthetic for the same duration but not operated on. Males were allowed to mature overnight into adults, isolated from hermaphrodites. For males requiring ChR2 assays, the OP50 food was supplemented with 50 mM ATR. Males were then subjected to standard mating assays and/or ChR2 assays (described below). To confirm that targeted cells had been killed in operated animals, each male was mounted on a slide after assays and examined at 6006 magnification on a Zeiss D1 compound microscope equipped with epi-fluorescence. AVA ablations were performed on males carrying the transgenic arrays fkEx71 or fkEx72; PVY and/or PVX ablations were performed on strains carrying fkEx32 or fkEx77.

Mating Behavior Assays
Mating assays (for Fig. 2, Fig. 3C and Fig. 5) were carried out based on procedures described in Liu et al. (2007) [55]. Twentyfour hours before assays, L4 males (maximum of 10) were placed together on a culture plate and allowed to mature overnight in the absence of hermaphrodites. The next day, individual males were placed on a mating lawn (a 5 mm diameter OP50 lawn, freshly prepared) containing 5 unc-64; lite-1 virgin 1-day old adult hermaphrodites. Their mating behavior was observed for 15 mins or until the male ejaculated, whichever occurred first. Each assay was digitally recorded using a Zeiss AxioCam HS digital camera and AxioVision software (release 4.7). These videos were analyzed for execution of the three motor behaviors presented in Fig. 2 and Fig. 5: % of successful contact responses, scanning speed on the non-vulva side and the frequency of lost tail contact during scanning. % Contact Response = 100 x [the number of times the male exhibits contact response/the number of times the male makes contact with a hermaphrodite via the rays]. A successful (complete) contact response requires both tail apposition and initiation of backward locomotion [12]. Scanning speed on the non-vulva side (mm/sec) = mean length of 1-day old adult unc-64; lite-1 hermaphrodites (i.e., 1044 mm)/the time required for a male to travel the length (sec). The average speed for a male was calculated from a random selection of 5 non-vulva sides scanned, or all non-vulva sides scanned if the number of sides completed was less than 5. Number of tail contacts lost = the number of times that a male lost tail contact with the hermaphrodite during the mating trial.

ChR2 and NpHR Assays
Relates to Fig. 3, Fig. 4 and Fig. S2. Except for -ATR controls, all strains were maintained on plates spread with OP50 E. coli supplemented with 50 mM ATR. Except during animal transfer or assays, plates were kept wrapped in foil. Twenty-four hours before assaying, five L4 males were placed on a plate freshly spread with 50 mM ATR+OP50 and allowed to mature overnight. Assays were performed on a Zeiss M2 Imager stereomicroscope equipped with epi-fluorescence. For the ChR2 assays, individual animals were exposed to three, evenly spaced, 500 msec flashes of blue light (470/40 nm wavelength), while moving forward. Assays were recorded using a Zeiss AxioCam HS camera and AxioVision software [12]. For PVY/PVX activation, strains carrying either array fkEx32, fkEx72 or fkEx77 were used. The presence of ChR2-YFP in both PVY and PVX was verified post-assay at 6006 magnification on a Zeiss D1 compound microscope equipped with epi-fluorescence.
To inhibit PVY and PVX during mating, strains of the following genotype were used: fkEx66 or fkEx67 [Pnlp-14(PVY+PVX)::NpHR-EYFP+pha-1(+)]; pha-1(e2123ts); him-5(e1490)V; lite-1(ce314)X. Strains were cultured with or without ATR as per ChR2 transgenic strains (+ATR and -ATR treatments, respectively in Fig. 3C). Twenty-four hours before assaying, five L4 transgenic males were placed on a plate freshly spread either with or without ATR. The next day, individual males were placed on a mating lawn with 5 unc-64; lite-1 1-day old virgin hermaphrodites. Males were allowed to initiate the vulva search with contact response and commence scanning. Males were then exposed to a 500 msec pulse of yellow light (540/25 nm) at 5 evenly spaced time intervals during scanning. The percentage of times a male paused when pulsed with light was scored. To test the effect of NpHR activation in transgenic males not engaged in mating, 1-day old, solitary virgin transgenic males (grown in the presence or absence of ATR) were exposed to three, evenly spaced 500 msec yellow light flash, while they were moving forward. The distance and direction travelled after the flash was measured as described below.

Measurement of the Distance Moved in Response to ChR2-mediated PVY+PVX Activation
Animal response after ChR2 activation was captured digitally as described above. The frame after the light flash in which response is first apparent (''start of response'' frame) was identified as well as the frame in which response ended (''end of response'' frame). In each frame, the animal's position in the frame was determined by measuring the distance from the animal's head (or tail) to the closest frame edge. The ''distance moved'' (in mm) = the position in the ''start of response'' frame -the position in the ''end of response'' frame, where a negative value corresponds to backward movement and a positive value corresponds to forward movement. Although each animal was exposed to three flashes, the first flash usually gave the most unambiguous response and was typically used for analysis. For animals that did not respond to the light flashes, the distance moved was determined using the time frames that corresponded to the average ''start of response'' and ''end of response'' times for the positive controls (in Fig. 3A this was M; in To calculate the % of males that backed, paused or moved forward for each population assayed in Fig. 3, Fig. 4 and Fig. S2, backing, pausing or forward movement were classified by the following criteria: backing = distance travelled is ,0 mm; pausing = distance is 0 to +2.5 mm; forward = distance.+2.5 mm.
All groups of observations for each measurement were tested for normality using the Shapiro-Wilks test. Most were found to be not normal, thus the Wilcoxon-Mann-Whitney.
ranksum test was used to compare two populations. Since up to 5 speeds per male were used in the scanning speed measurements, significance was determined by clustered bootstrapping on the difference between the observed and expected ranks.  Figure S2 The impact of acr-15 and nlp-14 mutations on PVY+PVX-induced reversal behavior. Males of the genotype indicated on the X-axis and carrying the Pnlp-14::ChR2-YFP transgene were subjected to artificial activation assays. See legend of Fig. 3 for details of graph and table format and statistical analyses. Except for the ''-ATR control'', all males were cultured and assayed in the presence of ATR. acr-15 has no significant impact on reversal behavior, either as a single mutation or in combination with acr-18 mutations, arguing that the absence of phenotype in acr-15 single mutants is not a consequence of functional redundancy with acr-18. In addition to cholinergic markers, PVY and PVX express the nlp(neuropeptide-like protein)-14 gene, which is predicted to encode neuropeptides with sequence similarity to orcokinin from Orconectes limosus (crayfish) [30]. nlp-14 mutants show impaired reversal response. acr-18; nlp-14 double mutants are phenotypically similar to acr-18 single mutants. This suggests that NLP-14 has a neuromodulatory role in PVY/PVX transmission and that cholinergic signaling (mediated by ACR-18-ACR-16-and UNC-29-containing receptors) is the rate-limiting factor. Treatments that were statistically different from wild type (WT) are indicated. Significance, *p,0.05; **p,0.005. (TIF) Video S1 C. elegans male locomotion during the vulva search of mating has a backward directional bias. Shown is a control (wild type) male executing all of the motor behaviors that make up the vulva search: contact response, scanning and turning. Related to Fig. 2

. (MOV)
Video S2 Males lacking the male-specific interneuron PVY are defective in vulva search locomotion. This PVYablated male fails to respond to contact, is unable to maintain backward movement and tail contact and exhibits slow backward locomotion. Related to him-5(e1490); lite-1(ce314)). At the movie start, the tail is at the bottom of the frame and the head at the top. The fluorescent spots on the body that appear during the light flash correspond to GFP expression in the coleomocytes (produced by the Punc-122::GFP co-transformation marker). The movie is played at reduced speed for clarity. Related to Fig. 3

. (MOV)
Video S4 A control male, expressing non-functional ChR2 in PVY+PVX, does not alter his locomotion in response to blue light. Shown is a transgenic control male of the same genotype as the male in Video S3 (genotype: fkEx32[Pnlp-14(PVY+PVX)::ChR2-YFP+Punc-122::GFP]; him-5(e1490); lite-1(ce314)), but grown in the absence of ATR. The tail is to the right; head to the left. The fluorescent spots on the body that appear during the light flash correspond to GFP expression in the coleomocytes (produced by the Punc-122::GFP co-transformation marker). The movie is played at reduced speed for clarity. Related to Fig. 3

. (MOV)
Video S7 unc-29; acr-16 acr-18 males fail to reverse in response to ChR2-mediated PVY+PVX activation. Shown is a triple mutant male expressing functional ChR2 in PVY+PVX (array fkEx32). In contrast to wild type (WT) control male behavior (Video S3), the triple mutant fails to reverse in response to the light flash and continues forward locomotion. The tail is to the right; head to the left. The fluorescent spots on the body that appear during the light flash correspond to GFP expression in the coleomocytes (produced by the Punc-122::GFP co-transformation marker). The movie is played at reduced speed for clarity. Related to Fig. 4