regions of interest
voltage-gated Ca2+ channel
Mammalian skeletal muscle contractions are initiated by all-or-none action potentials (APs) triggered by motoneurons. In C. elegans locomotory muscle, however, the characteristics of APs, the underlying ion channels, and their role in Ca2+ release are poorly understood. Here we show that C. elegans locomotory muscle fires spontaneous, all-or-none APs, which appear to be modulated by motoneuron activity. The upstroke of muscle APs requires Ca2+ entry through a voltage-gated Ca2+ channel (EGL-19), whereas the downstroke of the APs relies on a voltage-gated potassium channel as well as a Ca2+- and Cl−-activated potassium channel. AP-elicited elevations of intracellular Ca2+ concentration require both EGL-19 in the plasma membrane and the ryanodine receptor in the sarcoplasmic reticulum membrane. The discovery of all-or-none action potentials in C. elegans body-wall muscle brings the physiology of C. elegans much closer to that of other metazoans, and strengthens its utility as a model organism.
Although the neuromuscular system of C. elegans has been studied intensively, little is known about the properties of muscle action potentials (APs). By combining mutant analyses with in vivo electrophysiological recording techniques and Ca2+ imaging, we have established the fundamental properties and molecular determinants of body-wall muscle APs. We show that, unlike mammalian skeletal muscle APs, C. elegans muscle APs occur in spontaneous trains, do not require the function of postsynaptic receptors, and are all-or-none overshooting events, rather than graded potentials as has been previously reported. Furthermore, we show that muscle APs depend on Ca2+ entry through the L-type Ca2+ channel EGL-19 with a contribution from the T-type Ca2+ channel CCA-1. Both the Shaker K+ channel SHK-1 and the Ca2+/Cl−-gated K+ channel SLO-2 play important roles in controlling the speed of membrane repolarization, the amplitude of afterhyperpolarization (AHP) and the pattern of AP firing; SLO-2 is also important in setting the resting membrane potential. Finally, AP-elicited elevations of [Ca2+]i require both EGL-19 and the ryanodine receptor UNC-68. Thus, like mammalian skeletal muscle, C. elegans body-wall myocytes generate all-or-none APs, which evoke Ca2+ release from the sarcoplasmic reticulum (SR), although the specific ion channels used for AP upstroke and repolarization differ.
Action potentials (APs) are hallmarks of excitable cells. In mammalian skeletal muscle, Na+-dependent APs of short duration (less than a few milliseconds) underlie excitation–contraction coupling (Cooke & Grinnell, 1964; Redfern et al. 1970; Cruz et al. 1985; Catterall, 1986; Wolters et al. 1994; Cairns et al. 2003). APs are triggered by acetylcholine (ACh)-evoked end-plate potentials, and travel into transverse tubules where they activate the dihydropyridine receptor (DHPR), which is a CaV1.1/L-type voltage-gated Ca2+ channel (VGCC). The activation of DHPRs causes ryanodine receptor (RyR)-mediated Ca2+ release from the SR through protein–protein interactions between the DHPR and RyR (Schneider, 1994) rather than Ca2+-induced Ca2+ release, which generally occurs in cardiomyocytes (Cannell et al. 1995), some smooth muscle cells (Kotlikoff, 2003) and neurons (Verkhratsky & Shmigol, 1996). The SR appears to be the primary source of Ca2+ in skeletal muscle cells because blockade of Ca2+ entry does not have an obvious effect on the elevation of [Ca2+]i in response to membrane depolarization (Nakai et al. 1996). The SCN4A/NaV1.4 Na+ channel is the major Na+ channel expressed in innervated mammalian skeletal muscle (Trimmer et al. 1989), whereas the CLCN1/CLC-1 Cl− channel dominates the resting conductance of skeletal muscle and probably plays an important role in spike repolarization (Hudson et al. 1995; Jentsch et al. 1999).
C. elegans body-wall muscle is analogous to vertebrate skeletal muscle in structure and function (Moerman & Fire, 1997; Lecroisey et al. 2007); however, little is known about the property of APs and the mechanism of excitation–contraction coupling of C. elegans body-wall muscle cells. While it has been reported that graded APs occur in C. elegans body-wall muscle cells (Jospin et al. 2002), the general properties of body-wall muscle APs, the ion channels underlying the resting membrane potential and shaping AP waveform, and the processes of Ca2+ release are largely unknown.
As C. elegans does not have genes encoding homologues of known voltage-gated Na+ channels (Bargmann, 1998), APs in body-wall muscle cells most probably result from the activation of Ca2+ channels or other cation channels. Homologues of the pore-forming α1-subunit of three different vertebrate voltage-gated Ca2+ channels (VGCCs) are found in C. elegans, including UNC-2 (CaV2 or N/P/Q-type), EGL-19 (CaV1 or L-type) and CCA-1 (CaV3 or T-type) (Bargmann, 1998). In addition, C. elegans has NCA-1 and NCA-2 (Jospin et al. 2007; Yeh et al. 2008), which are homologues of the mammalian NALCN, a voltage-independent, non-selective cation channel (Lu et al. 2007). NCA-1 and NCA-2 resemble the pore-forming subunit of voltage-gated Ca2+ or Na+ channels in having four repeating domains with similar putative topology, but they are distinct in the primary sequence. EGL-19 and, possibly, CCA-1 are expressed in body-wall muscle cells (Lee et al. 1997; Steger et al. 2005), and EGL-19 is a major carrier of inward current, playing an important role in generating the reported graded APs in body-wall muscle cells (Jospin et al. 2002). The other three channels (UNC-2, NCA-1 and NCA-2) are primarily expressed in the nervous system with muscle expression detected only at the worm tail for NCA-2 (http://www.wormbase.org). In addition, C. elegans body-wall muscle cells express a variety of K+ channels (Kunkel et al. 2000; Yuan et al. 2000; Salkoff et al. 2001; Wang et al. 2001; Salkoff et al. 2005) and a RyR, which is encoded by the unc-68 gene (Maryon et al. 1998). Three of the identified K+ channels, including SHL-1 (Shal or Kv4 type), SHK-1 (Shaker or Kv1 type) and SLO-2 (Ca2+- and Cl−-gated) appear to be the predominant carriers of outward currents (Santi et al. 2003; Fawcett et al. 2006).
In the present study, we determined the electrical properties of APs in C. elegans body-wall muscle cells, analysed the role of specific ion channel gene products in generating and shaping APs, and determined how APs lead to an elevation of [Ca2+]i and muscle contraction. We found that body-wall muscle APs are all-or-none events rather than graded depolarizations as previously reported, and that unlike those of mammalian skeletal muscle fibres, body-wall muscle APs occur in the absence of neural input. We also report that SHK-1 and SLO-2 K+ channels contribute to body-wall muscle APs, and that trains of APs are associated with elevations of [Ca2+]i, which required the functions of both the EGL-19 Ca2+ channel and the UNC-68 ryanodine receptor channel.
Culturing of C. elegans
Worms were grown inside an incubator (21°C) on agar plates with a layer of OP50 Escherichia coli (Sulston & Hodgkin, 1988).
Recording of membrane potentials and currents
Adult hermaphrodite animals were immobilized and dissected as described previously (Wang et al. 2001; Liu et al. 2005). Briefly, an animal was immobilized on a glass coverslip by applying 3M Vetbond Tissue Adhesive (3M Company, MN, USA) along the dorsal side. Application of the adhesive was generally restricted to the middle portion of the animal, which allowed movements of the head and tail during the experiment. A longitudinal incision was made in the dorsolateral region. After clearing the viscera, the cuticle flap was folded back and glued to the coverslip, exposing the ventral nerve cord and the two adjacent muscle quadrants. A Nikon FN-1 microscope equipped with a 40× water-immersion objective and 15× eyepieces was used for viewing the preparation. Borosilicate glass pipettes with a tip resistance of 3∼5 MΩ were used as electrodes for current- and voltage-clamp recordings in the classical whole-cell patch-clamp configuration. Current- and voltage-clamp experiments were performed with a Multiclamp 700B amplifier (Molecular Devices, Sunnyvale, CA, USA) and the Clampex software (version 10, Molecular Devices). Data were sampled at a rate of 10 kHz after filtering at 2 kHz.
All the current-clamp experiments, except when specified otherwise, were performed without current injection. The holding potential was −60 mV in all voltage-clamp experiments. Outward currents were recorded by applying 1200 ms voltage steps from −60 to +70 mV at 10 mV intervals whereas inward currents were recorded by applying 200 ms voltage steps from −40 to +60 mV at 10 mV intervals. In experiments recording membrane potential changes and whole-cell outward current, the standard bath solution contained (in mm): NaCl 140, KCl 5, CaCl2 5, MgCl2 5, dextrose 11 and HEPES 5 (pH 7.2, 320 mOsm); and the standard pipette solution contained (in mm): KCl 120, KOH 20, Tris 5, CaCl2 0.25, MgCl2 4, sucrose 36, EGTA 5 and Na2ATP 4 (pH 7.2, 323 mOsm). In experiments recording whole-cell inward current, the standard bath solution contained (in mm): NaCl 120, KCl 5, CaCl2 5, MgCl2 5, dextrose 11, HEPES 5, tetraethylammonium chloride (TEACl) 20, and 4-aminopyridine 3 (pH 7.2, 323 mOsm); and the standard pipette solution contained (in mm): CsCl 140, MgATP 4, TEACl 5, EGTA 5, and Tris 5 (pH 7.2, 296 mOsm).
Imaging muscle Ca2+ transients
GCaMP2 (Tallini et al. 2006) was synthesized with codon optimization and introduction of two introns for improved expression in C. elegans. The synthesized GCaMP2 was cloned into a vector containing the myo-3 myosin promoter (Pmyo-3) (Okkema et al. 1993), and expressed in worms through a transgenic injection. Integrated transgenic worms (zw495 strain) expressing the Pmyo-3::GCaMP2 (wp962) were obtained following standard procedures (Evans, 2006), backcrossed 3 times with the wild-type, and subsequently crossed into the unc-68(r1162) strain (Maryon et al. 1996). Spontaneous [Ca2+]i changes in body-wall muscle cells expressing GCaMP2 were monitored by imaging GCaMP2 fluorescence intensity using an electron-multiplying CCD camera (iXonEM+885, Andor Technology, Belfast, Northern Ireland), a FITC filter (59222, Chroma Technology Corp., Bellows Falls, VT, USA), a light source (Lambda XL, Sutter Instrument, Novato, CA, USA) and the NIS-Elements software (Nikon). TTL signals from the camera were used to synchronize the recordings of Ca2+ transients and the membrane potential in some experiments.
A train of spikes was defined as five or more spikes with an interval between two consecutive spikes of no more than 2 s. To avoid complications by overlapping events and baseline drift, 10 APs with a clear baseline before the AP and a clear AHP following the repolarization were randomly chosen from the recordings of each cell for quantifying spike amplitude, 10–90% rise time, 90–10% decay time, width and AHP amplitude. The resting membrane potential of each cell was measured from an interval when there were no spikes and no large postsynaptic potentials. All quantification was performed using the Clampfit software (version 10, Molecular Devices).
For Ca2+ imaging data, a single muscle cell of each preparation was chosen as region of interest (ROI). The fluorescence intensity in the ROI over successive frames was first plotted as absolute fluorescence intensity over time using the NIS-Elements software (Nikon, Tokyo, Japan), and then converted to F/F0 using the MATLAB software (The MathWorks, Inc., Natick, MA, USA) running a custom module.
Data graphing and statistical analyses were performed with Origin (version 8, OriginLab, Northampton, MA, USA). Statistical comparisons were made either by unpaired t test or one-way ANOVA (followed by Bonferroni post hoc test) depending on whether the number of groups for comparison was two or more. All values are shown as mean ±s.e.m. P < 0.05 is considered to be statistically significant.
Body-wall muscle APs are all-or-none events
Body-wall muscle APs were recorded in the classical whole-cell patch-clamp configuration, using extracellular and pipette solutions that are routinely used for recording postsynaptic currents at the C. elegans neuromuscular junction (Richmond & Jorgensen, 1999). In the absence of current injection, muscle cells fired all-or-none overshooting APs spontaneously, which generally occurred in trains (Fig. 1A). The firing of APs typically lasted for several minutes (mean ∼4 min) after the whole-cell configuration was achieved. Subsequently, APs stopped firing rather than becoming graded. The disappearance of APs was accompanied by a rundown of inward currents (Fig. 1E), which appeared to be caused by dialysis of intracellular factors rather than deterioration of the preparation, as other cells could still fire APs normally after the first patch-clamped cell stopped firing. Therefore, we restricted our analyses to the period when AP firing could be observed.
A train of spikes was often preceded by a gradual elevation of the resting membrane potential (slow wave). Individual spikes within a train were usually followed by repolarization to near the resting membrane potential (Fig. 1A and B). We quantified spike peak amplitude, width (at 50% peak amplitude), 10–90% rise time, 90–10% decay time, and the amplitude of AHP, as were defined in Fig. 1B. In addition, we quantified the number of spikes per train as well as the duration of inter-spike intervals between consecutive events within a train (Fig. 1C and D). These analyses revealed several prominent features of body-wall muscle APs. First, all APs are overshooting all-or-none events that depolarize to approximately +20 mV. Second, the rate of rise of the AP is slower than the decay (repolarization). And third, the width of APs in C. elegans body-wall muscle (∼20 ms) is much longer than that in vertebrate skeletal muscle (<2 ms) (Cooke & Grinnell, 1964; Redfern et al. 1970; Cruz et al. 1985; Wolters et al. 1994; Cairns et al. 2003).
Our observation of all-or-none APs contradicts a previous report that APs in C. elegans body-wall muscle cells are graded events (Jospin et al. 2002). To exclude the possibility that this discrepancy resulted from differences in recording solutions, we also recorded APs using the bath and pipette solutions reported in the previous study. All-or-none overshooting APs were also observed with these solutions (Supplementary Fig. S1). The resting membrane potential measured with the two sets of solutions were also similar (−26.9 ± 0.5 mV with standard solutions; −27.5 ± 0.6 mV with the solutions of Jospin et al. 2002), and consistent with our previous report (−27.1 ± 0.8 mV) (Liu et al. 2006), but was substantially more negative than that reported by the previous study (−19.7 ± 1.8 mV) (Jospin et al. 2002).
The resting membrane potential in our study appeared to be near the threshold for AP firing because the slow waves that preceded APs were typically less than 10 mV above the resting membrane potential. To assess sensitivity of AP firing to the resting membrane potential, we analysed the effect of hyperpolarization (approximately −40 mV) on AP firing by injecting a small negative current. In all of the cells analysed (n= 3), the firing of APs stopped during the current injection and resumed after the current injection was terminated. A representative experiment is shown in Supplementary Fig. S2. Thus, rather than operating through graded potentials, C. elegans locomotory muscle cells generate all-or-none APs, and the rate of AP firing appeared to be very sensitive to membrane potential changes.
Body-wall muscle APs could occur after postsynaptic receptors were blocked
C. elegans body-wall muscle cells are innervated by cholinergic and GABAergic motoneurons (Chalfie & White, 1988; Wang, 2010). Cholinergic motoneurons release ACh to contract the muscle whereas GABAergic motoneurons release γ-aminobutyric acid (GABA) to relax the muscle. To determine whether AP firing requires neural input, we first analysed APs in the unc-49(e407) mutant, which lacks GABA receptors (Bamber et al. 1999), in the presence of the ACh receptor blocker d-tubocurarine (TBC). Under such experimental conditions, both cholinergic and GABAergic neuromuscular transmissions are blocked (Richmond & Jorgensen, 1999; Liu et al. 2005). Nevertheless, overshooting APs continued to occur with only altered resting membrane potential and spike amplitude (Fig. 2). These observations suggest that AP firing did not rely on neuromuscular transmission under the current experimental conditions.
Ca2+ channels underlie body-wall muscle APs
The lack of a candidate voltage-gated Na+ channel in the C. elegans genome (Bargmann, 1998), and the fact that voltage-dependent inward currents in C. elegans body-wall muscle cells are abolished by the Ca2+ channel blocker Cd2+ (Jospin et al. 2002) suggest that Ca2+ serves as the charge carrier in the depolarizing phase of all-or-none APs. To assess the role of extracellular Ca2+, we first tested the effect of adding Cd2+ (500 μm) to the standard extracellular solution, and found that this eliminated APs (Fig. 3A). We next tested the effects of different concentrations of extracellular Ca2+ on AP firing. All-or-none overshooting APs were always observed with 1 or 5 mm[Ca2+]o but never with 200 or 500 μm[Ca2+]o, and the amplitude of APs was significantly lower at 1 mm than at 5 mm[Ca2+]o (Fig. 3). These observations with different [Ca2+]o confirmed the importance of Ca2+ entry in generating muscle APs. Interestingly, non-overshooting plateau potentials were sometimes observed at 500 μm[Ca2+]o (not shown), suggesting that this [Ca2+]o was near but not above the threshold for overshooting APs, and probably insufficient for triggering fast repolarization. Adding Cd2+ to the extracellular solution also resulted in a significantly less negative resting membrane potential (−13.2 ± 0.7 mV) (Fig. 3), which presumably resulted from decreased activity of Ca2+-activated K+ channel(s) (e.g. SLO-2) that are important to setting the resting membrane potential (see below). However, reducing [Ca2+]o from 5 mm to 200 μm hyperpolarized the cells (Fig. 3B), although the mechanism underlying this effect is currently unclear.
We next determined which Ca2+ channel(s) carries the inward currents that underlie body-wall muscle APs. As null mutants of EGL-19, the predominant inward current carrier in C. elegans body-wall muscle cells (Jospin et al. 2002), are not viable (Lee et al. 1997) and nifedipine only blocks ∼50% of inward currents in C. elegans body-wall muscle cells (Jospin et al. 2002), we tested the effect of nemadipine-A, a putative specific blocker of the EGL-19 channel (Kwok et al. 2006). Application of nemadipine-A (3 μm) eliminated APs (Fig. 4A) as well as whole-cell inward currents (Supplementary Fig. S3) of body-wall muscle cells. To further confirm the importance of EGL-19, we analysed APs in egl-19(ad695), a putative gain-of-function (gf) mutant (Lee et al. 1997). Both AP width and decay time were greatly increased in this mutant compared with wild-type (Fig. 4), consistent with EGL-19 channels carrying inward current during the AP. Surprisingly, the amplitudes of both APs and AHPs were significantly decreased in egl-19(gf) mutant worms (Fig. 4), suggesting that the mutated channel may have other unusual properties.
In addition to EGL-19, we investigated the role of CCA-1, UNC-2, NCA-1, and NCA-2 channels in spike generation by analysing APs in null mutants, including cca-1(ad1650), unc-2(e55) and the double mutant nca-1(gk9);nca-2(gk5) (Shtonda & Avery, 2005; Bauer Huang et al. 2007; Jospin et al. 2007; Yeh et al. 2008). APs displayed normal depolarization and repolarization in the nca-1;nca-2 strain (Fig. 4). By contrast, APs in the unc-2 mutant showed increases in AHP amplitude and inter-spike interval relative to the wild-type (Fig. 4), while APs in the cca-1 mutant displayed lower amplitude, longer decay time, wider peak and more spikes per train (Fig. 4). These observations suggest that the T-type Ca2+ channel CCA-1 also probably plays a role in generating and shaping muscle APs. However, the effect of unc-2 mutation was probably indirect and resulted from deficient neuromuscular transmission because (1) UNC-2 expression in body-wall muscle cells was not detected (http://www.wormbase.org) and (2) neurotransmitter release at the C. elegans neuromuscular junction is greatly decreased in the unc-2(e55) mutant (Richmond et al. 2001).
K+ channels shape muscle APs
Our previous study with cultured C. elegans embryonic body-wall muscle cells showed that outward currents are mainly carried by three K+ channels, including the two voltage-gated channels SHL-1 and SHK-1, and the Ca2+/Cl−-gated channel SLO-2 (Santi et al. 2003). SHL-1 carries fast transient, whereas SHK-1 and SLO-2 carry sustained, outward currents (Santi et al. 2003). The importance of SLO-2 in carrying sustained outward currents was also confirmed by recording whole-cell currents from body-wall muscle cells in situ (Santi et al. 2003; Yuan et al. 2003). To determine whether these K+ channels are important in shaping muscle APs, we analysed APs in shl-1(ok1168), shk-1(ok1581) and slo-2(nf101). shl-1(ok1168) and slo-2(nf101) are putative nulls resulting from deletions (Yuan et al. 2003; Fawcett et al. 2006). shk-1(ok1581) was also a deletion mutant but the precise region of deletion was unclear. We sequenced the genomic DNA of this mutant and found a 634 bp deletion (9766598–9767231 on chromosome II) plus an 8 bp insertion (CTAAATAT), which resulted in a mutation of methionine 252 to isoleucine (based on amino acid number of the ZK1321.2c isoform, http://www.wormbase.org) followed by a stop codon. This mutation eliminates essentially all amino acid residues following the T1 domain (Wei et al. 1996).
shl-1(ok1168) did not show any significant change in AP parameters when compared with the wild-type (Fig. 5). In contrast, APs from shk-1(ok1581) showed increases in spike amplitude, width, decay time and AHP amplitude, and a decrease in the number of spikes per train (Fig. 5). The increase in spike amplitude suggests that SHK-1 is activated during the AP upstroke and plays an important role in initiating repolarization while the increase in spike width suggests that SHK-1 is important to repolarization. The larger AHP amplitude was presumably because of enhanced activation of Ca2+-activated K+ channels as a result of increased Ca2+ entry associated with spike broadening. The decrease in the number of spikes per train could be secondary to the increase in AHP amplitude. Thus, SHK-1 plays important roles in repolarization and in regulating the pattern of AP firing.
Like the shk-1 mutant, slo-2(nf101) showed an increase in both AP width and decay time compared with the wild-type. In addition, the slo-2 mutant exhibited a lower (less negative) resting membrane potential, a smaller AHP amplitude, a shorter inter-spike interval, and the occurrence of long-duration (>2 s) all-or-none membrane depolarizations, which may be referred to as plateau potentials (Lockery & Goodman, 2009) (Fig. 6). The plateau potentials frequently appeared as the last one or two events in a train of APs but sometimes occurred repeatedly (Fig. 6A). These observations suggest that SLO-2 regulates body-wall muscle excitability in several ways, including setting the resting membrane potential, facilitating spike repolarization, enhancing AHP and prolonging the inter-spike interval.
To assess the combined effects of SHK-1 and SLO-2, we built shk-1(ok1581);slo-2(nf101) double mutant worms and analysed APs in this strain. While the slo-2 mutant showed plateau potentials as well as typical APs, the double mutant showed exclusively plateau potentials (Fig. 6). The duration of plateau potentials was significantly longer in the double mutant (4.4 ± 0.4 s) than in slo-2 single mutant (2.8 ± 0.3 s) (P < 0.05, unpaired t test). A significant increase in AP amplitude was observed in the double mutant compared with wild-type (54 ± 2.2 mV versus 45.1 ± 0.7 mV) (Fig. 6A), which was also seen in the single mutant of shk-1 but not slo-2. Thus, the changes observed in the double mutant were consistent with the combined effects of the two single mutants.
The effects of shk-1 and shk-1:slo-2 mutations on whole-cell currents had not been analysed in body-wall muscle cells in situ, which could differ from the previously analysed cultured embryonic muscle cells in ion channel function and expression. To substantiate the roles of SHK-1 and SLO-2 in shaping APs, we compared the amplitude of whole-cell outward currents among the wild-type, shk-1(ok1581), slo-2(nf101) and the double mutant shk-1(ok1581):slo-2(nf101). Sustained outward currents were significantly decreased in either single mutant, and essentially absent in the double mutant (Fig. 7). Thus, SHK-1 and SLO-2 play key roles in shaping muscle APs by acting as major outward current carriers.
AP firing is associated with an elevation of intracellular [Ca2+] and muscle contraction
Elevation of [Ca2+]i is a key step initiating contraction in all myocytes. To determine how AP firing is related to Ca2+ mobilization and contraction, we analysed the temporal correlation between spike firing and [Ca2+]i in worms expressing the genetically encoded Ca2+ indicator GCaMP2 (Tallini et al. 2006) in body-wall muscle cells under the control of Pmyo-3 (Okkema et al. 1993). To improve expression in vivo, GCaMP2 was synthesized for C. elegans codon optimization and insertion of two introns. APs and Ca2+ transients were recorded simultaneously in vivo in dissected worms. APs recorded from muscle cells expressing GCaMP2 were variable in amplitude (Fig. 8A), probably a consequence of increased Ca2+ buffering by the intrinsic calmodulin domain of GCaMP2 (Tallini et al. 2006). Nevertheless, a correlation between AP firing and Ca2+ transients was observed, with a single Ca2+ transient peak often associated with a train of spikes, and smaller Ca2+ transients associated with fewer spikes (Fig. 8). Moreover, imaging experiments demonstrated that the elevations of [Ca2+]i were associated with muscle twitches (Supplementary Movie 1). A temporal correlation between Ca2+ transients and muscle contractions was established by measurements of the length of selected muscle cells in preparations in which Ca2+ transients had been measured (Fig. 9). Finally, we confirmed that Ca2+ transients could occur in the absence of neuromuscular transmission by recording Ca2+ transients in unc-49(e407) mutant worms in the presence of 0.5 mm d-TBC (Supplementary Fig. S4 and Supplementary Movie 2), which is in agreement with our observations of APs (Fig. 2).
The observations that APs and Ca2+ transients continued to occur in the absence of neuromuscular transmission were made with dissected worms in the presence of artificial extracellular and intracellular solutions, which raised the question as to whether body-wall muscle cells can indeed generate APs and Ca2+ transients spontaneously under physiological conditions. To answer this question, we recorded Ca2+ transients in intact worms of the wild-type as well as mutants that are deficient either in postsynaptic receptors or presynaptic release ability. Ca2+ transients were observed in all the strains that we analysed, including the wild-type, unc-49(e407), which lacks GABAA receptors (Bamber et al. 1999), unc-63(x37):acr-16(ok789), which lacks all acetylcholine receptors in body-wall muscle cells (Culetto et al. 2004; Francis et al. 2005; Touroutine et al. 2005) and unc-64(e246), which is severely defective in releasing neurotransmitters due to a mutation in the syntaxin gene (Wang et al. 2001) (Supplementary Fig. S5 and Supplementary Movies 3–6). These findings support the notion that body-wall muscle cells have intrinsic abilities to fire APs, since Ca2+ transients are associated with AP firing.
Both EGL-19 and UNC-68 are required for Ca2+ transients in muscle cells
Finally, we analysed the roles of EGL-19 and UNC-68 in the Ca2+ transients triggered by APs in body-wall muscle cells. These cytosolic Ca2+ transients could result from Ca2+ entry through VGCCs in the plasma membrane, Ca2+ release through RyRs in the SR membrane, or both. To determine whether EGL-19 and UNC-68 are required for the elevation of [Ca2+]i, we tested the effect of nemadipine-A on Ca2+ transients in wild-type worms, and analysed Ca2+ transients in unc-68(r1162), which is a null mutant of the single RyR gene in C. elegans (Maryon et al. 1996). We found that muscle Ca2+ transients were abolished in wild-type worms by nemadipine-A and never observed in unc-68 mutant worms (Fig. 10, Supplementary Movies 7 and 8). Thus, both the sarcolemmal Ca2+ channel EGL-19 and the sarcoplasmic reticulum RyR channel UNC-68 are necessary to produce muscle Ca2+ transients, which is similar to mammalian cardiac and striated muscles.
Regenerative membrane potential changes may be classified as APs, graded potentials, plateau potentials and intrinsic oscillations (Lockery & Goodman, 2009). APs are all-or-none events with a stereotypical waveform; graded potentials change amplitude and waveform according to the strength and waveform of the stimulus; plateau potentials are prolonged all-or-none depolarizations; and intrinsic oscillations are slow, cyclic alternations in the membrane potential (Lockery & Goodman, 2009). Our data indicate that C. elegans body-wall muscle cells generate all-or-none APs with a stereotypical waveform.
Several lines of evidence suggest that Ca2+ entry through the L-type Ca2+ channel EGL-19 plays a key role in generating muscle APs. First, APs were eliminated when the standard extracellular solution was modified by either reducing [Ca2+] or adding the Ca2+ channel blocker Cd2+. Second, blocking EGL-19 with nemadipine-A abolished APs whereas a gain-of-function mutation of egl-19 increased spike width and decay time. Third, overshooting spikes continued to occur under experimental conditions when Na+ flow could not possibly be responsible for them (details in the next paragraph). Besides EGL-19, CCA-1 might also play a role in generating or shaping muscle spikes, as the cca-1 mutant showed several significant changes in spike properties, including decreased spike amplitude, and increased spike width, decay time, and number of spikes per train. However, except for the decreased spike amplitude, the other changes are more easily interpreted as resulting from developmental changes caused by cca-1(lf), such as a compensatory increase in EGL-19 expression. Although further studies are needed to completely understand the roles of these two Ca2+ channels in body-wall muscle APs, our observations with both mutants are consistent with the previous reports of egl-19 and cca-1 expression in C. elegans body-wall muscle cells (Lee et al. 1997; Steger et al. 2005), and of EGL-19 being a predominant carrier of inward current in these cells (Jospin et al. 2002).
While Ca2+ entry played a key role in generating APs, extracellular Na+ was also implicated. Specifically, when [Na+]o was reduced from 140 mm to 4 mm, all-or-none overshooting APs became rare events and had smaller amplitudes (Supplementary Fig. S6A and B). The all-or-none overshooting APs observed under such experimental conditions were unlikely to be caused by Na+ influx because the Na+ equilibrium potential was −17.5 mV. The reduction of [Na+]o also caused a decrease in whole-cell inward currents without altering the current–voltage relationship, which was characteristic of high voltage-gated Ca2+ channels (Supplementary Fig. S6C). There are at least two potential mechanisms for the observed effects of low [Na+]o: (1) a permissive level of extracellular Na+ may be needed to effectively open VGCCs (e.g. EGL-19), and (2) EGL-19 and/or CCA-1 are not strictly selective for Ca2+. While mammalian VGCCs are highly selective for Ca2+ over Na+ (ratio >1000:1) (Tsien et al. 1987), it is unknown whether C. elegans VGCCs have a similarly high selectivity for Ca2+. Therefore, how extracellular Na+ contributes to AP firing and inward currents remains to be determined.
Loss-of-function mutation of either shk-1 or slo-2 delayed spike repolarization. In addition, SHK-1 dysfunction decreased the number of spikes per train while SLO-2 dysfunction caused lower (less negative) resting membrane potential, smaller AHP amplitude and shorter inter-spike interval. These observations suggest that SHK-1 facilitates repolarization and regulates AP firing pattern, and that SLO-2 plays several roles, including setting the resting membrane potential, facilitating AP repolarization, enhancing AHP and regulating the firing pattern. Facilitation of spike repolarization by SHK-1 appears to be a conserved function because spike broadening is also observed in mutants of the Drosophila Shaker K+ channel (Tanouye et al. 1981; Salkoff, 1985). In contrast, all the functions uncovered for SLO-2 were previously unknown for itself as well as its homologues in other species. Given that SLO-2 and its mammalian homologues (Slo2.1/Slick and Slo2.2/Slack) are widely expressed in neurons and muscle cells (Yuan et al. 2000; Bhattacharjee et al. 2002; Dryer, 2003; Santi et al. 2003; Bhattacharjee et al. 2005; Budelli et al. 2009), and that their activities may be modulated by a variety of factors, including intracellular Cl−, Ca2+, Na+ and pH (Yuan et al. 2000, 2003; Ruffin et al. 2008), neurotransmitters (Santi et al. 2006; Berg et al. 2007), nicotinamide adenine dinucleotide (NAD+) (Tamsett et al. 2009), oestradiol (Zhang et al. 2005), Fragile X mental retardation protein (Brown et al. 2010) and hypoxia/ischaemia (Mitani & Shattock, 1992; Yuan et al. 2003), the regulation of the resting membrane potential, spike waveform and firing pattern by SLO-2/Slo2 could have far reaching implications in cellular activities. It would be interesting to investigate whether SLO-2/Slo2 also plays an important role in shaping spike waveform or regulating spike firing pattern in mammals or in other cells of C. elegans.
Our observations suggest that AP-induced elevation of [Ca2+]i requires both egl-19 and unc-68 gene products; however, the relative contributions of extracellular Ca2+ entry and intracellular Ca2+ release, and the molecular linkage between these channels, remain to be established. The requirement for EGL-19 does not establish the importance of Ca2+ entry through it, since the major source of Ca2+ for muscle contractions could be the sarcoplasmic reticulum (SR), and the primary function of EGL-19 could be to activate the SR Ca2+-release channel UNC-68 through either Ca2+-induced Ca2+ release (Cannell et al. 1995; Verkhratsky & Shmigol, 1996; Kotlikoff, 2003) or direct protein–protein interactions between the L-type Ca2+ channel and ryanodine receptor (Schneider, 1994). Further investigations are needed to determine the role of Ca2+ entry and how UNC-68 may be activated in C. elegans body-wall muscle cells.
C. elegans body-wall muscle is controlled by cholinergic and GABAergic motoneurons, which contract and relax the muscle, respectively. The fact that spike trains continued to occur in the absence of neuromuscular transmission suggests that AP firing is probably an intrinsic property of body-wall muscle cells. Despite this intrinsic activity, motoneurons must play an important role in body-wall muscle AP firing under physiological conditions, through the modulation of muscle resting potential and the attendant effect on the AP firing pattern. This modulatory role of motoneurons is supported by pharmacological and genetic experiments. Neuromuscular blockade resulted in a decrease in the resting membrane potential and AP amplitude, suggesting tonic inhibitory control through GABAergic motoneurons. Moreover, mutations of UNC-2, the major Ca2+ channel controlling neurotransmitter release at the C. elegans neuromuscular junction (Richmond et al. 2001), resulted in an increase in inter-spike interval and AHP amplitude. While the mechanisms that underlie generation of the sinusoidal locomotory waveform are not well understood, the prevailing model enlists command interneurons and motoneurons as key players but body-wall muscle cells as simple actuators (de Bono & Maricq, 2005; Boyle & Cohen, 2008). This model does not explain the continued locomotory activities of mutant worms lacking both the levamisole-sensitive and nicotine-sensitive ACh receptors in body-wall muscle (Francis et al. 2005; Touroutine et al. 2005), however. Alternatively, the ‘neurocratic’ model proposes that propagation of electrical activity in muscle is the basis of locomotory waveform, and that the function of neurons is to modulate muscle activity (Crofton, 1971; Johnson & Stretton, 1980; Niebur & Erdos, 1988). In spite of the differences, these two models are not mutually exclusive and are consistent with a primary neuromodulatory role of C. elegans motoneurons.
The body-wall muscle APs reported here are distinct from those in C. elegans RMD interneurons and pharyngeal muscle cells. APs in RMD ring motor interneurons appear as membrane potential fluctuations between two stable levels (−70 and −20 mV) (Mellem et al. 2008), whereas those in pharyngeal muscle cells have a long duration (mean > 200 ms) with a plateau phase (Franks et al. 2002) that resembles APs generated by mammalian cardiac myocytes. The depolarization phase of APs in C. elegans pharynx also relies on EGL-19 and CCA-1 channels (Shtonda & Avery, 2005), which is consistent with the lack of Na+ channels in C. elegans, whereas the repolarization phase depends on the function of EXP-2, a voltage-gated K+ channel (Davis et al. 1999; Shtonda & Avery, 2005).
C. elegans body-wall muscle has been considered to be analogous to mammalian skeletal muscle (Moerman & Fire, 1997; Lecroisey et al. 2007). However, results of the present study suggest that APs in C. elegans body-wall muscle cells differ from mammalian skeletal muscle at least in three major ways. First, APs are myogenic in C. elegans body-wall muscle, but neurogenic in mammalian skeletal muscle; second, the time course of APs in C. elegans body-wall muscle is much longer than that in mammalian skeletal muscle; and third, APs in C. elegans body-wall muscle result from the activation of VGCCs rather than voltage-gated Na+ channels. The identities of ion channels important for spike repolarization also appear to be different between C. elegans body-wall muscle (K+ channels) and mammalian skeletal muscle (Cl− channels) (Hudson et al. 1995). Nevertheless, the finding that C. elegans body-wall muscle generates all-or-none APs brings the physiology of C. elegans muscle much closer to that of other metazoans, and strengthens its utility as a model organism.
Conception and design: Z.-W.W., P.L. and B.C. Data collection and analysis: P.L. did all the electrophysiological and Ca2+ imaging experiments, and all data analysis. Q.G. first observed all-or-none action potentials in C. elegans body-wall muscle cells. B.C. made the transgenic strain expressing Pmyo-3::GCaMP2. Data interpretation: Z.-W.W., P.L., B.C., M.I.K. and L.S. Article drafting and revisions: Z.-W.W. with inputs from the coauthors. Experiments were done at University of Connecticut Health Center. All authors approved the final version of the manuscript.
This study was supported by National Institutes of Health (5R01GM083049 and 1R01MH085927 to Z.-W.W., 5RO1DK065992 and GM086736 to M.I.K., and R01NS0661871-01 to L.S.). The authors wish to thank Erik Jorgensen for the unc-49 and nca-1:nca-2 strains, Janet Richmond for the unc-63:acr-16 strain, Ed Maryon for the unc-68 strain, Caenorhabditis Genetic Center for the unc-2, cca-1, egl-19, shl-1 and shk-1 strains, and Xi Bie for building slo-2:shk-1 double mutant and measuring solution osmolarities. The nca-1, nca-2, shl-1 and shk-1 strains were originally generated by the C. elegans Knockout Consortium.