Variations in K+ channel composition allow for differences in cell excitability and, at an organismal level, provide flexibility to behavioral regulation. When the function of a K+ channel is disrupted, the remaining K+ channels might incompletely compensate, manifesting as abnormal organismal behavior. In this study, we explored how different K+ channels interact to regulate the neuromuscular circuitry used by Caenorhabditis elegans males to protract their copulatory spicules from their tail and insert them into the hermaphrodite’s vulva during mating. We determined that the big current K+ channel (BK)/SLO-1 genetically interacts with ether-a-go-go (EAG)/EGL-2 and EAG-related gene/UNC-103 K+ channels to control spicule protraction. Through rescue experiments, we show that specific slo-1 isoforms affect spicule protraction. Gene expression studies show that slo-1 and egl-2 expression can be upregulated in a calcium/calmodulin-dependent protein kinase II-dependent manner to compensate for the loss of unc-103 and conversely, unc-103 can partially compensate for the loss of SLO-1 function. In conclusion, an interaction between BK and EAG family K+ channels produces the muscle excitability levels that regulate the timing of spicule protraction and the success of male mating behavior.
THE neuronal and muscular networks that execute behavioral responses to stimuli are ultimately maintained through cross talk by the various molecular signaling components. This signaling is essential for fine-tuning cell excitability levels to generate appropriate responses and is partially accomplished by a variety of K+ channels. Organisms have a large number of K+ channels that have multiple isoforms and ancillary subunits (Salkoff et al. 2006; Torres et al. 2007; Fodor and Aldrich 2009). In excitable cells, a few K+ channel types are responsible for the maintenance of depolarization; the rest are proposed to be modifiers, fine-tuning the particular cell or cell type for its specific function (Bargmann 1998; Santi et al. 2003).
The importance of these modifiers is highlighted by the multitude of human disorders that exist as a result of defective K+ channels. These disorders include epilepsy, which can be caused by dysfunction of a variety of channels (Schmitt et al. 2000; Du et al. 2005), long Q-T syndrome, one form of which can be caused by mutations in the HERG gene and leads to cardiac arrhythmias and sudden death (Sanguinetti et al. 1995), hypertension caused by kidney malfunction (Grimm and Sansom 2010), and erectile dysfunction (Werner et al. 2005).
The K+ channels, which fine-tune cell excitability, need to work together. When there is a defective or absent channel, the remaining K+ channels might compensate in an incomplete manner. However, the structure and function of the compensating channels likely differ from the conventional one, resulting in imperfect regulation of cell excitability. Determining how defective K+ channels change the dynamics of the remaining K+ channels will provide insights into how the integration of their functions maintains appropriate excitability levels.
To address how the regulation of cell excitability and behavior is coordinated by fine-tuning K+ channels, we used Caenorhabditis elegans male mating as a model system. The stereotyped steps of this behavior include contact response, backward locomotion and vulval scanning, turning, vulva sensing, repetitive spicule insertion attempts, complete spicule penetration, and sperm transfer. Males utilize a pair of copulatory spicules to breach the hermaphrodite vulva. The spicules must be maintained inside the male tail prior to mating and during mating until vulva penetration (Liu and Sternberg 1995). We have previously identified two K+ channels, ether-a-go-go (EAG) and EAG-related gene (ERG), that inhibit protraction until the vulva has been breached (Garcia and Sternberg 2003; LeBoeuf et al. 2007). However, these two K+ channels do not account for all the regulation necessary to coordinate spicule protraction. In this article we identify the big current (BK) SLO1 K+ channel as an additional regulator of male sex muscle excitability. Since these three K+ channels function in the mating circuit to regulate behavior, we explored their effects on each other and elucidated how their interplay maintains the cell excitability that allows a male to successfully sire progeny.
Materials and Methods
Strains and culture methods
The following strains were used in this study: unc-103(n1213) (Park and Horvitz 1986), unc-103(sy557) (Garcia and Sternberg 2003), and pha-1(e2123) (Schnabel and Schnabel 1990) on LGIII; unc-43(sy574) (LeBoeuf et al. 2007) on LGIV; him-5(e1490) (Hodgkin et al. 1979), slo-1(js379) (Wang et al. 2001), slo-1(rg432) (this work), egl-2(rg4) (LeBoeuf et al. 2007), and egl-2(n693) (Reiner et al. 1995) on LGV. C. elegans is a naturally hermaphroditic species with a low incidence of males; to generate a larger percentage of males, strains in this study contained the him-5(e1490) allele. Animals were maintained on NGM plates seeded with Escherichia coli strain OP50 at 20° (Brenner 1974).
Protraction constitutive assay
A total of 20–30 virgin L4 males were isolated on NGM plates with OP50. The males were allowed to develop into adults overnight and scored as positive for spicule protraction if at least one spicule partially extended from the cloaca. P values were determined using GraphPad Prism (version 4.03).
Identification of the slo-1(rg432) mutation
We sequenced all of slo-1, including the promoter region, from the genomic DNA of an unc-103(n1213); egl-2(rg4) strain with a lower-than-normal instance of the protraction constitutive (Prc) phenotype. We discovered that rg432 changes the sequence TGACATTTATATTATCATTTT to TGACATTTATTTTATCATTTT in the intronic region between exons 11 and 12 (exons numbered according to Wang et al. 2001).
pBK1 contains the slo-1 promoter and isoform slo-1(A2;B0;C1;D0) tagged with GFP and was provided by Michael Nonet (Washington University, St. Louis). To express slo-1(A2;B0;C1;D0):GFP in specific tissues, we removed the slo-1 promoter from pBK1 and added the XbaI restriction site in front of the cDNA via single-site mutagenesis with primers FpBK1XbaI and pBK1r to create plasmid pBL161. pBL161 was then cut with XbaI, blunt-ended, and Gateway vector conversion reading frame cassette (RfC) A (Invitrogen, Carlsbad, CA) was ligated into the site to create plasmid pBL176. The tissue-specific constructs were created using LR clonase (Invitrogen) to recombine the pBL176 destination vector with the vectors containing the tissue-specific promoters. pBL176 was recombined with pLR35 (Paex-3, pan-neuronal) (LeBoeuf et al. 2007), pLR21 (Punc-103E, sex muscles) (Reiner et al. 2006), and pLR92 (Pacr-8, body-wall muscles) (LeBoeuf et al. 2007) to create plasmids pBL179, pBL180, and pBL181, respectively.
We cloned slo-1(A2;B0;C0;D0) into plasmid pSX322YFP to generate pBL185cc. slo-1(A2;B0;C0;D0) was cloned from PCR-amplified cDNA using primers Fslo1cDNASphI and Rslo1cDNAXbaI. The cDNA was generated from poly(A) RNA using Qiagen’s (Valencia, CA) Oligotex mRNA Mini kit and Invitrogen’s SuperScript First-Strand Synthesis system for RT–PCR. Both the plasmid and PCR product were cut with SphI and XbaI and ligated together to generate pBL185cc. In this manner, we cloned nine different slo-1 full-length cDNA isoforms, and we discovered a new exon located between exons 2 and 3. We cut pBL185cc with NheI and ligated Gateway RfC A (Invitrogen) in front of slo-1(A2;B0;C0;D0) to create plasmid pBL192. pBL192 contained point mutations in the slo-1 cDNA that changed amino acids. To correct the slo-1 sequence, we cut pBL192 and a plasmid containing slo-1(A2;B0;C0;D0) provided by Brandon Johnson and Miriam Goodman (Stanford University, Palo Alto, CA) with BseRI. The short fragment from digested slo-1(A2;B0;C0;D0) was ligated to the long fragment from digested pBL192 to create pBL224. We recombined a PCR fragment of the slo-1 promoter generated with primers Fslo1p and Slo1pr from N2 genomic DNA with pDG15 using BP clonase (Invitrogen) to generate plasmid pBL153 (the same promoter region reported in Wang et al. 2001). We performed an LR reaction between pBL153 and pBL224 to generate plasmid pBL226.
A plasmid containing a Gateway RfC, slo-1(A1;B0;C0;D0), and GFP was created by cutting pBL176 and a plasmid containing slo-1(A1;B0;C0;D0) provided by Miriam Goodman with StuI and HpaI. The long restriction enzyme product from cutting pBL176 was ligated to the short restriction enzyme product from slo-1(A1;B0;C0;D0) to generate pBL198. pBL198 was recombined with pLR21 and pBL153 via LR clonase to create plasmids pBL204 and pBL198, respectively.
Plasmids containing mutated unc-103 genomic DNA were generated using single-site mutagenesis on plasmid pLR67, which contains unc-103 genomic DNA plus Gateway RfC C.1 (Invitrogen) (Reiner et al. 2006). The primers psy557A and phosrevsy557A were used to create pLR74 containing the sy557A mutation, and the primers psy557B and phosrevsy557B were used to create pLR75 containing the sy557B mutation. pLR74 was recombined with pLR21 and pLR28 (Punc-103F, sex neurons) (Reiner et al. 2006) using LR clonase to create plasmids pBL35 and pBL36, respectively. pLR75 was recombined with pLR21 and pLR28 using LR clonase to create plasmids pBL34 and pBL37, respectively.
pTG44 containing the unc-103E promoter driving egl-2 cDNA was constructed as previously described (LeBoeuf et al. 2007). We added the n693gf mutation by performing single-site mutagenesis on pTG44 with primers fegl2n698gf and regl2698gf to create plasmid pBL111 (LeBoeuf et al. 2011).
pBL160 was created via an LR reaction between pBL153 (Pslo-1) and pLR186 (Invitrogen Gateway RfC C.1:DsRed1-E5) (LeBoeuf et al. 2011).
Plasmids containing the construct of interest were injected into 1-day-old adult hermaphrodites of the appropriate strain following standard procedures (Mello et al. 1991). Either 50 ng/μl pBL66 or 100 ng/μl pBX1 were used as a transgenic marker (Granato et al. 1994; LeBoeuf et al. 2007). pUC18 was used as carrier DNA to complete each injection mixture to a final concentration of 200 ng/μl. For each plasmid injected, concentrations were as follows: 20 ng/μl was used for pBL34, pBL35, pBL36, pBL37, pBK1, pBL180, pBL179, pBL181, and pBL199 and 50 ng/μl was used for pBL111, pBL160, pBL226, and pBL204. Once stable transgenic lines were obtained, males from two or more independent transgenic lines were scored for the Prc phenotype following the procedure previously described.
To obtain a stable expression line to measure fluorescent levels in males carrying Pslo-1:DsRed1-E5, the transgene was integrated using trimethylpsoralen following standard procedures (Anderson 1995) to create transgenic line rgIs2.
Image acquisition and quantification
Males carrying the rgIs2 transgene were isolated at mid-L4. Images were either taken immediately, or the males were allowed to develop into adults on NGM plates containing E. coli OP50 for 1 or 2 days. Males were immobilized on 10% Noble agar (dissolved in water) pads with Polybead polystyrene 0.1-μm microspheres (Polysciences, Warrington, PA) and covered with a cover slip. Images were taken using an Olympus BX51 microscope with a ×40 objective. The image signal was split into two channels using the Dual View Simultaneous Imaging system by Photometrics (Tucson, AR). Pictures were taken by the Hamamatsu EM-CCD digital camera ImagEM using HC Image (version 184.108.40.206) (Bridgewater, NJ). All images were taken with the gain, sensitivity, and exposure time held constant. The one exception is at the 2-day time point, when the red channel was overexposed at the standard settings. One image was taken at the standard settings and a second image was taken when the gain was lowered so the red channel was no longer overexposed. Data from both images were used to obtain the mean gray level for the red channel (see below).
Images were analyzed using Hamamatsu Simple PCI (version 220.127.116.11). A region of interest (ROI) was placed over the male tail in both the red and green channels and in the background of both the red and green channels to get the background fluorescent level. The ROI was held to a constant size using the “ROI clone” tool. The mean gray level was recorded for each ROI, and the background fluorescence was subtracted. For 2 day-old adult males, the mean gray level was determined for the green channel under standard settings as well as both the green and red channel with the gain lowered. Since the green channel was not overexposed in either image, a ratio between the high gain and low gain image was obtained. This ratio was then combined with the low gain red mean gray level to approximate the standard gain red mean gray level. The analyzed images were not modified in any way. Statistical analysis was performed using GraphPad Prism (version 4.03).
Quantitative real-time PCR
Each strain of worms was cultivated on six 100-mm NGM plates containing E. coli OP50. Worms were harvested before they starved by washing the plates with M9 buffer. Worms were collected at 1000 rpm and the M9 and E. coli were removed. The worms were washed with an additional 1 ml of M9 to remove additional E. coli. For experiments where hermaphrodites were segregated from males, worms were first placed on a 20-μM Nitex nylon filter to remove larva, and then on a 35-μM Nitex nylon filter to separate males and hermaphrodites (Sefar Filtration, Depew, NY). Worms were divided into four 1.5-ml microcentrifuge tubes and 250 μl Tri Reagent (Sigma-Aldrich, St. Louis) and 0.5 mm zirconium oxide beads (Next Advance, Cambridge, MA) were added. The Bullet Blender (Next Advance) set for 3 min at speed 8 was used to break open the worms. The debris, Tri Reagent, and beads were spun down for 1 min at top speed in a microcentrifuge. The Tri Reagent and RNA were moved to a fresh RNase-free 0.5-microcentrifuge tube. Ten microliters of glycogen and 50 μl of chloroform were added, and the mixture was shaken by hand and centrifuged for 10 min at 12,000 rpm to separate the layers. The aqueous layer was transferred to a fresh 0.5-RNase-free microcentrifuge tube. Two ethanol precipitations were performed to remove any salts that might interfere with the ability to separate poly(A) RNA from total RNA. The quality of the total RNA was determined using a 2% agarose gel and the quantity was measured using a spectrophotometer. Poly(A) RNA was obtained using Ambion’s Poly(A) Purist kit (Applied Biosystems, Austin, TX). cDNA was made using the SuperScript First-Strand Synthesis system for RT–PCR (Invitrogen) with 200 ng of starting poly(A) RNA from each sample and replicated at least one time. Reference genes were determined as done in Vandesompele et al. (2002) and Hoogewijs et al. (2008). Thirteen reference genes were considered: act-1, gpd-3, F23B2.13, rrn-1.1, cdc-42, Y45F10D4F.3, ama-1, csq-1, eif-3.C, mdh-1, gpd-2, pmp-3, and tba-1. geNorm was used to identify the most stable reference genes (http://medgen.ugent.be/~jvdesomp/genorm/). The two most stable reference genes for comparison across K+ channel mutants were act-1 and mdh-1. Amplification was performed using Bio-Rad’s ssoFast EvaGreen on a Bio-Rad CFX96 Real-Time Detection system (Hercules, CA). act-1 and egl-2 were amplified using the primers reported in LeBoeuf et al. (2011). mdh-1, cdc-42, and pmp-3 were amplified using the primers reported in Hoogewijs et al. (2008). unc-103 was amplified using primers 2qPCRunc-103F and 2qPCRunc-103R. slo-1 was amplified using primers F1-qPCRslo1ex11 and qPCRslo1ex12-1R. The primers for slo-1 A1 were Fslo1a1ex9 and Slo1a1ex9r. The primers for slo-1 A2 were Fslo1a2ex10 and Slo1a2ex10r. The primers for slo-1 B0 were Fslo1b0 and Slo1b0ex14r. The primers for slo-1 B1 were Fslo1b1 and Slo1b1r. The primers for slo-1 C0 were Fslo1c0 and Slo1c0r. The primers for slo-1 C1 were Fslo1c1 and Slo1c0r. The primers for slo-1 D0 were Fslo1d0 and Slo1d0r. The primers for slo-1 D1 were Fslo1d0 and Slo1d1r2. The primers for acr-18 were qPCRacr-18F and qPCRacr-18R. The primers for gar-3 were 2qPCRgar-3F and 2qPCRgar-3R. The primers for unc-29 were qPCRunc-29F and qPCRunc-29R. The primers for unc-43 were qPCRunc-43F and qPCRunc-43R. The primers for cat-2 were qPCRcat-2F and qPCRcat-2R. The primers for let-363 were qPCRlet-363F and qPCRlet-363R. The primers for sir-2.1 were 2qPCRsir-2.1F and 2qPCRsir-2.1R. Samples were diluted 1:10 for all K+ channel sequences, 1:1000 for mdh-1, and 1:10,000 for act-1. All C(t) values were between 20 and 30. Each sample was done in triplicate, and each experiment was repeated at least once from newly made cDNA. Expression levels of each gene of interest were normalized to act-1 and mdh-1 and then reported as relative gene expression levels using CFX Manager Software Data Analysis (Bio-Rad version 1.6) according to the CT method (Livak and Schmittgen 2001). Statistical analysis was performed using GraphPad InStat (version 3.06). The results reported in Figure 3 are from one representative experiment for each gene tested.
The primer designations are as follows:
Fslo1cDNASphI: 5′- CGCGGCATGCGCTAGCATGGGCGAGATTTACTCGCCTTCGCAGTCG
Rslo1cDNAXbaI: 5′- CGCGTCTAGAAAAGTGTCGTTTGCCCGGCTCGTACTCCAGTC
A BK channel/slo-1 mutation results in increased male mating circuit excitability
C. elegans males possess copulatory spicules that must be held inside their tail until the hermaphrodite vulva is breached. Protractor and retractor muscles are attached to the base of the spicules and control their proper position. Once the appropriate mating cues have been received and integrated by both neurons and muscles, the protractors contract, forcing the spicules out of the tail and into the vulva (Sulston et al. 1980; Liu and Sternberg 1995; Liu et al. 2011). Permanent spicule protraction in the absence of mating cues interferes with successful sperm transfer and thus reproduction. This abnormal spicule phenotype is referred to as Protraction constitutive (Prc) and arises when mechanisms that tightly regulate the spicule protraction circuit begin to fail. The percentage of Prc males in a population is determined by isolating virgin larval males from hermaphrodites. The males are allowed to mature to adulthood, and the phenotype is scored 18 to 24 hr later. A male that displays at least partial protraction of one spicule is designated as displaying the Prc phenotype. We previously identified two members of the ether-a-go-go (EAG) K+ channel family, EAG and ether-a-go-go–related gene (ERG), that regulate spicule protraction (Garcia and Sternberg 2003; LeBoeuf et al. 2007). In C. elegans, EAG is encoded by egl-2 and in the male spicule protraction circuit is expressed in sex muscles (Weinshenker et al. 1999); males with a null mutation in the K+ channel display no obvious mating defects (0% of the males display the Prc phenotype) (Table 1) (LeBoeuf et al. 2007). ERG is encoded by unc-103 and is expressed in nearly all muscles and neurons (Reiner et al. 2006). In contrast to egl-2 mutants, males with a null mutation in unc-103 display spicule protraction defects (41% of the males display the Prc phenotype) (Table 1) (Garcia and Sternberg 2003). Mutating both unc-103 and egl-2 voltage-dependent K+ channel genes induce the Prc phenotype in 83% of males (Table 1) (LeBoeuf et al. 2011). While the penetrance of the Prc phenotype is much higher in the double mutant males than either single mutant, not all males in a population lacking both K+ channels displayed the Prc phenotype. This led us to ask whether other K+ channels are involved in maintaining cell excitability.
One likely candidate is the voltage- and calcium-activated BK channel slo-1, which is broadly expressed in neurons and muscles (Wang et al. 2001; Kim et al. 2009). slo-1 loss-of-function mutants exhibit many phenotypes, including impaired movement caused by defective neurotransmitter release, alcohol resistance, and muscle degeneration (Wang et al. 2001; Davies et al. 2003; Carre-Pierrat et al. 2006; Liu et al. 2007a). We found that 70% of males with an early nonsense mutation in slo-1 [allele js379, hereafter referred to as slo-1(lf); Wang et al. 2001] spontaneously protract their spicules in the absence of mating cues (Table 1), demonstrating that like EAG and ERG-like K+ channels, the BK channel also controls the excitability of male sex circuit components.
slo-1 has been shown in C. elegans and other systems to interact with calcium/calmodulin-dependent protein kinase II (CaMKII) (Hawasli et al. 2004; Liu et al. 2006, 2007a; LeBoeuf et al. 2007). unc-43 encodes CaMKII in C. elegans and regulates movement, defecation, and egg laying in addition to its role in male mating (Reiner et al. 1999). We previously reported that loss-of-function mutations in CaMKII cause C. elegans males to display the Prc phenotype. The sy574 allele in unc-43 changes a glycine to a glutamate in the putative substrate binding site of CaMKII. A total of 38% of unc-43(sy574) males displayed the Prc phenotype but exhibit no other obvious defects (Table 1) (LeBoeuf et al. 2007). To determine whether the frequency of the unc-43(sy574)–mediated Prc phenotype was related to reduced SLO-1 function, we created a double mutant containing slo-1(lf) and unc-43(sy574) to ask whether the K+ channel genetically acts downstream of CaMKII. In contrast to the single mutant males, we found that the incidence of the Prc phenotype in double mutant males was 96% [P value <0.0001 to slo-1(lf)] (Table 1). This synergistic effect suggests that while slo-1 might be a target of unc-43 in other behavioral contexts, it is not the only CaMKII target in the male mating circuit.
Specific isoforms of sex muscle-expressed BK channel/slo-1 can reduce abnormal protraction
Since slo-1 is involved in regulating the timing of male sex muscle contractions, we asked where it functions in the mating circuit. We obtained a construct (Pslo-1:slo-1(A2;B0;C1;D0)::GFP) containing the slo-1 isoform previously used to restore function in hermaphrodites with movement defects (Wang et al. 2001). slo-1 has four identified splice regions labeled A–D, therefore isoform A2;B0;C1;D0 includes the exons at A2 and C1 (Figure 1) (Johnson et al. 2011). Splice sites A–C were previously identified (Wang et al. 2001), whereas we identified D during cloning of full-length slo-1 cDNAs (Figure 1C) (Materials and Methods). We introduced Pslo-1:slo-1(A2;B0;C1;D0)::GFP into slo-1(lf) worms and discovered that spicule protraction was reduced from 61 to 38% (P = 0.02) (Table 2). We then generated tissue-specific constructs to express slo-1(A2;B0;C1;D0) in neurons (Paex-3), body-wall muscles (Pacr-8), and sex muscles (Punc-103E) (Reiner et al. 2006; LeBoeuf et al. 2007). slo-1(A2;B0;C1;D0) expression in neurons and body-wall muscles did not reduce the Prc phenotype, whereas expression in the sex muscles reduced the incidence of spontaneous protraction to 23% [P = 0.0005 to slo-1(lf)] (Table 2). Thus, slo-1 is partially regulating spicule protraction in the sex muscles.
Though slo-1(A2;B0;C1;D0) significantly reduces the instance of spontaneous protraction, a large percentage of the males (38%) expressing this isoform still displayed the mutant phenotype (Table 2). slo-1(A2;B0;C1;D0) is one isoform of at least 14 that exist in C. elegans (Wang et al. 2001; Johnson et al. 2011). We asked whether another isoform is capable of complete rescue of the mutant phenotype. We tested two additional isoforms of slo-1, slo-1(A2;B0;C0;D0) and slo-1(A1;B0;C0;D0). slo-1(A2;B0;C0;D0) does not have the exon at splice site C1 and was first reported in Wang et al. (2001). slo-1(A1;B0;C0;D0) exchanges one conserved hydrophobic region 38 amino acids in length in the C terminus with another 38-amino-acid sequence of high homology (Figure 1, A and B). When expressed from the slo-1 promoter, both slo-1(A2;B0;C0;D0) and slo-1(A1;B0;C0;D0) reduced spicule protraction (from 61% to 5 and 0%, respectively) (Table 2). Additionally, tissue-specific expression of slo-1(A1;B0;C0;D0) in the sex muscles via the unc-103E promoter resulted in a 9% instance of the Prc phenotype [P value <0.0001, compared to slo-1(lf)] (Table 2). In conclusion, the exon located at C1 reduces slo-1’s ability to inhibit premature spicule protraction and demonstrates that isoform specificity does play a role in restoring male mating behavior.
BK channel/slo-1(lf)–induced sex muscle spasms can be modified by EAG family K+ channels
Since mutations in BK/slo-1, ERG-like/unc-103, and EAG/egl-2 differentially regulate the excitability of the sex muscles, we asked whether simultaneous loss of multiple K+ channels would have an additive or even a synergistic effect on male mating circuit excitability. We made double mutants between slo-1(lf) and strains that carry large deletions of unc-103 and egl-2 [hereafter referred to as unc-103(0) and egl-2(0)] (Park and Horvitz 1986; Garcia and Sternberg 2003; LeBoeuf et al. 2007). Contrary to our initial expectations, both unc-103(0); slo-1(lf) and egl-2(0) slo-1(lf) double mutants displayed a decrease in the instance of the Prc phenotype [unc-103(0); slo-1(lf) and egl-2(0) slo-1(lf) dropped the frequency of the Prc phenotype to 51 and 30%, respectively, lower than the slo-1(lf)–induced 70% protraction] (Table 1). Thus, removing an ERG-like or an EAG K+ channel along with a BK channel from the male mating circuit lowers rather than increases sex muscle excitability. Males lacking all three K+ channels, an unc-103(0); egl-2(0) slo-1(lf) triple mutant, protracted their spicules 79% of the time, a frequency similar to the slo-1(lf) single mutant and the unc-103(0); egl-2(0) double mutant (Table 1). This suggests that when either egl-2 or unc-103 is removed from the slo-1 background, some aspect (transcriptional, translational, or post-translational) of the remaining K+ channels might be upregulated to compensate for the loss.
The nonadditive effect of combining the slo-1(lf) mutation with either unc-103(0) or egl-2(0) suggests the existence of mechanisms that sense the level of cell excitability and adjusts extant K+ channels to partially compensate for missing or defective K+ channels. To address this hypothesis, we utilized an allele described in an earlier report, a neomorphic loss-of-function allele (sy557) in the unc-103 ERG-like K+ channel gene. unc-103(sy557) contains two point mutations: sy557A, a H165N change in the linker region between transmembrane domains two and three, and sy557B, a W244R change in transmembrane domain five (Garcia and Sternberg 2003). unc-103(sy557) males displayed a Prc frequency of 75%, significantly higher than males completely lacking unc-103 (P = 0.0005, compared to unc-103(0) at 41%) (Table 1), suggesting that the mutant K+ channel gene adversely affects compensating molecules. Since the frequency of unc-103(sy557) males displaying the Prc phenotype was similar to unc-103(0); egl-2(0) double and slo-1(lf) single mutant males, we hypothesized that the sy557-encoded mutation(s) interferes with the compensating properties of EGL-2 and/or SLO-1.
To address which sy557 mutation confers the neomorphic dominant negative properties to UNC-103, we mutated a construct containing unc-103 genomic DNA to encode either the sy557A H165N or sy557B W224R change (Garcia and Sternberg 2003). We then expressed the mutated unc-103 sequences in unc-103(0) male sex muscles using promoter unc-103E [Punc-103E:unc-103(sy557A or B)] or sex neurons using promoter unc-103F [Punc-103F:unc-103(sy557A or B)] to determine whether transgenic mutated versions of unc-103 can interfere with compensating mechanisms and increase the Prc frequency (Reiner et al. 2006). Neither UNC-103(H165N) nor UNC-103(W224R) expressed individually in the sex muscles or neurons increased the rate of spontaneous spicule protraction (Table 2). However, expressing UNC-103(W224R), but not UNC-103(H165N), in both sex muscles and neurons significantly increased protraction from 26 to 62% (P = 0.008, Fisher’s exact test) (Table 2). Thus, disrupting unc-103 via a W244R change in transmembrane domain five can interfere with additional neural and muscle-compensating functions to consequently increase the instance of unregulated spicule protraction.
To determine whether the compensating functions that UNC-103(W224R) interferes with are related to EGL-2 and/or SLO-1 activity, we transgenically introduced the neomorphic dominant negative UNC-103(W224R) into egl-2(0) slo-1(lf) double mutant males. We then asked whether the frequency of the Prc phenotype is similar to unc-103(0); egl-2(0) slo-1(lf) triple mutants or higher. We hypothesized that if there is a genetic interaction between wild-type slo-1, egl-2 and unc-103, then UNC-103(W224R) would interfere with wild-type slo-1 and/or egl-2 activity and the males would display a Prc frequency of ∼70–80%. A testable prediction of this hypothesis is that if UNC-103(W224R) is transgenically introduced into egl-2(0) slo-1(lf) double mutants, then UNC-103(W224R) should only interfere with wild-type unc-103, since the other two K+ channels are already mutant; consequently the transgenic males should also display a Prc frequency of ∼70–80%. If the hypothesis is incorrect, and UNC-103(W224R) interferes with something other than egl-2 and slo-1 function, the transgenic males should display a Prc frequency greater than the unc-103(0); egl-2(0) slo-1(lf) triple mutant. As previously stated, egl-2(0) slo-1(lf) males were 30% Prc, whereas slo-1(lf) males were 70% Prc (Table 1); this decrease in the Prc frequency in the double mutant could be due to compensation by unc-103. We expressed UNC-103(W224R) in male sex muscles and neurons, individually and together, in egl-2(0) slo-1(lf) males. UNC-103(W224R) expressed in neurons of egl-2(0) slo-1(lf) males had no effect on spicule protraction (Table 2). However, UNC-103(W224R) expressed in the sex muscles was sufficient to return spontaneous spicule protraction to levels seen in slo-1(lf) mutants (72% of the males displayed the Prc phenotype) (Table 2). Expressing UNC-103(W224R) in both sex muscles and neurons was not significantly different from muscles alone (81% of the males displayed the Prc phenotype) (Table 2). In conclusion, UNC-103(W224R) appears to interfere with only unc-103 in an egl-2(0) slo-1(lf) background. Since the instance of spontaneous spicule protraction was similar to that seen in unc-103(sy557) and unc-103(0); egl-2(0) slo-1(lf) mutant males, unc-103(sy557) likely also interferes with egl-2 and slo-1. These data are consistent with the idea that there is a genetic interaction between slo-1, egl-2, and unc-103.
BK/slo-1 expression is increased in the absence of ERG/unc-103 but not EAG/egl-2
The genetic experiments discussed in the preceding sections suggested that BK channel/slo-1 function or expression is possibly upregulated in the absence of EAG family K+ channels unc-103 and egl-2. To ask whether slo-1 expression in the male sex muscles might be changed in mutant backgrounds, we transgenically expressed dsRed1-E5 from the slo-1 promoter as a proxy of promoter activity. dsRed1-E5 shifts its emission spectra from green to red over time, and we previously used it to differentiate fluorescent proteins that were newly expressed from older accumulated proteins (Terskikh et al. 2000; LeBoeuf et al. 2011). We expressed integrated Pslo-1:dsRed1-E5 in wild type, unc-103(0), unc-103(sy557), and egl-2(0) males. We then asked whether removing the K+ channels affected marker gene expression in the tails over the course of L4 development and 2 days of adulthood. In L4 wild-type males, we measured the marker gene expression in the male anal depressor muscle, an accessory to the spicule protractor muscles. Since the anal depressor muscle differentiates early in the embryo (Sulston et al. 1983), dsRed1-E5 expression was higher in the red channel than the green (Figure 2, A and D, P value <0.05 for red vs. green for all genotypes, Mann–Whitney t-test). However, the fluorescence intensity difference in the L4 anal depressor between the green and red channels changed when unc-103(0) was introduced into the genetic background (Figure 2A). Pslo-1:dsRed1-E5 expression in unc-103(0) male tails was increased in both channels relative to the wild type (Figure 2A). This suggests that in the absence of unc-103, the transgenic and possibly the endogenous slo-1 promoter is more active and indicates that more slo-1 gene product is made in an attempt to compensate. In contrast, Pslo-1:dsRed1-E5 expression was unchanged compared to wild type in egl-2(0) male tails (Figure 2A). Pslo-1:dsRed1-E5 expression was also unchanged in unc-103(sy557) males (Figure 2A), suggesting that unlike the unc-103(0) allele, the presence of this nonfunctioning UNC-103 protein does not promote an increase in slo-1 expression.
In day 1 and day 2 wild-type adult males, we measured marker gene expression in the fully developed male protractor muscles and the anal depressor accessory muscle. Pslo-1:dsRed1-E5 expression in the green channel was higher than in the red channel, due to the differentiation of the sex muscles (Figure 2, B and E, P value <0.05, Mann–Whitney t-test). There was no difference between the green and red channels in males lacking EGL-2 or UNC-103 K+ channels, though expression in the red channel was significantly higher than the green channel in unc-103(sy557) males (Figure 2B, P value <0.05, Mann–Whitney t-test). However relative to wild-type animals, there was a significant increase in marker gene expression when unc-103 was deleted (Figure 2B). In contrast, marker gene expression in egl-2(0) males was not significantly different from the wild type (Figure 2B). Therefore, similar to L4 males, the deletion of unc-103 caused an increase in transgenic and possibly endogenous slo-1 promoter activity but the loss of egl-2 had no effect. Pslo-1:dsRed1-E5 expression in unc-103(sy557) males was significantly higher in the red but not the green channel, likely due to the large variability in expression seen among these males. Individual males could have differing responses to the loss of excitability regulation induced by the unc-103(sy557) allele.
In day 2 wild-type adult males, expression in the red channel was much higher than in the green channel, due to the accumulation of dsRed1-E5 for 2 days (Figure 2, C and F). Similar to L4 and 1-day-old adult males, Pslo-1:dsRed1-E5 expression was higher in unc-103(0) mutants but unaffected in egl-2(0) mutants (Figure 2C). This is further evidence that a higher rate of slo-1 expression is maintained in unc-103(0) males as they age. In contrast, Pslo-1:dsRed1-E5 expression is lower in 2-day-old unc-103(sy557) males (Figure 2C), a reversal of the slight increases in expression in the younger adults. These data suggest that while slo-1 levels might be similar to wild type in young unc-103(sy557) adult animals, this rate is not maintained as they age.
Food deprivation can attenuate BK channel/slo-1(lf)–induced spicule protraction through EAG K+ channel/egl-2
In previous work, we demonstrated that starvation can suppress the excitability of the mating circuit in ERG-like K+ channel/unc-103 mutants through enhanced compensation by EAG/egl-2 function (Gruninger et al. 2006; LeBoeuf et al. 2007, 2011). Since we established that there is a genetic interaction between slo-1, egl-2, and unc-103, we asked whether food-deprived BK channel/slo-1 mutant males, like unc-103(0) animals, had a reduced frequency of the Prc phenotype via enhanced egl-2 compensation. C. elegans matures through four larval stages (L1–L4) to reach adulthood; the male sex muscles differentiate in the last stages of L4. We starved males by picking them at the late L4 stage and placed them on NGM plates lacking their food source, E. coli OP50. Males develop normally under these starvation conditions (Gruninger et al. 2006). After 18–22 hr, we scored whether the starved males displayed the spontaneous spicule protraction phenotype. We found that 31% of food-deprived slo-1(lf) males displayed the Prc phenotype, significantly lower than slo-1(lf) males on food (70% displayed the Prc phenotype, P value <0.0001) (Table 3). Thus, like unc-103 mutants, starvation is able to partially suppress slo-1(lf)–induced protraction.
We then asked whether starvation was able to inhibit protraction in males lacking both slo-1 and egl-2. We found that depriving egl-2(0) slo-1(lf) males of food had no effect on the instance of Prc (31 vs. 30%, P = 0.7) (Table 3), consistent with the idea that egl-2 mediates the effects of starvation on the excitability of the mating circuit. To support the idea that in the absence of food, enhanced egl-2 compensation can reduce slo-1(lf)–induced excitability defects, we expressed an egl-2 gain-of-function allele in the male sex muscles and asked whether the frequency of slo-1(lf)–induced spicule protraction can be further suppressed by starvation.
The egl-2(gf) allele induces egg retention in hermaphrodites as well as defecation and chemotaxis defects (Reiner et al. 1995; Weinshenker et al. 1995; Weinshenker et al. 1999). In the male mating circuit, egl-2(gf) suppresses muscle excitability when activated by starvation (LeBoeuf et al. 2011). We made a double mutant containing both slo-1(lf) and egl-2(gf) and found that the double mutant males were extremely constipated and in 91% of the males, the proctodeum, including the spicules, hemorrhaged from the cloacal opening.
slo-1 has a broad expression pattern in neurons and muscles outside the male mating circuit, whereas egl-2 is expressed in a few head neurons and intestinal muscles (Weinshenker et al. 1999; Wang et al. 2001). The slo-1(lf) and egl-2(gf) alleles appear to cause a synthetic effect resulting in chronic constipation and hemorrhaging. To circumvent this problem, we expressed egl-2(gf) cDNA only in the sex muscles of egl-2(0) slo-1(lf) males using the unc-103E promoter. A total of 66% of egl-2(0) slo-1(lf); rgEx253[Punc-103E:egl-2(gf)] males display abnormal spicule protraction, significantly lower than egl-2(gf) slo-1(lf) males and similar to slo-1(lf) males (91 and 70%, P values <0.005 and =0.7, respectively) (Table 3). Starvation reduced spicule protraction in egl-2(0) slo-1(lf); rgEx253[Punc-103E:egl-2(gf)] males to 7% (P value <0.05 to fed) (Table 3). Thus, egl-2 compensation can be stimulated by starvation to reduce sex muscle excitability, and this increase in egl-2 activity can reduce slo-1(lf)–induced muscle spasms.
slo-1(rg432) encodes an intronic point mutation that reduces sex muscle excitability
As we have previously stated, unc-103(0); egl-2(0) double mutant males display spontaneously protracted spicules at a frequency of 83%. In the course of working with the double mutant, we realized that a spontaneous genetic modifier (rg432) unknowingly got crossed into the strain. The modifier reduced the frequency of the unc-103(0); egl-2(0) double mutant Prc phenotype to 46% [P value <0.005, compared to unc-103(0); egl-2(0)] (Table 3). In addition, we discovered that starvation strongly suppressed the instance of the Prc phenotype for unc-103(0); egl-2(0) animals that contained the rg432 modifier [P value <0.005, compared to unc-103(0); egl-2(0)] (Table 3). We roughly mapped the genetic modifier (through conventional crosses, data not shown), and not surprisingly, it mapped close to slo-1. We sequenced slo-1 from the unc-103(0); egl-2(0) rg432 line and found an A-to-T change in an intron between exons 11 and 12 (Figure 1D). This mutation was not present in the normal unc-103(0); egl-2(0) line. When unlinked to egl-2(0) and unc-103(0), animals harboring the slo-1(rg432) allele appeared behaviorally wild type under fed or starved conditions (Table 3). In double mutant combinations, slo-1(rg432) had no effect on the frequency of the Prc phenotype of unc-103(0) or egl-2(0) males (Table 3). Since unc-103(0); egl-2(0) slo-1(rg432) triple mutant males displayed a similar Prc phenotype to unc-103(0) single mutant males (46 vs. 41%, P = 0.7, Fisher’s exact test) (Table 3), it was possible that the slo-1(rg432) allele facilitated increased SLO-1 compensation in the absence of egl-2.
It has recently been reported that a point mutation in a slo-1 intron can affect alternative splicing (Glauser et al. 2011). Since the slo-1(rg432) allele is located in an intron, we asked whether alternative splicing was affected. The slo-1 genetic locus contains four alternatively spliced regions, labeled A–D, and the rg432 mutation does not map to any of them (Figure 1D). We designed primers to distinguish between the differentially spliced exons and performed quantitative RT–PCR on cDNA obtained from wild-type and slo-1(rg432) nonstarved adult hermaphrodites or males. Total slo-1 expression was greatly increased from males to hermaphrodites, likely due to the difference in the number of excitable cells (sex-specific muscles and neurons) between the two sexes, and this was true for all isoforms (Figure 3, A and B). Similar to wild type, slo-1(rg432) males showed increased expression when compared to mutant hermaphrodites, and this effect was not isoform specific (Figure 3, A and B). slo-1 transcript levels were decreased in slo-1(rg432) hermaphrodites, with some splice sites being significantly affected while others were not (Figure 3, A and B). Importantly, slo-1 mRNA levels were significantly lower in slo-1(rg432) males, resulting in a much more pronounced effect on slo-1 expression levels by the mutant allele in males vs. hermaphrodites (Figure 3, A and B). Thus, the slo-1(rg432) allele results in a decrease in slo-1 transcript levels, which is most profound in males and is not isoform specific in adults.
Stable K+ channel transcription is dependent on BK channel/slo-1 and EAG K+ channel/egl-2
Our experiments discussed so far suggest that K+ channels in the male sex muscles can be upregulated to compensate for one another to control muscle excitability. The dsRed1-E5 marker gene analysis indicated that increases in gene expression via mRNA synthesis or stability could be one method of upregulation. To determine whether K+ channel compensation via mRNA synthesis or stability is more general and not just limited to the male spicule muscles, we measured global relative changes in K+ channel mRNA abundance in wild type and unc-103, egl-2, and slo-1 mutants. We isolated poly(A) mRNA from well-fed mixed gender and mixed staged whole worm extracts and performed quantitative RT–PCR experiments.
First, we measured global unc-103 and slo-1 transcript levels in egl-2(0) worms and found that K+ channel transcript levels were significantly decreased (Figure 4, A, C, and D). Thus, the relative normalcy of egl-2(0) worms cannot be explained by an increase in unc-103 and slo-1 transcript levels. Likewise, unc-103(sy557) worms displayed unaffected (egl-2) or decreased (slo-1 and unc-103) levels of transcript (Figure 4). However, egl-2 and slo-1 transcript levels were increased in unc-103(0) worms (Figure 4, B–D), indicating that gene expression is a possible mechanism of attempted compensation in these worms. Additionally, unc-103(sy557) males displayed a higher instance of the Prc phenotype than unc-103(0) males, a difference that could be due in part to the lack of K+ channel upregulation in unc-103(sy557) mutants.
We next measured K+ channel transcript levels in slo-1(lf) and slo-1(rg432) worms. While slo-1 transcript levels were unaffected in slo-1(rg432) mutants, egl-2 and unc-103 levels were decreased (Figure 4). The slo-1(rg432) allele promotes regulation of male sex muscle excitability, but it is not due to an increase in transcription or mRNA stability. slo-1(lf) mutants display a high level of the Prc phenotype, so it was unsurprising to see that unc-103 and egl-2 levels were significantly decreased (Figure 4). Also as expected, the levels were significantly lower than wild type in the slo-1(lf) mutant containing an early nonsense mutation (Figure 4, C and D), likely as a result of nonsense-mediated mRNA decay (Hodgkin et al. 1989; Maquat 2004; Longman et al. 2007). The primers used to detect slo-1 transcripts start at amino acid F641, considerably downstream from the Q251stop mutation in the loss-of-function mutant. There is little to suggest extensive read-through of the stop mutation, due to the severe defects shown by slo-1(lf) animals, indicating functional SLO-1 K+ channels are greatly reduced if not abolished (Wang et al. 2001; Carre-Pierrat et al. 2006).
Since the frequency of the Prc phenotype was significantly decreased in unc-103(0); egl-2(0) slo-1(rg432) as compared to unc-103(0);egl-2(0) males (Table 3), we asked how global slo-1 mRNA levels were affected in the triple and double mutants relative to wild type and the single mutants. In unc-103(0); egl-2(0) animals, the global slo-1 transcript abundance were similar to the wild type. This could be an additive consequence of the unc-103(0) mutation increasing slo-1 transcript levels, balanced with the effect of egl-2(0) decreasing them. Unexpectedly, slo-1 mRNA levels were drastically lower in unc-103(0); egl-2(0) slo-1(rg432) mutants relative to unc-103(0); egl-2(0) and slo-1(rg432) animals (Figure 4, C and D). This indicates that the lack of EGL-2 and UNC-103 activity has a synthetic interaction with the rg432 allele to globally reduce slo-1 mRNA abundance even more so than the rg432 allele alone. However, the behavior of unc-103(0); egl-2(0) slo-1(rg432) triple mutants was not worse than unc-103(0); egl-2(0) double mutants, but rather, resembles unc-103(0) single mutants. This suggests that the amount of slo-1 mRNA cannot be solely rate limiting in controlling behavior. Thus a complex interplay between the consequences of the three mutant alleles might have a pleotropic effect on other genes that control general neural and muscle excitability.
Finally, we asked whether a molecule that is known to pleotropically affect neural and muscle cell excitability also influences slo-1, egl-2, and unc-103 mRNA abundance. CaMKII is a calcium-activated protein and is known to influence gene expression (Kapiloff et al. 1991; Ramirez et al. 1997; Hughes et al. 2001; Zhang et al. 2004; Ronkainen et al. 2011). In C. elegans, loss-of-function mutations in CaMKII/unc-43 have profound deleterious effects on virtually every behavior of the worm (Reiner et al. 1999). The unc-43(sy574) allele we previously isolated causes males to display the Prc phenotype, but does not have any gross effect on other behaviors in males or hermaphrodites (LeBoeuf et al. 2007, 2011). We therefore asked whether the sy574 allele, which outwardly appears to affect one behavior, influences global slo-1, egl-2, and unc-103 mRNA abundance. We discovered that the transcript amount of all three K+ channels was decreased in animals that contain the unc-43(sy574) allele (Figure 4). This result indicates that despite the phenotypic mildness of the unc-43(sy574) allele, it has broad effects of mRNA levels, and that slo-1, egl-2, and unc-103 expression is directly or indirectly connected with CaMKII function.
slo-1(rg432) preferentially affects genes involved in regulating cell excitability in males
The above experiments do not explain how the slo-1(rg432) allele, which results in decreased slo-1 transcript levels in hermaphrodite and male populations and decreased egl-2 and unc-103 mRNA levels in mixed-staged populations, is able to reduce spicule protraction in unc-103(0); egl-2(0) males. Decreasing the pool of available transcripts should have a negative impact on the amount of protein produced, and consequently a cell’s ability to maintain its polarized state. One possibility is that, as a response to decreased slo-1 transcript levels, cells generally lower the amount of other components that promote cell depolarization, thus resulting in a more stable system. To test this hypothesis, we measured the transcript levels of genes involved in regulating the excitability of cells used in male mating behavior and also of genes involved in more general cell processes. We analyzed acr-18 [nicotinic acetylcholine receptor (nAChR) α-subunit], gar-3 (ACh receptor coupled to gαq), unc-29 (nAChR non–α-subunit), unc-43 (CaMKII), and cat-2 (tyrosine hydroxylase) as candidate genes that are enriched in excitable cells (Sulston et al. 1975; Fleming et al. 1997; Hwang et al. 1999; Lints and Emmons 1999; Reiner et al. 1999; Sze et al. 2000; Garcia et al. 2001; Mongan et al. 2002; LeBoeuf et al. 2007; Liu et al. 2007b, 2011). Neither acr-18 nor cat-2 transcript levels were increased in males as compared to hermaphrodites (Figure 3, C and D), although both were reduced in slo-1(rg432) males. While C. elegans males have many more excitable cells than hermaphrodites, only a few of these express cat-2, which could account for the lack of difference between sexes. acr-18 is expressed in many male-specific neurons and muscles (Liu et al. 2011); however, the acr-18 mRNA was slightly, but not significantly increased in males as compared to hermaphrodites. The acr-18 levels could be reduced in non–sex-specific cells in males, resulting in transcript levels that are not significantly different from hermaphrodites.
In contrast to acr-18 and cat-2 and similar to slo-1, the genes gar-3, unc-29, and unc-43 showed greatly increased transcript levels in the male; this increase was abolished in slo-1(rg432) worms (Figure 3, C and D). Thus, in response to the rg432 allele, cells appear to downregulate other molecules, along with the SLO-1 channel, involved in regulating cell excitability.
In addition to genes enriched in excitable cells, we also looked at the transcript levels of genes with more general expression patterns: let-363 (Tor-like kinase), sir-2.1 (histone deacetylase), cdc-42 (RHO GTPase), and pmp-3 [ATP-binding cassette (ABC) transporter] (Chen et al. 1993; McKim et al. 1993; Frye 2000; Hoogewijs et al. 2008). Relative to hermaphrodites, male transcript levels were increased for let-363 and sir-2.1 (Figure 3, C and D). mRNA levels for let-363, sir-2.1, and pmp-3, but not cdc-42, dropped in slo-1(rg432) males; however, the differences were not as pronounced as compared to transcripts enriched in excitable cells (Figure 3, C and D). In addition, transcript levels for let-363, sir-2.1, and pmp-3 were increased in slo-1(rg432) hermaphrodites, in contrast to the lower transcript levels of the analyzed excitable cell-enriched genes. Taken together, these data suggest that the slo-1(rg432) allele broadly affects the levels of mRNA encoded by genes that regulate cell excitability more so than genes with more general cell functions.
The excitability of neurons and muscles need to be fine-tuned for the optimal performance of complex behaviors. C. elegans male mating is a demanding behavior requiring coordinated feedback mechanisms to promote the multiple steps of mating. Failure of any one step leads to unsuccessful mating. For example, full spicule protraction occurs when the spicules penetrate the vulval slit and is inhibited until the correct cues are received from all inputs. A breakdown in regulation leads to promiscuous protraction at ectopic areas, thus reducing successful sperm transfer. Spicule protraction depends on the integration of signaling from neurons and muscles, including the SPC motor neuron and the dorsal and ventral protractor muscles. The SPC releases acetylcholine (ACh) onto the protractors, initiating contraction. Gap junctions between the protractors and other muscles and neurons in the circuit help reduce the excitability threshold of the protractors, so they can respond to the ACh released by the SPC neurons (Garcia et al. 2001; Liu et al. 2011). BK/SLO-1, ERG-like/UNC-103, and EAG/EGL-2 K+ channels attenuate circuit activity until sufficient cues have been received (Figure 5A).
An important contributor to regulating male mating circuit excitability is BK channel/slo-1. slo-1 mutants display ectopic spicule protraction, which can be rescued with specific slo-1 isoforms. Full rescue is achieved by shortening the distance between the two regulator of K+ channel conductance (RCK) sites. RCK sites are proposed to mediate calcium’s gating of the channel (Lee and Cui 2010; Yuan et al. 2010), a modification that is common across species (Fodor and Aldrich 2009). Similarly, isoform-specific rescue of flight was shown in Drosophila (Brenner et al. 2000). Across different species, age, cell type, and stress impact which isoforms of slo-1 are expressed (Tseng-Crank et al. 1994; Xie and McCobb 1998; MacDonald et al. 2006; Kim et al. 2010). In C. elegans SLO-1, inclusion of the exon at splice site C1 results in channels with slower activation when expressed in Xenopus oocytes (Johnson et al. 2011). Our data suggest a functional consequence for modifying this linker region. The subtle changes in slo-1 that promote the channel’s ability to inhibit precocious spicule protraction highlight the large impact that small changes in channel structure have on behavior.
In addition to identifying an isoform change that promotes reduction of sex muscle excitability, we identified an intronic point mutation [slo-1(rg432)] that significantly reduces slo-1 transcript levels and sex muscle excitability. slo-1(rg432) also reduces the transcript levels of genes involved in regulating cell excitability. This suggests that the compensatory response to reduced slo-1 levels is to diminish the levels of other genes involved in cell depolarization. Many human diseases are the result of intronic point mutations (Watanabe et al. 2000; Lamba et al. 2002; Keren et al. 2010); identifying ectopic genetic responses provides potential therapeutic routes to alleviate the effects of such diseases.
SLO-1 does not regulate male mating circuit excitability alone but functions with the EAG family K+ channels EAG/EGL-2 and ERG/UNC-103. By examining how these K+ channels impact one another, we elucidated their coordinated effort to regulate cell excitability and behavior. Approximately 70–80% of males lacking slo-1 display premature spicule protraction; this fraction does not change when both egl-2 and unc-103 are removed. However, removing only one results in males with reduced spontaneous muscle spasms. Thus, egl-2 or unc-103 can partially compensate for slo-1 deficiency. These three channels have also been shown to mediate synapse formation in Drosophila (Budnik et al. 1990; Berke et al. 2006; Lee and Wu 2010), highlighting the conserved mechanisms used to set up cell excitability and produce behavior.
Studying transcript levels in different K+ channel mutants indicates egl-2 stabilizes unc-103 and slo-1 mRNA. In egl-2(0) males, unc-103 and slo-1 mRNA levels are lower than wild type. Genetic interactions indicate egl-2 allows slo-1 to compensate for the loss of unc-103 function. unc-103(0); egl-2(0) males display a higher instance of abnormal spicule protraction than do unc-103(0) males. The difference is likely due to an upregulation of slo-1 in an unc-103(0) background. slo-1 mRNA stability increases in an unc-103(0) background, as does expression of a reporter gene from the slo-1 promoter. If the increase in transcription promotes protein synthesis, more EGL-2 and SLO-1 in the membrane could be compensating for UNC-103 deficiency. Additionally, the effect of the unc-103(sy557) neomorphic allele on egl-2 and slo-1 function provides further evidence of an interaction between the three K+ channels. unc-103(sy557) induces spicule protraction at a level similar to unc-103(0); egl-2(0) slo-1(lf) and the increased levels of slo-1 and egl-2 transcription seen in unc-103(0) males are reversed in unc-103(sy557) mutants. These data suggest that unc-103(sy557) negatively impacts egl-2 and slo-1, and conversely, these three K+ channels function interdependently in wild-type circuits. Finally, egl-2 needs to be absent for unc-103 to compensate for slo-1 deficiency: the instance of spontaneous spicule protraction drops 40% from slo-1(lf) males compared to egl-2(0) slo-1(lf) males. These results suggest SLO-1 activity is dependent upon EGL-2, while EGL-2 attenuates UNC-103 compensatory function.
In C. elegans, EAG and BK channels also interact to downregulate sex muscle excitability when the young adult males undergo a period of starvation, presumably to direct the males’ behavior from copulation and toward food acquisition (Gruninger et al. 2006; LeBoeuf et al. 2011). We previously reported that EAG K+ channels play a prominent role during food deprivation (LeBoeuf et al. 2007, 2011); here, we show that BK channels are also important. When slo-1(lf) males are food deprived, spontaneous spicule protraction is partially suppressed in an EAG K+ channel-dependent manner. Modifying EAG K+ channel function through a neomorphic genetic mutation enhances the effect of food deprivation to suppress slo-1(lf)–induced muscle spasms.
Modifying K+ channel function is one way for the circuit to regulate excitability; adjusting calcium levels is another. Calcium is a main transducer of excitatory signals and a circuit must maintain tight control of intracellular calcium levels. The Ca2+-sensitive BK channels provide one method of regulatory response; Ca2+-activated kinases are another. A candidate is CaMKII, which acts as a repressor of excitability in the male mating circuit (LeBoeuf et al. 2007). K+ channels in the circuit set up the delicate balance necessary for reproduction, and CaMKII may influence their transcription (Figure 5B). CaMKII’s role in regulating transcription is well documented (Li et al. 2010; Ely et al. 2011; Oruganti et al. 2011), especially in cardiac muscle (Zhang et al. 2004; Backs et al. 2006; Little et al. 2007). In C. elegans, it is possible CaMKII is regulating the transcription of slo-1, egl-2, and unc-103, since all three transcript levels are decreased in a CaMKII/unc-43(lf) mutant. CaMKII also acts through EAG K+ channels during starvation and both EAG and ERG K+ channels under standard conditions (LeBoeuf et al. 2007, 2011). However, there is genetic evidence to suggest that CaMKII acts through additional K+ channels in the male. Nearly 100% of unc-103(0); unc-43(lf) (LeBoeuf et al. 2007) and unc-43(lf); slo-1(lf) double mutant males (this article) spontaneously protract their spicules. However, only 79% of triple K+ channel mutant males display abnormal behavior. This indicates that there are additional unidentified molecules through which CaMKII is acting to regulate male sex muscle excitability.
We thank LaShundra Rodgers for technical assistance and Benjamin Russo, Daisy Gualberto, Liusuo Zhang, and James Midkiff for critical reading of the manuscript. C. elegans strains were provided by the Caenorhabditis Genetic Center, which is funded by the National Center for Research Resources, National Institutes of Health. This work was supported by the Howard Hughes Medical Institute.
Communicating editor: R. Anholt
- Received September 7, 2011.
- Accepted December 3, 2011.
- Copyright © 2012 by the Genetics Society of America