The Role of ATP and Purinergic Receptors in Taste Signaling

Abstract

This review summarizes our understanding of ATP signaling in taste and describes new directions for research. ATP meets all requisite criteria to be considered a neurotransmitter: (1) presence in taste cells, as in all cells; (2) release upon appropriate taste stimulation; (3) binding to cognate purinergic receptors P2X2 and P2X3 on gustatory afferent neurons, and (4) after release, enzymatic degradation to adenosine and other nucleotides by the ectonucleotidase, NTPDase2, expressed on the Type I, glial-like cells in the taste bud. Importantly, double knockout of P2X2 and P2X3 or pharmacological inhibition of P2X3 abolishes transmission of all taste qualities. In Type II taste cells (those that respond to sweet, bitter, or umami stimuli), ATP is released non-vesicularly by a large conductance ion channel composed of CALHM1 and CALHM3, which form a so-called channel synapse at areas of contact with afferent taste nerve fibers. Although ATP release has been detected only from Type II cells, it is also required for the transmission of salty and sour stimuli, which are mediated primarily by the Type III taste cells. The source of the ATP required for Type III cell signaling to afferent fibers is still unclear and is a focus for future experiments. The ionotropic purinergic receptor, P2X3, is widely expressed on many sensory afferents and has been a therapeutic target for treating chronic cough and pain. However, its requirement for taste signaling has complicated efforts at treatment since patients given P2X3 antagonists report substantial disturbances of taste and become non-compliant.

Taste buds, the sensory endorgans for the sense of taste are distributed widely throughout the oropharynx. Most taste buds reside in specific lingual papilla: fungiform papillae on the front of the tongue and foliate and vallate papillae on the sides and back of the tongue. In addition, taste buds occur on the arytenoid processes of the larynx. Lingual taste buds play an important role in appreciation of food quality while those on the larynx are more involved in reflex functions to protect the airways from erroneous entry of liquids and solids into the airways.

Regardless of location, taste buds comprise an assemblage of 35–100 specialized epithelial cells embedded in the oropharyngeal mucosa. About half of the cells in a taste bud, so-called Type I cells, serve a supportive function, much like astrocytes in the central nervous system (Kinnamon and Finger 2019). In this capacity, the Type I taste cells remove neurotransmitters (Bartel et al. 2006) and participate in ionic buffering (Dvoryanchikov et al. 2009) as well as separating the receptive elements of the taste buds one from another. The transducing cells of taste buds fall into two morphological classes: Type II and Type III.

These transducing cells of taste buds are not neurons, but rather are short, axonless receptor cells similar in this respect to hair cells of the inner ear. Similar to hair cells, taste cells release neurotransmitters from their basal aspects to activate the peripheral endings of primary sensory neurons of the system whose cell bodies reside in cranial ganglia of the facial, glossopharyngeal, and vagus nerves. These gustatory ganglion cells extend a central process into the primary gustatory nucleus in the medulla, nuc. solitary tract in mammals, where they release glutamate to activate the second-order neurons (Smeraski et al. 1996; Li and Smith 1997). Glutamate is also the peripheral neurotransmitter utilized by hair cells to activate the sensory nerve fibers (Usami et al. 2001), but is not the key neurotransmitter used by the taste transducing cells in taste buds.

The Type III cell is a neuron-like cell which responds to ionic taste stimuli including sour (H⁺) and some salts. In the case of sour, the protons directly permeate Otop1 channels in the apical membrane to depolarize the cell ultimately activating voltage-gated sodium channels to generate an action potential (Kinnamon and Finger 2019; Teng et al. 2019). The transduction mechanism for salty is less clear, in part because there is more than one mechanism involved – one for low concentrations of salt, which are appetitive and amiloride-sensitive, and others for high concentrations of salt, which are aversive and amiloride-insensitive (Oka et al. 2013). Both Type II and Type III cells likely transduce amiloride-insensitive salt, likely via different mechanisms. For Type II cells, recent data suggest chloride may play a role, although the precise mechanism that results in taste cell depolarization has not been identified (Roebber et al. 2019). At least 2 different mechanisms appear to contribute to amiloride-insensitive salt taste in Type III cells, but neither mechanism has been molecularly defined (Lewandowski et al. 2016). Whatever mechanism ultimately generates the action potential in Type III cells, as in neurons, the resulting strong depolarization opens voltage-gated Ca²⁺ channels which permit entry of Ca²⁺ into the basal compartment of the cell where typical synapses with synaptic vesicles reside. The Ca²⁺ influx permits fusion of the vesicles to the plasma membrane to release neurotransmitter (Vandenbeuch et al. 2010). The full panoply of neurotransmitters released at this synapse has yet to be delineated fully, but certainly includes serotonin and norepinephrine (Huang et al. 2005a, b, 2008, 2009, 2011a). But, as detailed below, neither serotonin nor norepinephrine appears to be necessary for transmission of taste information from taste buds to the taste nerves, but they may be involved in signaling between Type II and Type III cells (Roper 2021).

The Type II cells utilize a very different transduction cascade to effect release of neurotransmitter. These cells rely on G-protein coupled receptors to respond to sweet, umami, or bitter depending on the particular molecular receptor proteins expressed. Regardless of taste quality detected, the downstream signaling components are nearly identical involving PLCβ2, IP₃, the IP₃R3 receptor, TRPM5, and various voltage-gated Na⁺ channels, ultimately culminating in an action potential (Kinnamon and Finger 2019). Unlike Type III cells, Type II cells lack voltage-gated Ca²⁺ channels and so do not require Ca²⁺ influx from extracellular space for synaptic transmission. Rather, the action potential directly gates a hexameric ATP release channel, described in more detail below, consisting of CALHM1 and CALHM3 subunits (Taruno et al. 2013, 2021; Ma et al. 2018).

In addition to being required for transmission of bitter, sweet, and umami stimuli, the CALHM1/3 channels also mediate appetitive salt taste, which is amiloride-sensitive (AS) and involves the epithelial channel ENaC. Although ENaC subunits are rather widely expressed in all types taste cells, including Type I cells, only those cells that also express CALHM1/3 signal salt taste to the nervous system via release of ATP (Nomura et al. 2020). The cell type responsible for this AS salt taste appears to be a unique type of Type II cell that lacks expression of TrpM5. Instead, influx of Na⁺ via ENaC appears to be sufficient to depolarize the cell, activate the voltage-gated Na⁺ channels, and trigger ATP release from the voltage-activated CALHM1/3 channels (Taruno et al. 2021).

1. Neurotransmitters in Taste Buds

The neurotransmitter(s) implicated in transmission of taste information from taste transducing cells to nerve fibers has only become clear in the last 15 years. The first candidate neurotransmitter, identified in 1975 on the basis of fluorescence histochemistry, was serotonin (Nada and Hirata 1975). In all vertebrates studied to date (Nada and Hirata 1977; Barreiro-Iglesias et al. 2008; Kirino et al. 2013), a subpopulation of taste cells accumulate and presumably release serotonin. Taste-dependent release of serotonin was described only 40 years after the seminal Nada & Hirata study (Huang et al. 2005a, b) but, as detailed below, is not crucial for activation of the gustatory nerve fibers, as will be described in the next section.

Other neurotransmitters and neuromodulators have been localized to taste cells, including acetylcholine, GABA, noradrenaline, and the peptide GLP-1 (Kusakabe et al. 1998; Dvoryanchikov et al. 2007; Roper 2007; Cao et al. 2009; Dando et al. 2012; Huang and Wu 2015, 2018), but their role in neurotransmission appears limited. These neuroactive substances may, however, play a substantial role in modulation of taste cell and nerve responses (Kataoka et al. 2012; Roper 2013). As documented below, only one transmitter, ATP clearly meets all of the criteria for being a taste transmitter.

2. ATP as the Key Neurotransmitter in Taste Buds

The first suggestion that ATP may be important in neurotransmission came for the description of the ionotropic purinergic receptors, P2X2 and P2X3 in the gustatory nerves of taste buds (Bo et al. 1999). P2X2 and P2X3 belong to a 7-member class of ATP-gated monovalent cation channels that exist functionally as trimers. In the case of P2X2 and P2X3 – both homotrimers and heterotrimers exist. The presence of specific receptors in the postsynaptic element is but one criterion for a neurotransmitter. The others are: (1) presence of the presumed neurotransmitter in the presynaptic cells, (2) release of the substance upon stimulation, (3) activation of the postsynaptic partner by the neurotransmitter, and (4) clearance of the transmitter after stimulation – either by re-uptake or degradation.

Since ATP occurs ubiquitously in cells as a source of energy, it trivially satisfies the first of these criteria to be a neurotransmitter, i.e. presence in the presynaptic cell. The functional importance of ATP acting on P2X receptors in taste transmission was demonstrated both by knockout and by pharmacology. Single knockouts of P2X2 and P2X3, as well as a double knockout of both subunits had been developed and used to demonstrate a role for ATP in bladder pain (Cockayne et al. 2000). To determine if P2X2 and P2X3 play a role in taste transmission, we obtained the single and double knockout mice and tested whether behavioral and gustatory nerve responses to tastants were affected (Finger et al. 2005). Remarkably, in the P2X2/P2X3 double knockout mice, responses to all taste qualities were absent in both the chorda tympani and glossopharyngeal nerves, while responses to tactile and temperature information were unaffected (Fig. 1). Further, behavioral responses to all tastants other than acids (sour) were also dramatically reduced documenting the effectiveness of the knockout on taste. The residual avoidance of sour is likely due to collateral activation of polymodal nociceptors rather than through taste buds (Hallock et al. 2009; Ohkuri et al. 2012; Yu et al. 2020). Neither P2X2 nor P2X3 single knockout mice exhibited such severe taste deficits although both knockouts were impaired relative to wild-type controls (Finger et al. 2005). These findings suggested that ATP released from taste cells activated the postsynaptic nerve fibers via P2X2 and P2X3 receptors co-expressed in most gustatory ganglion cells (Dvoryanchikov et al. 2017).

Fig. 1.

Above: Representative recordings from the chorda tympani nerve of WT and P2X2/3 double-KO mice. Responses to tastants are eliminated while responses to cool temperature remain. Below: Bar graph comparing response magnitudes for WT (Blue) and KO (Red) animals to the array of taste and non-taste stimuli tested. Adapted from Finger et al. (2005)

One problem with knockout mice is that pleotropic effects can occur, especially when the knockout is global and not tissue specific. In that regard, it was shown that P2X2, in addition to being expressed on taste nerve fibers, is expressed in Type II taste cells, where it is involved in potentiating the release of ATP (Huang et al. 2011a, b). Indeed, the P2X2/3 DKO mice have reduced release of ATP compared to wild-type mice. Because of this concern and the concern that ATP release had not been detected from the Type III taste cells (Huang et al. 2007), a search began for alternative means to further investigate the role of ATP in the transmission of all taste qualities. At that time, as is discussed in more detail below, pharmacologists had developed P2X antagonists as a therapeutic intervention for chronic pain and cough. One of the antagonists, AF-353, is a membrane permeant antagonist of all P2X3-containing receptors – whether they be homomers or heteromers. To determine if AF-353 would phenocopy the P2X2/3 DKO mice, we applied the drug to the tongue of mice along with various tastants during chorda tympani nerve recording. Application of AF-353 to the tongue completely abolished responses to all tastants, including sour, similar to the findings in the P2XdKO mice (Fig. 2). Further, i.p. injection of AF-353 not only abolished all taste nerve responses, but also abolished behavioral preference to a synthetic sweetener, SC45647 (Vandenbeuch et al. 2015).

Fig. 2.

Effect of topical application of AF-353 on chorda tympani nerve responses. A. Representative integrated chorda tympani nerve response to different tastants before and after (red) application of 1.1 mM AF-353. Responses to all tastants were totally abolished after a 10 min treatment with AF-353. Responses start recovering 30 min after a rinse with water, denoting a reversible effect of the antagonist (not shown). Taste stimuli were applied for 30 s (bar beneath recording) and rinsed for 50 s with water. B. Percentage of neural response remaining after application of AF-353 at various concentrations on the tongue for 10 min. As all qualities were similarly affected, responses to all qualities were averaged (means ± SD) for each concentration of AF-353 applied to the tongue. Increasing the concentration of AF-353 proportionally decreased taste responses to all qualities. Modified, with permission, from Vandenbeuch et al. (2015)

To further demonstrate the role of ATP as a transmitter in the taste system, it was necessary to test whether ATP was released from taste buds with appropriate stimulation. Initial investigations showed that isolated taste epithelium released ATP when stimulated apically with bitter taste stimuli (Finger et al. 2005). Subsequently, two different teams showed measurable ATP release from individual Type II taste cells, i.e. those that transduce sweet, umami, or bitter taste qualities (Huang et al. 2007; Murata et al. 2010). Further, the amount of ATP released was directly related to the number of action potentials generated by the taste cell which is a measure of the magnitude of response of that cell (Romanov et al. 2007, 2008; Murata et al. 2010). Curiously, the Type II cells do not exhibit typical synaptic features complete with synaptic vesicles. Rather, at points of contact with the afferent nerves, Type II cells in mice show large mitochondria with tubular cristae closely apposed to the point of contact between taste cells and nerve fibers (Fig. 3) (Royer and Kinnamon 1988; Yang et al. 2020). These large mitochondria, termed “atypical” mitochondria appear to serve as a local reservoir for ATP (Romanov et al. 2018) which is released through large-pore, voltage-gated channels consisting of CALHM1 and CALHM3 heteromers (Taruno et al. 2013; Ma et al. 2018; Taruno et al. 2021). The pore size of this channel, 15–18 nm is sufficient to accommodate passage of hydrated ATP (Taruno 2018). The CAHLM1/3 type of synapse, recently named a “Channel Synapse” is unusual in lacking synaptic vesicles. Since the CALHM1/3 channel synapses are gated by voltage, they offer a means for regulating release of ATP in proportion to the level of activity of the receptor cell (Murata et al. 2010).

Fig. 3.

Semi-schematic diagram of a channel synapse from a Type II taste cell (blue) onto a terminal of gustatory nerve fiber (green). CALHM1/3 channels are embedded in the taste cell membrane at the point of contact, closely apposed to the large, “atypical” mitochondrion with tubular cristae. When (1) the taste cell fires an action potential, the strong depolarization (2) gates open the CALHM1/3 channels to release ATP into the synaptic cleft where (3) it activates P2X receptors on the afferent nerve fiber to generate a neural action potential. Modified from Romanov et al. (2018)

Another criterion for demonstration of a neurotransmitter is a means for inactivating or removing the neurotransmitter once released. A potential means for elimination of extracellular ATP, an ectoATPase, was first noted in the mid-1960s (Iwayama and Nada 1967; Nada and Iwayama 1969) but the molecular identity of the highly specific enzyme, NTPDase2, was not elucidated until 2006 (Bartel et al. 2006). NTPDase2 is highly specific for ATP over other nucleotides and will rapidly convert ATP to ADP which is ineffective in activating the P2X receptors on the afferent nerves although it can activate the P2Y receptors on taste cells (Huang et al. 2009) (Dando et al. 2012; Kataoka et al. 2012). Further, NTPDase2 is expressed along the membranes of Type I taste cells which envelop the other cell types and nerve fibers. Accordingly, the ectoATPase is well-positioned to eliminate any ATP released by either Type II or Type III cells.

Taken collectively, these studies show that ATP meets all the criteria for being a crucial neurotransmitter in the peripheral taste system. While taste-related release is clearly shown for the Type II taste cells (sweet, bitter, umami), no one has yet detected ATP release from Type III taste cells that transduce sour (Huang et al. 2007). Since functional P2X3-containing channels are necessary for successful transmission of sour taste information (Vandenbeuch et al. 2015), the source of the requisite ATP is unclear. As described above, activation of the Type III cells leads to release of serotonin via conventional vesicular type synapses. The released serotonin activates 5-HT3A receptors on the gustatory nerve fibers and contributes to the activation of these fibers (Fig. 4). However, serotonin alone is not sufficient to trigger the afferent fibers; co-activation of the P2X receptors is necessary (Larson et al. 2015, 2020). The source of the ATP for sour transmission is unclear, but it has not been detected from isolated Type III cells with biosensor cells expressing P2X receptors (Huang et al. 2007). Further, Type II cells are not the source of the ATP, since skn-1a knockout mice, which lack Type II taste cells, respond normally to sour stimuli (Larson et al. 2020). Other possible sources include the non-gustatory keratinocytes in the epithelium, but while keratinocytes are known to release ATP (Moehring et al. 2018) there is no evidence that they are stimulated by tastants. The source of the ATP required for Type III transmission remains one of the most important unanswered questions in the field.

Fig. 4.

The role of 5-HT3A signaling in sour taste transmission. (a) Representative chorda tympani nerve recording to various concentrations of citric acid before and 15 min after (red) i.p. injection of the 5-HT3 antagonist ondansetron (ODS; 1 mg/kg). (b) Average chorda tympani responses in WT mice to various stimuli before and after injection of ODS. The responses to acids were, in general, significantly smaller after ODS treatment. Similar results were observed with 5-HT3A knockout mice (data not shown). Data are presented as mean ± SEM. Modified from Larson et al. (2015)

3. Role of Purinergic Receptors in Intrabud Signaling

In addition to acting directly on nerve fibers via P2X receptors, ATP released from Type II cells stimulates purinergic receptors located on adjacent Type II and Type III cells to modulate ATP release from Type II cells (Roper 2013; Roper and Chaudhari 2017). ATP acts in an autocrine fashion to potentiate ATP release by stimulating P2X2 and P2Y1 receptors on Type II cells, causing an increase in intracellular calcium and an enhancement of ATP release in response to bitter, sweet, and umami stimuli. As mentioned above, knockout of P2X2 reduces ATP release as would be expected under these conditions (Huang et al. 2011b). However, ATP also acts in a paracrine fashion to reduce ATP release by acting on P2Y4 receptors on the Type III cells. Upon P2Y4 activation in response to ATP release from Type II cells, Type III cells release 5-HT, which subsequently binds to 5-HT1A receptors on Type II cells, causing an inhibition of further ATP release (Huang et al. 2009). One would expect that sour stimulation of Type III cells would also decrease responses to bitter, sweet, and umami stimuli since 5-HT would be released in response to sour stimuli. Since acids have non-specific effects on all taste cell types, this hypothesis was tested recently by expressing channelrhodopsin selectively in Type III cells and stimulating the tongue with blue light. Light stimulation decreased chorda tympani responses to bitter, sweet, and umami stimuli as well as to sour and salty stimuli, the latter presumably because the Type III cells were already activated by light and thus desensitized to further activation by sour and salty stimuli (Vandenbeuch et al. 2020).

4. Adenosine, a Product of ATP Hydrolysis, Modulates Sweet Taste via A2B Receptors

ATP is degraded first to ADP by NTPDase2 (Bartel et al. 2006) as well as other nucleosidases (Dando et al. 2012). The resulting ADP can be further degraded to adenosine by the action of ecto-5′-nucleotidase (CD73) expressed on Type III cells (Dando et al. 2012). The adenosine, in turn, can bind to adenosine receptors on gustatory neurons or taste cells to modulate taste responses. Two studies independently showed specific expression of the adenosine receptor Adora2b (A2B) on a subset of Type II cells in circumvallate taste buds (Dando et al. 2012; Kataoka et al. 2012). Both groups showed that A2B was expressed primarily in sweet-sensitive cells. Dando et al. (2012) used calcium imaging on a slice preparation of circumvallate taste buds to show that adenosine enhanced calcium responses and ATP release to sweeteners. Kataoka et al. (2012) showed that glossopharyngeal nerve responses to sweeteners were depressed in A2B knockout mice relative to controls. Further, they showed that the sweet receptors in posterior tongue that were modulated by adenosine were coupled to the G-protein Gα14 rather than gustducin, the main G-protein α subunit mediating sweet taste in anterior tongue as well as bitter taste throughout.

5. How Is Taste Function Altered in the Absence of NTPDase2?

As mentioned above, NTPDase2 is specifically expressed on the plasma membranes of the Type I glial-like taste cells, where it degrades the ATP released by the Type II cells to ADP. A specific knockout of NTPDase2 was developed by gene targeting methods and Entpd2-KO mice were characterized for changes in taste function (Vandenbeuch et al. 2013). All taste cell types were present in the knockout mice, including the Type I taste cells which normally express NTPDase2. Measurements of ATP release from the lingual taste epithelium of the knockout mice showed highly elevated levels of ATP in the epithelial tissues including the taste buds compared to their wild-type littermates. Chorda tympani and glossopharyngeal nerve recordings showed depressed responses to all taste stimuli, suggesting that the elevated ATP levels in the tissues desensitized the P2X receptors on the afferent nerve fibers resulting in decreased taste responses (Vandenbeuch et al. 2013). All taste qualities were affected, as would be expected since P2X3 is expressed on nearly all gustatory afferent nerve fibers including those that also express the 5-HT3 receptor (Larson et al. 2015).

6. Translational Implications

The P2X receptors, particularly P2X3, participate in widespread physiological activities including pain, cough, and urinary system functions. As a consequence, these receptors have become attractive pharmacological targets (Ford 2012). Several antagonists of P2X3 containing receptors (similar to AF-353) have been developed for the treatment of cough and have been very effective in diminishing chronic cough, although patient compliance has been an issue because, not surprisingly, these antagonists caused taste dysfunction (Abdulqawi et al. 2015; Smith et al. 2020). The P2X receptors on most cough and pain nerves are believed to be homotrimers of P2X3. In contrast, the geniculate ganglion neurons that innervate the tongue are likely heterotrimers of P2X3 and P2X2 since, in rodents, most ganglion cell neurons express both subunits (Dvoryanchikov et al. 2017). Calcium imaging of isolated geniculate ganglion neurons in mice showed that responses to ATP in the ganglion neurons differ in their sensitivity to the P2X3 antagonist AF-353 (Fig. 5) (Vandenbeuch et al. 2015). One population was completely blocked by I0 μM AF-353, while another population required 100 μM for a complete block of ATP responses, suggesting different composition of P2X subunits. Neurons isolated from P2X3 single knockout mice (presumably containing only P2X2 homotrimers) were completely unaffected by 100 μM AF-353. These data suggest that P2X receptors in all geniculate ganglion neurons contain at least one subunit of P2X3, since AF-353 applied to the tongue or injected i.p. blocks all taste responses. However, since the drug blocked some isolated neurons at lower concentrations than others, the P2X receptors in many ganglion neurons also contain at least one subunit of P2X2, which likely makes them less sensitive to the blockade by the AF-353 drug. Thus, since the majority of taste nerves in mice express P2X2/P2X3 heteromers, it seems reasonable that antagonists specific for P2X3 homotrimers might then spare taste function, since the heteromers would be less sensitive to the specific antagonist. Indeed, a new P2X3-selective antagonist, BLU-5937, does indeed ameliorate cough while sparing taste driven behaviors in rodents (Garceau and Chauret 2019). However, the rodents utilized all have dual expression of P2X2 and P2X3 in the taste nerves. For this approach to be effective in clinical situations, it is necessary that the taste neurons of humans be similar in stoichiometry to the rodents – i.e., contain at least one subunit of P2X2. Our preliminary data, however, suggest that taste nerves in most humans and primates do not express immunocytochemically-detectable P2X2 (Finger and High 2020). In keeping with this, more specific P2X3 antagonists may still show some taste disturbances (Friedrich et al. 2020). However, the full resolution of this issue requires further study.

Fig. 5.

A–C. Geniculate ganglion showing P2X2 (magenta) and P2X3 (green) immunoreactivity. Image is a maximum Z-projection of 12 optical sections through a ~ 16 μm tissue section. Scale bar = 100 μm. Brightness and contrast were adjusted linearly to preserve relative expression level information. D–E. Multiple populations of geniculate ganglion cells respond differently to AF-353. D. Change in fluorescence ratio of two ganglion cells in response to 10 μM ATP, 10 μM ATP with 10 μM AF-353 and 55 mM KCl. In the cell shown in the upper trace 10 μM AF-353 completely blocks the ATP response whereas it only blocks about 50% of the response in the cell shown in the lower trace. Drug application order was the same between top and bottom traces. E. Effect of various concentrations of AF-353 on ganglion cells of WT (circles; n = 19 cells), X2KO (diamonds; n = 7 cells), and X3KO; triangles; n = 7–9). WT cells were separated into two categories according to their response to ATP at 1 μM AF-353. Cells above the mean response were classified as “less sensitive” (closed circles) while cells below the mean response were classified as “more sensitive” (open circles). For WT, individual cells are represented as circles with straight lines connecting individual cells. For X2KO and X3KO symbols indicate means SEM. Asterisks indicate significance (p < 0.001 Mann–Whitney test between “more sensitive” and “less sensitive” cells). X2KO, P2X2KO; X3KO, P2X3KO; WT, wild-type. Adapted with permission from (Vandenbeuch et al. 2015)

7. Conclusions and Perspectives

Considerable evidence now exists that ATP is required for the transmission of all taste qualities to the nervous system. For the Type II taste cells, those that transduce sweet, bitter, or umami, ATP is released directly from the Type II cells via activation of large conductance CALHM1/3 channels. These channels in rodents are associated with a signaling complex that contains large atypical mitochondria in tight apposition to the neural membranes containing the P2X receptors. Action potentials in response to the taste stimuli activate the voltage-dependent CALHM1/3 channels to release ATP in a semi-quantal manner, causing activation of P2X2 and P2X3 on the sensory afferents and transmission of the taste information to the brainstem. For the Type III cells, the cells that transduce acids and some salts, the source of the ATP required for activation of sensory afferents is still unclear. Type III cells release serotonin in response to acid stimuli and the serotonin binds to 5-HT3a receptors on the sensory afferents, but serotonin alone is not sufficient to drive the afferents. ATP is also required since knockout or pharmacologic inhibition of 5-HT3a only partially reduces the nerve response, while P2X3 antagonists completely abolish the nerve response. The source of the ATP required is unclear – acid responses persist in mice that lack Type II taste cells and no one to date has measured ATP release from the Type III taste cells. This is clearly an important area of investigation for the future.

Another important unanswered question in the field is whether human taste nerves contain both P2X2 and P2X3 subunits. Antagonists selective for P2X3 homomers appear to be effective at reducing chronic cough in patients but cause minimal disturbance of taste. Yet preliminary data suggest that nerves innervating taste buds in humans express only P2X3 and not P2X2. Is this because human taste nerves contain only homomeric P2X3 receptors and if so, why do selective antagonists not block taste? Knowledge of the receptor stoichiometry in humans could address this question directly and help provide effective treatments for these conditions.