The Response to High CO Levels Requires the Neuropeptide Secretion Component HID-1 to Promote Pumping Inhibition

Summary

Carbon dioxide (CO2) is a key molecule in many biological processes. High levels of CO2 in patients with pulmonary diseases are associated with worse outcomes. However, mechanisms by which organisms sense and respond to high CO2 levels remain largely unknown. Using Caenorhabditis elegans as a model system, we found that exposure to high CO2 levels leads to a very rapid cessation in the contraction of the pharynx muscles. Further analysis revealed that the pharynx muscle response is controlled by dense core vesicle secretion from the BAG neurons in a hid-1-mediated pathway. This novel hid-1 pathway sheds new light on the physiological effects of high CO2 levels on animals at the organism-wide level.

Introduction

One of the fundamental features shared by most, if not all, living organisms is the ability to maintain levels of carbon dioxide (CO₂). Of particular importance is the ability of many animals to sense and respond to high levels of CO₂ by either attraction or aversion [1]–[5]. In mammals, high levels of CO₂ (hypercapnia) impair alveolar epithelial function of the lungs by activating the stress sensor AMPK, which leads to Na,K-ATPase endocytosis, impaired cell proliferation, and loss of distal lung epithelial function [6]–[10]. In addition, hypercapnia suppresses specific innate immune responses in Drosophila and mice, which increases mortality in a model of pneumonia and leads to changes in gene expression through the NF-kB pathway [11]–[14]. Cyclic AMP (cAMP) signaling also plays a role in the response of mammalian cells to elevated CO₂ levels [15]–[17]. The molecular pathways mediating the responses to hypercapnia are the focus of intensive research (see [11], [18] and review in [19]).

High levels of CO₂ quickly elicit an avoidance response in wild-type Caenorhabditis elegans animals via a cGMP signaling pathway [2], [4]. The cGMP-regulated avoidance response requires the CO₂- and oxygen (O₂)–sensing BAG neurons, in which the guanylyl cyclase receptor, gcy-9, controls the response to CO₂ [20]–[22]. Interestingly, the response to hypercapnia requires the ETS-domain transcription factor, ETS-5, which controls the expression of gcy-9 in the BAG neurons and plays a role in BAG neuron differentiation [20]–[22]. Recently, the thermosensory AFD neurons and the salt-sensing ASE neurons were also shown to participate in CO₂ sensing and avoidance [23]. These neurons, however, differ in their response kinetics to high levels of CO₂; whereas BAG neurons reach maximal activation within 30 s, ASE neurons reach maximal activation only after 2 min, and AFD neurons show intricate dynamics in which Ca²⁺ levels first drop and then increase to maximal levels after 2 min [23]. Interestingly, starved C. elegans do not avoid high CO₂ levels, nor do animals with defects in the daf-2 signaling pathway, which is an important regulator of the starvation response [2], [4].

In addition to avoidance, C. elegans exposed to high CO₂ levels show specific phenotypes independent of pH [24]. These include a smaller brood size, delayed development, reduced motility coupled with deterioration of striated muscle, and a significant increase in lifespan that is independent of known life-extending pathways [24].

Here we report that exposure of C. elegans animals to short (10 s) hypercapnia-inducing levels of CO₂ (≥5%) leads to a significant reduction in the rate of pharyngeal muscle contraction (pumping). Strikingly, this effect is independent of any currently known molecular pathways that regulate CO₂ avoidance or O₂ sensing. Specifically, through a forward genetic screen, we identified a novel participant in the response to CO₂, hid-1, that plays a role in continued pumping in the presence of high CO₂ levels. Moreover, we show that dense core vesicle secretion pathways in the BAG neurons contribute to the reduced pumping rate in response to high CO₂ levels.

Results

High levels of CO₂ significantly reduce the pumping rate of the pharynx

To investigate the effects of acute exposure of wild-type C. elegans (N2) to high levels of CO₂, we exposed 1-day-old adult animals grown on standard NGM plates in a small chamber to gas mixtures containing 21% O₂ and 5%, 10%, or 20% CO₂ at 22°C. In normal air (0.0391% CO₂), the rate of muscle contraction of the pharynx was ∼200 pumps/min. Within 10 s of exposure to 5% CO₂ balanced with 21% O₂ and 74% N₂, the pumping rate of the pharynx was reduced from ∼200 to ∼60 pumps/min (Figure 1A). Exposure to 10% and 20% CO₂ almost completely stopped the pumping of the pharynx (Figure 1A and Movie S1). After 2–3 min of continuous exposure to 10% CO₂, pumping rate recovered partially to ∼40 pumps/min, and after 5 min of continuous exposure to 10% CO₂ it recovered to ∼80 pumps/min (Figure 1B), suggesting a separate, existing mechanism that allows for a partial adaptation. Longer exposures of up to 30 min to 10% CO₂ did not result in full recovery of the pumping rate (Figure S1A). To test whether the effect on the pumping is mediated by a change in the pH of the growth medium due to high CO₂ levels, we measured the pumping rate of animals using NGM plates buffered to pH of 5.0 and 7.0 in addition to the normally used medium with a pH of 6.0. We did not find any differences between the animals in different growth mediums, both under normal air conditions and after exposure to 10% CO₂, which suggests that the effect on the pumping is probably not mediated by changes in pH (Figure S1B). This conclusion is supported by a recent finding that activation of CO₂-responsive neurons can occur independently of changes in extracellular or intracellular acidosis [25]. In addition, mutations in the carbonic anhydrase genes (cah-2, cah-5, and cah-6), which catalyze the conversion of CO₂ into bicarbonate, had no effect on the pumping rate (Figure S1C). This suggests that the conversion of CO₂ into bicarbonate is not necessary to induce the response of the pharynx. However, we cannot rule out the possibility of redundancy between the different carbonic anhydrase genes.

The response of the pharynx to high CO₂ levels was partially dependent on the nutritional state of the animal. Whereas “well fed” animals exposed to 10% CO₂ stopped pumping, animals starved for 4 h continued pumping at a rate of ∼60 pumps/min (Figure 1C). These starved animals exposed to 20% CO₂ stopped pumping, similar to wild-type animals (Figure 1C), suggesting a threshold effect of high CO₂ levels. Together, these data demonstrate that high CO₂ levels quickly affect muscle contraction of the pharynx, an effect that depends on the nutritional state of the animal.

The reduction in pumping is independent of CO₂ avoidance and O₂ sensing

C. elegans animals quickly withdraw when acutely exposed to CO₂. This response, known as CO₂ avoidance, is regulated by cGMP signaling [2], [4]. TAX-2 and TAX-4 are two subunits of a cGMP-gated ion channel required for normal chemosensory and thermosensory responses. C. elegans null mutants for either TAX-2 (tax-2[p691]) or TAX-4 (tax-4[p678]) do not avoid high CO₂. In addition, the insulin-IGF pathway mediates CO₂ avoidance, as daf-2 mutants show reduced CO₂ avoidance [2], [4]. The avoidance response also requires proper development of ciliated sensory neurons. Animals with mutations in osm-3 and che-10 have abnormal cilia as well as defective CO₂ avoidance. In C. elegans strains carrying mutations in daf-2, osm-3, or che-10 the CO₂ avoidance response is either reduced or absent [2], [4].

The effect of high CO₂ levels on the C. elegans pharynx is quick and robust, similar to the avoidance response. However, rather than CO₂ avoidance, tax-4(p678), daf-2(e1370), osm-3(n1540), and che-10(e1809) mutants show a significant reduction in pumping following exposure to 10% CO₂, similar to the reduction observed in wild-type (N2) animals (Figure 2A). Loss-of-function mutation of the neuropeptide Y receptor, npr-1, completely abolishes the CO₂ avoidance response by inhibiting the activity of the O₂-sensing URX neurons [3]. In our assay, exposing npr-1(ad609) animals to 10% CO₂ resulted in a response similar to that of wild-type animals (Figure 2A), suggesting that the high activity of the URX neurons in the animals with loss-of-function mutation in npr-1 does not regulate the pharynx response to CO₂. The gcy-9 gene encodes a receptor-type guanylyl cyclase and is a target of the ETS domain ETS-5 transcription factor. Both the ets-5 and gcy-9 genes are required for the CO₂ avoidance response [20]–[22]. When exposed to 10% CO₂, the gcy-9(tm2816), ets-5(tm1734), and ets-5(tm1755) mutants stopped pumping, similar to wild-type animals (Figure 2B), thus further demonstrating that CO₂ avoidance and acute CO₂-dependent pumping inhibition are mediated through independent pathways.

To determine whether molecular pathways shared by O₂ sensing mediate the response of the pharynx to elevated CO₂ levels, we tested strains with mutations in the gcy-31, gcy-33, gcy-35, or gcy-36 genes. These genes encode soluble guanylyl cyclases (sGC) that bind O₂ and sense decreases (gcy-31 and gcy-33) or increases (gcy-35 and gcy-36) in O₂ levels [26]. Exposing the C. elegans strains gcy-31(ok296), gcy-33(ok232), gcy-35(ok769), or gcy-36(db42) to 10% CO₂ resulted in pumping inhibition similar to that observed in wild-type animals (Figure 2C). These results suggest that the effect of high CO₂ levels on the pumping rate does not involve O₂ sensing.

HID-1 is required for the CO₂-dependent pumping response

To identify genes that are involved in regulating the pharynx response to 10% CO₂, we performed a forward genetic screen after ethyl methanesulfonate (EMS) mutagenesis. Specifically, we screened for mutant animals that do not stop pumping in response to 10% CO₂ (Movie S2). We screened the progeny of ∼1200 F1 animals and found three strains that continued pumping when exposed to 10% CO₂. One of these strains was further crossed to the Hawaiian strain, and deep sequencing was performed on DNA from recombinant F2 progeny. The region containing the mutant gene that enabled continuous pumping in 10% CO₂ was identified by searching for a low number of Hawaiian single-nucleotide polymorphisms (SNPs), as described elsewhere [27]. This mutant strain has a premature stop codon in a previously characterized highly conserved gene, hid-1. In 5% CO₂, unlike in wild-type animals, the pumping rate of animals with the isolated hid-1(yg316) allele was similar to the pumping rate in normal air conditions, and significant pumping continued after exposure to 10% CO₂, whereas in 20% CO₂ pumping was abolished (Figure 3A).

The effect of HID-1 on the response to high levels of CO₂ was specific to the pharynx, since hid-1 mutant animals still showed reduced egg laying (Figure S2) and a slower rate of development (data not shown), similar to wild-type animals exposed to high CO₂ levels [24]. Two other alleles of hid-1, sa722 and sa1058 [28], also showed continuous pumping when exposed to 10% CO₂ (Figure 3B). The change in pumping rate in response to CO₂ is specific to HID-1, since transgenic expression of HID-1::GFP under its own promoter, in either hid-1(sa722) or hid-1(yg316) strains (Figure S3), was sufficient to restore the normal reduced pumping rate in 10% CO₂ (Figure 3B). Together, these data suggest that hid-1 is required for the response of the pharynx to high levels of CO₂.

Other dense core vesicle secretion and maturation mutants are also involved in the inhibition of pumping by CO₂

Dense core vesicles (DCVs) secrete neuropeptides in peptidergic neurons [29]. HID-1 is associated with Golgi membranes by way of N-terminal myristoylation and is required for the sorting of DCVs, where it prevents sorting of peptide cargoes to lysosomes for degradation [30]–[32]. We hypothesized that HID-1 plays a role in the response of the pharynx to high CO₂ by regulating neuropeptide secretion. We tested this hypothesis by scoring pumping response to 10% CO₂ in mutants defective in other genes involved in neuropeptide secretion. The gene unc-31 encodes the C. elegans ortholog of CAPS (calcium-dependent activator protein for secretion), an important component of DCV exocytosis [33]. The gene egl-21 encodes the C. elegans ortholog of carboxypeptidase E, an important component in neuropeptide maturation [34]. Following exposure of unc-31(e928) or egl-21(n476) deletion strains to 10% CO₂, the pumping rate of the pharynx was significantly higher compared with that in wild-type animals exposed to the same concentration of CO₂ (Figure 4A). We also tested the role of synaptic vesicle secretion on the pumping response to 10% CO₂. The unc-13 gene is involved in synaptic vesicle secretion of neurotransmitters [35], [36]. The rab-3 gene is a Rab GTPase that affects the distribution of synaptic vesicle populations [37]. Exposure of unc-13(e1091) or rab-3(js49) mutant strains to 10% CO₂ showed pumping behavior similar to that of wild-type strains (Figure 4A). These data suggest that DCVs play an important role in mediating the response of the pharynx to high CO₂ levels and that compromising DCV secretion probably impairs the pumping response to high CO₂ levels.

Expression of HID-1 in the BAG neurons is sufficient to restore wild-type CO₂ response in hid-1 mutant strains

HID-1 is expressed in all neuron and gut cells of C. elegans [30]. To test whether inhibition of pharynx pumping in response to 10% CO₂ requires expression of hid-1 in the gut, neurons, or both, we used transgenic lines that express HID-1 fused to GFP driven by either the pan-neuronal promoter rab-3 or the gut-specific promoter ges-1. Expression of HID-1 under the rab-3 promoter in neurons of hid-1(sa722) background was sufficient to restore inhibition of the pharynx pumping almost to the levels shown by wild-type animals (Figure 4B). In contrast, expression of HID-1 under the ges-1 promoter in the gut of hid-1(sa722) had no significant effect on the response of the pharynx to 10% CO₂.

We next asked which subset of neurons is required for mediating the effect of high CO₂ levels on the pharynx. The nlp-3 gene is expressed in sensory neurons (ADF, ASE, ASH, AWB, ASJ, and BAG) as well as in pharyngeal neurons (I1, I2, I3, I4, M1, M3, and NSMR) (Figure S3) [38]. Transgenic expression of HID-1::GFP under the nlp-3 promoter in hid-1(sa722) background was sufficient to restore pharynx pumping inhibition after exposure to 10% CO₂ (Figure 4B). High levels of CO₂ activate the AFD neurons [23]. Surprisingly, transgenic expression of HID-1::GFP driven by a gcy-8 promoter in the thermosensory AFD of hid-1(sa722) background did not restore the CO₂-mediated pumping inhibition (Figure 4B), which suggests that the activation of the AFD neurons by high CO₂ levels is not sufficient to induce the peptidergic signaling that mediates the effect of high CO₂ levels on the pharynx. Among the sensory neurons expressing nlp-3 are the BAG and the ASE neurons, which are also known to respond to high CO₂ levels [23]. Transgenic expression of HID-1::GFP driven by the osm-6 promoter, which was expressed in ASE neurons (undetected in BAG neurons), did not restore the CO₂-mediated pumping inhibition (Figure 4B). We next tested the role of BAG neurons in the CO₂-dependent pumping inhibition of the pharynx. Transgenic lines expressing HID-1::GFP under the promoter of flp-17 showed expression in the BAG neurons (Figure S3). This expression was sufficient to fully restore the CO₂-dependent pumping inhibition (Figure 4B). Similarly, the expression of HID-1::GFP under the gcy-33 promoter in hid-1(sa722) background was specific to the BAG neurons (Figure S3) [21]. This expression was sufficient to fully restore the CO₂-dependent pumping inhibition (Figure 4B). Next, we ablated the BAG neurons in transgenic worms expressing HID-1::GFP under flp-17 and gcy-33 promoters in hid-1(sa722) background. We found that following the removal of the HID-1::GFP-expressing BAG neurons, the pumping in 10% CO₂ was similar to that of hid-1(sa722) animals (Figure 4C). These results suggest that the specific expression of HID-1 in the BAG neurons is sufficient to induce the CO₂-dependent pumping inhibition. We also ablated the BAG neurons in wild-type background using GFP driven by gcy-33 promoter as a marker. We found that following ablation of the BAG neurons, the pumping in response to 10% CO₂ was similar to that in HID-1-null animals (Figure 4C). These results suggest that the BAG neurons are required for the pumping inhibition. In addition, to test possible cross talk between the neuropeptide secretion pathway and the guanylyl cyclase receptor pathway, which is required for CO₂ avoidance, we measured the pharyngeal pumping rate of animals carrying both hid-1(sa722) and gcy-9(tm2816) mutations. The pumping rate was similar to that of hid-1(sa722) animals (Figure 4D). Moreover, in the same genetic background, transgenic expression of HID-1 in the BAG neurons, using the flp-17 promoter, restored the suppression of pumping in the presence of high CO₂ level (Figure 4D), further demonstrating that the response to CO₂ mediated by hid-1 is independent of the response to CO₂ mediated by gcy-9. We conclude that proper hid-1 activity in the BAG neurons is important to mediate the pumping inhibition by CO₂.

Discussion

In humans, high CO₂ levels have diverse effects on the lung epithelium, immunity, and muscle function. However, the effects of acute exposure of muscle cells to high CO₂ levels were unknown. In addition, recent studies suggest that mammals, like C. elegans, are able to sense elevated CO₂ levels, which is of broad physiologic significance.

CO₂ avoidance and CO₂-dependent reduced pharyngeal pumping are probably regulated via different pathways

Acute exposure of well-fed adult C. elegans animals to high CO₂ levels quickly reduces the pumping rate of the pharynx. This effect depends in part on the nutritional status of the animal, since starved animals exposed to 10% CO₂ in air continue to pump, albeit at a significantly slower rate. Our genetic data suggest that the effect of acute exposure to high CO₂ levels on the pumping rate is independent of the avoidance responses of C. elegans to high CO₂ levels. First, cGMP signaling is required for mediating the avoidance response, as mutations in the cGMP gated ion channel encoded by tax-2 and tax-4 completely disrupt the avoidance response [2], [4]. In contrast, the same mutation in tax-4 does not completely rescue the immediate response of the pumping rate to high CO₂ levels. Second, mutation in the insulin-like receptor encoded by daf-2 also disrupts the avoidance response [2], [4]. The pumping rate of daf-2 mutants under exposure to 10% CO₂ is dramatically reduced, like in the wild-type animals. The limited recovery of the pumping rate in daf-2 mutants at 10% CO₂ could be due to the effect of daf-2 on starvation regulating pathways [39], [40]. Third, interference with proper function of ciliated sensory neurons by mutations in osm-3 and che-10 also significantly changes the avoidance response. Again, in the pumping assay, similar mutations in these genes did not change the response of C. elegans to high CO₂ levels. In addition, mutations in ets-5 and gcy-9, which were previously shown to be required for the calcium response of the BAG neurons to CO₂, did not change the response of the pharynx to high CO₂ levels. Finally, the rescue of pumping by hid-1 in the BAG neurons was not affected by gcy-9 mutations.

C. elegans animals presumably interpret high CO₂ levels as a harmful cue that leads to avoidance and pumping inhibition. The ability of the animal to stop eating for several minutes probably allows it to avoid undesirable food. Surprisingly, although both the avoidance and the pumping responses to the same stressful cue are immediate, our genetic data suggest that different molecular pathways mediate the two responses to high CO₂.

The potential role of neuropeptides in the response of the pharynx to high levels of CO₂

Our genetic screen identified hid-1 as a regulator of the pumping response to high CO₂ levels, as mutations in the hid-1 gene blunted the response of the pharynx to high CO₂ levels. HID-1 is required for the neuropeptide secretion pathway [30], [32]. Indeed, mutations in other known genes in peptidergic signaling, unc-31 and egl-21, could also partially suppress the pharyngeal pumping suppression upon exposure to 10% CO₂. Neuropeptides are important signaling molecules in many physiological responses both in C. elegans and in other organisms. In C. elegans there are more than 250 neuropeptides that play a role in feeding and metabolism, and most neurons in C. elegans secrete neuropeptides [41]. Neuropeptides are also secreted from the intestine [38], and hid-1, an important peptidergic signaling gene, is expressed both in the nervous system and in the intestine [30], [32]. Neuropeptide signaling was previously shown to regulate pumping inhibition in the absence of food [42]. Specifically, unc-31 mutants demonstrate continuous pumping in the absence of food, unlike wild-type animals [42]. Since hid-1 also partially suppresses the inhibition of pumping in the absence of food (data not shown), we cannot completely rule out the possibility that hid-1 generally inhibits pumping and acts in parallel with CO₂.

The pharynx response to 10% CO₂ is probably mediated by several different neuropeptides, since pumping inhibition could not be inhibited by deletion of individual neuropeptide genes known to be overexpressed in the BAG neurons, including flp-10, flp-16, flp-27, nlp-1, flp-17, and nlp-14 (Figure S4). Neuropeptide secretion can only partially explain the response of the pharynx to high levels of CO₂, since none of the peptidergic signaling mutants we examined at 10% CO₂ (hid-1, unc-31, and egl-21) exhibited the pumping rate seen at normal air levels (Figure 4). Thus we cannot completely rule out the possibility that the effects of unc-31 and egl-21 are due to the other pathway(s) that must be acting in parallel with hid-1. This implies the existence of other, HID-1-independent mechanisms that must regulate the response of the pharynx to CO₂ levels. For example, it is possible that high CO₂ levels trigger other presynaptic inputs that mediate the effect on the pharynx in parallel with the peptidergic signaling, or that CO₂ has also a direct postsynaptic effect on the pharyngeal muscles that inhibits their normal function. Interestingly, such parallel pathways depend on the CO₂ levels, since hid-1 completely rescues the pumping inhibition at 5% CO₂ and fails to rescue the pumping at 20% CO₂ (Figure 3A).

The presence of HID-1 is specifically required in the BAG neurons

Using transgenic lines that express HID-1 either in the gut or in the nervous system we have determined that the hid-1 activity is specifically required in neurons to mediate the effect of high CO₂ levels on the pharynx. We used an AFD-specific promoter to show that hid-1 activity in the AFD neurons, which are activated by high CO₂ levels [23], is not sufficient to mediate the effect of high CO₂ levels on the pharynx. In contrast, transgenic expression of HID-1::GFP under the nlp-3 promoter is sufficient to restore CO₂-mediated pharynx pumping inhibition. Using the BAG-specific promoters flp-17 and gcy-33 and performing ablation experiments on the BAG neurons, we have further narrowed the CO₂ effect on the pharynx to the BAG neurons. Paradoxically, our genetic data (Figures 2 and 4) suggest the existence of different molecular pathways for the avoidance and the pharynx responses. However, the BAG neurons in the pharynx response are the same neurons that control the avoidance response. Interestingly, mutant animals that block CO₂-mediated calcium response in the BAG neurons still show normal pumping inhibition. The existence of such a pathway is especially surprising given that DCV secretion is expected to depend on an increase in calcium levels.

The physiological and molecular effects of high CO₂ levels, in both vertebrates and invertebrates, have been the focus of several recent studies [2], [4], [6], [7], [12], [14], [22], [24], [43]. However, the sensing mechanism of cells to high CO₂ levels is yet largely unknown. Soluble adenylyl cyclases are bicarbonate sensors in several organisms including mammals [15], [44], [45]. In C. elegans, which do not have soluble adenylyl cyclases, the soluble guanylyl cyclases GCY-31 and GCY-33 are important for eliciting CO₂ avoidance in the BAG neurons [23]. However, it is yet unknown whether the gcy genes are directly activated by either CO₂ or HCO₃⁻_. Our results show that neither GCY-31 nor GCY-33 are required for mediating the effect of high CO₂ levels on the pharynx (Figure 2C).

Our study sheds new light on the response of C. elegans to high CO₂ levels. It also shows that the CO₂-induced response is differentially regulated across different tissues. Furthermore, different levels of CO₂ lead to various outcomes in the same tissue. Deciphering the mechanisms underlying these fundamental pathways will hopefully help us to better understand the CO₂-induced responses that are activated in human diseases.

Materials and Methods

Strains

Worms were handled as described elsewhere [46]. The following strains were used in this study: N2 (wild type); CF1041, daf-2(e1370); CB3329, che-10(e1809); CX2948, tax-4(p678); PR802, osm-3(n1540); DA609, npr-1(ad609); CZ3714, gcy-31(ok296); CZ3715, gcy-33(ok232); CX6448, gcy-35(ok769); AX1296, gcy-36(db42); JT722, hid-1(sa722); JT1058, hid-1(sa1058); YG316, hid-1(yg316); YG2310, hid-1(yg316); jsEx896 [hid-1p::HID-1::GFP]; NM3017, hid-1(sa722) and lin-15(n765); jsEx896[hid-1p::HID-1::GFP]; NM3053, hid-1(sa722) and lin-15(n765); jsEx897[rab-3p::HID-1::GFP]; NM3139, hid-1(sa722) and lin-15(n765); jsEx909[ges-1p::HID-1::GFP]; YG2313, hid-1(sa722); ygEx317 [gcy-8p::HID-1::GFP]; YG2318, hid-1(sa722); ygEx318 [nlp-3p::HID-1::GFP]; YG2319, hid-1(sa722); ygEx319 [flp-17-p::HID-1::GFP]; YG2340, hid-1(sa722); ygEx320 [gcy-33-p::HID-1::GFP]; DA509, unc-31(e928); KP2018, egl-21(n476); YG2302, unc-13(e1091); YG2320, ets-5(tm1734); YG2321, ets-5 (tm1755); YG2322, gcy-9(tm2816); YG2323, gcy-9(tm2816) and hid-1(sa722); YG2324, gcy-9(tm2816) and hid-1(sa722); ygEx321 [flp-17-p::HID-1::GFP]; RB1340, nlp-1(ok1469); RB2575, flp-19(ok3587); VC2012, flp-27(gk3331); VC1108, nlp-14(ok1517)/szT1 X; RB1989, flp-10(ok2624); RB2275, flp-16(ok3085). All strains were obtained from the C. elegans Genome Center (CGC) or the National BioResourse Project (NBRP), except for CX2948, which was kindly provided by the De-Bono laboratory, and NM3017, NM3053, and NM3139, which were kindly provided by the Nonet laboratory [2], [30].

Measurement of pumping rate

A standard NGM plate covered with a lid-shaped chamber with inlet and outlet holes to allow gas flow was used to measure the pumping rate under different concentrations of CO₂ in air. The chamber was connected to a mechanical valve that controlled the humidified gas mixture entering the chamber. For all pumping assays, NGM plates were seeded with 20 µL of OP50 5 h before the experiment to allow normal feeding and to keep worms in a restricted area. A single 1-day-old adult worm was seeded on a plate just before the start of the experiment. Initially, normal air mixture (21% O₂, 79% N₂) flowed into the chamber and worms were allowed to adjust for 1 min. The number of pharynx muscle contractions was subsequently measured for 1 min under normal air conditions. Then the airflow was switched to a high-CO₂ gas mixture, and after 10 s the pharynx muscle contraction rate was measured again. To measure pumping rate after starvation, well-fed wild-type 1-day-old adult worms were collected using M9 buffer and washed four or five times in M9 buffer. Worms were then seeded on either NGM plates with no bacteria or NGM plates seeded with OP50, for 4 h prior to measurements. All pumping assays were performed at 22°C.

EMS screen and SNP mapping

The EMS mutagenesis was performed essentially as described elsewhere [46]. Briefly, wild-type (N2) worms in the L4 stage were exposed to 50 mM EMS in M9 buffer for 4 h and then transferred to fresh plates for 2–3 h (P₀) for recovery. After recovery, five P₀ animals were transferred again to an NGM plate and allowed to lay F1 progeny. Adult F1 animals were cloned onto individual NGM plates and their L4-adult F2 progeny where exposed to 10% CO₂. F2 worms that continued the pumping of the pharynx even after exposure to 10% CO₂ were isolated. In total, we scored the progeny of ∼1200 F1 animals. The isolated strains were outcrossed three times. The mutation was mapped as described elsewhere [27]. Mutant worms were crossed with the Hawaiian strain and F1 progeny were isolated. Then 44 F2 recombinants that continued the pumping after exposure to 10% CO₂ were isolated, and the DNA of their F3 and F4 offspring was extracted using a Gentra Puregene kit (Qiagen, cat. no. 158667). Whole genome sequencing was performed using the Applied Biosystems SOLiD 3 deep sequencing apparatus. The positions of the Hawaiian SNPs were mapped on the DNA of the yg316 strain. A 1.2-MB region in chromosome X that did not contain any Hawaiian SNP was found. Within this region a premature stop codon (W625X) in the coding sequence of hid-1 was found to cause the phenotype as described in the text.

Plasmid constructs, transgenes, and laser ablation

The NM1699 construct, which contains the hid-1 promoter driving the genomic hid-1 coding region fused to eGFP, was a kind gift from the Nonet laboratory [30]. The NM1699 construct was digested with KpnI and AatII to replace the native hid-1 promoter with various neuron-specific promoters. To drive AFD-specific expression an 800-bp fragment upstream of the gcy-8 start codon was amplified and subsequently digested with KpnI-AatII to generate pKS10. Similarly, to drive sensory and pharyngeal specific expression, a 700-bp fragment upstream of the nlp-3 start codon was amplified and digested with KpnI-AatII to generate pKS20. To drive BAG-specific expression a 3.4-kb fragment upstream of the flp-17 start codon was amplified and fused by PCR to HID-1::GFP from NM1699. In addition, to drive BAG-specific expression a 980-bp fragment upstream of the gcy-33 start codon was amplified and subsequently digested with KpnI-AatII to generate pKS30. All plasmids were verified by sequencing and microinjected to either JT722 or YG316 with an elt-2::GFP marker as described elsewhere [47].

Laser ablation was performed using an Andor Revolution XD confocal spinning disk system with a Nikon TiE inverted microscope equipped with a nitrogen pulsed laser and a 365-nm Micropoint dye cell. The microscope and laser were controlled by means of IQ software and the Micropoint Mosaic I System 85-75, respectively. The region of interest was set according to the size of the neuron cell body as revealed by the GFP marker. We used a frequency of 15 Hz, an energy range of 80%–90%, and 3–5 repeats in order to completely ablate the GFP marker in the neuron cell body.

Supporting Information

Attachment 1

Attachment 2

Attachment 3

Attachment 4

Attachment 5

Attachment 6