Abstract
Caenorhabditis elegans male mating provides a powerful model to study the relationship between the nervous system, genes, and innate sexual behaviors. Male mating is the most complex behavior exhibited by the nematode C. elegans and involves the steps of response, backing, turning, vulva location, spicule insertion, and sperm transfer. Because neuropeptides are important neural regulators of many complex animal behaviors, we explored the function of the FMRFamide-like neuropeptide (flp) gene family in regulating male copulation. We found that peptidergic signaling mediated by FMRF-amide like neuropeptides (FLPs) FLP-8, FLP-10, FLP-12, and FLP-20 is required for the sensory transduction involved in male turning behavior. flp-8, flp-10, flp-12, and flp-20 mutant males significantly increase repetition of substep(s) of turning behavior compared with wild-type males. Genes controlling neuropeptide processing and secretion in general, including egl-3, egl-21, ida-1, and unc-31, are also required for inhibiting repetitive turning behavior. Neuropeptidergic signaling adjusts the repetitiveness of turning independently of serotonergic modulation of the timing of turning. Surprisingly, the mechanosensitive touch receptor neurons are found to be part of the neural circuitry regulating male turning behavior, indicating the existence of functional dimorphisms in the nervous system with regard to sex-specific behaviors.
- Caenorhabditis elegans
- FMRFamide-like neuropeptide
- male sexual turning
- touch receptor neurons
- mechanosensory behavior
- sexually dimorphic function
Introduction
Sexual behavior is one of the most ancient and vital social behaviors exhibited among metazoans, yet remarkably little is known about its molecular basis. Caenorhabditis elegans copulation is comprised of an innate series of behavioral steps (Liu and Sternberg, 1995), raising the question of how these behaviors are genetically controlled. C. elegans male mating is a powerful model to study the relationship between the nervous system, genes, and behavior.
The adult male possesses 381 neurons, of which 87 are sex-specific. With the exception of four neurons in the head, the male-specific neurons are located in the copulatory structures in the tail (Sulston et al., 1980). Cell ablation experiments coupled with behavioral analysis have provided important insight into the neuronal basis of C. elegans mating behavior. In wild-type males, a stereotypic behavioral sequence includes response to hermaphrodite contact (requiring male-specific ray neurons), backing along the mate's body, turning at the mate's head or tail (requiring the three most posterior pairs of rays and the male-specific serotonergic CP neurons), location of vulva (requiring the male-specific hook neurons), spicule insertion (requiring male-specific cholinergic neurons of the spicules and postcloacal sensilla), and sperm transfer into the mate's uterus (requiring the male-specific SPV spicule neurons) (Loer and Kenyon, 1993; Liu and Sternberg, 1995; Barr and Sternberg, 1999; Garcia et al., 2001). C. elegans male sexual behaviors are also regulated by components of the core nervous system: spicule protraction is modulated by the NSM neurosecretory head neurons, and sperm transfer requires input from the ventral nerve cord motor neurons (Gruninger et al., 2006; Schindelman et al., 2006).
Neuropeptides modulate many complex animal behaviors, including feeding, nociception, circadian rhythm, reproduction, learning, and social behaviors (Storm and Tecott, 2005; Borgland et al., 2006; Sabatier, 2006; Saper, 2006; Hill and Oliver, 2007). The C. elegans genome contains at least 30 FMRFamide-like neuropeptide (flp) genes that are predicted to encode over 70 biologically active FMRF-amide like neuropeptides (FLPs) (Li, 2005) (S. Husson and L. Schoofs, personal communication). At least 50% of C. elegans neurons express one or more flp genes (Kim and Li, 2004), indicating their potential roles in the nervous system. We explored flp gene function in regulating C. elegans male sexual behaviors and found that peptidergic signaling modulates the turning step. FLP-8, FLP-10, FLP-12, and FLP-20 are required to suppress the repetitive turning phenotype in C. elegans males. Genes controlling neuropeptide processing and secretion in general, like egl-3, egl-21, ida-1, and unc-31, are also required for inhibiting repetitive turning behavior. Moreover, peptidergic signaling regulates repetitive turning behavior independently of serotonergic and dopaminergic signaling. Interestingly, the cellular site of action for egl-3 and flp-20 in mediating male turning behavior is in the mechanosensitive touch receptor neurons (TRNs) found in both males and hermaphrodites, indicating that the shared nervous system has sexually dimorphic functions. Finally, we demonstrate that the MEC-4/MEC-10 channel complex is required for inhibition of repetitive turning, suggesting that a mechanosensory stimulus regulates the turning step of male mating behavior.
Materials and Methods
Strains
Routine culturing of C. elegans was performed as described previously (Brenner, 1974). The following alleles were used: LGI, aex-5(sa23), bli-4(e937), kpc-1(gk8); LGII, cat-2(e1112), flp-4(yn35), tdc-1(n3420); LGIII, bas-1(ad446), eat-4(ky5), ida-1(ok409), unc-47(e307); LGIV, egl-21(n476), flp-10(pk367), him-8(e1489), unc-17(e245), unc-31(e169); LGV, egl-3(nr2090), flp-6(pk1593), flp-21(pk1601), him-5(e1490); LGX, flp-3(pk361), flp-8(pk360), flp-12(n4902), flp-19(pk1594), flp-20(pk1596), mec-4(e1611, d), mec-4(e1339, rf), mec-10(e1515, lf).
flp mutant isolation
flp deletion alleles were isolated by PCR screening of populations of EMS-mutagenized strains (Jansen et al., 1997). flp-6(pk1593) is a 1717 bp deletion that spans from −85 bp upstream of the ATG start codon to 32 bp downstream of the TAA stop codon (flanking sequences, tgcttcctgctttaaggattcttatctcata and ggctcacacatttaacttattttaaaatca). The deletion removes the entire coding sequences of the flp-6 gene. flp-8(pk360) is a 1414 bp deletion that removes the entire coding region and some 5′ and 3′ neighboring sequences of the flp-8 gene (flanking sequences, aatcaatttcagataaatctcagaacaagc and ataattttggcgtcatagtttttgatagaa). flp-10(pk367) is a 1883 bp deletion that spans 1.3 kb upstream of the start codon, exon 1, and 5′ part of intron 1 of the flp-10 gene (flanking sequences, tatttttctttgcttattctgcgttccatc and agtaaaaactatcttaataatccatttcca). flp-12(n4902) is a 1732 bp deletion that removes most of the flp-12 gene except exon 1 and adjacent 5′ part of intron 1 (flanking sequences, aaatgtaacactagacatagctctccatgt and caatttgaatttctgaataagagcccgcaa). flp-20(pk1596) is a 1343 bp deletion that removes the entire coding region and some 5′ and 3′ neighboring sequences of the flp-20 gene (flanking sequences, acgtcacggaattcttttaaaatgaatgga and ttagacgtattaaaagtaatcgttttggca). None of the above deletions disrupt the promoter region or coding sequence of neighboring genes. flp-19(pk1594) is a 1946 bp deletion that removes the 5′-UTR, the entire coding sequence, and 3′-UTR of flp-19 gene (flanking sequences, cattagcactatattccatagtgctccacc and atgcggtgagatttgttctgcatagacgga). It is important to note that the deletion also removes the 3′-UTR, the last exon, the last intron, and the final 17 bp of the penultimate exon of abl-1 gene. Each flp mutant strain was outcrossed at least 10 times.
Male mating behavior
Generating males.
To increase the frequency of males, the him-8 mutation was crossed into flp-6, flp-8, flp-12, flp-19, flp-20, egl-3, bli-4, aex-5, kpc-1, and ida-1 mutant strains, and the him-5 mutation was crossed into flp-10, flp-10; flp-12 flp-20 flp-8, mec-4(e1611), bas-1, and mec-4(e1611);bas-1 mutant strains. For other mutant backgrounds, males were generated through the heat shock method (Hodgkin, 1983).
Mating behavior assay.
Males were isolated at L4 stage, cultured at 15°C overnight, and the resulting young virgin adults were used for mating assays. Adult, mix-staged (∼1–2 d of age) unc-31 hermaphrodites were used as mates. Mating plates were prepared by inoculating 5 cm nematode growth medium plates with 10 μl of OP50 (outer diameter, 1.0) and culturing at room temperature (22–23°C) for 1 d. For mating behavior observations, 16–18 hermaphrodites were transferred to the mating plate and allowed to evenly distribute on the bacterial lawn. One male was added to the mating plate for each observation. The observation period was 3 min or until spicule insertion occurred. Behavioral observations were done at room temperature with a Leica (Heerbrugg, Switzerland) MZ12.5 stereomicroscope under 125× magnification. To score the repetitiveness of male turning behavior, we calculated Tavg (population average of maximum trials per turn exhibited by individual males of a particular genotype). Intergenotypic comparisons of Tavg were made using Kruskal–Wallis test (nonrepeated and nonparametric ANOVA test).
Transformation rescue of egl-3(nr2090) mutants
Plasmids were constructed by standard molecular cloning techniques, and sequences were verified when necessary. Plasmid KP509 contains the glr-1 promoter and 4.0 kb of egl-3 genomic coding region. Plasmid KP454 contains 4.2 kb endogenous egl-3 promoter, 4.0-kb egl-3 genomic coding fragment, 2.3 kb 3′-UTR, and an in-frame green fluorescent protein (GFP) fusion to the C terminus (Kass et al., 2001). Plasmid TL1115 and plasmid TL9284 both contain the same egl-3 genomic coding fragment excised from KP509, but under the control of the flp-8 and mec-7 promoter, respectively (plasmid sequence information available in supplemental material, available at www.jneurosci.org). Transgenic strains were generated by microinjecting different plasmid constructs at 30–50 ng/μl together with 60 ng/μl cotransformation marker coelomocyte::GFP (Miyabayashi et al., 1999) into egl-3;him-8 hermaphrodites. Male turning behavior was examined as described above.
Transformation rescue of flp-10(pk367) and flp-20(pk1596) mutants
Wild-type genomic flp-10 and flp-20 clones (promoter, coding region, 3′-UTR) were amplified with PCR using high-fidelity TaKaRa LA Taq polymerase. Primers FLP90 (5′-TAC TCG GCT AAT GAC TAG TGT T-3′) and FLP490 (5′-GAG AAC TCA AAG TCT TGA GC-3′) were used for flp-10 locus amplification. The purified flp-10(+) PCR product was injected at 33 ng/μl into flp-10;him-5 hermaphrodites. Primers FLP285 (5′-GGA AAC ATT GGT CGG GAG ATG-3′) and FLP485 (5′-TAG AAT AGG GGT CAA ATA GGA GG-3′) were used for flp-20 locus amplification. The purified flp-20(+) PCR product was injected at 17 ng/μl into flp-20;him-8 hermaphrodites. flp-8(+) or flp-12(+) genomic fragments were amplified in a similar manner, and the PCR products were injected at various concentrations (50 ng/μl, 5 ng/μl, 3 ng/μl, and 1 ng/μl) into flp-8;him-8 or flp-12;him-8 hermaphrodites, respectively. However, none of the 400+ F1 flp-8(+) or 100+ F1 flp-12(+) transgenic lines segregated the extrachromosomal array in the F2 generation, possibly because of toxic effects of the transgenes. Cotransformation marker coelomocyte::GFP (Miyabayashi et al., 1999) was injected at 50–100 ng/μl to achieve a total DNA concentration of ∼100 ng/μl. Male turning behavior was examined as described above.
TRN-specific rescue of flp-20(pk1596) mutants
Plasmid TL2137 was constructed by inserting flp-20(+) genomic coding sequence into Fire Lab Pmec-7 expression vector pPD52.102 using KpnI and EcoRV restriction sites. Plasmid TL2137 was microinjected into flp-20;him-8 hermaphrodite constructs at 35 ng/μl together with cotransformation marker coelomocyte::GFP (Miyabayashi et al., 1999) at 50 ng/μl. Male turning behavior was examined as described above.
Results
We used two strategies to study the role of neuropeptides in C. elegans male mating behaviors. First, we analyzed the behavior of mutants defective in flp genes (flp-3, flp-4, flp-6, flp-8, flp-10, flp-12, flp-19, flp-20, and flp-21), neuropeptide processing (egl-3 and egl-21), or neuropeptide secretion (ida-1 and unc-31). Second, we determined the cellular site of action of these neuropeptidergic components, which provided insight into the cellular and molecular basis of turning behavior.
The turning repetitiveness in wild-type populations
C. elegans male mating behavior includes the sequential steps of response to hermaphrodite contact, backing, turning, location of vulva, spicule insertion, and sperm transfer (Liu and Sternberg, 1995). Turning behavior can be dissected into three substeps: approaching the mate's body end while backing (a), initiating the turn with a small ventral tail arch (b), and completing the turn with a sharp ventral tail flexure and continued backing along the other side of the hermaphrodite (c) (Fig. 1A) (supplemental Movie 1, available at www.jneurosci.org as supplemental material).
The repetitiveness of male sexual turning behavior in wild-type populations. A, Turning behavior can be dissected into three substeps: approaching the mate's body end while backing (a), initiating the turn with a small ventral tail arch (b), and completing the turn with a sharp ventral tail flexure and continued backing along the other side of the hermaphrodite (c). Arrows indicate the typical behavioral sequence of a nonrepetitive turning (a-b-c). B, Plots of percentage of males against maximum number of trials per turn. The majority of wild-type males completed each turn in the first attempt before moving on to subsequent steps of mating (85% for him-5 and 88% for him-8). A low percentage of wild-type males required two or three trials before completing a turn (13% for him-5 and 10% for him-8). In rare instances, a wild-type male repeated three times for one turn (2% for both him-5 and him-8). The plot curves show that repetitive turns exist in the wild-type populations but at a low percentage of animals and with a low degree of repetitiveness. The number of animals observed per genotype is indicated in parentheses. Values shown are mean ± SEM.
The majority of wild-type males followed the turning behavioral sequence of “a-b-c” for each turn during mating (85% for him-5 and 88% for him-8) (Fig. 1B). A low percentage of wild-type males repeated a turn once or twice (13% for him-5 and 10% for him-8) (Fig. 1B). In this case, the male backed to the mate's body end and initiated a small ventral tail arch, but instead of proceeding to substep (c), he switched to moving forward along the mate's body. He traveled for one-eighth to one-fourth of the hermaphrodite body length before resuming backing and turning, exhibiting the behavioral sequence of “a-b-forwarding-a-b-c.” In rare instances, a wild-type male repeated a turn three times before turn execution and continued backing (2% for both him-5 and him-8) (Fig. 1B). We observed similar phenomena in heat-shock-generated N2 males (data not shown). Herein, we refer to two or more turning attempts as a repetitive turn. In summary, repetitive turns exist in wild-type populations but in a low percentage of animals and with a low degree of repetitiveness.
To quantify the turning repetitiveness for a given population, we developed a parameter, Tavg, which is the population average of maximum trials per turn. Tavg is 1.23 ± 0.06 for him-5 and 1.18 ± 0.04 for him-8 (no statistical difference between these two controls) (Fig. 2A). Because Tavg takes into account both the number of males exhibiting repetitive turnings and the maximum degree of repetitiveness exhibited by each male, Tavg can be used for intergenotypic comparisons of the repetitiveness of male sexual turning behaviors.
Disruption of flp signaling and neuropeptide release increases repetitive turning. A, Tavg values for flp-8, flp-10, flp-12, and flp-20 mutants, neuropeptide processing enzyme mutants egl-3(nr2090) and egl-21(n476), neuropeptide secretion mutants ida-1(ok409) and unc-31(e169), TRN genetically ablated mutants mec-4(e1611, d), and glutamatergic signaling mutants eat-4(ky5). All mutants are significantly different compared with wild type (p < 0.001). Tavg of flp quadruple mutants is significantly higher than that of flp single mutants (p < 0.01). B, Percentage of males exhibiting repetitive turning behavior in their first turn. All mutants are significantly different compared with wild type (p < 0.01 for eat-4 mutants, and p < 0.001 for other mutants). The number of animals observed per genotype is indicated in parentheses. Values shown are mean ± SEM.
FLP signaling pathways regulate male turning behavior
To determine whether C. elegans FMRF-amide like neuropeptides regulate male mating behavior, we analyzed the mating behavior of available flp mutant males (Table 1). flp-8(pk360), flp-10(pk367), flp-12(n4902), and flp-20(pk1596) single mutants exhibited a male repetitive turning phenotype, which is defined as Tavg significantly higher than the wild-type control (Fig. 2A). All other single flp mutant males successfully executed the mating behavioral steps of response, backing, turning, vulva location, and spicule insertion (Table 1).
flp gene expression patterna and mutant male mating behavior
The Tavg score is designed to evaluate the turning repetitiveness of a given population. However, one cannot easily deduce from Tavg how often males fail to initiate a turn at the first attempt or whether the repetitive turning phenotype is progressive (i.e., more severe with subsequent attempts). We therefore examined the percentage of males that exhibit repetitive turning behavior during their first turn. 10 ± 3% him-5 males and 5 ± 2% him-8 males failed to complete their first turn in one trial (Fig. 2B). There was no significant statistical difference between the two wild-type populations (p > 0.05), consistent with results using Tavg score.
Consistently, for flp-8, flp-10, flp-12, and flp-20 mutants, the percentages of males exhibiting repetitive turning behavior in their first turn were all significantly higher than corresponding wild-type controls (Fig. 2B).
Of the turning-defective mutants, flp-8 and flp-12 are expressed in a subset of male-specific neurons as well as neurons in the mechanosensory touch circuit (Table 1). To our surprise, flp-10 and flp-20 are not expressed in any male-specific neurons (Table 1). Rather, flp-10 is expressed in the PVR interneuron (White et al., 1986) and flp-20 in PVR and the touch receptor neurons AVM, ALM, PVM, and PLM (Chalfie et al., 1985), suggesting that the touch circuitry of the core nervous system contributes to turning behavior.
To establish a causal relationship between the flp mutation and the male turning phenotype, we performed transformation rescue experiments on flp mutants. The wild-type copy of flp-10 or flp-20 gene completely rescued the male repetitive turning phenotype of the corresponding mutant (Fig. 3), indicating that the abnormal male turning phenotype of flp-10 and flp-20 mutants is caused by a single gene mutation. We were unable to obtain stable transgenic lines expressing wild-type flp-8(+) or flp-12(+), perhaps because of toxic effects of the transgenes.
Transformation rescue of the flp-10 and flp-20 mutant repetitive turning phenotype. A, Distribution curves of “percentage of males” versus “maximum trials per turn” for him-5 wild-type control (WT), flp-10 mutant, and flp-10 rescue populations. The flp-10 mutant population has a plot curve significantly different from the WT population, whereas the flp-10 rescued population shows a plot curve similar to WT. Tavg of flp-10 mutants is 2.27 ± 0.22 (p < 0.001 compared with wild type), whereas Tavg is lowered to 1.31 ± 0.10 in flp-10 rescued males (p > 0.05 compared with wild type). B, Distribution curves of “percentage of males” versus “maximum trials per turn” for him-8 wild-type control, flp-20 mutant, and flp-20 rescue populations. The flp-20 mutant population has a plot curve significantly different from the WT population, whereas the flp-20 rescued population shows a plot curve similar to WT. Tavg of flp-20 mutants is 1.93 ± 0.12 (p < 0.001 compared with wild type), whereas Tavg is lowered to 1.23 ± 0.09 in flp-20 rescued males (p > 0.05 compared with wild type).
To examine genetic interactions, we constructed a flp-10; flp-12 flp-20 flp-8 quadruple mutant and analyzed the turning behavior. The male repetitive turning phenotype of the quadruple mutants was significantly more severe, with Tavg increasing from the 1.8–2.6 range in single mutants to 3.5 (Fig. 2A), indicating that these genes act in a partially redundant manner. Combined, these results demonstrate that multiple FLP peptidergic signaling pathways mediated by FLP-8, FLP-10, FLP-12, and FLP-20 function to regulate C. elegans male turning behavior.
The neuropeptide processing enzymes EGL-3 and EGL-21 and neuropeptide secretion regulators IDA-1 and UNC-31 are required for male turning behavior
Newly translated neuropeptide precursors are targeted to the endoplasmic reticulum-Golgi pathway for posttranslational modification, and the final mature products are stored in dense-core vesicles (DCVs) for secretion. Several genes involved in these processes have been identified and characterized in C. elegans, including the kex-2/subtilisin like proprotein convertase (PC) gene egl-3, the carboxypeptidase gene egl-21, the DCV cargo release regulatory gene ida-1, and the calcium-dependent activator protein for secretion gene unc-31 (Livingstone, 1991; Kass et al., 2001; Zahn et al., 2001; Jacob and Kaplan, 2003). Each neuropeptide processing and secretion component controls multiple behavioral programs. Although immunostaining and mass spectrometry studies have shown that egl-3 and egl-21 mutants have significantly reduced levels of mature FMRFamide-related peptides (Jacob and Kaplan, 2003; Husson et al., 2006), there have been no functional studies linking EGL-3, EGL-21, IDA-1, or UNC-31 to any of the C. elegans neuropeptides. Our newly identified male repetitive turning phenotype of the flp mutants provided an opportunity to genetically address the question.
egl-3, egl-21, ida-1, and unc-31 single mutants all displayed the repetitive turning phenotype exhibited by flp mutants. The Tavg scores of all four mutants are significantly higher than wild-type control (Fig. 2A), and the percentages of males exhibiting repetitive turning behavior in their first turn were also significantly higher than wild type (Fig. 2B). egl-3(nr2090) mutants have the most severe male repetitive turning phenotype (supplemental Movie 2, available at www.jneurosci.org as supplemental material). The Tavg for egl-3(nr2090) is 5.30 ± 0.31, and 78 ± 3% egl-3 males failed to complete their first turn in one trial (Fig. 2A,B). To confirm that the egl-3 repetitive turning phenotype is attributable to deletion of the egl-3 gene, we performed transformation rescues. Expression of wild-type egl-3(+) under its endogenous promoter rescued the male repetitive turning phenotype, and Tavg was decreased to 1.82 ± 0.15 (p > 0.05, compared with wild type) (Fig. 4A). Mutations in the other C. elegans PC genes bli-4, aex-5, and kpc-1 (Thacker and Rose, 2000) did not affect male turning behavior (data not shown), consistent with egl-3 acting as the PC that processes FLP-8, FLP-10, FLP-12, and FLP-20.
Effects of neuron-specific expression of egl-3 and flp-20 on the male repetitive turning phenotype of egl-3(nr2090) and flp-20(pk1596) mutants, respectively. A, Expression of wild-type egl-3 under its endogenous promoter rescued the repetitive turning phenotype, and Tavg is decreased to 1.82 ± 0.15 (p > 0.05, compared with wild type). Expression of wild-type egl-3 in all six TRNs using the mec-7 promoter partially but significantly rescued the egl-3 mutant repetitive turning phenotype. One representative rescued line is shown (Tavg, 2.17 ± 0.20; p < 0.001, compared with egl-3). Expression of wild-type egl-3 exclusively in PVM using the flp-8 promoter partially rescued the repetitive turning phenotype, although not as well as egl-3 TRN rescue. One representative rescued line is shown (Tavg, 3.23 ± 0.30; p < 0.001, compared with egl-3 mutants). Expressing egl-3(+) in command interneurons using the glr-1 promoter did not have any rescuing effect on the male repetitive turning phenotype (Tavg, 4.96 ± 0.51; p > 0.05 compared with the egl-3 mutant). B, Expression of wild-type flp-20 in TRNs using the mec-7 promoter rescued the flp-20 mutant repetitive turning phenotype. One representative rescued line is shown (Tavg, 1.24 ± 0.08; p > 0.05, compared with wild type). The behavioral data for egl-3 and flp-20 mutants are the same as shown in Figure 2A. The number of animals observed per genotype is indicated in parentheses. Values shown are mean ± SEM. WT, Wild-type control him-8(e1489).
EGL-3 acts in touch receptor neurons to regulate male turning behavior
To determine the egl-3 site of action and gain insight into the cellular basis of turning behavior, we performed tissue-specific rescue of the egl-3 male repetitive turning phenotype. egl-3 is expressed in male-specific ray neurons (data not shown) and some interneurons and sensory neurons of the shared nervous system, including the TRNs (Kass et al., 2001). Two of the flp genes (flp-10 and flp-20) with mutants exhibiting repetitive turning are not expressed in ray neurons but rather the touch circuitry. These results led us to hypothesize that neuropeptides may function in the shared nervous system to regulate turning.
To test this hypothesis, we used different cell-specific promoters to express wild-type egl-3(+) in select neurons. egl-3 is broadly expressed in the shared nervous system, and anti-egl-3 PC2 antibody staining identified a subset of egl-3-expressing neurons, which includes the polymodal ASH sensory neurons that mediate nose touch response, the TRNs ALM and AVM that mediate light body touch response, another TRN PVM with no assigned function in hermaphrodite, and the command interneurons AVB, PVC, and AVD that regulate forward and reverse locomotion (Kass et al., 2001). Although there has been no report of egl-3 expression in PLM touch receptor neurons, egl-3 might indeed be expressed in these TRNs. Additional experimentation is needed to clarify this ambiguity. We found that expression of wild-type egl-3 in all six TRNs using the mec-7 promoter (Hamelin et al., 1992) partially but significantly rescued egl-3 mutant repetitive turning phenotype. Tavg values were lowered from 5.30 ± 0.31 in egl-3 mutants to 2.17 ± 0.20 and 2.19 ± 0.19 in two rescued lines, respectively (p < 0.001 compared with egl-3) (Fig. 4A). The partial rescue may reflect a requirement of egl-3 function in other cell types, such as ray sensory neurons.
flp-8 is expressed in PVM neuron, and loss of flp-8 induces a repetitive turning phenotype in males (Fig. 2, Table 1). These observations strongly suggest that PVM, which has no known function in hermaphrodites (Chalfie et al., 1985), acquires a mechanosensory function in males. We tested this hypothesis by restoring wild-type egl-3 expression exclusively in PVM using the flp-8 promoter. Two transgenic lines were partially rescued for the repetitive turning phenotype, with Tavg lowered to 3.23 ± 0.30 and 3.94 ± 0.50, respectively (p < 0.001, compared with egl-3 mutants) (Fig. 4A). These results indicate that the PVM neuron contributes to male turning behavior in C. elegans.
Consistent with a role for peptidergic transmission in the TRNs, restoring flp-20(+) expression in TRNs using the mec-7 promoter completely rescued the male repetitive turning defect of flp-20 mutants. Tavg values were lowered from 1.93 ± 0.12 in flp-20 mutants to 1.24 ± 0.08 and 1.26 ± 0.10 in two rescued lines, respectively (p > 0.05, compared with wild type) (Fig. 4B). We conclude that egl-3-regulated peptidergic signaling in TRNs is required to mediate male turning. In contrast, expressing egl-3(+) in command interneurons with the glr-1 promoter (Brockie et al., 2001) does not have any rescuing effect on the male repetitive turning phenotype (p > 0.05 compared with egl-3) (Fig. 4A). Because command interneurons postsynaptically couple TRNs to the locomotion circuits, we conclude that the egl-3 repetitive turning phenotype reflects a defect in the TRNs rather than involving more downstream components.
The TRNs and MEC-4/MEC-10 DEG/ENaC channel inhibit repetitive turning behavior
To confirm the role of TRNs in mediating male turning, we examined turning behavior of males lacking these cells. The mec-4(e1611) gain-of-function mutation, or mec-4(d), causes constitutive activation of MEC-4/MEC-10 DEG/ENaC channel and TRN degeneration (Driscoll and Chalfie, 1991; Goodman et al., 2002). mec-4(d) mutants exhibited the male repetitive turning phenotype with Tavg 2.09 ± 0.11 (p < 0.001 compared with wild type) (Table 2), indicating that TRNs provide sensory input to the neural circuit regulating male turning behavior.
Turning behavior of TRN mechanosensory and neurotransmitter signaling mutants
To determine whether the MEC-4/MEC-10 mechanosensory complex plays a role in sexually dimorphic mechanosensation, we analyzed the male turning behavior of mec-4(e1339) reduction-of-function (rf) and mec-10(e1515) loss-of-function (lf) mutants (Chalfie and Sulston, 1981). e1339 was selected as the mec-4 reference allele and is not as strong as other loss-of-function mutations. Both mec-4(rf) and mec-10(lf) mutants displayed the male repetitive turning phenotype, with Tavg 1.65 ± 0.07 and 2.01 ± 0.12, respectively (both significantly different from wild type) (Table 2). The requirement for MEC-4/MEC-10 channels in male turning suggests the mechanosensitive nature of this sexually dimorphic behavior.
The repetitive turning defect of mec-4(d) males may be a direct consequence of no TRNs. Alternatively, the repetitive turning phenotype could reflect a more general problem with coordinated locomotion because of an altered balance in the activity of the AVB/PVC and AVA/AVD/AVE command interneurons. Two lines of evidence argue against the latter. First, mec-4(d) males respond normally to hermaphrodite contact. In contrast, unc-47 GABAergic mutant males, which have altered command interneuron signaling and a backward locomotion defect, respond poorly to hermaphrodite contact but do not exhibit repetitive turning (Table 2). Second, mec-4(d) males exhibit the backing–forwarding switch specifically after the turning initiation substep but not randomly as would be predicted for command interneuron signaling defect. We conclude that the TRNs provide sensory input to the male turning neural circuit via the MEC-4/MEC-10 mechanosensory receptors.
Glutamatergic neurotransmission is required for male turning behavior
C. elegans TRNs sense mechanical stimuli and provide input to the command interneurons via glutamatergic synaptic connections (Maricq et al., 1995; Brockie et al., 2001). To determine whether glutamate influences turning behavior, we assayed the eat-4(ky5) mutant. eat-4 encodes a sodium-dependent inorganic phosphate cotransporter required for glutamatergic neurotransmission (Lee et al., 1999). eat-4 mutants are defective in many behaviors, including chemotaxis, feeding, thermotaxis, and foraging. Interestingly, eat-4 mutants are also defective in responses to gentle touch mediated by the TRNs (Lee et al., 1999), which is the apparent site of action for egl-3 and several flp genes. We found that eat-4(ky5) mutants displayed the male repetitive turning phenotype (p < 0.001, compared with wild type) with a phenotypic severity similar to TRN genetically ablated mec-4(d) mutants (p > 0.05) (Table 2). The percentage of eat-4 mutant males exhibiting repetitive turning behavior in their first turn was also significantly higher than wild type (p < 0.01) (Fig. 2B). We conclude that glutamatergic signaling is required for male turning behavior.
We extended the survey of male turning behavior to include mutants defective in the classical neurotransmitters dopamine, serotonin, tyramine, octopamine, and GABA. Consistent with published reports, we found that 52% of the dopamine- and serotonin-deficient bas-1(ad446) mutant males exhibited a late-turn phenotype, and 43% of the dopamine-deficient cat-2(e1112) mutant males had a wide-turn phenotype (Table 2). A late turn is defined as when the male backs off the mate's body end and fails to initiate turning in a timely manner, whereas a wide turn is defined as when the male is unable to execute the sharp ventral tail coil and loses contact with mate's body after the turn (Loer and Kenyon, 1993). Tyramine- and octopamine-deficient tdc-1(n3420) mutants (Alkema et al., 2005) and vesicular GABA-transporting defective unc-47(e307) mutants (Eastman et al., 1999) were phenotypically wild type for male turning behavior (Table 2). However, none of the mutants displayed the repetitive turning phenotype (Table 2). These results indicate that dopamine, serotonin, tyramine, octopamine, and GABA do not individually contribute to turning repetitiveness.
TRNs mediate male sexual turning independently of ray neurons
Ray sensory neurons are required for the timing element of male turning. Males with the three most posterior pairs of rays ablated frequently swim off hermaphrodites instead of making turns in a timely manner (the late-turn phenotype) (Liu and Sternberg, 1995). Our finding that the TRNs inhibit the repetitiveness of male turning suggests that the mechanosensory input from TRNs is also an important regulatory factor for male turning (Fig. 5A). How do these two groups of sensory neurons function in the male turning circuitry? We propose two models (Fig. 5B). In the first “parallel” model, TRNs and ray sensory neurons independently sense hermaphrodite turning cues to coordinate the activities of their downstream neurons. In the second “linear” model, TRNs sense the change of the male's body posture resulting from tail curling (i.e., the motor output downstream of ray neurons acts as a proprioceptory signal for TRNs).
Neuronal circuitry of male turning behavior. A, Sensory neurons that regulate male turning behavior. An adult C. elegans male has six TRNs: ALML/R, AVM, PLML/R, and PVM (indicated in yellow). Each male also has nine pairs of bilaterally symmetric rays (indicated in green), each of which possesses two sensory neurons. Both groups of sensory neurons are required for male turning. This figure is modified from the study by Kaplan and Horvitz (1993). B, Models for male turning circuitry. In the parallel model (indicated by red arrows), TRNs and ray sensory neurons independently sense hermaphrodite turning cues to coordinate the activities of their downstream neurons. In the linear model (indicated by black arrows), TRNs sense the change of the male's body posture as a result of tail curling. Our data support the parallel model, in which two distinct sensorimotor circuits, the touch circuitry (left column) and the ray sensorimotor circuitry (right column), function in parallel to regulate male sexual turning. C, mec-4(d) males exhibit the repetitive turning phenotype (Tavg significantly higher than wild-type animals; p < 0.001) but not the late-turn phenotype. bas-1 males exhibit the late-turn phenotype but not the repetitive turning phenotype. In the mec-4(d);bas-1 double mutants, both turning phenotypes are intact. The black bar indicates Tavg, and the white bar indicates percentage of males exhibiting late turns. The number of animals observed per genotype is indicated in parentheses. Values shown are mean ± SEM. WT, Wild-type control him-5(e1490).
To distinguish between the two models, we examined the male turning behavior of mec-4(d);bas-1 double mutants. mec-4(d) single mutant males exhibited the repetitive turning phenotype but not the late-turn phenotype (Fig. 5C). bas-1 single mutant males presented the late-turn phenotype but not the repetitive turning phenotype (Fig. 5C). We separately calculated Tavg and the percentage of males exhibiting late turns for the mec-4(d);bas-1 double mutants and found both phenotypes were intact (Fig. 5C), supporting the parallel model. We conclude that TRNs function independently of ray neurons to mediate male sexual turning behavior.
Discussion
Neuropeptide regulation of male turning behavior
Neuropeptides are important neural regulators of many complex animal behaviors such as sleep, feeding, nociception, reproduction, and social behaviors (Storm and Tecott, 2005; Borgland et al., 2006; Sabatier, 2006; Saper, 2006; Hill and Oliver, 2007). FMRFamide-like neuropeptides constitute a major family of C. elegans neuropeptides (Li et al., 1999; Kim and Li, 2004). Although the neuronal expression patterns of the flp gene family have been characterized, limited evidence links neuropeptides to the regulation of genetically programmed behaviors. FLP-18 and FLP-21 have been identified as ligands differentially activating two variants of the neuropeptide Y receptor NPR-1 in the regulation of social feeding behavior (Rogers et al., 2003). Exogenously applied FLP-11A and FLP-13A inhibit pharyngeal muscle activity, whereas FLP-8, FLP-17A, and FLP-17B have excitatory effects (Papaioannou et al., 2005). Our finding that mutations in flp-8, flp-10, flp-12, and flp-20 cause a male turning defect implicates FLPs in regulating the most complex form of C. elegans behaviors: the male copulation ritual. In support of this finding, proprotein convertase mutant egl-3, carboxypeptidase mutant egl-21, and DCV secretion regulatory mutants ida-1 and unc-31 exhibit the same male turning defect. The repetitive turning defect of egl-3 mutants is much more severe than any of the flp single mutant or quadruple mutants, suggesting there are other unidentified neuropeptides that regulate male turning behavior. Combined, these results indicate that multiple FLP peptidergic signaling pathways regulate C. elegans male turning behavior.
The relationship between peptidergic and classical neurotransmitter signaling is an interesting topic. In the mammalian brain, the neuropeptide orexin A contributes to learning and addiction via modulation of glutamatergic excitability in dopaminergic neurons (Borgland et al., 2006). With respect to C. elegans male turning behavior and classical neurotransmitter signaling, we found that only the eat-4 glutamate signaling mutant exhibits the repetitive turning phenotype, suggesting that FLP and glutamate signaling coregulate male turning behavior. mec-4(e1611) TRN-ablated males have a Tavg value similar to that of flp-20 and eat-4 mutants (p > 0.05), hinting that FLP-20 and glutamate act in the same pathway to regulate TRN-command interneuron synaptic transmission. Dopaminergic and serotonergic signaling are also involved in turning behavior (Loer and Kenyon, 1993; Carnell et al., 2005) but in a circuitry and genetic pathway distinct from FMRFamide-like peptidergic signaling. Together, these results suggest that an intricate signaling network consisting of FLPs, dopamine, serotonin, and glutamate controls male sexual turning behavior.
C. elegans male sexual behavior and the turning step of male copulation
The innate sexual behavior of the C. elegans male involves a complex series of stereotyped steps that are evoked by multimodal sensory information provided by his mating partner. First, a long-range pheromone acts to attract males (Chasnov et al., 2007). Next, a short-range and short-lived hermaphrodite signal triggers male searching behavior (Lipton et al., 2004). A distinct, diffusible hermaphrodite cue influences male locomotion (Simon and Sternberg, 2002). Once in close proximity to a potential mate, the male initiates the copulatory behavioral program that includes the steps of response to hermaphrodite contact, backing, turning, location of vulva, spicule insertion, and sperm transfer. Each step is genetically separable and controlled by its own neural circuitry (Loer and Kenyon, 1993; Liu and Sternberg, 1995; Barr and Sternberg, 1999; Garcia et al., 2001; Schindelman et al., 2006). Each step likely involves distinct sensory cues from the mate. For example, the response and vulva location steps may involve chemical and mechanical signals from the hermaphrodite's body (J. Wang, M.M.B., and P. W. Sternberg, unpublished observation). The sperm transfer step requires the hermaphrodite's uterus (Emmons and Sternberg, 1997). The identity of the turning cue is not known but likely mechanosensory in nature, based on our findings that the TRNs regulate turning repetitiveness.
Male turning behavior involves spatial and temporal regulation to execute a sharp ventral tail coil (Loer and Kenyon, 1993; Liu and Sternberg, 1995). Herein, careful observation of turning behavior in wild-type males (n = 278) reveals a third criterion (i.e., turning repetitiveness) as gauged by Tavg (population average of maximum trials per turn exhibited by individual male). We propose that a low degree of turning repetitiveness (Tavg, ∼1.18–1.23 in wild-type males) may serve to ensure accurate timing for turning. The practice of increasing redundancy to enhance accuracy is used in several systems, including the control of behavioral performance by the rat motor cortex and in diagnostic reasoning by the human brain (Bloch et al., 2003; Narayanan et al., 2005). When a C. elegans male is engaged in mating behavior, his decision to turn is constantly challenged, particularly in the natural environment with a free-moving partner. A mechanism for fine-tuning and adjusting turning behavior is essential for successful copulation. In contrast, excessive repetitiveness of turning (the repetitive turning phenotype, as defined by Tavg value significantly higher than wild-type males) is likely to lower behavioral efficiency and hence put the male at a reproductive disadvantage.
We cannot rule out the possibility that repetitive turning phenotype is a consequence of a failure in the turning program. However, we think the latter possibility is unlikely based on three lines of evidence. First, repetitive turning behavior exists in the wild-type population. Second, bas-1(ad446) mutant males are defective in turning timing but do not have the repetitive turning phenotype. Third, both wild-type (Loer and Kenyon, 1993) and egl-3(nr2090) mutant males (T.L. and M.M.B., unpublished observation) produce sharp ventral tail coils when soaked in exogenous serotonin. Therefore, we favor the theory that repetitive turning is a normal behavior, and the repetitive turning phenotype results from an abnormal expansion of a wild-type behavior.
The neuronal regulation of male sexual turning behavior
The complexity of male turning behavior is reflected in its underlying neural circuitry. Several male-specific nervous system components have been identified previously as key components of the circuitry. At the sensory input level, the serotonergic ray neurons signal the male when to turn (Sulston and Horvitz, 1977; Loer and Kenyon, 1993; Liu and Sternberg, 1995). At the motor output level, CP motor neurons and diagonal muscles are required for the execution of ventral tail flex (Loer and Kenyon, 1993). Several turning-defective mutants further support the roles of these male-specific structures in mediating male turning. For instance, MSI-1, a neural RNA-binding protein involved in posttranscriptional gene regulation, is expressed in ray sensory neurons and CP neurons, and msi-1 mutant males have the wide-turn and late-turn phenotypes (Yoda et al., 2000). MAB-23 is a DM domain transcription factor necessary for the differentiation of sex-specific tissues, and mab-23 males are unable to adopt the appropriate turning posture because of a diagonal muscle differentiation defect (Lints and Emmons, 2002). The IP3 receptor gene itr-1 is expressed in CP8/9, and diagonal muscles and itr-1 mutant males also exhibit the wide-turn and late-turn phenotypes (Gower et al., 2005). Hence, the “ray sensorimotor circuit” is one element required for copulatory turning (Fig. 5B, right column).
Here, we demonstrate that the touch circuitry of the core nervous system (Chalfie et al., 1985; White et al., 1986) acts as a second element required for proper turning. First, we observed a repetitive turning phenotype in flp-10 and flp-20 mutants, yet flp-10 and flp-20 are not expressed in any male-specific neurons or muscles. We also showed that egl-3 neuropeptide processing enzyme mutant males exhibit the same repetitive turning defect. Additionally, restoring egl-3(+) expression in all six TRNs partially but significantly rescues the egl-3 mutant turning phenotype. In addition, flp-20 TRN-specific expression completely rescues the flp-20 mutant turning phenotype. Finally, we found that genetic ablation of TRNs in mec-4(d) animals or genetic disruption of the MEC-4/MEC-10 DEG/ENaC mechanosensory channel produces the male repetitive turning phenotype. These observations support the conclusion that mechanosensory TRNs of the shared nervous system are required to mediate the male sexual turning behavior. Also interesting is the fact that PVM-specific expression of egl-3(+) partially rescues the egl-3 mutant turning defect. These data provide the first genetic evidence that PVM neuron functions in the male mating behavioral program and exemplifies the functional dimorphism of the C. elegans nervous system.
Our model of male turning involves two distinct sensorimotor circuits acting in concert to control a behavioral outcome. We propose that the touch circuit coordinates backward movement with the tail turning event, which is controlled by the ray sensorimotor circuit (Fig. 5B). Cross talk between both circuits can be readily embraced into the current model. How the core and sex-specific nervous systems communicate and how gender-specific modifications of the shared nervous systems contribute to copulatory behaviors are not known. Reconstruction of the male C. elegans nervous system will enable dissection of the neural circuitries that generate complex sexual behaviors.
Footnotes
-
This work was supported by the American Heart Association (M.M.B.) and the National Institutes of Health (M.M.B., C.L.). We are grateful to Anne Hart for the suggestion to examine egl-3 mutant males; Joshua Kaplan, Tobias Zahn, John Hutton, and Mark Alkema for strains and reporters; Andrew Fire for vectors; Ronald Plasterk and Bob Horvitz for allowing us to screen their C. elegans deletion libraries; the Caenorhabditis Genetics Center and the Sanger Institute for other numerous reagents; Douglas Braun for excellent technical assistance; and Rene Garcia, Miriam Goodman, and Paul Sternberg for helpful discussions and comments on this manuscript.
- Correspondence should be addressed to Maureen M. Barr at her present address: Department of Genetics, Rutgers University, Piscataway, NJ 08854. barr{at}biology.rutgers.edu
References
