Cross Talk with the GAR-3 Receptor Contributes to Feeding Defects in Caenorhabditis elegans eat-2 Mutants

Precise signaling at the neuromuscular junction (NMJ) is essential for proper muscle contraction. In the Caenorhabditis elegans pharynx, acetylcholine (ACh) released from the MC and M4 motor neurons stimulates two different types of contractions in adjacent muscle cells, termed pumping and isthmus peristalsis. MC stimulates rapid pumping through the nicotinic ACh receptor EAT-2, which is tightly localized at the MC NMJ, and eat-2 mutants exhibit a slow pump rate. Surprisingly, we found that eat-2 mutants also hyperstimulated peristaltic contractions, and that they were characterized by increased and prolonged Ca2+ transients in the isthmus muscles. This hyperstimulation depends on cross talk with the GAR-3 muscarinic ACh receptor as gar-3 mutation specifically suppressed the prolonged contraction and increased Ca2+ observed in eat-2 mutant peristalses. Similar GAR-3-dependent hyperstimulation was also observed in mutants lacking the ace-3 acetylcholinesterase, and we suggest that NMJ defects in eat-2 and ace-3 mutants result in ACh stimulation of extrasynaptic GAR-3 receptors in isthmus muscles. gar-3 mutation also suppressed slow larval growth and prolonged life span phenotypes that result from dietary restriction in eat-2 mutants, indicating that cross talk with the GAR-3 receptor has a long-term impact on feeding behavior and eat-2 mutant phenotypes.

distinct types of contraction, termed pumping and peristalsis [reviewed in Avery and You (2012)]. Pumping occurs frequently (200/min), and it is a simultaneous contraction of muscles in the corpus and anterior isthmus that ingests bacteria into the lumen, and concentrates this material in the anterior region of the isthmus, while at the same time contraction of muscles in the terminal bulb crushes bacteria and expels the debris into the intestine ( Figure 1B). Peristalsis occurs relatively infrequently and always following a pump. Peristalsis is a wave-like contraction followed by rapid relaxation traveling through the pm5 muscle that transports a bolus of ingested bacteria from the anterior isthmus to the terminal bulb ( Figure 1C). The contractions in the pm5 muscle are remarkably complex. Three pm5 muscles are arranged around the circumference of the pharyngeal lumen and each of these cells extends through the entire length of the isthmus ( Figure 1A) (Albertson and Thomson 1976). During a pump, the anterior portions of the pm5 cells contract to open the lumen in the anterior isthmus, but the posterior halves of these cells remain relaxed. In contrast, during peristalsis, the anterior portion of pm5 remains relaxed, while a wavelike contraction followed by relaxation travels from the center to the posterior isthmus.
Rapid pumping is stimulated by the two MC neurons, which form synapses on the muscles near the junction of the pm4 and pm5 muscles (McKay et al. 2004), and laser ablation of the MCs results in a slow pump rate (Avery and Horvitz 1989).
The MCs stimulate an nAChR containing the non-a subunit EAT-2 that clusters at the MC synapses, and eat-2 mutants exhibit the same slow pump rate observed in MC-ablated animals (Raizen and Avery 1994;McKay et al. 2004). EAT-2 stimulation initiates an increase in cytoplasmic Ca 2+ concentration in the pharyngeal muscles that is mediated by the voltage-activated Ca 2+ channels CCA-1 and EGL-19, leading to muscle contraction Steger et al. 2005;Kerr 2006). MC does not synapse on the terminal bulb, but simultaneous contraction of the corpus and the terminal bulb is mediated by electrical coupling (Starich et al. 1996). In the absence of MC or EAT-2, pumping occurs slowly, and it is believed that these contractions are generated autonomously within the muscle.
Peristalsis is stimulated by the single M4 motor neuron that extends processes through the isthmus and synapses on the muscles in the posterior half of pm5 (Albertson and Thomson 1976), and ablation of M4 eliminates peristalsis (Avery and Horvitz 1987). The receptors that respond to M4 to produce a peristalsis are unknown, but our previous studies of M4defective mutants have implicated mAChRs (Ray et al. 2008;Ramakrishnan and Okkema 2014), while other studies suggest that peptide neurotransmitters secreted from other neurons may also be involved (Song and Avery 2012). The C. elegans genome contains three genes encoding mAChRs-gar-1, gar-2, and gar-3-but only gar-3 is expressed in pharyngeal muscles (Lee et al. 2000;Steger and Avery 2004;Dittman and Kaplan 2008) and the muscarinic agonist arecoline can stimulate pumping through GAR-3 (Steger and Avery 2004).
Here, we examine the roles that nAChRs and mAChRs play in peristalsis in wild-type animals, and mutants defective in ACh synthesis or response, using both direct observation and imaging of the genetically encoded Ca 2+ indicator (GECI) GCaMP3 (Tian et al. 2009). We find that mutants lacking endogenous ACh fail to pump or peristalse, indicating that pharyngeal muscle contractions are not generated autonomously. Further, treatment of these mutants with either nicotinic or muscarinic agonists stimulates both pumping and peristaltic contractions, and the response to these agonists depends on the nAChR subunit EAT-2 and the mAChR GAR-3, respectively. Surprisingly, in the absence of exogenous agonist, eat-2 mutants exhibit hyperstimulated isthmus muscle peristalses and increased isthmus muscle Ca 2+ concentrations, and this hyperstimulation depends on cross talk with the GAR-3 receptor. This relationship between EAT-2 and GAR-3 affects feeding behavior throughout the life of the animal, as gar-3 mutation suppresses the slow larval growth and prolonged life span phenotypes resulting from dietary restriction in eat-2 mutants (Avery 1993a;Lakowski and Hekimi 1998). ace-3 acetylcholinesterase mutants similarly exhibit hyperstimulated isthmus peristalses that are dependent on gar-3, and we hypothesize that, in the absence of EAT-2 or ACE-3, unbound ACh spills over from the synapse and stimulates extrasynaptic GAR-3 in the isthmus muscles.

Analysis of feeding behavior and drug studies
To analyze pharyngeal muscle contractions, L1 larvae hatched in the absence of food were suspended in 5 ml of M9 buffer containing OP50 and imaged on a 2% agarose pad under a coverslip. For drug treatments, either arecoline [catalog number (Cat#) CAS: 300-08-3; Acros Organics] or nicotine (Cat# N5260-25G; Sigma [Sigma Chemical], St. Louis, MO) was included, and animals were treated for 15 min prior to adding a coverslip. Individual N2 or mutant animals that pumped were recorded at 25 frames/sec for 1 min (cha-1 mutants were recorded for 5 min) using a Zeiss ([Carl Zeiss], Thornwood, NY) AxioImager microscope with an MRm camera and ZEN Software. For each genotype and drug treatment, the feeding behavior was analyzed in at least five animals (500 pumps/per animal). Time-lapse images were exported and processed using Fiji/ImageJ (Schindelin et al. 2012) to generate time-lapse movies. Acquisition times were exported from ZEN and quantifications were performed using Microsoft Excel.

Calcium imaging and microscopy
Young adult animals were incubated in 5 ml of 20 mM serotonin (Cat# H7752-5G; Sigma) for 10 min on a 2% agarose pad, and then immobilized using 1.5 ml polystyrene 0.10 mm microspheres and a coverslip (Cat# 00876; Polysciences, Warrington, PA). GCaMP3 imaging was performed on a Zeiss AxioImager microscope using a Q-Imaging Rolera EM-C 2 Figure 2 Muscarinic and nicotinic receptor agonists stimulate pharyngeal muscle contractions in cha-1 mutants. The percentage of wild-type (wt) or cha-1(ok2253) L1 animals exhibiting pharyngeal muscle pumping and peristalses, either untreated or treated with the indicated concentrations of arecoline or nicotine. Animals were visualized for 5 min each and the number of animals observed (n) is indicated.
EMCCD camera. Time-lapse movies were captured at 25-30 frames/sec using ZEN software and 14-bit TIFF images were exported. Animals pumping with a frequency of ,100 pumps/min were analyzed. For quantification, the pharynx was straightened with CellProfiler (Kamentsky et al. 2011) and aligned using the StackReg Fiji plugin (Thevenaz et al. 1998). The isthmus was cropped, and two regions of interest (ROI) were drawn as 10 pixel (2.54 mm)-wide lines in the center and posterior isthmus. Fluorescence measurements were analyzed using custom Matlab scripts or using Microsoft Excel. Total fluorescence was measured within the ROI and normalized for each GCaMP3 peak using the formula: normalized DF = (Fmax 2 Fmin)/Fmin*100% (where Fmax is the maximum fluorescence of the GCaMP3 peak and Fmin is the minimum fluorescence immediately before the GCaMP3 increase). Peak duration was quantified as a width of GCaMP3 fluorescence at half-peak height. Rise time was quantified as the time that it took for each fluorescence peak to reach its maximum. A Student's t-test was used to compare GCaMP3 fluorescence measurements between different genotypes and boxplots were generated using the Matlab boxplot command. Peak delay was calculated as the time between the maximum increase in GCaMP3 fluorescence in the center and posterior isthmus using the Matlab diff command. Average baseline fluorescence levels in the entire pharyngeal isthmus were measured in a representative image during an interpump period acquired at a time point early in time-lapse image sets of each animal using Fiji/ImageJ. GFP and DIC images were captured using a Zeiss Axiocam MRm and Zeiss Zen software. To characterize ceh-19b Prom ::gfp expression, Z-series images were collected. Maximum intensity Z projections were produced using Fiji/ImageJ (Schindelin et al. 2012). ceh-19b Prom ::snb-1::gfp expression was faint, and Z-series were collected in adult animals immobilized in 10 mM NaN 3 and false colored in Fiji/ImageJ using the Rainbow RGB look up table.
animals were allowed to grow for an additional day to quantify slow-growing adults. For each genotype, two plates were set up with 50 embryos each.

Life span assay
Life span assays were performed according to a previously published protocol (Sutphin and Kaeberlein 2009). Assays were performed starting with 30 L4 worms fed with UV-killed OP50 Escherichia coli at 20°, and dead or missing animals were replaced by animals of the same age. For each experiment, triplicate plates of each genotype were assayed, and representative results of three or more experiment are shown. Kaplan-Meier survival analysis was used to generate survival plots and calculate median life span, and for statistical analysis (GraphPad Prism 5).

Data availability
Strains and plasmids are available upon request. Supplemental Material, Figure S1 contains measurements of baseline GCaMP3 fluorescence intensity in wild-type and mutant strains. Figure S2 contains tracings of GCaMP3 fluorescence intensity from representative wild-type and mutant animals. Movies S1-S9 include time-lapse images of wild-type and mutant C. elegans. Table S1 contains measurements of pharyngeal muscle contractions in acr mutant C. elegans. Tables S2-S52 contain raw data of pharyngeal muscle contractions (Tables S2-S13) and GCaMP3 fluorescence intensities (Tables S14-S52) derived from time-lapse images. Supplemental material available at Figshare: https://doi.org/10.25386/genetics. 6160355.

Results
Pumping and isthmus peristalsis can be stimulated through both nicotinic and muscarinic receptors We are interested in the signaling mechanisms that produce pharyngeal muscle contractions and productive feeding. M4 and the MCs are cholinergic motor neurons that express choline acetyltransferase encoded by cha-1 (Raizen et al. 1995;Ramakrishnan and Okkema 2014;Pereira et al. 2015), and strong cha-1 mutants have reduced pharyngeal contractions and arrest as severely uncoordinated L1s (Rand 1989;Avery and Horvitz 1990). Previous studies have shown that stimulation of nAChRs with nicotine or mAChRs with arecoline could induce pharyngeal muscle contractions, but these studies did not specifically examine peristalses in the absence of endogenous ACh (Avery and Horvitz 1990;Raizen et al. 1995;Ramakrishnan and Okkema 2014).
To extend this work, we examined the effect of nicotine or arecoline treatment on cha-1(ok2253) mutant L1 animals. cha-1(ok2253) is a previously uncharacterized null allele containing a 1.7-kb deletion that eliminates the catalytic histidine (His341) required for ACh synthesis, and homozygous mutants arrest as paralyzed L1s with a coiled appearance (C. elegans Deletion Mutant Consortium 2012). Untreated cha-1(ok2253) homozygotes completely lacked pharyngeal muscle contractions, but mutant animals treated with either nicotine or arecoline exhibited both pumping and peristalsis ( Figure 2 and Movies S1-S4). These results demonstrate that ACh is necessary for pumping and peristalsis, indicating that Figure 3 eat-2 mutants are insensitive to exogenous nicotine. Percentage of wild-type or eat-2(ok3528) L1 animals exhibiting pharyngeal muscle pumping and peristalses, either untreated or treated with the indicated concentrations of nicotine. ** indicates significantly different from untreated animals (P , 0.0001), and the bar indicates significant difference between wild-type animals treated with increasing concentrations of nicotine. Animals were visualized for 5 min each and the number of animals observed (n) is indicated. pharyngeal muscles do not have autonomous contractile activity. Further, activation of either nAChRs or mAChRs can stimulate these contractions. Notably, we observed occasional animals treated with either agonist that exhibited peristaltic contractions in the isthmus without a preceding pump (2/10 animals treated with 10 mM arecoline and 2/40 animals treated with 5 mM nicotine), consistent with our previous observations that peristalsis can be uncoupled from pumping by directly stimulating receptors in the isthmus muscles (Ramakrishnan and Okkema 2014). In addition, while higher arecoline concentrations increased the percentage of animals exhibiting pharyngeal muscle contractions, higher nicotine concentration progressively decreased the percentage of animals exhibiting pharyngeal muscle contractions. This decrease may result from hyperstimulation of nAChRs, as has been previously reported (Avery and Horvitz 1990), although we did not observe the tetanic contractions previously observed in animals treated with nicotine.
EAT-2 and GAR-3 mediate the response to agonists The results above indicate that activation of either mAChRs or nAChRs can stimulate pumping and peristaltic contractions in the pharynx. To identify the receptors involved in these responses, we asked if mutants affecting receptors expressed in the pharyngeal muscles become insensitive to agonists.
To examine the role that mAChR signaling plays in pharyngeal muscle contractions, we characterized these contractions in gar-3(gk305) null mutants (Liu et al. 2007). gar-3(gk305) mutants are viable, look healthy, and grow normally, and pharyngeal muscle contractions in gar-3(gk305) L1s were similar to those in wild-type worms (Table 1). Neither the rate of pumping, nor the frequency or duration of peristalsis, was significantly different between wild-type and gar-3(gk305) animals. In comparison, gar-3(gk305) animals were almost completely insensitive to exogenously applied arecoline (Table 1). Wild-type animals treated with arecoline exhibited a strongly reduced pump rate, and each of these pumps was followed by a peristalsis that was prolonged compared to untreated animals. In contrast, gar-3(gk305) mutants were largely unaffected by arecoline. The pump rate was unaffected in gar-3(gk305) and, although these animals did exhibit a small decrease in the duration of peristaltic contractions when treated with arecoline, we note that this change is the opposite of the prolonged peristalses observed in arecolinetreated wild-type animals. Thus, GAR-3 mediates the response to arecoline, but loss of this receptor does not have a strong effect on pharyngeal muscle contractions in the absence of this drug.
We then focused on eat-2, which encodes the only nAChR subunit known to function in the pharyngeal muscles (Raizen et al. 1995;McKay et al. 2004). To test if eat-2 is necessary to respond to nicotine, we compared wild-type animals and eat-2(ok3528) mutants treated with nicotine at the L1 stage. Like cha-1 mutants, wild-type animals treated with increasing concentrations of nicotine exhibited a dose-dependent decrease in pharyngeal muscle contractions (Figure 3), but eat-2 animals were not significantly affected by this treatment. Thus, the pharyngeal muscle response to nicotine depends on EAT-2-containing nAChRs.

eat-2 mutants exhibit prolonged pumps and peristalses
To understand how EAT-2 affects muscle contractions, we examined three strong eat-2 mutants at the L1 stage. eat-2(ad465) and eat-2(ad1116) are point mutations introducing an early stop codon and affecting a splice site, respectively (McKay et al. 2004) (WormBase WBVar00000089), while eat-2(ok3528) contains a 614-bp deletion predicted to cause a frameshift mutation upstream of the transmembrane domains (C. elegans Deletion Mutant Consortium 2012). As expected, the rate of pumping in all three mutants was significantly reduced compared to in wild-type worms (Table 1) (Avery 1993a;Raizen et al. 1995). However, the duration of peristaltic contractions in the posterior isthmus was unexpectedly increased up to nearly fourfold, with relaxation of the posterior region of the isthmus muscles particularly delayed (Movie S5). The percent of pumps that were followed by a peristalsis was also increased in eat-2 mutants, but as the pump rate was decreased in these animals, the absolute rate of peristalsis was similar in wild-type animals and eat-2 mutants. Notably, eat-2(ok3528) mutants exhibited the strongest phenotypes, which, together with a molecular lesion in this allele, suggests that eat-2(ok3528) is a null allele.
We next examined MC in eat-2(ok3528) mutants using a ceh-19b Prom ::gfp reporter (Feng and Hope 2013), and found that MC morphology is similar in wild-type and eat-2(ok3528) adults (Figure 4). The MC cell bodies were located in the anterior bulb of the pharynx and these cells extended processes posteriorly into the pharyngeal isthmus. Varicosities in these processes that mark synaptic sites were observed in both strains at the junction between the pm4 and pm5 muscles, and in the anterior region of the isthmus. We then looked specifically at the locations of synapses in MC using a snb-1::gfp fusion expressed using the ceh-19b promoter. SNB-1::GFP is a functional synaptobrevin that marks synaptic vesicle clusters in presynaptic cells (Nonet 1999). In both wild-type adults and eat-2(ok3528) mutants, synapses were visible in the MC process beginning near the junction of the pm4 and pm5 muscle cells, and extending into the anterior region of the isthmus ( Figure 5). eat-2(ok3528) mutants exhibited a small but not statistically significant decrease in the number of vesicle clusters compared to wild-type (wildtype 7.0 6 2.40 clusters, n = 8; eat-2(ok3528) 4.9 6 1.0; P = 0.09, n = 13). Thus, the prolonged peristalses in eat-2 mutants do not result from morphological or synaptic defects in the MCs. Notably, using both of these reporters, we observed the MC axon extending into the anterior isthmus further than has been previously reported (Albertson and Thomson 1976) and forming synapses directly on the pm5 muscles.
Taken together, these results indicate that loss of the EAT-2 nAChR subunit results in slow pumping and prolonged peristalses that occur after nearly every pump. Paradoxically, the eat-2 mutant defects in peristalsis are opposite to those in pumping, indicating that, while wild-type EAT-2 stimulates rapid pumping, it also limits the duration of peristaltic contractions in the isthmus muscles.

gar-3 mutation suppresses the peristalsis defects in eat-2 mutants
Pharyngeal muscle contractions in eat-2 mutants are strikingly similar to those of wild-type animals treated with arecoline (Table 1), suggesting that some of the eat-2 mutant phenotypes are related to mAChR signaling. Since GAR-3 is the mAChR responding to arecoline in the pharynx, we examined pharyngeal muscle contractions in eat-2(ok3528); gar-3(gk305) double mutants.
We found that gar-3 mutation partially suppressed the prolonged peristalses observed in eat-2 single mutants ( Table 1). The duration of peristalses in eat-2(ok3528); gar-3(gk305) double mutants was strongly reduced compared to eat-2(ok3528) single mutants, although they were still longer than those in wild-type animals or gar-3(gk305) single mutants. In contrast, the reduced pumping frequency and increased frequency of peristalsis observed in eat-2(ok3528) single mutants was not suppressed in eat-2(ok3528); gar-3(gk305). Thus, eat-2 loss-of-function affects pumping and peristalsis by different mechanisms, and only the prolonged peristalsis phenotype depends on cross talk with the GAR-3 receptor.
Ca 2+ transients in the pharyngeal isthmus parallel the contractile phenotypes in eat-2 and gar-3 mutants To characterize isthmus muscle excitation in wild-type animals and mutants, we constructed strains expressing the GECI GCaMP3 using the pharyngeal muscle-specific myo-2 promoter (Okkema et al. 1993;Tian et al. 2009). GCaMP3 fluorescence increases in muscle cells as cytoplasmic Ca 2+ concentrations increase during excitation contraction coupling [reviewed Bers (2002)]. Similar to previous analyses of GECIs in the pharynx, we characterized adult animals treated with serotonin to stimulate pharyngeal muscle contractions and focused on animals pumping slowly to resolve individual excitation events (, 100 pumps/min; , 1.67 Hz) (Shimozono et al. 2004;Kerr 2006). Baseline fluorescence levels in the isthmus muscles were similar in the strains examined ( Figure S1) and representative traces of time-lapse fluorescence are shown in Figure S2.
We observed changes in Ca 2+ concentration in the isthmus muscles that were very dynamic ( Figure 6 and Movie S6). As reported in previous studies, we observed that wild-type animals displayed an increase in Ca 2+ throughout the central isthmus during pumping, followed by a longer and delayed increase in Ca 2+ in the posterior isthmus during peristalsis (Figure 7, A and B and Movie S6) (Shimozono et al. 2004). We found that this Ca 2+ signal during peristalsis exhibited a wave-like increase and decrease that traveled in an anterior to posterior direction through the posterior isthmus, and resembled the progression of the open pharyngeal lumen during peristalsis (Movie S7). While Ca 2+ waves have not been previously described in the C. elegans pharynx (Shimozono et al. 2004), similar waves of Ca 2+ can be observed in individual cardiac muscle cells, which are generated by progressive release of Ca 2+ released from adjacent sites in the sarcoplasmic reticulum [reviewed in Stuyvers et al. (2000)]. This mechanism may also underlie the wave-like contraction of pm5 during peristalsis.
To compare Ca 2+ transients in the center and posterior isthmus in wild-type animals to those in various mutants, we quantified for each pump and peristalsis the normalized change in fluorescence (DF/F o ), the width of each peak at 50% maximum fluorescence (peak duration), and the time that it took for fluorescence to reach its maximum (rise time) ( Figure 7C and Table 2). In addition, we calculated the time delay between the maximum increase in GCaMP3 fluorescence in the center and posterior isthmus (peak delay).
We found changes in GCaMP3 fluorescence in the pharyngeal isthmus that paralleled many of the defects we observed in peristaltic contractions in various mutants (Figure 8 and Table 2). eat-2(ok3528) mutants, which have prolonged peristalses, also exhibited significantly increased DF/F o , peak duration, and rise time, in both the center and posterior isthmus. In addition, the peak delay was increased twofold (Movies S8 and S9). In comparison, gar-3(gk305) mutants, which have normal duration of peristalses, exhibited DF/F o and rise times that were similar to those of wild-type animals, with a small decrease in DF/F o only in the posterior isthmus, as well as a small decrease in peak duration in both the center and posterior isthmus. However, the most striking change in these animals was a strongly decreased peak delay between the center and posterior isthmus. Finally, eat-2(ok3528); gar-3(gk305) double mutants, which suppress the prolonged peristalses observed in eat-2(ok3528) single mutants, exhibited decreases in DF/ F o , peak duration, rise time, and peak delay compared to eat-2(ok3528) single mutants, although only DF/F o was reduced to wild-type levels. Taken together, these results show that the increases in level, duration, and rise time of Ca 2+ , in both the center and posterior isthmus, best correlate with the duration of peristaltic contractions. Loss of eat-2 increases these Ca 2+ signals and produces prolonged peristalses that depend on wild-type gar-3.

ace-3 mutants exhibit prolonged peristalses that depend on GAR-3
It is surprising that eat-2 loss leads to a reduced pump rate, while at the same time hyperstimulating peristalsis and increasing the cytoplasmic Ca 2+ concentration in the isthmus muscles. One hypothesis for these unexpected effects is that, in eat-2 mutants, ACh released from MC spills over from the synapses and stimulates GAR-3 receptors located in the isthmus muscles. To examine this hypothesis, we characterized pharyngeal muscle contractions in ace mutants defective in acetylcholinesterases that normally hydrolyze ACh and have increased unbound ACh in the synapse (Johnson et al. 1988;Combes et al. 2000). ace-3(dc2) mutants exhibited significantly prolonged peristalses, which were similar to those of eat-2 mutants (Table 1). In comparison, ace-2(g72); ace-1(p1000) double mutants exhibited only mildly prolonged peristalses. Both of these strains also exhibited a small decrease in the pump rate, but this was not significantly different from wild-type, indicating that MC can effectively stimulate pumping in ace mutants.
To ask if the prolonged peristalses were suppressed by gar-3 mutation, we examined ace-3(dc2); gar-3(gk305) double mutants and found that these animals exhibited peristalses that were similar to wild-type (Table 1). Thus, the prolonged peristalses in ace-3 mutants depend on cross talk with the wildtype GAR-3 receptor, which is similar to what we observed in eat-2 mutants. gar-3 suppresses the slow growth and extended life span of eat-2 mutants As gar-3 mutation suppresses the peristalsis defects in eat-2 mutants, we wanted to determine if this mutation also affects the persistent feeding defects in eat-2 mutants. eat-2 mutants exhibit slow larval growth and a prolonged adult life span, which result from dietary restriction (Avery 1993a;Lakowski and Hekimi 1998). We found that gar-3 mutation at least partially suppressed both of these phenotypes.
Thus, gar-3(gk305) suppresses both the slow larval growth and extended life span phenotypes of eat-2(ok3528) mutants, and we suggest that this suppression is the result of improved isthmus peristalsis and feeding.

Discussion
In this work, we show that cross talk between the nAChR EAT-2 and the mAChR GAR-3 affects peristaltic muscle contractions in the isthmus of the C. elegans pharynx. Under normal circumstances, GAR-3 has a relatively minor role in a Fifteen wild-type OK1020 young adults were recorded and an average of 130 pumps were analyzed. b Six eat-2(ok3528) young adults were recorded and an average 56 pumps were analyzed. These animals exhibited increased normalized DF, peak duration, and rise time in both the center and posterior isthmus (P , 10 27 ), and an increased peak delay (P , 10 24 ) compared to wild-type. c Five gar-3(gk305) young adults were recorded and an average of 31 pumps were analyzed. These animals exhibited a decreased peak delay (P , 10 25 ) compared to wildtype. d 12 eat-2; gar-3 young adults were recorded and an average of 94 pumps were analyzed. These animals exhibited decreased normalized DF, peak duration, rise time, and peak delay compared to eat-2(ok3528) single mutants (P , 10 25 ). these contractions, but in eat-2 mutants, GAR-3 can stimulate increased cytoplasmic Ca 2+ levels and peristalses that are prolonged compared to those of wild-type animals. We suggest that this GAR-3-dependent stimulation of peristalsis results from ACh spillover from synapses between MC and the pharyngeal muscles in eat-2 mutants, and, consistent with this suggestion, acetylcholinesterase mutants similarly exhibit prolonged peristalses that are dependent on GAR-3. Cross talk with GAR-3 contributes to the slow larval growth and extended life span phenotypes observed in eat-2 mutants, indicating that this cross talk has a long-term impact on C. elegans feeding. eat-2 mutants exhibit different effects on pumping and peristalsis eat-2 mutants have a reduced pump rate, indicating that EAT-2 plays an excitatory role in pumping (Avery 1993b;Raizen et al. 1995). In comparison, eat-2 mutants exhibit prolonged peristalses, indicating that EAT-2 does not excite isthmus muscle contractions. Rather, it is necessary for rapid relaxation. These peristalsis phenotypes are similar to previous observations demonstrating that eat-2 mutation leads to prolonged depolarization and contraction of the terminal bulb muscles during pumping (Steger and Avery 2004).
The EAT-2 receptor is specifically localized at the sites where MC forms synapses near the junction of the pm4 and pm5 pharyngeal muscles, and it is necessary for MC to directly stimulate rapid pumping (McKay et al. 2004). Because MC does not synapse on either the posterior isthmus or the terminal bulb (Albertson and Thomson 1976;McKay et al. 2004), the prolonged contraction of these muscles is likely an indirect effect of loss of EAT-2-containing receptors.

GAR-3 stimulates prolonged peristalsis in eat-2 mutants
The prolonged peristalses in eat-2 mutants depend on the GAR-3 receptor. gar-3 mutation partially suppressed the prolonged peristaltic contraction of eat-2 mutants, as well as the increased and prolonged Ca 2+ transients in the isthmus muscles of these animals. This suppression demonstrates that GAR-3 receptor activation contributes to the eat-2 mutant peristaltic phenotypes. Previous studies demonstrated that gar-3 mutation similarly suppressed the prolonged contraction of the terminal bulb muscles in eat-2 mutants (Steger and Avery 2004). In comparison, gar-3 single mutants have relatively minor defects in peristalsis. The duration of peristalsis was similar to wild-type animals. However, these mutants exhibited a sharp decrease in the peak delay between Ca 2+ transients in the center and posterior isthmus, and a decreased DF/F o in the posterior isthmus. Thus, in wild-type animals, GAR-3 contributes to the spatially and temporally dynamic changes in Ca 2+ concentration in the isthmus muscles, but this did not produce recognizable changes in the duration of peristalsis. These observations demonstrate that GCaMP3 may be a more sensitive assay for muscle excitation.
In contrast to its effect on peristalsis, gar-3 mutation did not suppress the reduced pump rate observed in eat-2 mutants (Steger and Avery 2004). Thus, the pumping and peristalsis defects in eat-2 mutants arise via different mechanisms. The reduced pump rate results directly from a loss of EAT-2 excitatory activity stimulated by the MC neuron, while the prolonged peristalsis results from excitation that depends on GAR-3.
ACh spillover may produce prolonged peristalses in eat-2 mutants Extrasynaptic GAR-3 has previously been shown to respond to humoral ACh (Chan et al. 2013) and we suggest that in eat-2 mutants, ACh spillover from synapses between MC near the junction of the pm4 and pm5 muscles similarly activates GAR-3 receptors in pm5. Activation of slow-acting, metabotropic GAR-3 stimulates increased and prolonged Ca 2+ transients in the isthmus, which lead to prolonged peristalsis. GAR-3 is known to regulate Ca 2+ -dependent processes in the pharyngeal muscles and increased GAR-3 signaling produces prolonged pharyngeal muscle contractions during pumping (Steger and Avery 2004). Because gar-3 mutation only partially suppresses the prolonged peristalsis in eat-2 mutants, other mechanisms must also contribute to it. For example, voltage-activated K + channels such as EXP-2 might not be efficiently activated in eat-2 mutants for rapid repolarization of the isthmus muscles (Davis et al. 1999;Shtonda and Avery 2005).
At cholinergic synapses, unbound ACh is normally degraded by acetylcholinesterases located at the synapse [reviewed in Rotundo (2003)]. However, ACh released at mammalian NMJs can coactivate mAChRs to stimulate vasodilation in nearby blood vessels (Welsh and Segal 1997), although the physiological significance of this stimulation is uncertain (Hong and Kim 2017). ACh spillover has also been observed when quantal content is increased or when acetylcholinesterases are inhibited (Stanchev and Sargent 2011;Petrov et al. 2014). We have similarly observed that C. elegans ace-3 mutants with reduced acetylcholinesterase activity produce prolonged peristalses that are dependent on GAR-3.
While MC morphology and synapse positioning is normal in eat-2 mutants, we cannot rule out that other defects in synapse formation or remodeling occur in these animals. Interestingly, loss of the g-subunit of the nAChR leads to a diffuse distribution of acetylcholinesterase clusters during the development of mouse diaphragm muscles (Liu et al. 2008), and a similar defect in acetylcholinesterase localization could underlie signaling from the MC to the pharyngeal isthmus muscles in eat-2 mutants.
Increased cytoplasmic Ca 2+ levels in muscle cells have been observed in a number of human neuromuscular disorders, including myotonic dystrophies, Duchenne muscular dystrophy, as well as in ColQ and slow-channel myasthenic syndromes [reviewed in Vallejo-Illarramendi et al. (2014), Engel et al. (2015)]. In particular, myotonic dystrophies are initially characterized by the delayed relaxation of muscles after contraction, and this resembles the prolonged peristalsis that we have observed in eat-2 and other mutants. Thus, Figure 9 gar-3 mutation suppresses the slow growth and prolonged life span of eat-2 mutants. (A) Bar graphs indicating the percent animals of the indicated genotypes reaching adulthood by day 5 or day 6 at 20°(n = 100 for each genotype). (B) Adult survival curves for animals of the indicated genotypes. Averages of representative triplicate assays performed on plates containing UV-killed E. coli as a food source. Nearly identical results were obtained in assays with living E. coli. eat-2 mutants may provide a new model to characterize the immediate effects of Ca 2+ dysregulation in muscle cells.