Initiation of male sperm-transfer behavior in Caenorhabditis elegans requires input from the ventral nerve cord

Background The Caenorhabditis elegans male exhibits a stereotypic behavioral pattern when attempting to mate. This behavior has been divided into the following steps: response, backing, turning, vulva location, spicule insertion, and sperm transfer. We and others have begun in-depth analyses of all these steps in order to understand how complex behaviors are generated. Here we extend our understanding of the sperm-transfer step of male mating behavior. Results Based on observation of wild-type males and on genetic analysis, we have divided the sperm-transfer step of mating behavior into four sub-steps: initiation, release, continued transfer, and cessation. To begin to understand how these sub-steps of sperm transfer are regulated, we screened for ethylmethanesulfonate (EMS)-induced mutations that cause males to transfer sperm aberrantly. We isolated an allele of unc-18, a previously reported member of the Sec1/Munc-18 (SM) family of proteins that is necessary for regulated exocytosis in C. elegans motor neurons. Our allele, sy671, is defective in two distinct sub-steps of sperm transfer: initiation and continued transfer. By a series of transgenic site-of-action experiments, we found that motor neurons in the ventral nerve cord require UNC-18 for the initiation of sperm transfer, and that UNC-18 acts downstream or in parallel to the SPV sensory neurons in this process. In addition to this neuronal requirement, we found that non-neuronal expression of UNC-18, in the male gonad, is necessary for the continuation of sperm transfer. Conclusion Our division of sperm-transfer behavior into sub-steps has provided a framework for the further detailed analysis of sperm transfer and its integration with other aspects of mating behavior. By determining the site of action of UNC-18 in sperm-transfer behavior, and its relation to the SPV sensory neurons, we have further defined the cells and tissues involved in the generation of this behavior. We have shown both a neuronal and non-neuronal requirement for UNC-18 in distinct sub-steps of sperm-transfer behavior. The definition of circuit components is a crucial first step toward understanding how genes specify the neural circuit and hence the behavior.


Background
To execute a complex behavior, an animal must perceive and integrate information from various sources, process this information and perform the appropriate task. In many animals, one such complex behavior is mating, as different aspects of this behavior must be coordinated to ensure sexual reproduction. Male mating behavior in C. elegans affords the opportunity to study a complex multistep behavior in an organism with a relatively simple nervous system.
When mating, C. elegans males exhibit a highly reproducible series of behavioral steps, culminating in the fertilization of hermaphrodite eggs [1][2][3][4]. Briefly, when the male comes in contact with a hermaphrodite, the male responds by placing the ventral side of his tail against the hermaphrodite, and moves backward along the body, scanning for the vulva. If the male's tail reaches the end of the hermaphrodite before encountering the vulva, the male turns and continues scanning the other side. When the vulva is located, the male ceases backward motion, prods with his copulatory spicules, inserts them into the vulva, and transfers sperm into the uterus [3][4][5][6]. Although the steps of reproductive behavior have been described in many organisms, relatively little is known about the genes that specifically control these behaviors [7].
We and others have begun molecular genetic analyses of the C. elegans male response [8][9][10], turning [4,[11][12][13], vulval location [8][9][10]14], spicule insertion [12,[15][16][17], and most recently, sperm-transfer [12] behaviors. By dissecting each step, their eventual integration as a complex series of overlapping sub-behaviors can be understood, and provide insight into nervous system function. In addition, the study of each individual step, each with its own unique aspects, may provide knowledge about the regulation of behavior not gotten from studying the other steps of mating. For example, the sperm-transfer sub-behavior in C. elegans is slower than any of the other mating steps, and does not occur instantaneously upon spicule insertion; therefore, its analysis will further our understanding of behavior.
To generate a behavior, an animal must employ its nervous system to perceive and integrate stimuli, and often coordinate this information with non-neuronal targets necessary for its proper execution. What is known about the neuronal contribution to sperm-transfer behavior comes from systematic ablation of the sexually dimorphic neurons in the male to assign them function during mating behavior [3]. Each of the two male spicules contains three neurons: two sensory neurons, designated SPV and SPD, and the motor neuron SPC [18]. Ablation of the SPV spicule neurons leads to premature sperm transfer at the vulva; therefore, the SPV neurons regulate the timing of sperm transfer by acting as a negative regulator [3]. Less is known about non-neuronal contribution to sperm-transfer behavior. Gower et al have recently shown that signaling mediated by inositol 1,4,5-triphosphate is required for efficient sperm transfer and propose that the site of action for this signaling is the male somatic gonad [12].
Here we present an analysis of the sperm-transfer step of male mating behavior. Direct observation suggests that sperm transfer comprises four sub-steps: initiation, release, continued transfer, and cessation. In a genetic screen for mutations affecting successful copulation, we isolated mutants in the sperm transfer process. Of the mutants we isolated, unc- 18(sy671), the focus of this work, has defects in two distinct sub-steps of sperm transfer: initiation and continued transfer.
UNC-18 is a member of the SM family of highly conserved proteins involved in membrane-trafficking pathways, particularly in regulated exocytosis [19]. The mammalian unc-18 homologue (Munc-18) has been reported to be expressed in neurons, as well as in non-neuronal cell types [20][21][22][23]. In C. elegans, UNC-18 expression and function has been previously reported only in neuronal cells [24][25][26]. Here we show that, while the UNC-18 site of action for initiation involves neuronal expression, the continuation of transfer involves non-neuronal expression of UNC-18 in the male gonad.

Steps of transfer and timing
As a first step toward understanding the molecular genetic processes underpinning sperm-transfer behavior and its regulation in C. elegans, and to establish a context to characterize sperm transfer mutants, we further defined this behavior. Our observation of wild-type sperm transfer and subsequent genetic screening (see below) has defined at least four steps in the transfer process: initiation, release, continued transfer, and cessation ( Figure 1).
Once a male has inserted his spicules into the hermaphrodite's vulva, there is movement of sperm out of the seminal vesicle, which we call 'initiation'. The initiation step, from tonic insertion of spicules to sperm exit from the seminal vesicle, takes 14.4 ± 0.9 seconds (n = 33) (see Additional file 1). No gross changes in morphology were observed, except for the exit of sperm from the seminal vesicle, through the 'valve region', and into the tube formed by the vas deferens.
The sperm leave the seminal vesicle, travel through the vas deferens, and are released from the animal through the cloaca, a sub-step termed 'release' (see Additional file 2). The anal sphincter is contracted at this time, presumably to open the cloaca [18]. Release of sperm takes 17.2 ± 6.5 Mating and the sub-steps of sperm transfer Figure 1 Mating and the sub-steps of sperm transfer. (A) A male with his tail positioned over a hermaphrodite's vulva after response, backing, turning, and vulva location. The inset shows the male tail after spicule insertion. Arrowhead denotes the tip of the spicules. The bracket denotes the approximate portion of the male shown in schematic in B. (B) Time (t) is equal to 0 when the spicules are tonically inserted into the vulva. The time it takes from t = 0 to completion of each sub-step is shown to the right of that step. Continued transfer was not timed, as it is simply the difference between the time from release to the cessation of transfer. For all times given, the minimum number of animals tested is 33. Note: 'valve' region is shown as open throughout this schematic as no morphological changes can be detected using the microscope we used to analyze sperm-transfer behavior. seconds (n = 34) from the time of insertion, and failure at this step causes sperm to collect in the vas deferens. Failure to release is rarely observed in wild-type animals; however, if a male's spicules fall out of the vulva while transferring sperm, sperm will sometimes collect in the vas deferens. Sperm continue to exit the seminal vesicle in a genetically separable step (see below) that we call 'continued transfer' (see Additional file 2). When all sperm that exit the seminal vesicle have been released, the spicules retract into the male. This is the 'cessation' sub-step and the termination of sperm transfer (see Additional file 3). The entire process from spicule insertion to cessation takes 90.2 ± 18 seconds (n = 33).
Liu and Sternberg reported that it took approximately 4 seconds, compared to our reported 90 seconds, to transfer sperm [3]. We attribute the discrepancy between the results to the microscope used for behavioral analysis (see Methods). Because the majority of sperm to be transferred are ejaculated in the beginning of the transfer process, it is difficult to observe without the use of a high-magnification dissecting microscope that as the behavior progresses, the male still continues to intermittently transfer a few sperm.

Screen for mating behavior mutants
To begin to answer fundamental questions about how the sub-steps of sperm transfer are regulated, we performed a genetic screen designed to isolate males defective for this behavior. Briefly, plg-1(e2001d) males will extrude a gelatinous substance that puts a 'plug' over the hermaphrodite vulva after sperm transfer [27]. The plg-1(e2001d) allele is a dominant allele derived from a wild C. elegans isolate that was introgressed into the N2 wild-type strain. By screening for the lack of copulatory plugs in a clonal population of EMS mutagenized worms, we were able to isolate mutants defective in all steps of mating (see Methods). Secondary screening of putative mutants by observation of mating behavior allowed us to determine those mutants with defects in sperm transfer.
To isolate only those with non-developmental mating defects, we screened through the mutants we isolated for defects in morphology, focusing on those with wild-type anatomy. One recessive mutant isolated, sy671, is defective in the initiation step of sperm transfer. While sy671 mutant males will insert their spicules into the hermaphrodite vulva and remain tonically inserted for 82.9 ± 17 seconds (n = 14) (not statistically different by one-way analysis of variance (ANOVA) from wild type), they rarely initiate sperm transfer (Table 1). A second recessive mutant isolated, sy672, allows initiation and release, but is defective in the continuation of sperm transfer. sy672 males transfer very few sperm, although sperm are clearly visible in the seminal vesicle, and sy672 males keep their spicules tonically inserted during the transfer behavior for 62.6 ± 12 seconds (n = 15). Although the duration that sy672 males remain with their spicules inserted in the hermaphrodite vulva is less than wild type (significant using ANOVA), they should transfer significantly more sperm, as the bulk of transfer should have already occurred by this time. The sy672 mutant allows us to genetically separate the continued-transfer step from initiation and release steps. For the remainder of this study, we will focus on the analysis of sy671. sy671 acts downstream or in parallel to SPV As mentioned above, the SPV neurons regulate sperm transfer by inhibiting release until the spicules have penetrated the vulva: SPV-ablated wild-type males prematurely transfer sperm before spicule insertion [3]. Because sy671 has the opposite phenotype (failure to initiate sperm transfer after spicule insertion) we ablated the SPV neurons in the sy671 mutant to see if this could suppress the initiation defect. In particular, we ablated the B.β cell, the precursor of both SPV neurons (SPVL, SPVR) ( Figure 2). Although the B.β ablation also eliminates four of 12 structural cells (two sheath and two socket), the sheath cells are not necessary for spicule morphology, and the function of the socket cells is redundant [28]. In addition, ablation of the B.β cell causes the same defect in sperm transfer as the ablation of the SPV alone [3]. We found that ablation of the B.β cell (SPV neurons) in the sy671 mutant did not suppress the initiation defect (Table 2). This observation suggests that the defect in our mutant lies downstream or in parallel to the SPV regulation in the sperm-transfer pathway.

Molecular characterization of sy671
To understand the sperm-transfer initiation defect in sy671 males, we cloned this locus. We employed singlenucleotide polymorphism (SNP) mapping [29] to map sy671 to a 120-kb interval between the polymorphisms F26A10:1544 and C47C12:6306 (see Methods; Figure 3A and 3B). Injection of cosmids and PCR fragments narrowed the sy671 locus to a 7.2-kb region containing F27D9.1 (previously identified as the unc-18 locus). We sequenced the predicted coding and non-coding regions of unc-18 from the sy671 mutant, and found that the sy671 mutant contains a G→A missense mutation in an unc-18 exon, which changes the amino-acid Arg476 to His (see Methods).
In C. elegans, unc-18, which encodes a member of the Sec1/Munc-18 family of proteins, is required in neurons for synaptic-vesicle exocytosis [30,31]. For exocytosis to occur, secretory vesicles must dock, be primed, and then fuse with the plasma membrane [32][33][34]. UNC-18 has been proposed to function in each of these steps of exocytosis, based on work in various systems, and may also act as a trafficking factor for other proteins necessary for exocytosis [21,[35][36][37][38]. Although UNC-18 has been postulated to act in various steps of the exocytosis pathway, in C. elegans it has been shown that there is a reduction of docked vesicles at the active zone in unc-18 mutants, indicating that UNC-18 functions at least as a facilitator of vesicle docking in C. elegans [39].
Mutations in unc-18 cause locomotor defects and acetylcholine accumulation, and therefore UNC-18 has been implicated in the release of acetylcholine [24,25,40]. UNC-18 interacts with syntaxin, a protein believed to mediate fusion of synaptic vesicles to the plasma membrane and hence neurotransmitter release [26,41]. While most alleles of unc-18 have severe locomotor defects, sy671 males have only mild locomotor defects (data not shown). Although the amino-acid change in sy671 (Arg476 to His) occurs in a residue conserved among UNC-18 orthologs (data not shown), this residue lies outside the syntaxin-binding region [37].
We were intrigued by the unique position of the molecular lesion in the sy671 allele of unc-18. All reported sequenced alleles of unc-18 have a molecular defect in either exon 4 or exon 9 [31]; however, sy671 has its defect in exon 8 ( Figure 3C). Of the other unc-18 mutants, only two, md1401 and md1264, move sufficiently well to allow sperm transfer to be assayed. Although their movement is indistinguishable from sy671, both mutants are able to initiate transfer (data not shown). This difference among alleles leads us to speculate that the mutation in sy671 has some specificity for sperm transfer through a mechanism different from the function of UNC-18 in vesicle release.
To address if a perturbation of vesicle release is the likely defect in sy671 males, we assayed males that were transheterozygous for sy671 and for a severe allele of unc-18, b403. The unc-18(b403) allele has a single amino-acid substitution (exon 9) that causes the accumulation of acetylcholine, the inability to bind syntaxin, and a severe locomotor defect, all characteristics of defective vesicle release [31]. If sy671 does have a defect in a yet uncharac- The B.β lineage Figure 2 The B.β lineage. The B.β lineage gives rise to both SPV neurons, left (L) and right (R), as well as two sheath (sh) cells and two socket (so) cells. Mapping and cloning of unc-18(sy671) terized UNC-18 function, we would expect that when placed in trans to an allele that perturbs vesicle release through a single amino-acid substitution, we would see compensation for the sperm-transfer defect. However, the b403 allele fails to complement sy671 for the initiation of transfer (0/16 animals initiated sperm transfer), and therefore it seems likely that the mutation in sy671 also alters vesicle release.

unc-18 is expressed in the male gonad and nervous system
Previous characterization of UNC-18 expression using polyclonal antisera showed that UNC-18 was present in ventral-cord motor neurons and some unidentified head neurons in the adult hermaphrodite [25]. To replicate the expression in the hermaphrodite and determine expression in the male, we fused the unc-18 upstream sequences used in the mutant-rescue experiment to a YFP reporter. In transgenic animals, this reporter (unc-18::YFP) recapitulated the previously observed expression in the hermaph-rodite. In the male, unc-18::YFP showed the same expression in the ventral cord and head, but surprisingly also had strong expression in the gonad. The expression was strongest in the 'valve region', a non-neuronal tissue that lies at the junction of the seminal vesicle and the vas deferens ( Figure 4). Expression was also present in most, if not all neurons in the male tail ( Figure 5).

Site of action: initiation
To define the site of action of unc-18 in sperm-transfer behavior, we created a YFP-tagged version of UNC-18 under the regulation of the unc-18 upstream regulatory region (unc-18::UNC-18::YFP), which was able to rescue the sy671 sperm transfer initiation defect (Table 1). This construct was also able to confer the null allele, unc-18(e81), with the ability to initiate sperm transfer (Table  1). Next we expressed the UNC-18::YFP functional protein under different tissue-specific, transcriptional-control regions to define its site of action in sperm-transfer behav-unc-18 expression in the male gonad and ventral nerve cord ior, beginning with the two most probable candidates: the gonad and the spicule neurons.
Because unc-18 expression in the male gonad coincides with the region where sperm exit from the seminal vesicle to the vas deferens, this seemed a likely candidate for the site of action of unc-18 in the initiation of sperm-transfer behavior. An equally attractive candidate is the SPV spicule neurons in the male tail, as they are known regulators of sperm transfer.
To express unc-18 under a gonad-specific promoter, we used a portion of the described itr-1 gene [42]. It was previously reported that two introns and one exon of the itr-1 gene (including a 2.1-kb fragment that includes the Cterminal six codons, the 3' untranslated region, and downstream sequences) gave expression in the intestine, pharyngeal isthmus, and gonad of hermaphrodites [42].
Through further dissection of this region, called the pC promoter, we were able to drive expression (using only intron 2) in the seminal vesicle, the valve region, and the vas deferens of the male ( Figure 6A-C). This expression coincides with that described for the entire pC promoter in the male [12]. We call this element int2itr-1. We then expressed unc-18 under the control of the itr-1 gonadal enhancer (int2itr-1::UNC-18::YFP), but were not able to rescue the mutant phenotype in sy671 (Table 1). To express unc-18 under a spicule neuron promoter, we used 3 kb of 5' upstream sequence of the gpa-1 gene that drives expression in all three types of spicule sensory neurons, including the SPV neurons [28] (Jiang LI, Mendel JE and Sternberg PW, unpublished observations; Figure 6D-F).
After initial testing of possible gonadal and spicule neuron contributions, we chose to express unc-18 more broadly to see if unc-18 was at least acting in neurons to initiate the transfer of sperm. We used a pan-neuronal promoter (unc-119) [43] to drive expression of unc-18 (unc-119::UNC-18::YFP) in neurons ( Figure 6G-I), and also used a body-wall muscle promoter (myo-3::UNC-18::YFP) to rule out non-specific rescue effects [44]. The neuronal expression of unc-18 was able to rescue the initiation defect (Table 1). These results suggest that although unc-18 is expressed in the male gonad, its expression in this tissue is not required for the initiation of sperm transfer, but it is required in neurons. To confirm this result, we used the 5' regulatory region of the aex-3 gene, another pan-neuronal promoter, to express unc-18 (aex-3::UNC-18::YFP) [45]. As expected, this transgene was able to rescue the sperm-transfer initiation defect in sy671 ( Table 1). The negative result using gpa-1::UNC-18::YFP to drive expression in the spicule neurons does not exclude an unc-18 contribution from these neurons; therefore, we decided to further define the site of action of UNC-18 in this behavior using various neuronal promoters.
UNC-18 was initially described to be expressed in all ventral-cord motor neurons [25]. The ventral cord has been divided into eight classes of motor neurons: four classes that innervate ventral muscles (VA, VB, VC, VD) and four classes that innervate dorsal muscles (DA, DB, DD, AS) [46][47][48][49]. The D-type ventral cord neurons (VD and DD) are gamma-aminobutyric acid (GABA)ergic and the others (VA, VB, VC, DA, DB, AS) are cholinergic [49,50]. The VC cholinergic neurons are hermaphrodite-specific, while in males, the homologous cells divide to give the malespecific CP and CA motor neurons [18,47].
To test whether unc-18 was acting exclusively in either cholinergic or GABAergic ventral cord motor neurons to initiate sperm transfer, we drove expression specifically in unc-18 expression in the developing male tail  [51]. Expression in only the GABAergic neurons was unable to rescue our mutant (Table 3).
To express unc-18 in all cholinergic neurons in the sy671 mutant, we initially tried using promoters from both unc-17 [52] and cho-1 [53] to drive expression. However, the use of either of these promoters (unc mating behavioral defects in the steps prior to sperm transfer, and therefore the worms could not be assayed for their ability to transfer sperm. These defects, not seen in non-transgenic worms, could be caused by overexpression of UNC-18 or possible titration of promoter regulatory proteins caused by our transgenic arrays. Next we tried the acr-2 promoter, which drives expression in the VA, VB, DA, and DB cholinergic ventral-cord motor neurons [54]. The acr-2 construct, acr-2::UNC-18::YFP, when expressed in sy671 mutants, was able to rescue initiation (Table 3). Because acr-2 is expressed in cholinergic neurons throughout the ventral cord and not in the male spicule sensory neurons (data not shown), these ventralcord motor neurons are strong candidates for contributing to the initiation of sperm transfer.
To further test ventral-cord input for the initiation of sperm-transfer behavior and to further pinpoint the site of action to a specific subset(s) of cells, we used additional promoters to attempt to rescue the sy671 defect. We initially tested two classes of cholinergic ventral-cord motor neurons, the A-type neurons, VA and DA (necessary for backward movement) and the B-type neurons, VB and DB (necessary for forward movement) [55]. We used the unc-4 upstream sequences to drive expression in the A-type neurons [56] and the acr-5 promoter to drive expression in B-type neurons [57]. By expressing unc-18 under the acr-5 regulatory region, we were able to rescue the sy671 initiation defect (Table 3), but were unable to do so with the unc-4 promoter ( Table 3). This result suggests that unc-18 in the B-type neurons (VB and/or DB) could be necessary for initiation, again with the exception of the neurons outside the ventral cord that express acr-5.
We determined that acr-5 is expressed in the male-specific ray sensilla (rays 2, 3, 4, 6, and 9) using the acr-5 regula-tory region to drive expression of DsRed2 (acr-5::DsRed2). The expression in the rays appears to be in the A-type ray neurons, as acr-5 expression in the tail does not overlap with pkd-2::GFP, which is expressed in the B-type ray neurons (data not shown) [10]. acr-5 also shows expression in a cell in the right preanal ganglion, which we have tentatively identified as PGA or PVZ. However, input from the acr-5 expressing cells in the male tail in initiating transfer is very unlikely, as there is no overlap with acr-2 expression in the tail ( Figure 7A-L), yet both promoters driving unc-18 expression can rescue the initiation defect in unc-18(sy671) mutants. The overlap in expression of acr-2 and acr-5 is in the ventral cord ( Figure 7M-P), making a strong case for it being necessary for the initiation of sperm transfer. Although these data also suggest it is specifically the Btype motor neurons in the ventral cord that are relevant for the initiation of sperm transfer, further experiments indicate that the male-specific CA ventral-cord motor neurons are stronger candidates (see below).
In the adult male ventral cord we noticed that acr-5::YFP was expressed not only in the B-type neurons, but also in additional neurons ( Figure 8A-C). Because the male has additional ventral cord neurons, the CPs and CAs, we thought that these may account for the additional cells showing expression, and it may be these cells that are necessary for initiation of sperm transfer. Because the CP neurons are necessary for efficient turning behavior [4], we wanted to test if the acr-5-expressing cells were CP neurons. We used a strain carrying a tph-1::GFP transgene that drives expression in the CP neurons [58], and crossed it into animals expressing acr-5::DsRed2. The additional cells in the male ventral cord that express acr-5::DsRed2 did not overlap with the CP-expressing cells, but in many cases these cells were anterior to the CP neurons, suggesting that they could be CA neurons [18] ( Figure 8D-F).
Because there are no reported markers for the CA neurons, we tested the possible contribution of these neurons to sperm-transfer initiation by assaying lin-39 mutant males.
To further corroborate that the unassigned ventral cord motor neurons expressing acr-5 in the wild-type males could be CA neurons, we analyzed the expression of acr-5::GFP in lin-39 mutant males. As expected, we saw fewer cells expressing acr-5::GFP in the ventral cord of lin-39 mutant males ( Figure 8G-I).

Site of action: continued transfer
Although the sy671 mutants can be rescued for the initiation of sperm transfer using a transgene to drive neuronal expression of UNC-18 in a subset of ventral cord cholinergic neurons (most likely the CA neurons), these animals did not transfer as many sperm as wild type. We observed that by using the unc-119 promoter to drive unc-18 expression, the animals would initiate, and usually the sperm that were released from the seminal vesicle were released from the animal, but the animals transferred much less sperm and thus appeared to be defective in the continuation of transfer. This observation raised the possibility that UNC-18 functions in two separate sub-steps of the sperm-transfer process (both the initiation and the continuation of transfer), and that neuronal expression of unc-18 was responsible for the initiation of transfer, while the gonad expression may be necessary for continued transfer.
To formally test this possibility, we hypothesized that if we could add back both neuronal and gonadal expression acr-2::YFP and acr-5::CFP expression overlaps in the ventral cord, but not in the tail of the male of unc-18, using different 5' control regions, we should restore wild-type sperm transfer to the unc-18(sy671) mutants. To do this, we again used our neuronal (unc-119::UNC-18::YFP) construct, and remade our gonadal transgene with a CFP marker (int2itr-1::UNC-18::CFP) to facilitate the scoring of worms carrying both transgenes.
To quantify the amount of sperm transferred, we devised a 'continuation assay' (Methods). This assay uses the percentage of cross progeny as a proxy for the amount of sperm transferred.
As a control for the effectiveness of our continuation assay, we tested unc-18(sy671) mutants carrying a rescuing transgene (unc-18::UNC-18::YFP) to determine if they could sire a wild-type percentage of cross progeny. Not only did we observe that these transgenic animals appeared to transfer a similar amount of sperm to that of wild type, but as expected they sired an equivalent percentage of cross progeny (see Additional file 6). This construct also was able to completely restore wild-type sperm transfer in the unc-18 null mutant (see Additional file 6). In addition, as expected based on observation of sperm transfer, animals carrying the unc-119-driven transgene sired much fewer progeny than wild type in our assay (see Additional file 6), because they initiated transfer and released sperm, but did not continue to transfer. In addition, lines carrying only gonadally expressed UNC-18 (int2itr-1::UNC-18::CFP) did not sire cross progeny (data not shown), consistent with an inability to initiate transfer. We then tested if we could restore wild-type sperm transfer by expressing unc-18 in both neuronal (unc-119::UNC-18::YFP) and gonadal (int2itr-1::UNC-18::CFP) tissues. Indeed, when unc-119 and int2itr-1 promoter elements are used to drive unc-18 expression simultaneously in the sy671 mutant, sperm transfer is restored completely (see Additional file 6). These transgenes are lost at some frequency during meiotic cell divisions [61], and if either transgene is lost, wild-type sperm transfer no longer occurs in that animal (see Additional file 6): if the unc-119::UNC-18::YFP transgene is lost, animals do not initiate, and if the int2itr-1::UNC-18::CFP transgene is lost animals will initiate, but no longer continue to transfer and hence sire fewer crossprogeny. The results from the continuation assays support our hypothesis that UNC-18 functions in two different tissues to mediate two sub-steps of sperm transfer.
Considering that the acr-5 regulatory region we used does not drive detectable expression in the gonad, expression in the gonad may not be absolutely required, but important for proper continued transfer.
One caveat for the use of the continuation assay as a proxy for sperm transfer is that the sperm being transferred must be fertile. Mature spermatozoa that are capable of fertilization develop from spermatids following ejaculation by the male [62]. As it is possible to activate spermatids in vitro to test whether they are competent to mature (through the addition of exogenous proteases [62]), we isolated and activated sperm from unc-18(sy671) males.
Spermatids from unc-18(sy671) males were capable of developing into mature spermatozoa and there appeared to be similar quantities of spermatids released upon dissection compared with wild-type males (data not shown). This observation suggests that there is no sperm formation or activation defect in this mutant.
Although the sperm from unc-18(sy671) males are capable of being activated in vitro, we cannot rule out the possibility that the vas deferens in these mutant males does not release the secreted material necessary to activate the sperm in vivo. Because UNC-18 is involved in regulated exocytosis, such a secretory function in the vas deferens could be postulated. Although this defect would account for a decrease in cross progeny, we do observe a coincident decrease in sperm being transferred. This observation means that even if there is a defect in seminal release, there must be some feedback to the gonad to tell it to stop the continuation of transfer. In either case, UNC-18 is necessary for the proper execution of this step.

Discussion
Understanding how genes function within neural circuits and networks to coordinate a behavioral output is a central focus in neurogenetics. We are using C. elegans mating behavior as a model to address how genes control the ability of an animal to execute a stereotyped behavior. During C. elegans mating behavior, the male must perceive sensory cues, integrate and process information, and generate the appropriate motor responses in order to successfully mate. Here we have defined the motor outputs (sub-steps), one of the molecules (UNC-18), and parts of the circuitry (site of action) involved in C. elegans male sperm transfer, the last step in mating behavior.
Our results indicate that there are at least four sub-steps in the sperm-transfer behavior in C. elegans: initiation, release, continued transfer, and cessation. Unlike the other steps of mating behavior, which occur almost instantaneously upon stimulation, the sub-steps of sperm transfer occur over a 90-second period. This observation suggests that a slower acting signal-transduction pathway(s) may be involved in aspects of this behavior.
To gain insight into genetic control of the individual substeps of sperm-transfer behavior, we screened for mutants in this process. In a screen for sperm-transfer-defective mutants, we isolated two mutants, sy671, defective in sperm-transfer initiation, and sy672, defective in the continued transfer of sperm. We cloned and characterized sy671 and determined it to be an allele of unc-18. UNC-18 is a critical component of the vesicle exocytosis machinery. We infer that the sperm-transfer defects of unc-18(sy671) result from a failure in exocytosis. unc- We used unc-18(sy671) as a tool to further define the circuitry necessary for initiation of sperm-transfer behavior by determining in which tissue/cells UNC-18 acts during this process. We have shown that neuronal expression of unc-18 in the ventral nerve cord is required for initiation. Through a combination of transgene and mutant analyses, the input from the ventral-nerve cord is possibly from the male-specific CA motor neurons. Also possible, and not ruled out by our experiments, is that the B-type motor neurons are responsible for the initiation of sperm transfer.
How is the male's ventral nerve cord communicating with the gonad to initiate sperm transfer? Initially it was reported that there was no innervation of the C. elegans male gonad. However, reconstruction of the male nervous system has shown neuronal connections to the gonad, for example, through the SPC spicule motor neurons [63]. The connections of the male-specific CA motor neurons are not yet reconstructed, and it will be interesting to see if they, or the B-type ventral-cord motor neurons, innervate the gonad. In Drosophila melanogaster, the release of sperm and seminal fluid has been shown to be controlled by a small group of male-specific cholinergic abdominal neurons that innervate male internal reproductive organs [64]. It will be interesting to see if C. elegans parallels the configuration in Drosophila.
Although restoration of neuron-specific UNC-18 function was able to rescue the initiation defect in sy671 males, we determined that animals transferred very few sperm, appearing to initiate, but not continue transfer. Through a series of experiments, we were able to show that non-neuronal expression of unc-18 in the male gonad is required for correct continued transfer. Non-neuronal expression of unc-18 had not been previously reported in C. elegans; however, in other systems, the UNC-18 homologue, Munc-18, is known to function in non-neuronal cells such as the pancreatic β-cells and also chromaffin cells [20][21][22][23].

Expression of reporter constructs in the male
In these cells, Munc-18 is involved in the exocytosis of dense-core vesicles that mediate the release of hormones and possibly other humoral factors. Molecules potentially involved in non-neuronal exocytosis, in particular unc-64 (syntaxin), have been described as expressed in non-neuronal tissue in C. elegans [41].
The mechanism underlying the ability of gonadally expressed UNC-18 to rescue the continuation of transfer remains to be explored. Garcia et al found that the male somatic gonad, the germ line, and the connection between the vas deferens (somatic gonad) and the cloaca helps to sustain prolonged spicule protraction [15]. Because sustained spicule protraction is necessary for proper continuation and cessation of sperm-transfer behavior, the coordination between sperm transfer and spicule protraction suggests that multiple sensory inputs and feedback are necessary for the proper execution of this behavior, especially from the gonad. Exploring these mechanisms in the future will enhance our understanding of the regulation of these behavioral outputs.

Conclusion
Our work provides further understanding of how sperm transfer in C. elegans is achieved. We have separated the behavior of sperm transfer into four sub-steps: initiation, release, continued transfer, and cessation. In a genetic screen, we isolated an allele of unc-18 defective for the initiation sub-step of sperm transfer (sy671), and we determined that the site of action for UNC-18 in the initiation of transfer was in the ventral-cord motor neurons, most likely the male-specific CA neurons. In addition, we have also assigned a role for UNC-18 in the gonad for the proper completion of another sub-step of sperm transfer, the continuation of transfer. We have shown not only that UNC-18 is required in two distinct tissue types for two sub-steps of sperm-transfer behavior (initiation and continued transfer), but also that the SPV neurons that inhibit premature sperm transfer act either upstream or in parallel to the requirement for unc-18 in initiation. Further analysis of sperm transfer will hopefully provide novel information about the genetic control of behavior. Because many parasitic nematodes require sperm transfer for reproduction [65], the sub-steps of sperm transfer are potential targets for new antehelminthics.

Laser ablation
We used standard laser-ablation protocols to kill the B.β cell in the male tail [71]. Staged L3 males were mounted, ablated, and allowed to recover and develop into young adults before assay. For sham-operated males, we kept animals on agar pads with the same concentration of sodium azide as the operated males.

Isolation of mating-behavior mutants
To isolate mutants, we mutagenized PS3696 hermaphrodites using ethyl methanesulfonate (EMS) [66], and placed P0 animals on individual plates. From each plate, three F1 animals were pooled on a new plate, and 10 individual animals from the F2 generation were placed onto individual plates. The F3 populations were scored for mating plugs. The him-5 mutation increases the incidence of males in the population from <1% to 33%, and the plg-1 allele causes males to secrete a mucus plug over the vulva after sperm transfer [27] [67].
Using this approach, approximately 95% of the initial F3 populations were eliminated as either plug-positive or sterile, leaving 5% enriched for mating-defective mutants. These 'plug-less' lines were subsequently assayed by observing mating behavior directly, and assigned a corresponding defect. From the 1400 haploid genomes screened, we isolated two sperm-transfer-defective mutants, sy671 and sy672. The sy671 animals were defective in initiation, while sy672 animals were defective for continued transfer. sy671 animals were out-crossed at least three times with N2 strains before subsequent assays.

Mapping and cloning of sy671
For SNP mapping we used the Hawaiian CB4856 strain as the source of the polymorphisms [29]. To analyze the SNPs, we used the recommended primer pairs [72] to amplify the genomic regions of the recombinants for the following markers: In addition, the region between ZK154 and F26A10 on chromosome X did not have any SNP data; therefore, we sequenced three intergenic 1.5-kb regions in this interval from the CB4856 strain to find polymorphisms. Two of the three sequenced fragments revealed the new SNP markers C47C12:6306 and F27D9:6408. Primers and information about these markers are detailed below.
Our initial mapping populations included 18 mutant and six wild-type F3 populations (total of 48 chromosomes analyzed) from a cross between the Hawaii strain (CB4856) and sy671. These populations were used to map sy671 to the X chromosome, between the markers ZC449 and F11A1. The numbers adjacent to the marker shown below are the number of recombinant chromosomes/ total analyzed for that marker: For generation of additional recombination events in our interval of interest, we generated the triple mutant lon-2(e678) unc-18(sy671) egl-15(n484) for further SNP mapping. To help avoid potential modifiers of mating behavior from the Hawaii strain, we intragressed the CB4856 sequence into the lon-2 egl-15 interval of N2 by crossing CB4856 with the lon-2 egl-15 double mutant and selecting a non-Lon non-Egl F3 population (three times). The intragressed CB4856 strain was checked with internal SNP markers and then crossed with lon-2 (e678) unc-18(sy671) egl-15(n484). In total, 99 Lon non-Egl recombinants were selected, made homozygous, and used for mapping sy671 to a five-cosmid interval.

Mutant rescue
All cosmids, including the rescuing cosmid F27D9, were injected into unc-18(sy671) at 20 ng/µl with myo-2::GFP (5 ng/µl) as a co-injection marker and pBSKS(175 ng/µl) as a carrier. In addition, a rescuing 7.2-kb PCR fragment was amplified using the following primers F27D9P5F (5'-TCGTGACGATCTAGAAGTGGCATTCC-3') and F27D9P5R (5'-TTGGCTTCTCAACGTGGAATGACTGG-3') and purified using the QiaQuick PCR purification kit. The PCR fragment was injected at the same concentration as the cosmid. Transgenic animals were recognized by myo-2::GFP expression in the pharynx. Rescue was also achieved using a 5-kb fragment containing 2.6 kb of 5' upstream promoter sequence and the entire genomic region of unc-18.

Observations of mating behavior
We observed mating behavior with a Zeiss M2-Bio dissecting microscope, up to 660× magnification. The mating behavior of mutant or control males was observed with sluggish unc-31 adult hermaphrodites. All males were isolated from sibling hermaphrodites at the L4 stage and were kept on fresh plates in groups before observation [10,15]. For the mating assay, a virgin adult male (18-24 hrs post L4 lethargus) was placed on a 10-20 mm bacterial lawn with multiple unc-31 hermaphrodite adults. The hermaphrodites used were between 3 and 4 days post L4, as these are partners with which spicule insertion occurs almost instantaneously [15], facilitating analysis. After tonic spicule insertion, a stopwatch was started. Time points were taken when the sperm were released into the hermaphrodite (release) and when the spicules were retracted after sperm transfer (cessation). Because the time difference between initiation and release is generally of the order of 1-2 seconds, a Sony DFW-V500 color digital camera was used to record this behavioral interval from a separate group of animals, as a stopwatch was not suitable. These videos were later analyzed using iMovie (Apple).
When testing animals specifically for sperm-transfer initiation, animals were assayed until the first tonic spicule insertion and transfer was noted. If an animal did not insert their spicules within 10 minutes, that animal was not scored.
We observed that older sy671males (48 hours post-L4) would initiate sperm transfer more often than the 24-hour post-L4 sy671 males used in our assays. Although these males initiate more often, they will not transfer as many sperm as wild type.

Continuation assays
L4 unc-31 hermaphrodites were staged and used 3.5 days later for mating with isolated 18-24-hour post L4 virgin adult males. After an individual mating event, the hermaphrodite was moved to a fresh plate. The progeny sired (both cross and self) were tallied to determine the percentage of cross progeny. Percentage of cross progeny is the number of cross progeny divided by the total number of progeny. transgenes were scored as L4 males to determine if either YPF or CFP fluorescence was lacking in these animals. Animals completely lacking either Punc119::unc-18::yfp or Pint2itr-::unc-18::cfp were identified using a Zeiss M2Bio stereofluorescence microscope with the appropriate filter. These animals were then assayed for mating behavior, and then loss of fluorescence confirmed using conventional fluorescence microscopy (Zeiss Axioskop) after assay.

Sperm isolation and in vitro activation
Individual males were placed in a 7-µl drop of sperm media [74] containing 10 mg/ml PVP and 200 µl/ml pronase [62]. Dissections to release sperm were accomplished by cutting in front of the tail with a hypodermic needle [74], and released spermatids were observed as they became activated to form spermatozoa within 5 minutes on a Zeiss Axioskop with DIC optics.
The following were constructed as for unc-119::UNC-18::YFP (above), except the primers for amplification of the appropriate promoter regions were unique: The unc-119 reporter (unc-119::YFP) was created by digesting unc-119::UNC-18::YFP with PstI and SmaI to drop out the unc-18 coding region, followed by removal of the 3' overhang with T4 polymerase and re-ligation.
acr-5::YFP; acr-2::YFP; aex-3::YFP, and int2itr-1::YFP were all constructed by excising the unc-18 region from the appropriate construct above with PstI and SmaI, following which the resulting vector fragment was blunt-ended with T4 polymerase and the vector relegated. acr-5::GFP was constructed by swapping the yfp from acr-5::YFP with a nuclear localized version of gfp from pPD95.70 (a gift of A. Fire).
The acr-5::DsRed2 reporter was constructed as follows. The acr-5 promoter was removed by linearization of acr-5::UNC-18::YFP by digestion with NotI, followed by Klenow treatment and subsequent digestion with PstI, creating a blunt/PstI promoter fragment. This fragment was ligated into a vector containing DsRed2 followed by the unc-54 3' UTR (PSX77-DsRed2 courtesy of S. Xu) that was prepared by digestion with HindIII, Klenow-treated, and subsequently cut with Sse8387 I.
Because the promoter region of cho-1 [53] was not stable in pSOAunc-18, we excised the BssHII fragment containing the ATG of unc-18 through the end of the unc-54 3' UTR from pSOAunc-18, treated it with T4 polymerase, and blunt-end cloned it into the EcoRV site in pBR322. We then PCR amplified the cho-1 promoter with the primers 5'-TCCCCGCGGATACTTCAATTATTCCCGTCTTAC-CAGG-3' and 5' '-TCCCCGCGGTAAGATAGAG-GAACTCGCAAGATGTCG-3', digested with SacII, and cloned into the SacII site carried over to pBR322 and checked the orientation.

High-resolution light microscopy
Cell anatomy was observed with Nomarski optics, and GFP expression was analyzed by conventional fluorescence microscopy (Zeiss Axioskop) using a Chroma Technology High Q GFP long-pass filter set (450 nm excitation, 505 nm emission). CFP and YFP were visualized using a Chroma Technology CFP filter set (31044v2; exciter D436/20, emitter D480/40, beamsplitter 455dclp) and a YFP set (41029; exciter HQ 500/20, emitter HQ520lp, beamsplitter Q515lp). Photographs were taken with a digital camera using Improvision Openlab software to obtain images, and images processed using Adobe Photoshop version 7.0.