If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Skirball Institute of Biomolecular Medicine, Molecular Neurobiology Program and Department of Cell Biology, New York University Medical Center, New York, New York 10016Department of Pharmacology, University of Cambridge, Cambridge, CB2 1PD, United Kingdom
Pew Scholar in the Biomedical Sciences. To whom correspondence may be addressed: Skirball Inst. of Biomolecular Medicine, Molecular Neurobiology Program and Dept. of Cell Biology, NYUMC, New York, NY 10016
Animals from diverse phyla possess neurons that are activated by the product of aerobic respiration, CO2. It has long been thought that such neurons primarily detect the CO2 metabolites protons and bicarbonate. We have determined the chemical tuning of isolated CO2 chemosensory BAG neurons of the nematode Caenorhabditis elegans. We show that BAG neurons are principally tuned to detect molecular CO2, although they can be activated by acid stimuli. One component of the BAG transduction pathway, the receptor-type guanylate cyclase GCY-9, suffices to confer cellular sensitivity to both molecular CO2 and acid, indicating that it is a bifunctional chemoreceptor. We speculate that in other animals, receptors similarly capable of detecting molecular CO2 might mediate effects of CO2 on neural circuits and behavior.
Carbon dioxide is detected by animals as an environmental cue that indicates the presence of prey, hosts, or mates and is also detected as an internal cue that reflects the metabolic state of the organism (
). Understanding the molecular mechanisms of neuronal CO2 chemosensitivity is therefore a central question in the study of many physiological and behavioral processes.
In solution, where it would be sensed by neurons, CO2 generates multiple chemical species. CO2 reacts with water to form carbonic acid, which almost instantly dissociates to produce protons and bicarbonate ions. The CO2 hydration reaction can occur rapidly in biological systems because of the catalytic action of carbonic anhydrase (CAH)
). CO2-sensing neurons, therefore, encounter CO2 in equilibrium with the major products of its hydration: protons and bicarbonate ions. In many cases, neurons that respond to increases in CO2 levels have been found to respond to one or the other of these CO2 metabolites. For example, CO2-sensitive areas of vertebrate respiratory centers and the amygdala are highly sensitive to changes in extracellular pH (
). The sensitivity of many different types of neurons to CO2 metabolites is the basis for the hypothesis that the effects of CO2 on neural physiology are mediated by CO2 metabolites rather than by CO2 itself.
The sensory nervous system of the nematode Caenorhabditis elegans offers an excellent opportunity for the study of molecular mechanisms used by neurons to detect CO2. C. elegans possesses a pair of CO2-sensing neurons, the BAG neurons, which mediate acute avoidance of CO2 by adults and attraction to CO2 by Dauer larvae (
). To sense CO2, BAG neurons require a cGMP signaling pathway. The BAG cell-specific receptor-type guanylate cyclase GCY-9 and heteromeric TAX-2/TAX-4 cyclic nucleotide-gated ion channels are required both for cellular responses to CO2 and for BAG cell-dependent behaviors (
). The GCY-9 cyclase and CNG channels likely constitute the core of the molecular machinery that endows BAG neurons with CO2 chemosensitivity; expression of GCY-9 in sensory neurons that use cGMP signaling is sufficient to mediate calcium responses to CO2 stimuli (
Although components of a transduction pathway that mediates CO2 sensing by BAG neurons have been identified, it was not known whether BAG neurons are principally tuned to detect CO2 or CO2 metabolites. Unlike many chemosensory neurons of C. elegans, BAG neurons are completely contained within the animal and not in contact with the external environment (
). It is therefore not possible to determine whether BAG neurons sense CO2 or CO2 metabolites by simply exposing intact animals to these stimuli. In such an experiment, the intrinsic tuning properties of the neuron would be convolved with the relative permeability of the cuticle and hypoderm to different chemical cues. To identify the specific chemical cue that activates BAG neurons, we studied isolated BAG neurons in culture, using methods that allow both monitoring of cell physiology and control of the extracellular and intracellular environments.
CO2-responsive BAG neurons are located in the head and extend ciliated processes toward the nose (Fig. 1A). These processes are not in contact with the external environment (
). We observed that the receptor-type guanylate cyclase GCY-9, which is an essential component of the sensory transduction apparatus in BAG neurons and is likely to function as part of a receptor complex (
), is enriched in the terminus of the BAG cell neurite (Fig. 1, B and C). The CO2 transduction apparatus of BAG neurons is therefore in a cellular compartment that is not in direct contact with the external environment and is experimentally inaccessible. To determine the chemical tuning of BAG neurons, we therefore sought to isolate BAG neurons and study their sensitivity to CO2 and other stimuli in vitro.
We cultured cells from C. elegans embryos expressing the ratiometric calcium indicator YC3.60 in BAG neurons (Fig. 1, A, adults, and D, late stage embryos) and could readily find YC3.60-expressing neurons in culture (Fig. 1E). Approximately 1% of adherent cells were BAG neurons, and we observed a similar frequency for other sensory neurons when cultures were prepared from different transgenic strains (Fig. 1F). Although BAG neurons in situ are bipolar (Fig. 1A), not all BAG neurons extended two processes in vitro; they did, however, extend neurites more frequently than did unidentified neurons (Fig. 1G).
Isolated BAG neurons responded to a brief depolarizing stimulus (100 mm KCl) as shown by BAG cell calcium responses (Fig. 2A). We next stimulated BAG neurons with 33 mm NaHCO3 in equilibrium with 10% atmospheric CO2, pH 7.2, to determine whether they functioned in vitro as sensors for CO2 or its metabolites. A majority of cells (69%) responded to this stimulus and displayed large calcium responses (Fig. 2B). The chemosensitivity of cultured BAG neurons was similar to the sensitivity of BAG neurons in situ, which we previously measured: the EC50 of CO2 for activation of isolated BAG neurons was 0.4% CO2 (Fig. 2C), compared with 0.9% CO2 for activation of BAG neurons in situ (
). In vitro, however, AFD neurons did not respond to CO2, although they did respond to depolarization (Fig. 2D). Similarly, neither oxygen-sensing URX neurons (Fig. 2E) nor unidentified neurons displayed robust CO2 responses in vitro (Fig. 2F). Intrinsic CO2 chemosensitivity is therefore specific to BAG neurons. We stimulated BAG neurons with a panel of compounds that either are chemically similar to CO2, activate CO2-responsive neurons of other organisms, or are gases with known roles in physiological signaling. None of these compounds activated BAG neurons, suggesting that BAG neurons are narrowly tuned to detect CO2 or its metabolites (Fig. 2G).
), was not required for CO2 responses of isolated BAG neurons (Fig. 3, D–F). We observed that BAG neuron calcium responses were reversibly blocked by the nonselective voltage-gated calcium channel (CaV) blocker CdCl2 (Fig. 3, G and H), and a similar effect was observed with nemadipine-A, which targets the sole L-type CaV expressed by C. elegans, EGL-19 (
) (Fig. 3, I and J). These data demonstrate a critical role for L-type CaVs in CO2 sensing by BAG neurons. Disrupting intracellular calcium stores by blocking the SERCA calcium pump with thapsigargin did not significantly affect BAG neuron function (Fig. 3, K and L), suggesting that release of calcium from intracellular stores plays little or no role in activation of BAG neurons by CO2.
We next sought to assay activation of BAG neurons using a method that was independent of calcium imaging. BAG neurons express a number of neuropeptides, including FLP-17, which modulates activity of the egg laying system (
). We tested whether a FLP-17::Venus fusion would allow us to measure evoked exocytosis of neuropeptides from BAG neurons in response to CO2 stimuli. In adults, FLP-17::Venus was in the cell soma and puncta in the posterior neurite (Fig. 4A). In cultured BAG neurons FLP-17::Venus was similarly distributed (Fig. 4B). Both depolarization and CO2 stimuli evoked a stepwise decrease in FLP-17::Venus fluorescence in the neurites of cultured BAG neurons (Fig. 4C). We noted that the average destaining response of BAG neurons to CO2 stimuli was less than the destaining caused by depolarization (Fig. 4C) and found that this was because some BAG neurons that released peptide in response to KCl failed to release peptide in response to CO2. When we analyzed the destaining of neurons that responded to both KCl and CO2 (55% of the neurons tested), we found that the destaining responses to these two stimuli were of comparable magnitude (Fig. 4D). CO2-evoked peptide release required the CO2 transduction pathway: gcy-9 mutant BAG neurons displayed KCl-induced destaining, but did not respond to CO2 (Fig. 4, E and F).
By multiple criteria, therefore, isolated BAG neurons in vitro retain their function as CO2 chemosensors. We therefore proceeded to use BAG neurons in culture to determine whether they are principally tuned to CO2 itself or the products of CO2 hydration: protons and bicarbonate ions. We first tested whether BAG neurons are proton sensors by exposing them to sequential CO2 (10%) and pH 6.5 acid stimuli. A majority of cells responded robustly to CO2, but not to acid (24 of 36). However, some BAG neurons (12 of 36) responded to both stimuli (Fig. 5, A and B). The magnitude of the calcium responses increased with increasing proton concentrations over a range of pH 7 to pH 6 with half-maximal responses observed at pH 6.7 (Fig. 5C). Like CO2 sensitivity, BAG neuron acid sensitivity required the receptor-type guanylate cyclase GCY-9 (Fig. 5D). Unidentified neurons in culture failed to respond to acid (data not shown), indicating that acid sensitivity is not widespread among C. elegans neurons. Thermosensory AFD neurons were also insensitive to acid (Fig. 5E), but expression of GCY-9 conferred acid sensitivity to a fraction of AFD neurons (Fig. 5, F and G). Together, these data indicate that acid sensing and CO2 sensing by BAG neurons are mediated by a common transduction pathway.
Although BAG neurons can be activated by acid, it remained unclear whether the acid sensitivity of BAG neurons is the mechanism by which they detect CO2. If BAG neurons detect CO2 via acid, how is it that CO2 better activates these cells? One possibility is that BAG neurons might respond to changes in intracellular pH. Although CO2 might readily permeate cell membranes to generate protons intracellularly, protons generally require transport mechanisms to cross cell membranes. This model demands that CO2 stimuli change intracellular pH. We measured the intracellular pH of BAG neurons during presentation of acid and CO2 stimuli and found that although the intracellular pH of BAG neurons rapidly responded to extracellular acid, 10% CO2 did not affect intracellular pH (Fig. 5, H and I). The activation of BAG neurons by CO2, therefore, can occur independently of either extracellular or intracellular acidosis.
If BAG neurons sense CO2 in an acid-independent manner, alkaline solutions containing CO2 should also activate BAG neurons. We tested CO2 solutions at different pH to determine whether this was the case. A fixed amount of bicarbonate in solution at pH 7.2 generates 4.5-fold more CO2 that it does at pH 7.9 (Fig. 6A, inset). A pH 7.2 bicarbonate solution predicted to contain the EC50 for CO2 (90 μm) evoked large calcium responses from BAG neurons in culture, whereas the same solution at pH 7.9 (20 μm CO2) did not (Fig. 6A). Increasing the bicarbonate concentration at pH 7.9 so as to generate 90 μm CO2, we observed calcium responses indistinguishable from those evoked by a 90 μm CO2 stimulus at pH 7.2 (Fig. 6A). Importantly, weak alkalinization (pH 7.2 versus 7.9) did not inhibit BAG neuron sensitivity to KCl-evoked depolarization; by contrast, depolarization-induced calcium responses were increased by mild alkaline conditions (Fig. 6, B and C). Thus BAG neuron responses to CO2 solutions are proportionate to the concentration of molecular CO2 and are unaffected by increased pH. Furthermore, these experiments indicate that BAG neurons are not activated by bicarbonate, which has been proposed to mediate some cellular responses to CO2 (
). To further confirm that BAG neurons sense molecular CO2, we determined whether CAH is required for CO2 sensing. Some CO2-responsive neurons require CAH, which catalyzes the hydration of CO2, suggesting that these neurons are tuned to detect either protons or bicarbonate, not CO2 itself (
). Inhibiting CAH using two chemically distinct inhibitors of carbonic anhydrase, methazolamide (Fig. 6, D and E) and punicaligin (Fig. 6, F and G), did not affect BAG neuron CO2 responses, further suggesting that BAG neurons directly detect molecular CO2.
Previously, we showed that CNG channels and the receptor-type guanylate cyclase GCY-9 constitute the core of the CO2 sensory transduction apparatus in BAG neurons (
). Expression of GCY-9 is instructive for CO2 sensitivity and confers upon AFD neurons CO2 sensitivity that is independent of their ability to detect temperature stimuli (Fig. 7, A–D). To determine whether GCY-9 itself mediates detection of molecular CO2, we stimulated transgenic AFD neurons that express GCY-9 with bicarbonate solutions at neutral or alkaline pH. GCY-9-expressing AFD neurons, like BAG neurons, displayed calcium responses proportional to the concentration of CO2 (Fig. 7E). These data support a model in which GCY-9 functions as a receptor for molecular CO2 and show that GCY-9 detects CO2 independently of acid (Fig. 7F).
The remarkable tuning of C. elegans BAG neurons to molecular CO2 distinguishes them from previously characterized CO2 chemoreceptor neurons. CO2-responsive neurons have long been thought to principally detect either protons or bicarbonate produced by the hydration of CO2 (
). For example, in the vertebrate central nervous system, CO2-responsive neurons of the amygdala and respiratory centers display acid sensitivity, which in the case of amygdala neurons has been attributed to expression of the acid-sensing ion channel ASIC1a (
). Vertebrate gustatory neurons that are activated by CO2 are also thought to do so via protons generated by CO2 hydration. These neurons constitute a subset of acid-sensitive taste neurons and require an extracellular CAH for CO2 sensing (
). Likewise, a subset of mammalian olfactory sensory neurons also responds to CO2 in a CAH-dependent manner. These neurons, unlike CO2-responsive gustatory neurons, are proposed to principally detect bicarbonate (
). That vertebrate GC-D neurons and C. elegans BAG neurons both use receptor-type guanylate cyclases that likely function as chemoreceptors might reflect divergence of an ancient mechanism for CO2 sensing based on cyclic nucleotide signaling.
There are instances of CO2 chemosensitivity that suggest that other chemotransduction systems might, like the GCY-9 system, detect molecular CO2. CO2-responsive neurons have been found in sensilla of the insect gustatory and olfactory systems. These sensilla are activated by CO2, not acid (
), consistent with the hypothesis that their associated chemoreceptor neurons are tuned to detect molecular CO2. However, the neurons within these sensilla have only been studied in situ where it is not possible to control their extracellular environment. It is possible that the insensitivity of the chemosensory neurons in these sensilla to external acid results from a permeability barrier that prevents acid or bicarbonate from diffusing into the sensilla. If these neurons are tuned to detect molecular CO2, they will likely use receptor signaling mechanisms different from those used by C. elegans BAG neurons. Insect chemoreceptors are members of an insect-specific family of multipass transmembrane proteins. It is likely that the insect gustatory receptor that mediates CO2 sensitivity is a member of this family, and the receptor molecule that mediates olfactory detection of CO2 has already been identified as such (
), which is expressed by astrocytes in areas of the medulla that respond to hypercapnia. ATP released at the ventral surface of the medulla in response to hypercapnia, as well as the changes in ventilation rate, are both diminished by inhibiting connexins (
). The effects of Cx26 mutation on the respiratory motor program remain to be determined, however, and in vivo genetic manipulations are needed to confirm a role for Cx26 as a CO2 sensor in the rodent nervous system.
Although BAG neurons are robustly activated by molecular CO2, we found that they can also be activated by acid. The receptor-type guanylate cyclase GCY-9 mediates sensitivity to both acid and molecular CO2, and our data suggest that GCY-9 is activated independently by either stimulus. Sensitivity of a receptor to multiple stimuli, including protons, has been described in other contexts, such as the cation channel TRPV1 in mammalian sensory neurons (
). Indeed, the acid sensitivity of a receptor signaling system that mediates CO2 sensing might match cellular responses to CO2 to the internal state of an organism. Acidosis is a hallmark of metabolic stress, during which pH homeostasis might be especially sensitive to perturbations caused by increased environmental CO2. The intrinsic acid sensitivity of BAG neurons might, for example, mediate enhanced CO2 avoidance behavior during periods of such stress. Another mechanism by which CO2 sensing might be regulated by the metabolic state of the animal is a functional connection between neurons that monitor internal oxygen concentration and the BAG neurons. CO2 avoidance behavior is modulated by hypoxia (
Why are there so many distinct mechanisms for detecting CO2, not only with respect to cellular receptors, but also in terms of the chemical cue detected? Neurons tuned to detect molecular CO2 might play fundamentally different roles in animal physiology from cells that respond to CO2 metabolites. The principal products of CO2 metabolism, protons and bicarbonate, are substrates for a large number of redundant cellular systems that buffer and transport these ions. Only when the capacity of these buffers and transporters is exceeded will cells and tissues experience changes in pH or bicarbonate concentration. Neurons that detect protons and bicarbonate might therefore be considered tuned to the failure of pH and bicarbonate homeostasis and might mediate responses to these physiological stresses. By contrast, neurons that directly detect molecular CO2 would be able to detect changes in environmental or internal CO2 levels that would fail to significantly alter pH or bicarbonate levels. Mechanisms that detect molecular CO2 might therefore permit neural circuits that control host-finding behaviors and the respiratory motor program to detect low concentrations of CO2 that would otherwise be heavily buffered. In the context of respiratory control, such a mechanism would allow the respiratory motor program to respond to increased CO2 levels in the absence of acidosis. Indeed, some CO2-sensititive neurons implicated in respiratory control have been shown to respond to pH-neutral CO2 stimuli (
), suggesting that such a mechanism exists in vertebrates. It is an intriguing possibility that these neurons express a receptor for molecular CO2 that is analogous or homologous to GCY-9 and functions in central circuits to control respiratory rhythms.
We thank Sonya Aziz-Zaman and Rouzbeh Mashayekhi for creating some plasmids used in this study and Mitchell Chesler and the Ringstad laboratory for many helpful comments and discussions.
Out of thin air. Sensory detection of oxygen and carbon dioxide.
Bianchi, L., Driscoll, M., (2006) Culture of embryonic C. elegans cells for electrophysiological and pharmacological analyses. Wormbook (The C. elegans Research Community, ed) 10.1895/wormbook.1.122.1, www.wormbook.org