Human blood group antigen H is not the specific marker for type I cells in the taste buds

We examined the localization of human blood antigen H (AbH) and its correlation with other cell type markers in the taste buds of circumvallate papillae of the adult rat. Immunoreactivity for AbH was localized in the membrane of two cell populations in the taste buds: in spindle-shaped cells extending from base to the apical portion of the taste buds as well as in round-shaped cells at the basal portion of the taste buds. Quantitative analysis revealed that approximately 47.8%, 24.4%, and 14.6% of cells within the taste buds displayed AbH-, alpha-gustducin- or protein gene product 9.5 (PGP 9.5)-immunoreactivity, respectively. Approximately 16.3% and 6.6% of AbH-immunoreactive taste bud cells displayed alpha-gustducin- or PGP 9.5-immunoreactivity, respectively. Although previous studies proposed that AbH immunoreactivity was specific for type I cells (dark cells or supporting cells), the present results indicate that AbH immunoreactivity is also present in some type II cells (alpha-gustducin immunoreactive cells) and type III cells (PGP 9.5-immunoreactive cells).     

Immunohistochemical identification of cells expressing steroidogenic enzymes cytochrome P450scc and P450 aromatase in taste buds of rat circumvallate papillae

The present study demonstrated for the first time the localizations and patterns of expression of key enzymes for steroidogenesis, cytochrome P450 side-chain-cleavage (P450scc), and P450 aromatase in the taste buds of rat circumvallate papillae, using immunoblot analyses and immunohistochemistry. Immunoblot analyses showed that proteins with a molecular weight close to that of rat adrenal cytochrome P450scc and a molecular weight close to that of rat ovary cytochrome P450 aromatase were present in the rat circumvallate papillae. In immunohistochemistry, antibodies against cytochrome P450scc and P450 aromatase yielded the labelings of a subset of taste bud cells. In the double immunolabeling of P450scc and alpha-gustducin or phospholipase C beta2(PLCbeta2), which were considered as markers of a majority of type II cells, P450scc was co-expressed in a subset of alpha-gustducin or PLCbeta2, but did not co-express neural adhesion molecule (NCAM), a marker of major type III cells. Further double immunolabeled studies showed that P450 aromatase was co-expressed in a subset of alpha-gustducin or PLCbeta2, but did not co-express PGP9.5, a marker of a majority of type III cells. The selective localization of cytochrome P450scc and P450 aromatase strongly suggests that estrogen biosynthesis from cholesterol might occur in a subset of type II cells of the rat taste buds. Although the full significance of estrogen in the taste bud function is not yet understand, estrogen appears to be an important regulator of taste transduction, as is the case with ATP (Finger et al., 2005), which further supports the centrality of taste cells in the life of taste buds.     

Espin cytoskeletal proteins in the sensory cells of rodent taste buds

Espins are multifunctional actin-bundling proteins that are highly enriched in the microvilli of certain chemosensory and mechanosensory cells, where they are believed to regulate the integrity and/or dimensions of the parallel-actin-bundle cytoskeletal scaffold. We have determined that, in rats and mice, affinity purified espin antibody intensely labels the lingual and palatal taste buds of the oral cavity and taste buds in the pharyngo-laryngeal region. Intense immunolabeling was observed in the apical, microvillar region of taste buds, while the level of cytoplasmic labeling in taste bud cells was considerably lower. Taste buds contain tightly packed collections of sensory cells (light, or type II plus type III) and supporting cells (dark, or type I), which can be distinguished by microscopic features and cell type-specific markers. On the basis of results obtained using an antigen-retrieval method in conjunction with double immunofluorescence for espin and sensory taste cell-specific markers, we propose that espins are expressed predominantly in the sensory cells of taste buds. In confocal images of rat circumvallate taste buds, we counted 21.5 ± 0.3 espin-positive cells/taste bud, in agreement with a previous report showing 20.7 ± 1.3 light cells/taste bud when counted at the ultrastructural level. The espin antibody labeled spindle-shaped cells with round nuclei and showed 100% colocalization with cell-specific markers recognizing all type II [inositol 1,4,5-trisphosphate receptor type III (IP₃R₃)_, α-gustducin, protein-specific gene product 9.5 (PGP9.5)] and a subpopulation of type III (IP₃R₃, PGP9.5) taste cells. On average, 72%, 50%, and 32% of the espin-positive taste cells were labeled with antibodies to IP₃R₃, α-gustducin, and PGP9.5, respectively. Upon sectional analysis, the taste buds of rat circumvallate papillae commonly revealed a multi-tiered, espin-positive apical cytoskeletal apparatus. One espin-positive zone, a collection of ∼3 μm-long microvilli occupying the taste pore, was separated by an espin-depleted zone from a second espin-positive zone situated lower within the taste pit. This latter zone included espin-positive rod-like structures that occasionally extended basally to a depth of 10–12 μm into the cytoplasm of taste cells. We propose that the espin-positive zone in the taste pit coincides with actin bundles in association with the microvilli of type II taste cells, whereas the espin-positive microvilli in the taste pore are the single microvilli of type III taste cells.     

Jacalin and peanut agglutinin (PNA) bindings in the taste bud cells of the rat: new reliable markers for type IV cells of the rat taste buds

Lectin histochemistry of Jacalin (Artocarpus integrifolia) and peanut agglutinin (PNA), specific lectins for galactosyl (beta-1, 3) N-acetylgalactosamine (galactosyl (beta-1, 3) GalNAc), was applied to the gustatory epithelium of the adult rat. In the ordinary lingual epithelium, Jacalin and PNA labeled the cell membrane from the basal to granular cell layer. They also bound membranes of rounded-cells at the basal portion of taste buds, but the number of PNA labeled cells was smaller than that of Jacalin labeled cells. There was no apparent difference in the binding patterns of Jacalin and PNA among the taste buds of the lingual papillae and those of the palatal epithelium. Occasionally, a few spindle-shaped cells were labeled with Jacalin, but not with PNA. Double labeling of Jacalin and alpha-gustducin, a specific marker for type II cells, revealed that Jacalin-labeled spindle-shaped taste cells were immunonegative for alpha-gustducin. Spindle-shaped cells expressing protein gene product 9.5 (PGP 9.5) immunoreactivity lacked Jacalin labeling. During the development of taste buds in circumvallate papillae, the binding pattern of Jacalin became almost identical from postnatal day 5. The present results indicate that rounded cells at the basal portion of the taste buds cells (type IV cells) bind to Jacalin and PNA, and these lectins are specific markers for type IV cells of the rat taste cells.     

Immunohistochemical localisation of regulatory neuropeptides in human circumvallate papillae

The occurrence and distribution of neuropeptide-containing nerve fibres in the human circumvallate papillae were examined by the peroxidase–antiperoxidase immunolocalisation method using surgical specimens that had not been subjected to radiotherapy, and the abundance of neuropeptide-containing fibres was expressed as the percentage of total nerve fibres demonstrated by protein gene product (PGP) 9.5 immunoreactivity for a quantitative representation of these peptidergic fibres. Substance P (SP) and calcitonin gene-related peptide (CGRP) immunoreactive (IR) nerve fibres were densely distributed in the connective tissue core of the circumvallate papillae, and some SP and CGRP-IR fibres were associated with the taste buds. A moderate number of vasoactive intestinal polypeptide (VIP)-IR fibres and a few galanin (GAL)-IR fibres were also seen in the connective tissue core and subepithelial layer. There were, however, no VIP-IR or GAL-IR fibres associated with the taste buds. Neuropeptide Y (NPY)-IR fibres were few and were associated with the blood vessels. Within the epithelium of the circumvallate papillae, no peptidergic fibres were found, although a number of PGP 9.5-IR fibres were detected. The abundance of SP, CGRP, VIP, and GAL-IR fibres expressed as the percentage of total PGP 9.5 IR fibres was 25.35±3.45%, 22.18±3.26%, 10.23±1.18%, and 4.12±1.05%, respectively. The percentage of NPY-IR fibres was below 3%. In a deeper layer of the papillae, a few VIP, GAL, and NPY-IR ganglion cells were found, and VIP immunoreactivity was detected in a few cells of the taste buds. There was no somatostatin, leucine enkephalin, or methionine enkephalin immunoreactivity in the circumvallate papillae. These results suggest that the dense SP and CGRP-IR fibres within the connective tissue core of the human circumvallate papillae may be involved in the deep sensation of the tongue.     

Protein gene-product 9.5 in developing mouse circumvallate papilla: comparison with neuron-specific enolase and calcitonin gene-related peptide

The present study was made to investigate the ontogeny of protein gene-product 9.5 (PGP 9.5)-like immunoreactivity (-LI) in the developing mouse circumvallate papilla (CVP), and its distribution was compared to that of neuron-specific enolase (NSE) and calcitonin gene-related peptide (CGRP). In adult CVP, PGP 9.5-LI was observed in the subgemmal nerve plexus; some thin PGP 9.5-like immunoreactive (-IR) nerve fibers penetrated taste buds and apical epithelium. PGP 9.5-LI was also observed in the spindle-shaped cells in taste buds, and a small number of round- or oval-shaped ganglionic cells in the lamina propria. The distribution of NSE-LI was comparable to that of PGP 9.5-LI. CGRP-LI was observed in the nerve fibers only; distribution of CGRP-IR nerve fibers was similar to that of PGP 9.5-IR nerve fibers, although the number of CGRP-IR nerve fibers was smaller than that of PGP 9.5-IR nerve fibers. At least six developmental stages were defined with regard to the developmental changes in the distribution of PGP 9.5-LI from embryonic day (E) 12 to adulthood: Stage I (E12–13) — a dense nerve plexus of PGP 9.5-IR nerve fibers was detected in the lamina propria beneath the core of newly-formed papilla. Stage II (E14–16) — thin PGP 9.5-IR nerve fibers penetrated the apical epithelium, and a few round-shaped cells in the apical epithelium also displayed PGP 9.5-LI. Stage III (E17–18) — thin PGP 9.5-IR nerve fibers penetrated the inner lateral epithelium of the trench. Stage IV [Postnatal day (P) 0–3] many PGP 9.5-IR nerve fibers penetrated the outer lateral epithelium of the trench; later in this stage, taste buds appeared. Stage V (P5–10) — a small number of PGP 9.5 IR cells in the taste buds appeared, and their number increased gradually. Stage VI (PI4-adult) — the number of PGP 9.5-IR taste cells increased and reached the adult level, while the number of PGP 9.5-IR nerve fibers decreased. The development of NSE-LI was similar to that of PGP 9.5-LI. CGRP-IR nerve fibers were detected at E12 in the lamina propria, and the development of the intraepithelial CGRP-IR nerve fibers was similar to that of PGP 9.5-IR nerve fibers. The present results indicate that invasion by nerve fibers of the epithelium of lingual papillae occurs in a complex manner, and that these nerve fibers may participate in the formation of the taste buds.     

Patterns of immunoreactivity specific for gustducin and for NCAM differ in developing rat circumvallate papillae and their taste buds

α-Gustducin and neural cell adhesion molecule (NCAM) are molecules previously found to be expressed in different cell types of mammalian taste buds. We examined the expression of α-gustducin and NCAM during the morphogenesis of circumvallate papillae and the formation of their taste buds by immunofluorescence staining and laser-scanning microscopy of semi-ultrathin sections of fetal and juvenile rat tongues. Images obtained by confocal laser scanning microscopy in transmission mode were also examined to provide outlines of histology and cell morphology. Morphogenesis of circumvallate papillae had already started on embryonic day 13 (E13) and was evident as the formation of placode. By contrast, taste buds in the circumvallate papillae started to appear between postnatal day 0 (P0) and P7. Although no cells with immunoreactivity specific for α-gustducin were detected in fetuses from E13 to E19, cells with NCAM-specific immunoreactivity were clearly apparent in the entire epithelium of the circumvallate papillary placode, the rudiment of each circumvallate papilla and the developing circumvallate papilla itself from E13 to E19. However, postnatally, both α-gustducin and NCAM became concentrated within taste cells as the formation of taste buds advanced. After P14, neither NCAM nor α-gustducin was detectable in the epithelium around the taste buds. In conclusion, α-gustducin appeared in the cytoplasm of taste cells during their formation after birth, while NCAM appeared in the epithelium of the circumvallate papilla-forming area. However, these two markers of taste cells were similarly distributed within mature taste cells.     

Mash1-expressing cells could differentiate to type III cells in adult mouse taste buds

The gustatory cells in taste buds have been identified as paraneuronal; they possess characteristics of both neuronal and epithelial cells. Like neurons, they form synapses, store and release transmitters, and are capable of generating an action potential. Like epithelial cells, taste cells have a limited life span and are regularly replaced throughout life. However, little is known about the molecular mechanisms that regulate taste cell genesis and differentiation. In the present study, to begin to understand these mechanisms, we investigated the role of Mash1-positive cells in regulating adult taste bud cell differentiation through the loss of Mash1-positive cells using the Cre-loxP system. We found that the cells expressing type III cell markers—aromatic L-amino acid decarboxylase (AADC), carbonic anhydrase 4 (CA4), glutamate decarboxylase 67 (GAD67), neural cell adhesion molecule (NCAM), and synaptosomal-associated protein 25 (SNAP25)—were significantly reduced in the circumvallate taste buds after the administration of tamoxifen. However, gustducin and phospholipase C beta2 (PLC beta2)—markers of type II taste bud cells—were not significantly changed in the circumvallate taste buds after the administration of tamoxifen. These results suggest that Mash1-positive cells could be differentiated to type III cells, not type II cells in the taste buds.     

The survival of vagal and glossopharyngeal sensory neurons is dependent upon dystonin

The vagal and glossopharyngeal sensory ganglia and their peripheral tissues were examined in wild type and dystonia musculorum mice to assess the effect of dystonin loss of function on chemoreceptive neurons. In the mutant mouse, the number of vagal and glossopharyngeal sensory neurons was severely decreased (70% reduction) when compared with wild type littermates. The mutation also reduced the size of the circumvallate papilla (45% reduction) and the number of taste buds (89% reduction). In addition, immunohistochemical analysis demonstrated that the dystonin mutation reduced the number of PGP 9.5-, calcitonin gene-related peptide-, P2X3 receptor- and tyrosine hydroxylase-containing neurons. Their peripheral endings also decreased in the taste bud and epithelium of circumvallate papillae. These data together suggest that the survival of vagal and glossopharyngeal sensory neurons is dependent upon dystonin.     

Effects of streptozotocin-induced diabetes on taste buds in rat vallate papillae

Some studies have documented taste changes in patients with diabetes mellitus (DM). In order to understand the relationships between taste disorders caused by DM and the innervation and morphologic changes in the taste buds, we studied the vallate papillae and their taste buds in rats with DM. DM was induced in these rats with streptozotocin (STZ), which causes the death of beta cells of the pancreas. The rats were sacrificed and the vallate papillae were dissected for morphometric and quantitative immunohistochemical analyses. The innervations of the vallate papillae and taste buds in diabetic and control rats were detected using immunohistochemistry employing antibodies directed against protein gene product 9.5 (PGP 9.5) and calcitonin gene-related peptide (CGRP). The results showed that PGP 9.5- and CGRP-immunoreactive nerve fibers in the trench wall of diabetic vallate papillae, as well as taste cells in the taste buds, gradually decreased both intragemmally and intergemmally. The morphometry revealed no significant difference in papilla size between the control and diabetic groups, but there were fewer taste buds per papilla (per animal). The quantification of innervation in taste buds of the diabetic rats supported the visual assessment of immunohistochemical labeling, that the innervation of taste cells was significantly reduced in diabetic animals. These findings suggest that taste impairment in diabetic subjects may be caused by neuropathy defects and/or morphological changes in the taste buds.     

Immunolocalization of SNARE proteins in both type II and type III cells of rat taste buds

Double immunohistochemistry of soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptor (SNARE) proteins [synaptosomal-associated protein of 25 kDa (SNAP-25), syntaxin and vesicle-associated protein-2 (VAMP-2)], and specific cell markers of taste buds cells [alpha-gustducin and phospholipase Cbeta2 (PLCbeta2) for type II cells; neural cell adhesion molecule (NCAM) for type III cells] was applied to gustatory epithelia of the rat circumvallate papillae. All three SNARE proteins were present in some elongated taste buds cells as well as intra-, peri- and subgemmal nerve fibers. Double immunohisotochemistry revealed that nearly all alpha-gustducin and PLCbeta2 immunoreactive cells expressed SNAP-25, syntaxin, and VAMP-2. A majority of NCAM immunoreactive cells showed immunoreactivity for these SNARE proteins. These results indicate that these synapse-associated proteins (SNAP-25, syntaxin and VAMP-2) are present in both type II cells and type III cells. Moreover, more than 50% of intragemmal cells containing SNARE proteins showed immunoreactivities for alpha-gustducin, PLCbeta2, and NCAM, suggesting the possible presence of transitional cells having histochemical properties of both type II and type III cells.     

Differentiation of the lingual and palatal gustatory epithelium of the rat as revealed by immunohistochemistry of alpha-gustducin.

We used alpha-gustducin, a taste-cell-specific G protein to investigate the onset of taste transduction and its relation to the development of the palatal and lingual taste buds. Frozen cryostat and paraffin sections were prepared from the palatal and lingual gustatory epithelium of the rat from birth till postnatal day 21 (PN 21d). At PN 1-7d, alpha-gustducin-immunoreactive solitary ovoid or bipolar cells were scattered among the oral epithelium either horizontally along the oral surface or vertically oriented between the basal lamina and oral surface. In the circumvallate and foliate papillae, these cells became wrapped in alpha-gustducin-immunonegative cells surrounded by an extracellular space forming a bud-like structure. Simultaneously, different stages of typical taste buds were recognized, but alpha-gustducin was only expressed in some neonatally developed pored buds. At PN 1d, alpha-gustducin was expressed in pored taste buds with a relatively higher frequency recorded in the soft palate as compared with the nasoincisor, circumvallate, and foliate papillae. The immunoreactive cells were spindle shaped with elongated processes extending from the base to the pore of the taste buds. During the second week, the solitary cells could no longer be recognized while the total counts of immunoreactive cells within the taste buds gradually increased. We argue that taste transduction is essentially required from the time of birth and can be fulfilled by both of the solitary chemosensory cells, which are immunoreactive for alpha-gustducin and scattered in the oral epithelium, and the taste cells within the mature taste buds. Moreover, the onset of taste transduction accomplished by the palatal taste buds developed earlier than that achieved by taste buds in the circumvallate and foliate papillae.     

Immunohistochemical detection of neurotrophin-3 and -4, and their receptors in mouse taste bud cells

Neurotrophin-3 (NT3) and neurotrophin-4 (NT4) affect the survival and maintenance of central and peripheral neurons. Using an immunohistochemical method, we examined whether the taste bud cells in the circumvallate papillae of normal mice expressed NT3, NT4, and their respective receptors TrkC and TrkB, and if so, what type of cells in the taste buds expressed them. Double immunostaining for either of them and PGP 9.5, NCAM, or gustducin was used to determine which cell types expressed which neurotrophins and receptors. Normal taste bud cells expressed NT3, NT4, and the TrkB receptor, but not TrkC. The percentage of NT3-immunoreactive cells among all taste bud cells was 89.0%, that of NT4-immunoreactive cells, 58.6%, and that of TrkB-immunoreactive cells, 80.8%. Almost none of the NT4-immunoreactive cells were reactive with anti-PGP 9.5 or the anti-NCAM antibody, but they could be stained with anti-gustducin, revealing that NT4-immunoreactive cells were contained only in the type-II--and possibly type-I--cell population. On the other hand, NT3-, and TrkB-immunoreactive cells included type-III cells, together with type-II, -I, and basal cells, because they were positive for PGP 9.5 and gustducin. We conclude that NT4 may exert trophic actions on all types of taste bud cells by binding to their TrkB receptors, and NT3 may also have a similar, though negligible role.     

The distribution of PGP9. 5, BDNF and NGF in the vallate papilla of adult and developing mice

The development and innervation of vallate papillae and taste buds in mice were studied using antibodies against the neuronal marker, protein gene product 9.5 (PGP 9.5), and against nerve growth factor (NGF) and brain-derived neurotrophic factor (BDNF). PGP 9.5 immunohistochemical studies revealed that the earliest sign of median vallate papilla formation was an epithelial bulge at embryonic day 13 (E13), and at E14, a dense nerve plexus was found within the connective tissue core of the papilla. Thin nerve fibers penetrated the apical and medial trench wall epithelium of the papilla at E16 and a few of these began to invade the lateral trench wall epithelium at E17. At postnatal day 1 (P1), the newly formed taste buds were recognizable and a small number of PGP 9.5-immunoreactive (IR) cells appeared on the medial trench wall epithelium. The number of PGP 9.5-IR taste bud cells then increased gradually and reached the adult level at postnatal week 2. PGP 9.5 immunoreactivity increased systematically with age. NGF and BDNF immunoreactivity was first seen at the boundary between the columnar cells in the apical epithelium of the developing vallate papilla at E13, then in the medial and lateral trench walls at E15 (BDNF) or E18 (NGF). At P1, BDNF immunoreactivity was exclusively present in the newly formed taste buds of the medial trench wall. The number of BDNF-IR taste bud cells then increased gradually, reaching the adult level at P7. Similar degrees of NGF and BDNF immunoreactivity were seen in the developing vallate papilla. In the present study, we found that the vallate papilla was formed prior to its innervation, and we propose that initiation of papilla formation does not require any direct influence from the specific gustatory nerve. We also suggest that neurotrophins in the early developing vallate papillae might act as local tropic factors for the embryonic growth of nerve fibers to induce differentiation of the taste buds.     

Distribution of taste buds on fungiform and circumvallate papillae of bovine tongue

The distribution of taste buds on the fungiform and circumvallate papillae of the cow tongue has been determined. The two tongues studied were from Holstein-Friesian cows four to six years of age; they contained 14,765 and 21,691 taste buds, respectively. The tip of the tongue is well supplied with fungiform papillae, and the posterior portion contains the circumvallate papillae. The midportion of the tongue contains relatively few taste papillae. The fungiform papillae contained 1,580 and 1,838 taste buds on the two tongues, respectively, and the circumvallate papillae were estimated to contain 13,185 and 19,853 taste buds. The highest concentration of taste buds therefore occurs in the circumvallate papillae; these relatively few papillae contain approximately 90% of the taste buds. On a circumvallate papilla, taste buds are found only on the papillary sidewall, with none either on the apical surface of the papilla or on the outer wall of the moat.     

Maturation of taste buds on the soft palate of the postnatal rat

Taste bud distribution on the soft palate and within three types of tongue papillae (fungiform, foliate, and circumvallate) were examined histologically in the rat at different postnatal ages. After paraffin embedding, serial sections (10 microm) were made and stained by HE, and digitized images of each section were examined. The existence of a taste pore was used to identify mature taste buds. At birth, 53% (68 of 127 observed) of the taste buds on the soft palate, but only 14% (14 of 110 observed) within fungiform papillae, contained a taste pore. One week after birth, the number of mature taste buds increased rapidly, resulting in 90% of soft palate taste buds and 80% of fungiform taste buds containing taste pores. In contrast, no taste buds with pores were observed at birth within foliate and circumvallate papillae; however, at two weeks after birth 52% (71 of 132 observed) of the foliate and 68% (180 of 267 observed) of the circumvallate taste buds examined contained taste pores. These results suggest that taste buds within the soft palate play an important role in the detection of nutrients in the neonatal rat.     

Taste buds and nerve fibers in the rat larynx: an ultrastructural and immunohistochemical study

We investigated the rat laryngeal taste buds and their innervation by electron microscopy and immunohistochemical methods. Taste buds were densely arranged in the surface facing the laryngeal cavity of the epiglottis, the aryepiglottic fold, and the cuneiform process of the arytenoid cartilages. The cells of the buds were classified into types I, II, III, and basal cells, the ultrastucture of which was almost the same as that previously reported in lingual taste buds. The type III cells that had synaptic contacts with nerve fibers were considered to be sensory cells. Immunohistochemical analysis revealed thick calbindin D28k-immunoreactive fibers and thin varicose fibers immunoreactive for calcitonin gene-related peptide or substance P in and around the taste bud. Serotonin-immunoreactive cells were also observed here. The results revealed the innervation pattern of laryngeal taste buds to be the same as that in lingual taste buds. Carbonic anhydrase (CA) is known to catalyze the hydration of CO2 and dehydration of H2CO3, and seems to be essential in CO2 reception. Immunoreactivity for CAI was detected in slender cells and that for CAIII was observed in barrel-like cells in the laryngeal taste buds. The pH-sensitive inward rectifier K+ (Kir) channel in the cell membrane may be involved in CO2 reception as well. CAII-reactive cells were also reactive to Kir4.1, PGP 9.5 and serotonin. Our results indicated that CAII and Kir4.1 are located in type III cells of the laryngeal taste buds, and supported the idea that the buds may be involved in the recognition of CO2.     

Three-dimensional structure of the gustatory cell in the mouse fungiform taste buds: a computer-assisted reconstruction from serial ultrathin sections

Taste buds in fungiform papillae of the mouse were examined with transmission electron microscopy and computer-assisted, three-dimensional reconstructions from serial ultrathin sections. In accord with observation by Murray (1971), four distinct cell types, type I, II, III and basal cells, were identified. Of these, only the type III cell made synaptic contacts with nerve terminals and contained both small, clear vesicles and dense-cored granules. The former vesicles were synaptic-type and accumulated in the cytoplasm just below the synaptic in membrane thickenings. This finding clearly indicates a sensory function for the type III cell. One to three type III cells were identified within a taste bud. The type III cell had at most eight synapses with nerve terminals. One nerve fiber making two synapses with the type III cell was occasionally observed in its terminal region.