Synapses between gustatory receptor cells and primary sensory afferent fibers transmit the output signal from taste buds to the CNS. Several transmitter candidates have been proposed for these synapses, including serotonin (5-HT), glutamate, acetylcholine, ATP, peptides, and others, but, to date, none has been unambiguously identified. We used Chinese hamster ovary cells stably expressing 5-HT2C receptors as biodetectors to monitor 5-HT release from taste buds. When taste buds were depolarized with KCl or stimulated with bitter, sweet, or sour (acid) tastants, serotonin was released. KCl- and acid-induced 5-HT release, but not release attributable to sweet or bitter stimulation, required Ca2+ influx. In contrast, 5-HT release evoked by sweet and bitter stimulation seemed to be triggered by intracellular Ca2+ release. These experiments strongly implicate serotonin as a taste bud neurotransmitter and reveal unexpected transmitter release mechanisms.
It has long been known that serotonin stimulates gustatory sensory afferent fibers when that biogenic amine is injected into the tongue near taste buds (Esakov et al., 1983), suggesting that 5-HT may be a taste bud neurotransmitter (for review, see Nagai et al., 1996). A number of findings have subsequently supported that notion. For example, serotonin is present in a subset of taste cells in a wide variety of species (Whitear, 1989; Delay et al., 1993; Kim and Roper, 1995). In mammals, many of the taste cells that synapse with nerve fibers are serotonergic: they take up the serotonin precursor 5-hydroxytryptophan (5-HTP) and are immunopositive for 5-HT (Takeda, 1977; Kim and Roper, 1995; Yee et al., 2001). Reverse transcription-PCR data indicate that 5-HT1A and 5-HT3 receptors are expressed in taste tissue, with 5-HT1A receptors found in taste cells and 5-HT3 receptors found in sensory primary afferent fibers (Kaya et al., 2004). Finally, there is indirect evidence from autoradiographic studies showing that, when depolarized, amphibian taste cells release serotonin (Nagai et al., 1998). However, despite these findings, one of the most important of the canonical criteria for identifying synaptic neurotransmitters, namely detecting its release from stimulated synapses, has yet to be unambiguously established in taste buds for serotonin or for any other candidate. We used novel biosensor cells to detect the secretion of transmitter candidates from taste buds and report that taste stimulation elicits serotonin release.
Parts of this work have been published previously in abstract form (Huang et al., 2005).
Materials and Methods
Biosensor cells. Chinese hamster ovary (CHO) cells expressing 5-HT2C receptors (Berg et al., 1994) were seeded into 35-mm-diameter culture dishes. To produce biosensors, cells were suspended in HBSS containing 0.125% trypsin and collected in a 15 ml centrifuge tube after terminating the reaction with 2% FBS. Dispersed CHO cells were loaded with 4 μm fura-2 AM for 1 h at room temperature. An aliquot of fura-2-loaded cells was transferred to a recording chamber and viewed with an Olympus Optical (Tokyo, Japan) IX70 inverted microscope to test their responses to bath-applied 5-HT, KCl, cycloheximide, acetic acid, and saccharin. Sequential fluorescence microscopic images were recorded with a longpass emission filter (≥510 nm) with the cells excited at 340 nm, followed by 380 nm, and the ratios were calculated with Imaging Workbench version 5 software (INDEC Biosystems, Mountain View, CA) (data shown are these ratios, labeled F340/F380 in Figures 1, 2, 3).
Isolated taste buds. We removed the lingual epithelium containing taste papillae from the tongues of adult C57BL/6J female mice by injecting a mixture of 1 mg/ml collagenase A (Roche Diagnostics, Indianapolis, IN), 2.5 mg/ml dispase II (Roche Diagnostics), and 1 mg/ml trypsin inhibitor (Sigma, St. Louis, MO) directly under the epithelium surrounding taste papillae. The peeled epithelium was bathed in Ca-free solution for 30 min at room temperature, and isolated taste buds were drawn into fire-polished glass micropipettes with gentle suction. Taste buds were transferred to a shallow recording chamber that had a glass coverslip base. The coverslip was coated with Cell-Tak (BD Biosciences, San Jose, CA) to hold the taste buds firmly in place. Taste buds were perfused with Tyrode's solution (in mm): 140 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES, 10 glucose, 10 Na-pyruvate, and 5 NaHCO3, pH 7.4 (310-320 Osm).
Stimulation. Taste buds were stimulated by bath perfusion of KCl (50 mm; substituted equimolar for NaCl), cycloheximide (10-100 μm), sodium saccharin (2-20 mm), aspartame (1-10 mm), and acetic acid (8 mm). All of the stimuli were made up in Tyrode's solution and applied at pH 7.2, except for acetic acid, which was applied at pH 5.
Immunohistochemistry. Isolated taste buds were fixed for 10 min in 4% paraformaldehyde in PBS. Taste buds were then rinsed three times in PBS and incubated for 2 h in PBS containing 0.3% Triton X-100, 2% normal donkey serum, and 2% bovine serum albumin. Taste buds were incubated with rabbit polyclonal anti-serotonin antibodies (1:1000; catalog #S5545; Sigma) for 60-90 min at room temperature. Thereafter, taste buds were washed three times in PBS, incubated for 1 h in Alexa Fluor 594-conjugated donkey anti-rabbit secondary antibodies (1:500; Molecular Probes, Eugene, OR), and then washed again three times in PBS.
We produced biosensor cells for detecting serotonin by using CHO cells transfected with high-affinity 5-HT receptors. CHO-K1 cells stably expressing 5-HT2C receptors (Berg et al., 1994) were suspended in buffer and loaded with the Ca2+- sensitive dye fura-2. Fura-loaded CHO/5-HT2C cells responded robustly to bath-applied serotonin. The threshold for activation was ∼3 nm, and EC50 was 9 nm (Fig. 1A). This compares with the EC50 values of 15 and 29 nm reported previously for Ca2+ mobilization in fura-loaded CHO cells transiently transfected with 5-HT2C receptors (Akiyoshi et al., 1995, 1996). Biosensor cells respond to ATP (endogenous purinoreceptors), as well as to serotonin. Thus, we used mianserin, a 5-HT2 receptor antagonist, to differentiate responses to 5-HT from activation of endogenous (ATP) receptors (see below). Responses to 5-HT were reversibly blocked by 10 nm mianserin, but ATP responses were unaffected (Fig. 1B). Fura-loaded CHO/5-HT2C cells did not respond with Ca2+ signals to carbachol or glutamate up to 1 mm (data not shown). Furthermore, CHO/5-HT2C cells did not respond to depolarization with bath-applied KCl (50 mm; substituted for NaCl) nor to bath-applied taste stimuli-cycloheximide (100 μm), an aversive (bitter) taste compound for rodents, or saccharin (20 mm), a preferred (sweet) taste compound (Fig. 1C). The biosensor cells also did not respond to acetic acid stimulation (in humans, sour; 8 mm, pH 5; data not shown). Last, CHO/5-HT2C cells maintained their responses to bath-applied serotonin even if Ca2+ in the medium was replaced with Mg2+, consistent with the coupling of 5-HT2C receptors to intracellular Ca2+ release mechanisms (Fig. 1C). In short, CHO/5-HT2C cells were sensitive and reliable biosensors for serotonin, serotonergic responses could be differentiated from those produced by activating endogenous purinergic receptors by applying mianserin, and biosensor cells did not respond to taste stimuli or depolarization with KCl.
Subsequently, we isolated taste buds from vallate papillae of the mouse tongue. Serotonergic cells could be observed readily by immunostaining taste buds isolated for recording (Fig. 2A), verifying that their 5-HT content was maintained throughout the isolation procedure. Using fura-2-loaded taste buds, we verified that, as shown previously (Richter et al., 2004), ∼50% of taste bud cells are depolarized by bath-applied KCl (50 mm). Also, as shown previously (Bernhardt et al., 1996), cells in isolated taste buds responded to bath-applied bitter tastants (here, 10-100 μm cycloheximide) and saccharin (2-20 mm). Parenthetically, different taste cells responded to cycloheximide, saccharin, or KCl depolarization, but we did not investigate this systematically. Others have shown in more intact preparations of taste buds that many taste cells are responsive to two or more taste qualities (Gilbertson et al., 2001; Caicedo et al., 2002). Taste buds did not respond to low concentrations of bath-applied serotonin (3-10 nm).
We transferred an aliquot of suspended, fura-2-loaded CHO/5-HT2C cells into a chamber containing freshly isolated mouse taste buds adhering to the glass coverslip base of the chamber. A biosensor cell was drawn onto a fire-polished glass micropipette with gentle suction and maneuvered to an isolated taste bud (Fig. 2B). Mere physical contact between the biosensor cell and the taste bud did not elicit a response, nor did bath perfusion with buffer alone. However, perfusing the chamber with KCl, cycloheximide, saccharin, or acetic acid evoked rapid and repeatable biosensor responses (Figs. 2C-E, 3A-C). When a first attempt to record a response was unsuccessful, repositioning the biosensor sometimes revealed signals evoked by stimuli. To date, we have obtained successful biosensor recordings from 31 of 105 taste buds when testing KCl depolarization. Eight of those 31 taste buds also yielded biosensor responses to stimulation with cycloheximide or saccharin (one of these taste buds produced a biosensor response to cycloheximide and saccharin alike). In a separate series of experiments, 6 of 13 taste buds showed responses to acetic-acid stimulation; these same taste buds also showed biosensor responses to KCl depolarization. Responses were abolished if the biosensor cell was withdrawn a few micrometers from the isolated taste bud and returned when the biosensor was repositioned to the same location. These tests identified the taste bud as the source of the release and suggested that release sites were spatially restricted within the taste bud, consistent with the presence of only a limited number of serotonergic cells (Fig. 2A).
Biosensor responses were generally small but were enhanced significantly by incubating lingual epithelium for 30 min with 500 μm 5-HTP, the immediate precursor to 5-HT, before isolating taste buds (Fig. 2, compare C, D). Furthermore, responses to stimuli were reversibly blocked by 1 nm mianserin (Fig. 2F shows data for KCl depolarization). Because mianserin specifically blocked serotonergic responses (Fig. 1) and because 5-HTP is a precursor for serotonin, these observations verified that the biosensor cell was detecting 5-HT released from taste buds and not responding to other compounds, such as ATP, had they been secreted. We also stimulated isolated mouse taste buds with aspartame, an artificial sweetener for humans but one that is generally not preferred by and does not evoke taste nerve responses in mice (Inoue et al., 2001). Figure 2E shows that aspartame, even at 10 mm, is ineffective, validating further that the biosensor responses are produced by taste stimulation per se.
Last, we tested whether serotonin release was Ca2+ dependent. Magnesium (3 mm) was substituted for Ca2+ (2 mm) in the bath, and taste buds were stimulated with KCl and tastants, as above. In the case of KCl depolarization and acid stimulation, replacing Ca2+ with Mg2+ rapidly and reversibly blocked or significantly reduced serotonin release (Fig. 3A,B), consistent with influx of Ca2+ through depolarization-gated Ca channels and consistent with known synaptic mechanisms. Surprisingly, serotonin release elicited by cycloheximide (Fig. 3A) or saccharin (Fig. 3B) was not affected by replacing Ca2+ with Mg2+. Bitter and nonsugar sweeteners stimulate intracellular Ca2+ release in taste cells via a signaling cascade involving phospholipase C β2 (PLCβ2) and IP3 (Bernhardt et al., 1996; Spielman et al., 1996). Thus, a likely source of Ca2+ for transmitter release elicited by these compounds was an intracellular store. Because pharmacological interventions designed to test this notion would also affect the CHO/5-HT2C biosensor, we used a different approach to answer this question. We isolated taste buds from PLCβ2 null mutant mice (Jiang et al., 1997) and tested their ability to release serotonin after stimulation. Taste buds from PLCβ2 null mice responded to bath-applied KCl, showing a normal release of serotonin (Fig. 3D), as above. However, we were unable to detect serotonin release evoked by cycloheximide or saccharin from taste buds of the mutant mice, consistent with the source of Ca2+ for transmitter release being intracellular stores (via a PLCβ2 signaling cascade). Specifically, we obtained successful recordings from 16 of 56 taste buds from the knock-out mice when testing KCl depolarization; none of these taste buds responded to cycloheximide or saccharin.
This report presents data that indicate that serotonin is one of the neurotransmitters released by taste cells in response to gustatory stimulation. Furthermore, the results suggest that certain stimuli evoke neurotransmitter release in response to Ca2+ derived from intracellular stores. The postsynaptic target(s) for the serotonin remain to be determined. Serotonin may excite primary sensory afferent fibers, as proposed long ago by a number of investigators (Takeda, 1977; Esakov et al., 1983) (for review, see Nagai et al., 1996). Serotonin may also function as a paracrine hormone and act on cells within the taste bud, as discussed by Fujita et al. (1988), Ewald and Roper (1994), and more recently by Kaya et al. (2004). Our experiments were not designed to distinguish between these possibilities, both of which remain open questions.
The approach we used to detect neurotransmitters, a sensitive biosensor, has been used previously. For example, catfish horizontal cells were used to detect glutamate release from goldfish retinal bipolar cells (Tachibana and Okada, 1991) and from lizard cone photoreceptors (Savchenko et al., 1997). Myocytes have been pressed against Chinese hamster ovary fibroblasts to detect acetylcholine release (Morimoto et al., 1995). Researchers used human embryonic kidney 293 (HEK293) cells expressing the E-prostanoid receptor EP1 to monitor the release of prostaglandin E2 from macula densa cells isolated from rabbit kidneys (Peti-Peterdi et al., 2003). That group subsequently used cultured mouse mesangial cells to detect ATP released from macula densa cells (Komlosi et al., 2004). We are not aware of any group using biosensors to identify serotonin as a transmitter.
The incidence of successful serotonin release sites (e.g., for KCl stimulation, 31 of 105 taste buds) is low, presumably attributable to the relatively few numbers of serotonergic cells present in taste buds (Fig. 2), to the necessity that the taste bud was oriented such that serotonergic cells were positioned at the periphery and accessible to the biosensor, and to the requirement that the biosensor had to be placed quite near to a release site. Movement of the biosensor of only a few micrometers from a successful recording site would immediately abolish the responses, emphasizing the spatial sensitivity for placing the biosensor. It is possible that surveying a more extensive panel of taste stimuli and alternative depolarizing agents would increase the incidence of recording serotonin release sites.
Neurotransmitter release from taste cells seems to depend on two sources of Ca2+. For depolarizing stimuli (here, KCl and acid taste stimulation) (cf. Richter et al., 2003, 2004), Ca2+ influx was necessary and sufficient to stimulate serotonin release. For bitter and sweet tastants (cycloheximide and saccharin) that stimulated G-protein-coupled receptors, transmitter release persisted in the absence of extracellular Ca2+ (substituted with Mg2+). There are three possible explanations for this. (1) Replacing Ca2+ with Mg2+ in the bath did not totally eliminate Ca2+, and residual Ca2+ in the chamber, albeit low, might have sufficed to elicit transmitter release (Piccolino and Pignatelli, 1996). (2) Transmitter release in taste buds is truly Ca2+ independent, and, for example, is mediated by Na-dependent transporters, such as that which occurs at certain retinal synapses (Schwartz, 2002). (3) Transmitter release evoked by tastants is triggered by Ca2+ from internal stores. Our finding that exchanging Ca2+ for Mg2+ totally abolished depolarization-evoked release is inconsistent with the first possibility. There is a report of the serotonin transporter SET in taste buds, but it was not believed to be related to transporter-mediated transmitter release (Ren et al., 1999). Instead, our findings with the PLCβ2 knock-out mice support the last explanation. Previous studies have reported that Ca2+ derived from intracellular stores can modulate (Emptage et al., 2001; Yang et al., 2001; Simkus and Stricker, 2002; Galante and Marty, 2003) (but see Carter et al., 2002) or possibly elicit (Llano et al., 2000) neurotransmitter release at synapses. Taste bud cells may represent an excellent model for studying transmitter release triggered by Ca2+ from intracellular stores.
Note added in proof. Hayashi et al. (2004) recently reported generating biosensor cells for ATP using pheochromocytoma 12 or HEK293 cells expressing P2X2 purinergic receptors.
This work was supported by National Institutes of Health-National Institute on Deafness and Other Communication Disorders Grant DC006077 (S.D.R.). We thank Dr. K. Berg for the generous gift of CHO/5-HT2C cells.
Correspondence should be addressed to Dr. Stephen Roper, Department of Physiology and Biophysics, University of Miami School of Medicine, 1600 Northwest 10th Avenue, Miami, FL 33136. E-mail:.
Y.-J. Huang's present address: Department of Anatomy, College of Medicine, Kaohsiung Medical University, Kaohsiung 807, Taiwan.
K.-S. Lu's present address: Department of Anatomy and Cell Biology, College of Medicine, National Taiwan University, Taipei 100, Taiwan.
I. Plonsky's present address: The Institute for Information Transmission Problems, Russian Academy of Science, 19 Bolshoy Karetny per., Moscow 101447, Russia.
J. E. Baur's present address: Department of Chemistry, 214 Julian Hall, Illinois State University, Normal, IL 61790. D. Wu's present address: Department of Genetics and Developmental Biology, University of Connecticut Health Center, 263 Farmington Avenue, Farmington, CT 06030.
Copyright © 2005 Society for Neuroscience 0270-6474/05/250843-05$15.00/0