Of the multiple neurotransmitters and neuropeptides expressed in the mammalian taste bud, serotonin remains both the most studied and least understood. Serotonin is expressed in a subset of taste receptor cells that form synapses with afferent nerve fibers (type III cells) and was once thought to be essential to neurotransmission (now understood as purinergic). However, the discovery of the 5-HT1A serotonin receptor in a subset of taste receptor cells paracrine to type III cell suggested a role in cell-to-cell communication during the processing of taste information. Functional data describing this role are lacking. Using anatomical and neurophysiological techniques, this study proposes a modulatory role for serotonin during the processing of taste information. Double labeling immunocytochemical and single cell RT-PCR technique experiments documented that 5-HT1A-expressing cells co-expressed markers for type II cells, cells which express T1R or T2R receptors and release ATP. These cells did not co-express type III cells markers. Neurophysiological recordings from the chorda tympani nerve, which innervates anterior taste buds, were performed prior to and during intravenous injection of a 5-HT1A receptor antagonist. These experiments revealed that serotonin facilitates processing of taste information for tastants representing sweet, sour, salty, and bitter taste qualities. On the other hand, injection of ondansetron, a 5-HT3 receptor antagonist, was without effect. Collectively, these data support the hypothesis that serotonin is a crucial element in a finely-tuned feedback loop involving the 5-HT1A receptor, ATP, and purinoceptors. It is hypothesized that serotonin facilitates gustatory signals by regulating the release of ATP through ATP-release channels possibly through phosphatidylinositol 4,5-bisphosphate resynthesis. By doing so, 5-HT1A activation prevents desensitization of post-synaptic purinergic receptors expressed on afferent nerve fibers and enhances the afferent signal. Serotonin may thus play a major modulatory role within peripheral taste in shaping the afferent taste signals prior to their transmission across gustatory nerves.
Citation: Jaber L, Zhao F-l, Kolli T, Herness S (2014) A Physiologic Role for Serotonergic Transmission in Adult Rat Taste Buds. PLoS ONE 9(11): e112152. https://doi.org/10.1371/journal.pone.0112152
Editor: Hiroaki Matsunami, Duke University, United States of America
Received: May 21, 2014; Accepted: October 13, 2014; Published: November 11, 2014
Copyright: © 2014 Jaber et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper.
Funding: This work was supported by the National Institute on Deafness and Other Communication Disorders, National Institutes of Health (www.nidcd.nih.gov; 5R01DC00401). Partial funding for Open Access provided by The Ohio State University Open Access Fund. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The view of how the taste bud operates has changed dramatically over the last two decades. Once considered a passive recognition unit, the taste bud is now known to be a complicated sensory end-organ composed of elaborate networks of autocrine and paracrine communication pathways that significantly process the gustatory sensory information prior to signaling the central nervous system. These findings have led to the classification of gustatory transduction mechanisms in the taste bud into early and late events . Early transduction events occur between receptor activation by tastant molecules and the resulting depolarization of the taste receptor cell (TRC). Late transduction mechanisms, on the other hand, describe the processing of information among cells of the taste bud by excitatory and inhibitory feedback mechanisms which ultimately shape the neural discharge. A large number of neurotransmitters, neuropeptides, and their corresponding receptors are expressed in defined patterns across the varying cell types of the taste bud, typically referred to as types I, II, and III. Examples include neurotransmitters, such as serotonin, norepinephrine, GABA, and acetylcholine, and neuropeptides, such as cholecystokinin, neuropeptide Y, and vasoactive intestinal peptide. Late transduction events may shape peripheral gustatory signaling through mechanisms that include lateral inhibition, gain modulation, and adaptation. Thus, single TRCs are influenced not only by apical receptors activated by taste stimuli but also through basolateral receptor activation.
Of the multiple neurotransmitters expressed in the taste bud, serotonin ironically remains the best studied yet least understood. Serotonin is expressed in a subset of TRCs (type III cells) which form classic synapses with afferent nerve fibers in a large number of species including mouse, rat, rabbit, and monkey , , , , , . These cells also express the candidate sour receptor PKD2L1 . Largely because of this classic synaptic morphology, it was long assumed that serotonin was essential to transmission of gustatory information to the central nervous system. ATP is now widely acknowledged as the main gustatory neurotransmitter within the taste bud, acting on P2X receptors on afferent nerve fiber terminals , . ATP is released from type II cells (cells which express T1R and T2R receptors) in response to tastant stimulation . Release occurs in a calcium-independent but voltage-dependent manner through ATP-release channels. The identity of these channels has been suggested to be connexin or pannexin hemichannels ,  or a newly identified release channel, CALHM1 . Additionally, ATP may participate in cell-to-cell-communication through the activation of P2Y and P2X receptors expressed on TRCs , , , , . Hence, ATP release from type II cells may not only stimulate afferent nerve fibers but additionally stimulate type III cells via cell-to-cell communication by activation of purinergic receptors , . This stimulation may then result in serotonin release from these cells .
The physiological consequence of this released serotonin remains equivocal. Interestingly, the first demonstration of cell-to-cell communication with the taste bud involved serotonin . At the time, the report that TRCs responded to serotonergic stimulation was unexpected given its putative role as a neurotransmitter to the afferent fiber . In rat posterior TRCs, application of exogenous serotonin results in inhibition of ionic currents that cumulatively would render the cell less responsive to stimuli , . These effects were mimicked by 5-HT1A but not by 5-HT3 receptor subtype agonists. They were additionally blocked by a 5-HT1A receptor subtype antagonist. Subsequently, this observation has been extended to include a number of species , , ,  which similarly implicates the 5-HT1A receptor. In rat posterior taste buds, serotonergic TRCs and the 5-HT1A receptor have been demonstrated to be expressed in a paracrine manner  suggesting that release of serotonin from type III cells acts to inhibit a subset of neighboring cells. The identity and function of these cells are unknown.
In one scenario, serotonin could be important in transmitting the presence of sour stimuli to ATP-releasing type II cells. Type III cells respond to sour stimuli but do not themselves release ATP; however, blocking purinergic transmission blocks all taste qualities including sour . Therefore, one possibility is that serotonin acts to relay sour information to purinergic type II TRCs which express the 5-HT1A receptor. In another scenario, serotonin may provide a negative feedback pathway onto type II cells that influences ATP release since type III cells additionally express purinergic receptors. In this scenario, serotonin might influence all taste qualities, not just sour. In fact, serotonin is released from TRCs in response to multiple tastant qualities in mouse taste buds . This anatomical arrangement of purinergic and serotonergic crosstalk between type II and type III cells, suggests that serotonin release from type III cells may influence ATP release from type II cells.
The present study explores the hypothesis that serotonin modulates transmission of taste information to the central nervous system by operating during late transduction mechanisms. To investigate serotonin’s role in the peripheral taste system, anatomical, molecular, and physiological techniques were employed to explore the hypothesis that paracrine cell-to-cell communication within the taste bud between the serotonergic TRCs and the 5-HT1A receptor expressing cells plays a physiologically significant role in late transduction mechanisms. Immunocytochemistry and single cell reverse-transcriptase-PCR techniques were employed to phenotype the 5-HT1A-expressing cells with respect to co-expression pattern of this receptor relative to distinct phenotypic markers associated with specific cells types and with the transduction of specific tastes. Additionally, whole nerve neurophysiological recordings were employed to study the effect of blocking serotonergic transmission on gustatory afferent nerve responses. We conclude that serotonin facilitates gustatory responses during late transduction mechanisms by exerting an inhibitory action on type II cells of the bud which serves to regulate ATP release from these cells and thus prevents desensitization of purinergic receptors on the afferent nerve fiber.
Materials and Methods
Serotoninergic signaling mechanisms in the rat taste bud were studied with a combination of histological, molecular, and physiological approaches. Adult male Sprague–Dawley rats served as subjects for all experiments. All procedures were approved by The Ohio State University's Institutional Animal Care and Use Committee (Protocol ID 2009A0026-R1).
The expression pattern of the 5-HT1A receptor throughout the oral cavity and its co-localization with TRC phenotypic markers were studied using single and double labeled immunofluorescent techniques. Subjects were sacrificed by decapitation after first achieving surgical level of anesthesia with a Ketamine (91 mg/ml)/Acepromazine (0.09 mg/ml) mixture (0.09 ml/100 gm BW). Harvested tissue included anterior tongue, nasoincisor duct, palate, foliate, and circumvallate papillae. For experiments using anti-serotonin antibodies, animals were preloaded with a 5-HT precursor 5-hydrxytryptophan (80 mg/kg, i.p. injection 1 hr prior to sacrifice). Tissue was immersion fixed in Bouin’s fixative (5 hrs, 4°C), paraffin embedded, and sectioned at 4 micron thickness. Conventional and TSA amplified immunocytochemistry (ICC) were conducted as previously described using fluorescent microscopy . For double labeling immunocytochemistry with primary antibodies raised in different species, sections were blocked in 10% normal serum of species matching the secondary fluorescent antibody (1 hr, RT). Sections were then incubated in primary antiserum (150 µL per slide, diluted in PBS containing 2% normal serum) at the appropriate dilutions. Antibody dilutions were empirically adjusted to minimize cross reactivity and background staining. Descriptions of primary antibodies are presented in Table 1. The following day, sections were rinsed in PBS and incubated with Cy3- or FITC-conjugated secondary antibody (150 µL per slide, 1∶400 in PBS containing 1.5% normal serum, 1 hr, RT) in the dark. Slides were then washed in PBS (3×10 min) and incubated in the second primary antiserum at the appropriate dilution (24 hr, 4°C). The following day, slides were washed in PBS and incubated with the second secondary antibody (1∶400, 1 hr) in the dark, washed in PBS, mounted in Fluoro-Gel, and observed under a fluorescent microscope equipped with a digital camera (Nikon EFD-3, Japan).
When it was not possible to acquire two primary antibodies raised in different species, an indirect immunofluorescence double-labeling protocol was employed which allowed detection of two antigens with primary antibodies raised in the same species by altering the dilution and detection method for each antigen with either conventional or Tyramide Signal Amplification (TSA Biotin System, PerkinElmer, Waltham MA, USA). One primary antibody is used at very high dilution with TSA detection whereas the second primary is used with standard dilution and detection methods. Streptavidin-conjugated IgG Fab fragment (instead of a fluorophore-conjugated one) is used to detect the first primary antibody so that the second primary antibody can be used at very high dilution thus minimizing interference or cross-reaction between the first primary antibody and the second secondary antibody . This method also prevents the first secondary antibody from reacting with the second primary antibody.
Rehydrated deparaffinized tissue sections were treated with a 0.5% hydrogen peroxide in methanol (30 min) to eliminate endogenous peroxidase activity. Sections were subsequently incubated in PBS containing 10% normal serum (1 hr, RT) to reduce nonspecific antibody binding. Primary antiserum (diluted in PBS containing 2% normal serum) was applied and slides were housed in a closed moist chamber (36 hr, 4°C). After PBS washing, sections were incubated with biotin-streptavidin-conjugated IgG Fab fragment (1∶800 in PBS containing 1.5% normal serum, 1 hr, RT) and processed according to the kit’s instructions (NEN Life Science Products, Boston, MA, USA). Slides were mounted in Fluoro-Gel and observed under the fluorescent microscope.
Control slides were included in every experiment to ensure that fluorescent signals did not arise from cross-reactivity of secondary antibodies with the inappropriate primary antibody. To control for false positives and non-specific binding in single-label ICC experiments, two slides were included in each experiment; the first slide omitted the primary antibody and the second slide omitted the secondary antibody. Similarly, for double-label ICC experiments (including TSA experiments), four slides were included; the first slide lacked the first primary, the second lacked the second primary, the third lacked the second primary, and the fourth slide lacked the second secondary antibody. In all control slides, cells were expected to lack positive staining. As a criterion for acceptable labeling, the emitted signal of the secondary antibody was expected to be visible only in the emission spectrum matching its fluorophore. If color bleeding to the emission spectrum of the alternate secondary antibody was visible, adjustments were made in the dilutions of the primary antibody.
Data were analyzed by counting individual taste buds and TRCs. To avoid counting the same cell more than once, only every fourth consecutive section on each slide was chosen for data analysis. This ensured that the sections being considered for cell counting were separated by at least 12 µm, greater than the diameter of a taste cell. Taste buds were analyzed twice to confirm the cell count. Only cells that showed clear apical and/or perinuclear labeling were considered immunopositive. All cell counts were performed under 40X magnification. A limitation for direct microscopic cell counting was that it only accounted for stained cells within the visible plane of the 4 µm-thick section and it did not fairly represent all labeled cells within the actual tri-dimensional frame of the taste bud. Antigen-expressing cells cannot be presumed to be uniformly distributed throughout the taste bud. Consequently, consecutive tissue sections would most likely show discrepancies in the number of labeled cells. This method therefore serves as an approximate representation of the actual number of immunopositive cells. Digital photos of immunofluorescent cells were captured under the 40X objective lens by a digital camera (Photometrics CoolSNAP. Tucson, AZ), processed using MetaMorph software (Molecular Devices LLC, Sunnyvale, CA, USA), and subsequently imported into imaging software (Canvas Illustrator; ACD systems, USA).
Single Cell RT-PCR
Individual TRCs were phenotyped using single cell RT-PCR reactions. RNA was extracted from single TRCs dissociated from circumvallate papillae, used as template in a two-round cDNA synthesis process, followed by standard PCR reactions with primers for specific cell markers. Single dissociated TRCs were collected under microscopic examination from a total of ten animals. Each TRC was carefully suctioned into a glass pipette (∼8 µm-diameter) in less than 20 nanoliters of ECF solution. The pipette tip was examined at 100X magnification using an inverted bifocal microscope to verify the presence of a single TRC. The majority of harvested TRCs had a morphological appearance consistent with type II cells, as indicated by a circular soma and short cellular processes. The pipette solution containing the single cell was expelled into a PCR tube containing 3 µl of RNA extraction solution. The cell was lysed by vigorous vortex mixing (1 minute RT).
RNA was first amplified in a linear fashion prior to cDNA synthesis using the MessageBOOSTER cDNA Synthesis from Cell Lysates Kit (EPICENTRE Biotechnologies, Madison, WI, USA). Prior to cDNA synthesis, template RNA was desalted and purified from potential genomic DNA contamination using RNase-free DNase. First-strand cDNA was reverse-transcribed from template RNA by adding T7-oligo(dT) primer, annealing at 65°C for 5 min and then incubating the reaction with MMLV reverse transcriptase at 37°C for 60 min. Second-strand cDNA synthesis was then performed by adding DNA polymerase and incubating at 65°C for 10 min. The reaction was terminated by heating at 80°C for 3 min. cRNA was then transcribed from cDNA by incubating the reaction with RNA polymerase at 42°C for 4 hrs. RNase-free DNase was subsequently added to remove any remaining DNA (37°C for 15 min). The transcribed RNA was purified and desalted using the RNeasy MinElute Cleanup Kit (Qiagen, Inc., Valencia, CA, USA). RNA was washed and centrifuged in RNA spin columns with a succession of buffers and ethanol solutions (RLT, RPE, 100° and finally 80° ethanol), eluted in a final volume of 14 µL nuclease-free H2O, concentrated in an Eppendorf Vacufuge (3 to 8 µL final volume), and then used in the second round of 1st-strand cDNA synthesis with random primers and MMLV reverse-transcriptase (37°C, 1 hr). Finally, RNase H was incubated with the reaction for 20 min at 37°C to eliminate any remaining RNA and the reaction was terminated at 95°C for 2 min.
The expression of specific phenotypic markers in a single TRC was investigated using custom-designed forward and reverse primers. Primer sequences are described in Table 2. Primer sets included 5-HT1A and taste transduction receptor molecules T1R3, T2R9, α-gustducin, PKD2L1, and GAD. Three sets of 5-HT1A primers corresponding to three different sequences within the 5-HT1A gene were tested in whole taste papilla and single cell RT-PCR (termed 5-HT1A[i], 5-HT1A[ii], and 5-HT1A[iii]). These primers were tested in a group of 10 single cells extracted from the CV papillae of five animals (group “A”), along with primers for α-gustducin, T1R3, and T2R9.
The standard PCR reaction mixture contained 10 µL iQ SYBR Green Supermix (Bio-Rad Laboratories, Hercules, CA, USA), 1 µL of template cDNA, 1 µL each of 100 µM forward and reverse primers, and 7 µL nuclease-free water. Template cDNA that was produced from total bud RNA was diluted at 1∶5 for RT-PCR; cDNA produced from a single TRC RNA was used undiluted. The PCR profile was 3 min at 95°C (initialization step, one cycle), 15 sec at 95°C (denaturation step, 40 cycles), 30 sec at 58°C (annealing step, 40 cycles), 30 sec at 72°C (extension step, 40 cycles), and 30 sec at 72°C (final elongation step, 1 cycle). PCR products were separated by gel electrophoresis on a 1.5% agarose gel containing 0.5 µg/ml ethidium bromide. Each lane was loaded with a mixture containing 5 µl of PCR product, 4 µl double-distilled H2O and 1 µl gel loading buffer (Invitrogen 10X Blue Juice, Life Technologies, NY, USA). Bands were measured relative to a 100 bp DNA ladder (Lonza SimplyLoad DNA ladders, Rockland, ME, USA).
PCR reactions were optimized on cDNA synthesized from RNA extracted from whole posterior papillae. In addition to performing whole taste papilla RT-PCR experiments as positive control, primers for cytokeratin 8 (CK8), a protein ubiquitously expressed across all mature TRCs, were included in every single-cell RT-PCR run as a positive control. Negative control reactions, to ensure that the amplified product did not result from genomic or extraneous DNA, included reactions lacking primers or DNA template. The latter contained a small aliquot of ECF (in lieu of the DNA template) sampled adjacent to the collected cell. Single cells were only included in the data set if they satisfied criteria for both positive and negative controls.
Chorda tympani neural recordings
Chorda tympani neurophysiological recordings were performed as previously described . Subjects, weighing between 160 and 300 gm, were anesthetized with 50 mg/kg pentobarbital (Nembutal 50 mg/ml, Ludbeck Inc, USA) and supplemented with 0.05 ml every thirty minutes or as needed to maintain surgical level of anesthesia. The animal was secured with a non-traumatic head holder and tracheotomized. The left external jugular vein was catheterized with 15 cm polyethylene tubing (2 mm internal diameter) filled with heparinized saline solution. Body temperature was maintained at 36–37°C. Animals were sacrificed at the end of the recording session by decapitation while under surgical level of anesthesia.
The chorda tympani was exposed from the point where it leaves the tympanic bulla caudally to the point where it separates from the lingual nerve rostrally. It was freed of surrounding fascia and connective tissue and severed at the caudal end. The peripheral end of the nerve was carefully desheathed of its epineurium. Neural recordings were made between two platinum iridium electrodes using an AC preamplifier (A-M systems Inc. Model 3000) at a gain of 500X and filtered between low pass and high pass bandwidths of 3 kHz and 100 Hz respectively. Signals were full-wave rectified and integrated with a time constant of 0.2 sec on a wide band AC to DC converter (GRASS systems, model 7P3B) and displayed on an ink recording chart.
Stimuli were presented to the anterior tongue via a Pasteur pipette. Gustatory stimuli (0.5 M NaCl, 1.0 M sucrose, 0.03 M quinine HCl, or 0.01 N HCl) were prepared with reagent grade chemicals in distilled water. The chosen concentrations of taste solutions produced greater than half-maximal integrated whole nerve responses based on results from previous CT recording studies , , . Stimuli were delivered in the following sequence: NaCl, sucrose, quinine, and then HCl, with a two-minute interval between consecutive stimuli.
Stability criteria of neural responses had to be demonstrated prior to drug delivery. These criteria included a steady baseline between stimuli, consistent response magnitude to three sets of taste stimuli delivery, and overall consistency in the shape of the neural responses. Once these criteria were satisfied, experimental drug or control vehicle was administered into the jugular vein in the form of a single 300 µL bolus injection delivered over a period of 5 seconds. Nerve responses to taste stimuli were monitored and recorded in the periods before, and at least up to one hour following the injection to examine the immediate, as well as the long term effects of the injected solution.
The tonic portion of the integrated neural response was selected for data analysis. The amplitude was measured in arbitrary units between baseline and the midpoint on a line drawn through the response plateau. Magnitudes of the three responses taken just before drug injection were averaged and their mean was compared with that of the averaged amplitude of the first three responses for that stimulus taken immediately after the drug injection. Post-injection data were normalized to the pre-injection mean, thus the normalized value of post-injection that had a value of 1.0 indicated no change in mean-response. A mean value above 1.0 signified an increase in post-injection response magnitude and a mean below 1.0 signified a decrease. Relative data were plotted as response means ± standard errors.
Statistical analysis was performed with Shapiro-Wilk’s normality test with the following null hypothesis (Ho): the population has a normal distribution. The test resulted in p values that were always greater than 0.05, which implied that the alternative hypothesis should be rejected. Since the goal was to compare the means of two variables and test if the average difference is significantly different from zero, the paired T-test was chosen for statistical analysis with the following null hypothesis (Ho): mean response-amplitudes corresponding to two different stages of recording were not significantly different. A two-tailed significance (2-tailed p value) that was ≥0.05 signified that the null hypothesis is valid and that the difference between the two means being compared was not statistically significant.
Phenotypic expression patterns of the 5-HT1A receptor in taste buds
Cellular and molecular experiments were performed to explore the expression pattern of the 5-HT1A receptor in taste buds across anterior and posterior gustatory fields and to examine the co-expression patterns of 5-HT1A–expressing TRCs with known TRC-type phenotypic markers and taste receptor molecules.
5-HT1A expression throughout the oral cavity.
While previous work in posterior taste buds has demonstrated the expression of 5-HT and the 5-HT1A receptor in a paracrine manner, no data exist on 5-HT1A expression in anterior gustatory fields. To investigate these fields, taste buds from the fungiform papillae (FP), nasoincisor duct (NID), and soft palate were examined using a single-label immunocytochemical approach with either a 5-HT or a 5-HT1A antibody. Clearly labeled 5-HT or 5-HT1A immunopositive cells were observed using immunofluorescence in taste buds of FP and the NID (Figure 1). In all cases, 5-HT-immunopositive TRCs featured tall and slender cell bodies and relatively small nuclei, characteristics of type III cells. On the other hand 5-HT1A-immunopositive TRCs were relatively shorter and with well-formed roundly-shaped soma, as commonly observed in type II cells. In fungiform papillae, an average of about two 5-HT-immunopositive TRCs (1.83±0.03; n = 3 animals) and about two 5-HT1A-immunopositive TRCs (1.97±0.05; n = 3) were detected per cross-sectioned taste bud. In taste buds of the NID, an average of one (1.1±0.1; n = 3) 5-HT1A-immunopositive cell was detected per cross-sectioned taste bud. 5-HT1A immunopositive cells were not observed in taste buds of the soft palate (n = 3). To examine the co-expression pattern of 5-HT and 5-HT1A in the anterior field, double label immunocytochemistry experiments were performed on fungiform and NID taste buds. In all instances of these experiments, immunopositive TRCs to each antibody were observed in non-overlapping subpopulations of TRCs (Figure 1) in taste buds of fungiform papillae and the nasoincisor duct. These data suggest that the 5-HT1A receptor is distributed in the anterior field in a manner similar to foliate and circumvallate taste buds. This receptor may not be expressed in the soft palate.
Illustrated is single labeling ICC overlaid on a bright field image for a single taste bud from the anterior tongue (top). In double labeling experiments, paracrine expression of 5-HT and 5-HT1A was observed in taste buds of both the fungiform papillae and in the NID. Expression of the 5-HT1A receptor was not observed in taste buds of the palate (not illustrated). Scale bars are 15 microns.
5-HT1A expression and cell type.
To investigate the cell type expression of the 5-HT1A receptor, double label immunocytochemistry experiments were conducted on taste buds from anterior and posterior gustatory fields with the phenotypic cell type markers gustducin, a marker of a subset of type II cells, or NCAM, a marker of type III cells in rat.
A majority of 5-HT1A-immunopositive TRCs co-expressed gustducin in anterior and posterior fields. These immunopositive cells displayed emblematic morphology of type II cells (Figure 2). In anterior fields, about 82% of all 5-HT1A cells co-expressed gustducin (n = 44 cells from 56 cross sectioned taste buds) whereas in posterior fields, about 70% of 5-HT1A cells co-expressed gustducin (n = 91 cells from 92 taste buds). Data are summarized in Table 3. On the other hand, fewer gustducin-positive cells co-expressed 5-HT1A in anterior fields whereas comparable numbers were expressed in posterior fields. About 37% of gustducin cells co-expressed 5-HT1A in anterior taste buds and 75% in posterior taste buds. 5-HT1A immunopositive cells were not observed to co-express NCAM. In 180 cross sectioned taste buds from posterior field, 135 5-HT1A-immunopositive cells and 175 NCAM immunopositive cells were observed. None displayed overlapping expression. Since reliable NCAM immunostaining could not be produced in anterior tongue tissue, data were only obtained from posterior tongue tissue. Taken together, these results confirm 5-HT1A expression in type II cells, agreeing with our previous suggestion , and exclude its expression in type III cells.
5-HT1A-expression extensively overlapped with expression of alpha-gustducin. Illustrated (top) is an example from foliate papillae (FOL). On the other hand, no overlapping expression was observed for 5-HT1A and NCAM, illustrated at bottom with tissue from the circumvallate papillae (CV). Scale bars are 15 microns.
Since type II cells are heterogeneous, additional rat type II phenotypic markers were tested . These included cholecystokinin 8 (CCK-8), neuropeptide Y (NPY), and the NPY-1R receptor. 5-HT1A expressing cells in fungiform taste buds significantly co-expressed CCK-8 (Figure 3, Table 3). Seventy-five percent of 5-HT1A positive cells co-expressed CCK-8 in anterior taste buds (36 of 42 immunopositive TRCs) and 70% of 5-HT1A positive cells co-expressed CCK-8 in posterior cells (57 of 81 immunopositive TRCs). Not all CCK-8 positive cells expressed 5-HT1A. Most anterior CCK-8 cells co-expressed 5-HT1A (85%; 36 of 42 cells) whereas fewer posterior CCK-8 cells co-expressed 5-HT1A (41%; 57 of 138 TRCs). A similar expression pattern was noted for NPY. NPY-expressing TRCs are a subset of CCK cells (i.e., all NPY cells co-express CCK) so strong overlapping expression with 5-HT1A was expected. In anterior fields, 60% of 5-HT1A cells co-expressed NPY (36 of 60 TRCs) whereas 84% (59 of 70 TRCs) did so in the posterior field. Similar to CCK results, most NPY cells in anterior fields co-expressed 5-HT1A (75%; 36 of 48 TRCs) whereas fewer did so in posterior taste buds (41%; 59 of 142 TRCs). NPY-1R, which is expressed paracrine to NPY cells, similarly had strong overlapping expression with 5-HT1A. In the anterior field, about half of the 5-HT1A cells expressed NPY-1R (58%; 56 of 96 TRCs) with greater co-expression observed in posterior taste buds (85%; 59 of 70 TRCs). Conversely, in anterior TRCs, 75% of NPY-1R TRCs co-expressed 5-HT1A (36 of 48 TRCs) compared to 42% (59 of 142 TRCs) in posterior taste buds. Thus, peptide co-expression patterns with 5-HT1A confirmed previous gustducin results and were highly consistent the known co-expression patterns with CCK, NPY, and NPY-1R among themselves.
Immunocytochemical double labeling examples are presented for 5-HT1A and CCK8 in fungiform (FP) and foliate (FOL) papillae, for 5-HT1A and NPY in fungiform and foliate papillae, and for 5-HT1A and NPY-1R in foliate papillae. Since these neuropeptides are known to be expressed in type II cells, these co-expression patterns corroborate expression of 5-HT1A in type II cells. Scale bars are 15 microns.
Finally, a third set of phenotypic markers, which display less type II cell-specific expression in rat, were examined: synaptosomal-associated protein 25 (SNAP-25) and glutamic acid decarboxylase (GAD). In the rat, SNAP-25 is expressed predominantly in type II cells, with a minor degree of co-expression in type III cells . The observed co-expression between 5-HT1A and SNAP-25 resembled that of 5-HT1A and gustducin (Figure 4; Table 3). In posterior taste buds, most of the 150 5-HT1A-immunopositive TRCs (89±1%; n = 134) co-expressed SNAP-25, which was expressed in 162 TRCs. Similarly, a large majority of the SNAP-25 cells (134 of 162 cells) co-expressed 5-HT1A. Results showed varying degree of overlap in the expression of 5-HT1A with GAD (Figure 4). Only 34 of the 135 5-HT1A expressing cells (18% ±2) co-expressed GAD, which was expressed in a total of 122 cells.
Each is robustly expressed in type II cells of rat taste buds. Examples from circumvallate (CV) taste buds are illustrated. In each, strong overlapping patterns of immunopositive cells were observed. Scale bars are 15 microns.
Co-expression patterns of the 5-HT1A receptor with the tested phenotypic markers were statistically compared between posterior and anterior gustatory fields (posterior tongue vs. anterior tongue/NID respectively). Using the paired samples T-test, data revealed no significant differences between posterior and anterior fields for the co-expression of 5-HT1A with gustducin (p = 0.062), NPY (p = 0.058), NPY-1R (p = 0.053), CCK-8 (p = 0.119), or GAD (p = 0.151). This suggests a similar pattern of co-expression for 5-HT1A in posterior tongue and anterior tongue/NID (Table 3).
5-HT1A expression and taste receptor molecules.
Phenotypic characterization of the 5-HT1A expressing TRC was continued with co-expression pattern with members of the T1R and T2R families of taste receptor molecules using single cell RT-PCR experiments on 51 TRCs cells isolated from the CV papillae of ten animals. The bitter receptor T2R9, the sweet/umami receptor T1R3, the sour receptor PKD2L1, the phenotypic markers gustducin and GAD were examined. Cytokeratin 8 (CK8), expressed in all mature TRCs but not in neighboring epithelial cells, was included as an internal positive control. All primers were optimized in separate RT-PCR experiments (n = 6) on cDNA reverse transcribed from total RNA extracted from individually excised whole CV papillae (see Table 2 for primer sequences). Negative control experiments were performed to ensure that PCR product did not arise from extraneous DNA. In one negative control, nuclease-free water was substituted for template in the RT-PCR experiment (water control). In another, ECF was collected from a region adjacent to a dissociated cell to test for extraneous DNA that might be present in the bathing medium (cell-free control). In other negative controls, primers were omitted. Gel analysis of PCR products from all negative control reactions showed no visible bands.
In the first group of experiments, ten single TRCs were randomly chosen from dissociated CV papillae and tested with three 5-HT1A primers sets (5-HT1A[i], 5-HT1A[ii], 5-HT1A[iii]), each corresponding to a different region of the 5-HT1A gene, along with primers for α-gustducin, T1R3, T2R9, and CK8. Four cells were negative for CK8 PCR product and were excluded. Of the remaining six cells, four showed PCR product of correct size for all three 5-HT1A primers, confirming the specificity of these three primer sets. Three of the four cells expressing 5-HT1A also expressed T2R9. The remaining cell expressed neither taste receptor molecule.
In subsequent experiments, one of the 5-HT1A primers, 5-HT1A[iii], was chosen and phenotypic expression markers were expanded to include primer sets for CK8, gustducin, T1R3, T2R9, PKD2L1 and GAD on 35 individually isolated TRCs. Twenty-five cells were CK8-positive and included in data analysis. Combined data (Table 4) indicated that among all 51 single TRCs tested, 31 were CK8-positive, and thus were included in the data analysis. Of the CK8-positive TRCs, twelve showed specific bands for 5-HT1A. Of these twelve cells, ten expressed taste receptor molecules; four co-expressed T2R9; four co-expressed T1R3; two co-expressed T2R9 and T1R3; and two expressed neither taste receptor molecule. On the other hand, of the twelve T2R9-expressing TRCs observed, six did not express 5-HT1A, and of ten T1R3-expressing TRCs, four TRCs did not co-express 5-HT1A. Thus while almost all 5-HT1A expressing TRCs expressed members of the TR families, the converse was not true. 5-HT1A-expressing TRCs may represent a unique subpopulation of T1R and T2R expressing TRCs.
5-HT1A co-expression patterns with single cell RT-PCR appropriately mirrored those observed with immunocytochemistry. For example, of the 12 5-HT1A and 14 gustducin expressing TRCs, 7 co-expressed both proteins. Of the 6 GAD-positive cells observed, only two TRCs co-expressed 5-HT1A and GAD. Of the two PKD2L1-positive cells observed in the 25 tested cells, none expressed either 5-HT1A or other type II TRC-specific markers. As well, the expression pattern of multiple type II cell markers within a single TRC served as internal control. Of twelve TRCs that expressed T2R9, ten co-expressed its linked G-protein α-gustducin. Further, co-expression between gustducin and T1R3 was observed in five out of thirty-one TRCs. This pattern of co-expression agrees with previous studies that reported these two molecules co-localized in mouse TRCs , . Of the six GAD-positive cells, three co-expressed T1R3, one co-expressed T2R9, and two co-expressed gustducin. These results agree with previous studies that report GAD co-expression with gustducin and T1R3 in mouse taste buds , , . Overlap was not detected between the type III TRC marker PKD2L1 and any of the markers associated with type II cells, including 5-HT1A. Table 4 illustrates data from single cell RT-PCR experiments, and Figure 5 illustrates sample gel electrophoresis data from the RT-PCR products of seven different single TRCs.
PCR products from seven individual taste receptor cells (sc1, sc5, sc7, sc12, sc20, sc21, and sc28) to seven separate PCR reactions are illustrated. The PCR reactions include positive control cytokeratin 8 (CK8), the 5-HT1A receptor, the sweet/umami taste receptor molecule T1R3, the bitter receptor molecules T2R9, the sour taste receptor PKD2L1, and the phenotypic markers alpha-gustducin (GUS) and glutamic acid decarboxylase (GAD). For the most part, cells which expressed the 5-HT1A receptor also co-expressed members of the T1R or T2R receptors. On the other hand, cells which expressed the sour receptor PKD2L1, which are type III cells, were not observed to express 5-HT1A. These data are in agreement with immunocytochemical data that 5-HT1A expression occurs predominately in type II cells.
On the other hand, unexpected co-expression patterns were also observed. Among the 32 CK8-positive TRCs, two co-expressed T1R2 and T2R9 (out of the total of the twelve T2R9- and ten T1R3-expressing cells). This was an unexpected outcome since the general consensus supports the notion that those two molecules are expressed in non-overlapping populations of TRCs , , , , . While it is conceivable that genetic contamination may have occurred during the isolation process of single TRCs, recent studies have reported co-expression in T1R and T2R in a minor subpopulation of TRCs in zebrafish  and in the solitary chemoreceptor cells in mice . Therefore, the co-localization of T2R and T1R in two cells in the current experiments may represent a small overlapping subpopulation T1R and T2R expressing TRCs.
Chorda tympani neural recordings
To explore the physiological role of serotonin in the taste bud, chorda tympani (CT) neural responses were measured to lingually-applied tastants before and after intravenously injected serotonergic receptor antagonists. Control experiments were performed to determine vehicle effects and confirm the access of intravenously applied agents to taste buds.
Control saline injections.
To examine vehicle effects on chorda tympani responses to taste stimuli, 300 µL single bolus jugular injection of saline was performed in eleven rats. Post-injection tastant responses were normalized to pre-injection response magnitudes recorded over the 20–30 minutes prior to injection. No change in baseline integrated neural activity was observed in any experiments during saline injection. Data averaged from all eleven experiments demonstrated no significant change of the neural response to any tested taste qualities when post-injection response amplitudes either immediately after (first three responses post injection) or up to one hour after saline injection (final 3 responses) were compared with pre-injection values (last 3 responses prior to injection; Figure 6). When pre-saline and post-saline mean response amplitudes were compared, the paired-samples T-test resulted in p-values greater than 0.05 (p = 0.085, 0.063, 0.191 and 0.193 for NaCl, sucrose, quinine and HCl, respectively).
Sample neural responses to 0.5 M NaCl, 0.03 M Quinine, 1.0 M sucrose, or 0.01 N HCl before and after a single bolus injection (300 µM) of saline are illustrated (upper left). Data are summarized in the histogram (upper right). On the other hand, intravenous injection of ATP produced transient neural responses which could be blocked by prior injection of the purinergic antagonist PPADS (lower left). Intravenously administered PPADS was also effective in blocking lingually applied tastant responses. Data summarizing the antagonistic effect of PPADS on ATP injection or tastant lingual stimulation are illustrated (histograms middle and lower right, respectively).
Activation and antagonism of purinergic receptors.
As an additional test that intravenously injected drugs could predictably affect gustatory neural responses, purinergic receptors were tested. As expected, injection of intravascular ATP produced neural excitations. ATP administered at three different concentrations (0.75, 1.5 or 3 mg ATP/kg BW, 300 µl) at 6 minute intervals, uniformly caused transient neural excitations (n = 3). Responses were bell-shaped and displayed dose-dependent magnitudes (Figure 6). Accompanying neural excitations, there was a transient dose-dependent increase in respiratory rate that lasted between 10 and 20 seconds produced by the transient ATP-induced increase in pulmonary artery pressure , .
The non-selective P2 purinergic receptor antagonist, PPADS, was highly effective in blocking neural responses produced by either intravascular injections of ATP or lingually applied tastants. Following injection of 10 mg/kg PPADS, mean CT response amplitude to 1.5 mg/kg ATP declined by 97% (n = 3; p<0.001). Responses recovered to half of their pre-injection amplitudes 50 minutes after injection. At 20 mg/kg, PPADS virtually abolished CT responses to ATP (98.8% inhibition, n = 3, p = 0.001). Responses recovered to almost half (42%) of pre-injection magnitude 62 minutes after PPADS injection. Additionally, 10 mg/kg PPADS (n = 3) resulted in significant reversible inhibitions of CT responses to all taste qualities (Figure 6). Immediately after injection, CT responses decreased 53% in mean amplitude to NaCl (p = 0.003), 82% to sucrose (p = 0.001), 79% to quinine (p = 0.001), and 75% to HCl (p = 0.001). PPADS at 20 mg/kg (n = 3) resulted in larger decreases in neural responses: (87% NaCl (p = 0.001), 85% sucrose (p = 0.002), 97% quinine (p = 0.001), and 97% HCl (p = 0.001)). Responses to all taste qualities displayed significant recovery within 45 minutes after 10 mg/kg PPADS injection and 60 minutes after 20 mg/kg injection. Collectively, these positive control experiments with ATP and PPADS provided confidence in the physiological preparation.
Effect of blocking 5-HT1A on CT responses to gustatory stimulation.
Chorda tympani responses to lingual taste stimuli were recorded before and after single bolus injections of one of two tested 5-HT1A receptor antagonists, WAY 100635 (10, 25, 100, or 200 µg/kg BW) or NAD-299 (175, 350 or 750 µg/kg BW).
Intravascular injections of WAY-100635 significantly reduced both phasic and tonic portions of the CT responses to all four tested taste stimuli. (Figure 7, upper left). Maximal drug effects were observed within the first ten minutes post-injection followed by a gradual recovery in response magnitude. On average, maximum recovery occurred 45 and 65 minutes post-injection. All tested concentrations of WAY-100635 inhibited neural responses. At 100 µg/kg (n = 8), CT responses were inhibited by 31% for NaCl (p = 0.009), 50% for sucrose (p = 0.025), 46% for quinine (p = 0.002), and 36% for HCl (p = 0.005; Figure 7, histogram, upper right). Each post-injection inhibition was significant compared to saline controls (p<0.05).
Sample responses are illustrated prior to (PRE), immediately following (AFTER), and 45 to 60 minutes following (RECOVERY) a single bolus intravenous injection of 100 µM WAY100635 (upper left). Summarized data are presented in the histogram (upper right) for these four stimuli. In the lower figures, summarized data for four concentrations of WAY100635 are presented for each tastant and compared to vehicle (saline) injection.
To confirm the above findings, a second 5-HT1A antagonist, NAD-299, was tested. Similar to WAY-100635, NAD-299 (175, 350, and 700 µg/kg) produced inhibition in neural responses to all gustatory stimuli (Figure 8, upper left, 700 µg/kg NAD-299). At each concentration, NAD-299 produced significant inhibitions of responses to all taste qualities with one exception (NaCl response amplitudes were reduced, but not significantly, post-NAD-299 175 µg/kg injection; Figure 8, lower graphs). Recovery from inhibition peaked between 40 and 65 minutes post-injection, similar to WAY-100635 treatment. Limited recovery was observed with NaCl responses following 175 µg/kg dosage, and NaCl, sucrose, and quinine responses following the 350 µg/kg dosage.
Summarized data of 700 µg/kg NAD-299 are presented and concentration data for each tastant are presented to three concentrations of NAD-299 as well as vehicle injection. Unlike the 5-HT1A antagonists tested, the 5-HT3 antagonist ondansetron was without noticeable effects on the magnitude of the chorda tympani neural responses (upper right).
Injecting NAD-299 at 175 µg/kg (n = 5) significantly reduced the CT response to sucrose by 68% (p = 0.018), to quinine by 60% (p = 0.004), and to HCl by 37% (p = 0.008). At 350 µg/kg NAD-299 (n = 6), significant reductions were observed in the CT response to NaCl (27%; p = 0.014), sucrose (35%; p = 0.025), quinine (34%; p = 0.023), and to HCl (37%; p = 0.034). At 750 µg/kg NAD-299 (n = 6), inhibition was at 56% for NaCl (p = 0.001), 59% for sucrose (p = 0.001), 49% for quinine (p = 0.002), and 33% for HCl (p = 0.05).
Correlation between drug dose and the degree of inhibition was tested using Pearson’s correlation test. There was a statistically significant correlation between WAY-100635 dosage and neural inhibitions for sucrose (r = 0.987; p = 0.006), but not for NaCl (r = 0.803 p = 0.099), quinine (r = 0.536; p = 0.232), or HCl (r = 0.738; p = 0.131). Analysis of dose-effect relationship for NAD-299 revealed significant correlation between dosage and neural inhibitions for NaCl (r = 0.999; p = 0.015), but not for sucrose (r = −264; p = 0.415), quinine (r = −4.21; p = 0.362), or HCl (r = −0.866; p = 0.167). Although the amount of inhibition caused by either WAY-100635 or NAD-299 did not always statistically correlate to dosage, a clear pattern of significant inhibitions, followed by a recovery of neural responses, was observed after each administration.
These data suggest that 5-HT1A-mediated serotonergic activity is involved in the processing of sweet, sour, bitter and salt tastes and that, since antagonism of the 5-HT1A receptor produced consistent neural inhibition to all taste qualities, the net effect of this serotonergic activation facilitates neural responses.
Effect of 5-HT3 antagonism on CT responses to gustatory stimulation.
To examine a possible role for the 5-HT3 receptor, CT responses to gustatory stimuli were measured before and after single bolus injection of a 1000 µg/kg ondansetron, a 5-HT3 receptor antagonist. Data averaged from 10 separate experiments showed that no significant change in mean CT response amplitudes to any taste qualities up to 65 minutes post ondansetron injection. Paired-samples T-test for pre-injection and post-injection mean response amplitudes resulted in p-values greater than 0.05 (p = 0.295, 0.368, 0.396 and 0.081 for NaCl, sucrose, quinine and HCl respectively; Figure 8, top right). The only observed significant decline occurred to sucrose responses between 55 and 65 minutes post-injection. Given that this decline occurred long after drug injection, it is more likely it arose from other non-specific factors. Overall, no major modulatory role for the 5-HT3 receptor in neural processing of the tested gustatory stimuli was observed.
Taken together, findings from these nerve recording experiments demonstrate that: (1) systemically injected reagents are able to reach lingual tissue and target their corresponding receptors; (2) ATP is the main neurotransmitter of the taste bud and blocking its P2 receptors significantly reduces the bud’s neural output to the chorda tympani; (3) serotonin seems to act as a universal facilitator to all taste qualities; (4) this action of 5-HT appears to be mediated through its 5-HT1A receptor subtype. In summary, serotonin’s 5-HT1A receptors are expressed in a subset of type II TRCs that include sweet-, bitter-, and umami- processing cells. The physiological action of serotonin is to facilitate chorda tympani responses to sweet, bitter, salty, and sour taste stimuli through activation of 5-HT1A receptors in type II TRCs.
Collectively, data in this study are the first to provide strong evidence for the functional role of the neurotransmitter serotonin in the peripheral gustatory system. We propose serotonergic transmission within the taste bud modulates taste signals at the TRC level during late transduction processes initiated by tastant stimulation. The expression pattern of serotonin and the 5-HT1A receptor suggest a functional circuitry with serotonin release from type III cells acting on the 5-HT1A receptor on a subset of type II cells to modulate ATP release preventing purinergic receptor desensitization on afferent fibers, thus enhancing the neural signal. In support of this hypothesis are anatomical data outlining a negative feedback circuitry between type III and type II cells and physiological data demonstrating functional evidence based on peripheral nerve responses. We suggest that activation of the 5-HT1A receptor may regulate ATP release by modulating activity of the ATP-releasing channel, potentially through PIP2 resynthesis. We have observed, using patch clamp analysis, that serotonin slows the restoration of activity to potassium channels after adaptation by inhibiting PIP2 resynthesis . PIP2 resynthesis is required to restore activity to many classes of ion channels, including connexion and pannexin channels. Hence, activation of the 5HT1A receptor in TRCs may modulate ATP release by slowing the time course of the restoration of functional activity to ATP-release channels, thus reducing desensitization at purinergic receptors and in turn maintaining the gustatory neural signal to the brain.
In addition to the 5-HT1A receptor, the taste bud expresses the ionotropic 5-HT3 receptor. At present, it is thought that this serotonergic receptor subtype is expressed on afferent gustatory nerve fiber terminals within the taste bud, rather than on TRCs . The role of the 5-HT3 receptor in processing of peripheral gustatory information remains unknown and manipulation of the 5-HT3 receptor using an electrophysiological assay failed to produce a measurable functional modulation. While it had long been assumed that serotonin’s main action in the taste bud was as a neurotransmitter to the afferent nerve fiver, serotonin’s main action within the taste bud may be to act functionally as a gain modulator of the afferent neural output, acting primarily on TRCs within the bud via 5-HT1A receptors rather than on gustatory afferent nerve fibers, the presumed location of 5-HT3 receptors in the taste bud.
Phenotypic expression patterns of the 5-HT1A receptor in taste buds suggest it acts on cells which release ATP
Although 5-HT1A is known to be expressed in the taste buds of several species, its functional role remains uncertain. 5-HT1A was first discovered pharmacologically in the rat taste bud  followed by immunocytochemical and mRNA expression data . Functional studies have been limited to patch clamp analysis of 5-HT1A expressing TRCs. In these studies, an inhibition of several ion currents have been reported including calcium-activated potassium current, sodium current, and calcium current , , . Knowledge of 5-HT1A activation at the single cell level only marginally predicts its function in peripheral gustatory physiology since the taste bud is a complex processing unit with multiple levels of cell-to-cell communication . Knowing the cell type in which the 5-HT1A receptor is expressed would be useful next step. Although serotonin-releasing cells are well known to be type III cells, and the 5-HT1A expressing cell is known only to be paracrine to type III cells. Our earlier work suggested these cells to be type II cells . Since type II cells release ATP , , , inhibition of these cells by serotonin would be functionally significant.
Our data provide strong evidence that the 5-HT1A receptor is expressed in type II cells in the rat taste bud. General morphology, co-expression with multiple type II markers, and exclusion from a type III marker are all consistent with type II cell expression. Combined data from double-label ICC from anterior and posterior tongue, and single cell RT-PCR from posterior tongue, showed that the 5-HT1A receptor is consistently co-expressed with phenotypic markers associated with type II cells (GUST, GAD, NPY, NPY-1R, CCK-8, SNAP, and GAD), including receptors for sweet and umami (T1R), and bitter (T2R) taste qualities. Further, immunocytochemical and single cell data showed that 5-HT1A does not co-express with PKD2L1 and NCAM, phenotypic markers linked to type III cells. The association of NCAM with markers associated with type III cells has been well documented in rodent taste buds , . As well, PKD2L1, a member of the transient receptor potential channel (TRP) family and candidate sour taste receptor, is expressed exclusively in a subset of type III cells . Finally, without testing a type I marker, expression of the 5-HT1A receptor in the type I cell remains an open question. However, expression in type I cells may be unlikely. The morphology of 5-HT1A immunopositive cells, while not a definitive criteria by itself, is inconsistent with type I cell expression. Characteristically, 5-HT1A immunopositive cells display large round soma which is characteristic of type II cells. At present type I expression cannot be excluded.
Collectively, 5-HT1A phenotyping experiments suggest that this receptor is not expressed in all type II cells. For example, 5-HT1A was not expressed in all gustducin-positive cells or in all T2R-positive cells. Gustducin is thought to be expressed in all cells that process bitter taste , and all gustducin-positive cells have the ability to release ATP . Therefore, the incomplete pattern of co-expression between gustducin/T2R and 5-HT1A observed in the present study suggests that not all bitter-transducing cells are directly targeted by serotonergic modulation. Hence, subsets of bitter-transducing TRCs may contribute differentially to the neural response; i.e., not all TRCs which express bitter transduction machinery may equally activate the afferent nerve. Bitter-sensitive TRCs may functionally differentiate based on a number of parameters such as cell age, subset of T2R molecules expressed, or proximity to an afferent terminal. Additionally, TRCs may only express a full repertoire of required molecular transduction units at a defined window during cell lifespan. Whether all TRCs expressing T2R molecules contribute equally or differentially to afferent nerve activation is, at present, an open question. The possibility that serotoninergic modulation operates on only a cohort of bitter receptor cells which in turn contribute differentially to the neural response remains an interesting, though unproven, hypothesis. If true, the subset of T2R-expressing TRCs that additionally express 5-HT1A would represent a functionally distinct group of cells in the bud capable of strong influence over the afferent neural response.
Along this line, the co-expression pattern of the 5-HT1A receptor with neuropeptides and neuropeptide receptors, all known to be expressed in type II cells, suggests that serotonin may participate in the complex network of neuropeptide signaling within the taste bud. Within the taste bud, cholecystokinin, the CCK-1 receptor, NPY, and vasoactive intestinal polypeptide are predominantly co-expressed in a subset of type II cells which are paracrine to another subset of type II cells expressing the NPY-1 receptor. CCK and NPY, which are co-expressed, act in an autocrine stimulatory manner and a paracrine inhibitory manner, respectively . In the present study, immunocytochemical results revealed that 73% of the TRCs expressing NPY and 66% of those expressing NPY-1R co-express 5-HT1A, a pattern consistent throughout taste buds in posterior tongue, anterior tongue, and the nasoincisor ducts. These findings imply that in late transduction mechanisms, 5-HT1A-expressing cells themselves may be subject to complex modulation, such as by the neuropeptide signaling system, and hence there may be multiple layers of modulation influencing ATP release within the taste bud.
A functional role for serotonin in the taste bud
Electrophysiological data demonstrate that chorda tympani neural responses to tastant stimuli are significantly diminished by administration of a 5-HT1A antagonist. Responses were reduced in magnitude with little effect on temporal properties, i.e. phasic and tonic responses. These observations are consistent with reduced neurotransmitter release from the TRC without alteration of other mechanisms, such as adaptation. Positive control experiments demonstrated that intravenous agents reach the taste bud as evidenced by injection of ATP and by strong inhibition of all taste responses to the ATP antagonist PPADS. Inhibition was observed with two 5-HT1A antagonists and was not replicated by either vehicle injection or by a 5-HT3 antagonist. Further, our data are in lines with those who have previously used this technique . We believe that the combination of three positive control experiments, one negative control experiments, and the repeatability of two 5-HT1A have produced a strong and reliable electrophysiological data set.
In total, these observations are consistent with the notion that the 5-HT1A receptor plays a facilitating role in taste transduction mechanisms of multiple taste qualities. Because serotonin’s action at the cellular level is inhibitory, serotonin acts during transduction by a process similar to disinhibition. We suggest the site of this disinhibition is regulation of ATP release preventing purinergic receptors desensitization. As a neurotransmitter in the taste bud, ATP activates P2X2/P2X3 purinoceptors on afferent nerve endings to transmit gustatory signals . P2X3 channels display high rates of desensitization , , . By regulating ATP release through its release channels, such desensitization would be prevented, thus facilitating the neural response. Additionally ATP has neuromodulatory roles within the taste bud including both positive autocrine and paracrine feedback mechanisms , . ATP may activate P2Y1 and P2X2 receptors on type II cells ,  via autocrine routes, thereby auto-enhancing and triggering its own release respectively. ATP may also activate P2Y receptors expressed on type III cells in a paracrine manner, including serotonergic cells, ultimately triggering serotonin release .
Serotonin can also be directly released from type III cells upon their stimulation with sour stimuli . Serotonin is released from TRCs in response to multiple tastant qualities in mouse taste buds , which agrees with our electrophysiological results of serotonin’s influence on multiple taste qualities. Subsequently, serotonin activates 5-HT1A receptors expressed in a group of type II cells through a negative paracrine feedback, inhibiting those cells and the release of ATP. Supporting this notion is the observation that in mice taste-evoked ATP release from type II cells was inhibited by serotonin and enhanced by blocking 5-HT1A with WAY-100635 . Recent quantitative studies of rodent taste buds have documented contain that about 14% of taste receptor cells in a bud are serotonergic; hence, a typical bud of 50 to 100 cells would contain an average of 7 to 13 cells serotonin cells . Hence, it is easy to envision that serotonin could exert significant and measurable modulation of the neural response, particularly with the close apposition of type II cells with a single type III cell. Therefore, blocking the 5-HT1A-receptor-mediated serotonergic transmission in the taste bud would be logically assumed to disinhibit ATP release, and thus enhance the stimulus for synaptic P2X purinoceptors.
The identity of the ATP release channel is not known with certainty. Purinergic neurotransmission in the taste bud does not require traditional synapses, as found on type III cells , , . Instead release is nonvesicular through channels whose candidates include connexins, pannexin-1, and a newly discovered release channel, CALHM1 , , . As CALHM1 knockout mice do not have impaired responses for all taste qualities, multiple ATP-release mechanisms may exist. Whether activation of the 5HT1A receptor targets some or all of these channels is not known. Several studies have shown that ion channels , , , , ,  including connexin and pannexin hemichannels ,  are regulated by PIP2 levels. We are unaware if the sensitivity of CALHM1 to phospholipids has yet been tested. Data from preliminary patch-clamping work have also demonstrated that pharmacologically altering PIP2 re-synthesis modulates the excitability of TRCs . The cellular inhibition caused by the activation of 5-HT1A receptors reduces PIP2 resynthesis , ,  and, consequentially, would inhibit the activity of connexin/pannexin channels, resulting in a reduction in ATP release.
Finally, if the 5HT1A receptor is involved in modulating the peripheral neural response to taste stimuli, one might predict that alterations of the serotonergic system might result in gustatory perceptual changes. For example, in humans, circumstances in which serotonin is altered, such as depression, are associated with taste disturbances. Along this line, there are a few relevant examples of experiments in human and rodent models with the selective serotonin reuptake inhibitor (SSRI) paroxetine. Taste thresholds to sweet and bitter in healthy humans with paroxetine were reduced . These observations are consistent with increased serotonin concentration in the taste bud which might result in greater taste acuity reflected as either lower thresholds and/or a shift in perceptual intensity. In rodents, behavioral experiments with paroxetine in brief access test and operant condition paradigms have been conducted. In a brief access taste test, paroxetine did not affect concentration-dependent licking to taste stimuli but produce a decrease in the number of trials that water-deprived rats initiated when quinine was the stimulus and that non-deprived rats initiated when sucrose was the stimulus . Systemically increased serotonin did not affect consummatory behavior but, at least in some conditions, did attenuate appetitive behavior. The effects observed in the present paper would agree with the lack of effect on consummatory behavior since 5-HT1A receptor activation influences modulation rather than detection. In an operant conditioning task, water-deprived rats had lower breakpoints for water and quinine when injected with paroxetine compared with vehicle . The authors conclude that a general systemic increase in serotonergic activity decreases the reward value of both preferred and avoided fluids under different motivational states and suggest these findings to be consistent with those from the brief access taste test. For the most part, these studies are complicated by the inability to delineate a primary site of action of serotonin since such treatments affect both central and peripheral serotonergic systems. Nevertheless, the results display an internal consistency across animal and human models that altering serotonin levels in the taste bud may influence peripheral taste detection and that this alteration may correlate higher serotonin levels with overall higher taste acuity.
In summary, the present study strongly supports the notion that serotonin is a crucial player in a finely tuned feedback mechanism involving 5-HT1A receptors, ATP, and the P2X2/P2X3 purinoceptors. Data suggest that serotonin’s primary role in the mammalian taste bud is to modulate taste signals prior to the final neural output and this modulation likely occurs by controlling the release of the gustatory neurotransmitter ATP from 5HT1A-expressing type II cells. Serotonin thus plays a role in the complex set of late gustatory transduction mechanisms that take place at the cellular level within the taste bud during the processing of taste information.
Conceived and designed the experiments: LJ FlZ TK SH. Performed the experiments: LJ FlZ TK. Analyzed the data: LJ FlZ TK SH. Contributed reagents/materials/analysis tools: LJ FlZ TK SH. Contributed to the writing of the manuscript: LJ SH.
- 1. Adler E, Hoon MA, Mueller KL, Chandrashekar J, Ryba NJ, et al. (2000) A novel family of mammalian taste receptors. Cell 100: 693–702.
- 2. Antunes VR, Braga VA, Machado BH (2005) Autonomic and respiratory responses to microinjection of ATP into the intermediate or caudal nucleus tractus solitarius in the working heart-brainstem preparation of the rat. Clin Exp Pharmacol Physiol 32(5–6): 467–72.
- 3. Baryshnikov SG, Rogachevskaja OA, Kolesnikov SS (2003) Calcium signaling mediated by P2Y receptors in mouse taste cells. J Neurophys 90: 3283–3294.
- 4. Beidler LM (1971) Taste receptor stimulation with salts and acids. In: Beidler LM, editor. Handbook of sensory physiology. New York: Springer-Verlag. 200–220.
- 5. Bo X, Alavi A, Xiang Z, Oglesby I, Ford A, et al. (1999) Localization of ATP-gated P2X2 and P2X3 receptor immunoreactive nerves in rat taste buds. Neuroreport 10: 1107–1111.
- 6. Borroto-Escuela DO, Romero-Fernandez W, Pérez-Alea M, Narvaez M, Tarakanov AO, et al. (2012) The existence of FGFR1-5-HT1A receptor heterocomplexes in midbrain 5-HT neurons of the rat: relevance for neuroplasticity. J Neurosci 32: 6295–303.
- 7. Bystrova MF, Yatzenko YE, Fedorov IV, Rogachevskaja OA, Kolesnikov SS (2006) P2Y isoforms operative in mouse taste cells. Cell Tissue Res 323: 377–382.
- 8. Cao Y, Zhao FL, Kolli T, Hivley R, Herness S (2009) GABA expression in the mammalian taste bud functions as a route of inhibitory cell-to-cell communication. PNAS 106: 4006–4011.
- 9. Cameron SJ, Malik S, Akaike M, Lerner-Marmarosh N, Yan C, et al. (2003) Regulation of epidermal growth factor-induced connexin 43 gap junction communication by big mitogen-activated protein kinase1/ERK5 but not ERK1/2 kinase activation. J Biol Chem 278(20): 18682–18688.
- 10. Chen K, Yan J, Suo Y, Li J, Wang Q, et al. (2010) Nutritional status alters saccharin intake and sweet receptor mRNA expression in rat taste buds. Brain Res 14(1325): 53–62.
- 11. Cook SP, McCleskey EW (1997) Desensitization, recovery and Ca2+-dependent modulation of ATP-gated P2X receptors in nociceptors. Neuropharmacology 36: 1303–1309.
- 12. DeFazio RA, Dvoryanchikov G, Maruyama Y, Kim JW, PereiraE, et al (2006) Separate population of receptor cells and presynaptic cells in mouse taste buds. J Neurosci 26: 3971–3980.
- 13. Delay RJ, Kinnamon SC, Roper SD (1997) Serotonin modulates voltage-dependent calcium current in Necturus taste cells. J Neurophysiol 77: 2515–2524.
- 14. Ewald DA, Roper SD (1994) Bidirectional synaptic transmission in Necturus taste buds. J Neurosci 14: 3791–3804.
- 15. Fedorov I, Rogachevskaja OA, Kolesnikov SS (2007) Modeling P2Y receptor-Ca2+ response coupling in taste cells. Biochim Biophys Acta 1768: 1727–1740.
- 16. Finger TE, Danilova V, Barrows J, Bartel DL, Vigers AJ, et al. (2005) ATP signaling is crucial for communication from taste buds to gustatory nerves. Science 310: 1495–1499.
- 17. Florea L, Di Francesco V, Miller J, Turner R, Yao A, et al. (2005) Gene and alternative splicing annotation with AIR. Genome Res 15(1): 54–66.
- 18. Fujimoto S, Ueda H, Kagawa H (1987) Immunocytochemistry on the localization of 5-hyroxytryptophan in monkey and rabbit taste buds. Acta Anat (Basel) 128: 80–83.
- 19. Getchell TV (1997) Communication within the bud. NeuroReport 8: iii.
- 20. Hayato R, Ohtubo Y, Yoshii K (2007) Functional expression of ionotropic purinergic receptors on mouse taste bud cells. J Physiol 584: 473–488.
- 21. He ML, Koshimizu TA, Tomić M, Stojilkovic SS (2002) Purinergic P2X2 receptor desensitization depends on coupling between ectodomain and c-terminal domain. Mol Pharmacol 62: 1187–1197.
- 22. Heath TP, Melichar JK, Nutt DJ, Donaldson LF (2006) Human taste thresholds are modulated by serotonin and noradrenaline. J Neurosci 26: 12664–12671.
- 23. Herness MS (1985) Neurophysiological and Biophysical Evidence on the Mechanism of Electric Taste. J Gen Physiol 86: 59–87.
- 24. Herness S (2012) The Neurobiology of Gustation: Taste buds and transduction processes. In: Johnson L, editor. Physiology of the Gastrointestinal Tract, 5th edition. Oxford: Elsevier Press. 1(26): 741–768.
- 25. Herness MS, Chen Y (1997) Serotonin inhibits calcium-activated K+ current in rat taste receptor cells. NeuroReport 8: 3257–3261.
- 26. Herness MS, Chen Y (2000) Serotonergic agonists inhibit calcium-activated potassium and voltage-dependent sodium currents in rat taste receptor cells. J Membr Biol 173: 127–138.
- 27. Hilgemann DW, Feng S, Nasuhoglu C (2001) The complex and intriguing lives of PIP2 with ion channels and transporters. Science STKE re19.
- 28. Hoon MA, Adler E, Lindemeier J, Battey JF, Ryba NJ, et al. (1999) Putative mammalian taste receptors: a class of taste-specific gpcrs with distinct topographic selectivity. Cell 96: 541–551.
- 29. Hoon MA, Northup JK, Margolskee RF, Ryba NJ (1995) Functional expression of the taste specific G-protein, alpha-gustducin. Biochem J 309(Pt 2): 629–636.
- 30. Huang AL, Chen X, Hoon MA, Chandrashekar J, Guo W, et al. (2006) The cells and logic for mammalian sour taste detection. Nature 442: 934–938.
- 31. Huang YA, Dando R, Roper SD (2009) Autocrine and paracrine roles for ATP and serotonin in mouse taste buds. J Neurosci 29: 13909–13918.
- 32. Huang Y-J, Maruyama Y, Dvoryanchikov G, Pereira E, Chaudhari N, et al. (2007) The role of pannexin 1 hemichannels in ATP release and cell-cell communication in mouse taste buds. PNAS 104: 6436–6441.
- 33. Huang YJ, Maruyama Y, Lu KS, Pereira E, Plonsky I, et al. (2005) Mouse taste buds use serotonin as a neurotransmitter. J Neurosci 25: 843–847.
- 34. Imendra KG, Fujiyama R, Miyamoto T, Okada Y, Sato T (2000) Serotonin inhibits voltage-gated sodium current by cyclic adenosine monophosphate-dependent mechanism in bullfrog taste receptor cells. Neurosci Lett 294: 151–154.
- 35. Imendra KG, Miyamoto T, Okada Y, Toda K (2002) Serotonin differentially modulates the electrical properties of different subsets of taste receptor cells in bullfrog. Eur J Neurosci 16: 629–640.
- 36. Ishikawa M, Iwamoto T, Nakamura T, Doyle A, Fukumoto S, et al. (2011) Pannexin 3 functions as an ER Ca2+ channel, hemichannel, and gap junction to promote osteoblast differentiation. JCB 193: 1257–1274.
- 37. Kakinohana O, Hefferan MP, Miyanohara A, Nejime T, Marsala S, et al. (2012) Combinational spinal GAD65 gene delivery and systemic GABA-mimetic treatment for modulation of spasticity. PLoS One 7(1): e30561.
- 38. Kataoka S, Toyono T, Seta Y, Ogura T, Toyoshima K (2004) Expression of P2Y1 receptors in rat taste buds. Histochem Cell Biol 121: 419–426.
- 39. Kaya N, Shen T, Lu SG, Zhao FL, Herness S (2004) Expression and localization of serotonin receptor subtypes in rat taste buds. Am J Physiol 286: R649–R658.
- 40. Kinnamon SC, Finger TE (2013) A taste for ATP: neurotransmission in taste buds. Frontiers in Cellular Neuroscience 7: 264.
- 41. Kim DJ, Roper SD (1995) Localization of serotonin in taste buds: A comparative study in four vertebrates. J Comp Neurol 353: 364–370.
- 42. Kishi M, Emori Y, Tsukamoto Y, Abe K (2001) Primary culture of rat taste bud cells that retain molecular markers for taste buds and permit functional expression of foreign genes. Neuroscience 106: 217–225.
- 43. Kokrashvili Z, Mosinger B, Margolskee RF (2009) T1r3 and alpha-gustducin in gut regulate secretion of glucagon-like peptide-1. Ann N Y Acad Sci 1170: 91–94.
- 44. Liu D, Liman ER (2003) Intracellular Ca2+ and the phospholipid PIP2 regulate the taste transduction ion channel TRPM5. Proc Natl Acad Sci 100: 15160–15165.
- 45. Ma H, Yang R, Thomas SM, Kinnamon JC (2007) Qualitative and quantitative differences between taste buds of the rat and mouse BMC Neuroscience. 8: 5.
- 46. Max M, Shanker YG, Huang L, Rong M, Liu Z, et al. (2001) TaslRS, encoding a new candidate taste receptor, is allelic to the sweet responsiveness locus Sac. Nat Genet 28: 58–63.
- 47. Margolskee RF, Dyer J, Kokrashvili Z, Salmon KS, Ilegems E, et al. (2007) T1R3 and gustducin in gut sense sugars to regulate expression of Na+-glucose cotransporter 1. Proc Nat Ac Sc 104(38): 15075–15080.
- 48. Mathes CM, Gregson JR, Spector AC (2013) The selective serotonin reuptake inhibitor paroxetine decreases breakpoint of rats engaging in a progressive ratio licking task for sucrose and quinine solutions. Chem Senses 38: 211–220.
- 49. Mathes CM, Spector AC (2011) The selective serotonin reuptake inhibitor paroxetine does not alter consummatory concentration-dependent licking of prototypical taste stimuli by rats. Chem Senses 36: 515–526.
- 50. McLaughlin SK, McKinnon PJ, Margolskee RF (2002) Gustducin is a teast-cell-specific G protein closely related to the transducins. Nature 357: 563–569.
- 51. Minisclou C, Benavides J, Claustre Y (1995) Activation of 5-HT1A receptors inhibits carbachol-stimulated inositol 1,4,5-trisphosphate mass accumulation in the rodent hippocampus. Neurochem Res 20(9): 977–983.
- 52. Montmayeur JP, Matsunami H (2002) Receptors for bitter and sweet taste. Curr Opin Neurobiol 12: 1–8.
- 53. Murata Y, Yasuo T, Yoshida R, Obata K, Yanagawa Y, et al. (2010) Action potential-enhanced ATP release from taste cells through hemichannels. J Neurophysiol 104(2): 896–901.
- 54. Nada O, Hirata K (1975) The occurrence of the cell type containing a specific monoamine in the taste bud of the rabbit’s foliate papilla. Histochem 43: 237–240.
- 55. Ni YG, Panicker MM, Miledi R (1997) Efficient coupling of 5-HT1A receptors to the phospholipase C pathway in Xenopus oocytes. Brain Res Mol Brain Res 51(1–2): 115–22.
- 56. Obata H, Shimada K, Sakai N, Saito N (1997) GABAergic neurotransmission in rat taste buds: immunocytochemical study for GABA and GABA transporter subtypes. Brain Res Mol Brain Res 49(1–2): 29–36.
- 57. Ohmoto M, Matsumoto I, Yasuoka A, Yoshihara Y, Abe K (2008) Genetic tracing of the gustatory and trigeminal neural pathways originating from T1R3-expressing taste receptor cells and solitary chemoreceptor cells. Mol Cell Neurosci 38(4): 505–17.
- 58. Ohmoto M, Okada S, Nakamura S, Abe K, Matsumoto I (2011) Mutally exclusive expression of Gαia and Gα14 reveals diversification of taste receptor cells in zebrafish. J Comp Neurol 519: 1616–1629.
- 59. Raymond JR, Mukhin YV, Gettys TW, Garnovskaya MN (1999) The recombinant 5-HT1A receptor: G protein coupling and signalling pathways. Br J Pharmacol 127(8): 1751–64.
- 60. Ren Y, Shimada K, Shirai Y, Fujimiya M, Saito N (1999) Immunocytochemical localization of serotonin and serotonin transporter (SET) in taste buds of rat. Molec Brain Res 74: 221–224.
- 61. Rettinger J, Schmalzing G (2004) Desensitization masks nanomolar potency of ATP for the P2X1 receptor. J Biol Chem 279: 6426–6433.
- 62. Romanov RA, Rogachevskaja OA, Bystrova MF, Jiang P, Margolskee RF, et al. (2007) Afferent neurotransmission mediated by hemichannels in mammalian taste cells. EMBO Journal 26: 657–667.
- 63. Romanov RA, Rogachevskaja OA, Khokhlov AA, Kolesnikov SS (2008) Voltage dependence of ATP secretion in mammalian taste cells. J Gen Physiol 132(6): 731–44.
- 64. Saez JC, Berthoud VM, Branes MC, Martinez AD, Beyer EC (2003) Plasma membrane channels formed by connexins: their regulation and functions. Physiol Rev 83: 1359–1400.
- 65. Sato T, Beidler LM (1983) Dependence of gustatory neural response on depolarizing and hyperpolarizing receptor potentials of taste cells in the rat. Comp Biochem Physiol A Comp Physiol 75(2): 131–7.
- 66. Shindler KS, Roth KA (1996) Double immunofluorescent staining using two unconjugated primary antisera raised in the same species. J Histochem Cytochem 44: 1331–1335.
- 67. Shingai T, Beidler LM (1985) Response characteristics of three taste nerves in mice. Brain Res 335(2): 245–9.
- 68. Simon SA, Liu L, Erickson RP (2003) Neuropeptides modulate rat chorda tympani responses Am J Physiol Regul Integr Comp Physiol 284: R1494–R1505, 2003.
- 69. Smith DV, Akeson RA, Shipley MT (1993) NCAM expression by subsets of taste cells is dependent upon innervation. J Comp Neurol 336(4): 493–506.
- 70. Suh BC, Hille B (2005) Regulation of ion channels by phosphatidylinositol 4,5-bisphosphate. Curr Opin Neurobiol 15: 307–378.
- 71. Suh BC, Hille B (2008) PIP2 is a necessary cofactor for ion channel function: How and why? Annu Rev Biophys 37: 175–195.
- 72. Taruno A, Vingtdeux V, Ohmoto M, Ma Z, Dvoryanchikov G, et al. (2013) CALHM1 ion channel mediates purinergic neurotransmission of sweet, bitter and umami tastes. Nature 495: 223–226.
- 73. Thomas T, Marshall JM (1993) The role of adenosine in hypoxic pulmonary vasoconstriction in the anaesthetized rat. Exp Physiol 78(4): 541–543.
- 74. Uchida T (1985) Serotonin-like immunoreactivity in the taste bud of the mouse circumvallate papilla. Jpn J Oral Biol 27: 509–521.
- 75. Yee CL, Yang R, Bottger B, Finger TE, Kinnamon JC (2001) “Type III” cells of rat taste buds: Immunohistochemical and ultrastructural studies of neuron-specific enolase, protein gene product 9.5, and serotonin. J Comp Neurol 440: 97–108.
- 76. Uchida S, Hara K, Kobayashi A, Funato H, Hobara T, et al. (2010) Early life stress enhances behavioral vulnerability to stress through the activation of REST4-mediated gene transcription in the medial prefrontal cortex of rodents. J Neurosci 30: 15007–15018.
- 77. Ueda K, Ichimori Y, Okada H, Honma S, Wakisaka S (2006) Immunolocalization of SNARE proteins in both type II and type III cells of rat taste buds. Arch Histol Cytol 69(4): 289–296.
- 78. Zhang Y, Hoon MA, Chandrashekar J, Mueller KL, Cook B, et al. (2003) Coding sweet, bitter, and umami tastes: different receptor cells sharing signaling pathways. Cell 112(3): 293–301.
- 79. Zhao FL, Herness S (2009) Resynthesis of phosphatidylinositol 4,5-bisphosphate mediates adaptation of the caffeine response in rat taste receptor cells. J Physiol 587(2): 363–377.
- 80. Zhao FL, Shen T, Kaya N, Lu SG, Cao Y, et al. (2005) Expression, physiological action, and co-expression patterns of NPY in rat taste bud cells. PNAS 102: 11100–11105.
- 81. Zhao GQ, Zhang Y, Hoon MA, Chandrashekar J, Erlenbach I, et al. (2003) The receptors for mammalian sweet and umami taste. Cell 115(3): 255–266.