Report

Genetic Tracing Shows Segregation of Taste Neuronal Circuitries for Bitter and Sweet

See allHide authors and affiliations

Science  29 Jul 2005:
Vol. 309, Issue 5735, pp. 781-785
DOI: 10.1126/science.1110787

Abstract

The recent discovery of mammalian bitter, sweet, and umami taste receptors indicates how the different taste qualities are encoded at the periphery. However, taste representations in the brain remain elusive. We used a genetic approach to visualize the neuronal circuitries of bitter and sweet tastes in mice to gain insight into how taste recognition is accomplished in the brain. By selectively expressing a transsynaptic tracer in either bitter- or sweet and/or umami-responsive taste receptor cells, and by comparing the locations of the tracer-labeled neurons in the brain, our data revealed the potential neuronal bases that underlie discrimination of bitter versus sweet.

The gustatory system is primarily devoted to a quality check of food, while at the same time detecting nutrients and avoiding toxic substances. The initial step in taste perception takes place at the apical end of taste receptor cells, tightly packed into taste buds of the oral epithelium. The cells express taste receptors, which are responsible for detecting and distinguishing among sweet, bitter, salty, sour, and umami stimuli (1). In mammals, bitter and sweet and/or umami are the two main taste modalities evoking aversion and attraction, respectively. Humans also express pleasure for sweet taste but displeasure for bitter taste. On the other hand, mammals learn to reject a tastant if this tastant is associated with subsequent visceral malaise (2). Therefore, it is likely that the mammalian gustatory system is an excellent system to address the question of how emotion interacts with cognition and memory. To decipher rationally the underlying molecular, cellular, and system mechanisms, it is first necessary to understand and to compare precisely the contrastive neuronal circuitries that process and integrate the information of aversive and attractive taste modalities in the whole brain.

Bitter tastants are detected by members of a family of 30 different G protein-coupled receptors (GPCRs), the T2Rs (3-5). Sweet and umami tastes are substantially mediated by a small family of three GPCRs (T1R1, T1R2, and T1R3). T1R2 and T1R3 combine to function as a sweet receptor, whereas T1R1 and T1R3 form the umami receptor, which detects glutamate (6, 7). Sweet, umami, and bitter receptors appear to be expressed in distinct populations of taste cells that operate independently of each other to trigger taste recognition (6, 8-10). The receptor cells are innervated by afferent fibers that transmit information to the gustatory cortex through synapses in the brain stem and thalamus (11). How is taste information processed in the central nervous system, while it is discriminated and as it evokes the emotional and behavioral responses such as aversion and attraction? We applied a genetic approach to visualize the neuronal circuitries of bitter and sweet-umami taste by using the taste receptor genes and the plant lectin WGA as molecular tools. Injected lectin proteins are an effective tracer for transsynaptically delineating the wiring patterns in the central nervous system (12-14). Furthermore, the genetic approach using the WGA transgene, expressed under the control of specific promoter elements, is a powerful tool for tracing selective and functional neuronal circuitries originating from a specific type of neuron (15, 16).

We prepared transgenic mice in which the transsynaptic tracer WGA, C-terminally truncated and fused by the fluorescent protein (tWGA-DsRed), was coexpressed with selected taste receptors. The transgene for tracing the bitter taste neuronal circuitries is shown in Fig. 1A. We selected the promoter element of the mT2R5 gene, reported as a receptor for cycloheximide (5), to drive tWGA-DsRed expression in bitter receptor-expressing cells. In phospholipase Cβ2 (PLCβ2)-deficient mice, which lack sweet, amino acid, and bitter taste reception, the PLCβ2 transgene, expressed under the control of the mT2R5 promoter, rescued the response to multiple bitter compounds, but not to sweet or amino acid taste (8, 10). The findings support not only that taste receptor cells are not broadly tuned across these modalities, but also that the mT2R5-expressing cells coexpress the multiple T2Rs and are capable of responding to a broad array of bitter compounds. The transgene was constructed by connecting the promoter element of mT2R5, the fusion constructs of mT2R5 with green fluorescent protein (mT2R5-GFP) and tWGA-DsRed, which intercalated the internal ribosome entry site (IRES), and the polyadenylation signal (Fig. 1A). To test whether a part of the transgene works, Cos7 cells and HEK293 cells were transiently transfected with the construct, in which the upstream region of mT2R5 was replaced by the cytomegalovirus (CMV) promoter (Fig. 1B). The fusion proteins of mT2R5-GFP and tWGA-DsRed were expressed in these cells (Fig. 1B), which suggested that the transgene produces a bicistronic mRNA from which mT2R5-GFP and tWGA-DsRed are independently translated. Furthermore, the tWGA-DsRed fusion protein was associated with the intracellular granule-like structures of cultured cells. The mT2R5-GFP fusion protein was localized in the cell surface membrane (Fig. 1B). In addition, fura-2 calcium imaging of HEK293 cells, cotransfected with T2Rs-GFP and Gα15, revealed that fusion of GFP to the C terminus of T2Rs does not affect the receptor function (17).

Fig. 1.

Generation of the mT2R5-WGA mouse to visualize bitter taste neuronal circuitries. (A) Schematic diagram indicating the structure of the transgene to trace bitter taste neuronal circuitries. Blue boxes represent the homologous regions, found in the 5′ upstream sequences of mT2R5 and human T2R10. (B) Expression of mT2R5-GFP and tWGA-DsRed in cultured cells. Subcellular distribution of mT2R5-GFP and tWGA-DsRed, transiently expressed under the control of the CMV promoter in Cos7 cells and HEK293 cells, was directly visualized by the GFP and DsRed fluorescence. The expression levels of mT2R5-GFP and tWGA-DsRed were monitored with immunoblotting by using antibodies against GFP and WGA, respectively. (C) In situ hybridization demonstrated concordance in the expression pattern of the endogenous mT2R5 gene (red) and the transgene (green). (D) Direct fluorescence detection of mT2R5-GFP and tWGA-DsRed, expressed in taste receptor cells. Arrows indicate the transgene-expressing taste buds. (E) Spatial distribution of tWGA-DsRed in coronal sections of the mT2R5-WGA mouse brain, clarified by direct fluorescence detection (right). Darkfield images at the same magnification (middle) and at the lower magnification (left) were also shown. The distance to the posterior end of the fasciculus retroflexus (pfr) was calculated and denoted in each section.

First, to verify the expression patterns of the endogenous mT2R5 and the transgene in the transgenic mice (mT2R5-WGA), we performed in situ hybridization using the 3′ untranslated region of mT2R5 and the GFP-coding region as probes for double-label fluorescent detection. The cells expressing the endogenous mT2R5 receptor also expressed the transgene (Fig. 1C). Next, we analyzed the expression patterns of the transgene at a protein level by direct fluorescence detection of mT2R5-GFP and tWGA-DsRed in taste buds. mT2R5-GFP and tWGA-DsRed were rarely expressed in fungiform papillae. The few fungiform taste buds that express both mT2R5-GFP and tWGA-DsRed appear clustered at the posterior region of the front half of the tongue surface (Fig. 1D). Taste buds expressing mT2R5-GFP and tWGA-DsRed were detected in circumvallate papillae, foliate papillae, and the palate epithelium (Fig. 1D), consistent with the expression patterns of T2Rs (3, 4).

In the mT2R5-WGA mouse brain, we analyzed the spatial distribution of tWGA-DsRed, which is expressed in the specific taste receptor cells and transferred to the neurons. Anatomical and physiological data have shown the nuclear relays and their connecting pathways in the central gustatory system (11-14, 18-20), although the exact dimensions and internal organization for each taste modality remain unclear. Here, tWGA-DsRed was located in the posterior part of the solitary tract nuclei (Sol), the pontine parabrachial nuclei (PB), the thalamic gustatory area (Gus), and the gustatory cortex (DI) (Fig. 1E). Labeling in the most posterior part of the solitary tract nuclei might be associated with a projection from T2R-expressing cells of the gut. DsRed fluorescence was also observed in the amygdala and the olfactory cortex (fig. S1C). Neither mT2R5-GFP fluorescence nor in situ hybridized tWGA mRNA was detected in those parts of the brain.

To compare the neuronal circuitries of bitter and sweet-umami taste, we selectively expressed mT1R3-GFP and tWGA-DsRed in sweet-umami-responsive taste cells (mT1R3-WGA) under the control of the specific promoter element of the sweet-umami taste receptor mT1R3. The transgene is shown in Fig. 2A. The identical promoter element drives the transgene expression in perfect concordance with the endogenous mT1R3 gene (6). In mT1R3-WGA mice, mT1R3-GFP and tWGA-DsRed were coexpressed in subsets of taste receptor cells located in fungiform, foliate, circumvallate, and Geschmackstreifen (taste stripes) taste buds (Fig. 2B). tWGA-DsRed was located in the anterior part of Sol, PB, Gus, and DI (fig. S2). To characterize the cell type of the transgene-expressing cells in taste buds, we examined the cellular distribution of tWGA-DsRed in comparison with α-gustducin, ubiquitin carboxyl terminal hydrolase (PGP-9.5), and serotonin (5-HT). Taste cells are characterized by expressions of these marker proteins and are divided into three classes (types I, II, and III) (21). Both mT1R3-GFP and tWGA-DsRed were detected in taste cells with and without α-gustducin (Fig. 2C). Furthermore, mT1R3-GFP and tWGA-DsRed were coexpressed in subsets of both PGP-9.5-immunoreactive cells and 5-HT-immunoreactive cells (Fig. 2C). These results suggest that, at least, subsets of both the type II cells and type III cells with obvious synapses onto afferent nerve fibers expressed mT1R3-GFP and tWGA-DsRed at a protein level.

Fig. 2.

Generation of the mT1R3-WGA mouse to visualize sweet-umami taste neuronal circuitries. (A) Schematic diagram indicating the structure of the transgene to trace sweet-umami taste neuronal circuitries. (B) Direct fluorescence detection of mT1R3-GFP and tWGA-DsRed, expressed in taste receptor cells. Arrows indicate the transgene-expressing taste buds. (C) Confocal images showing localization of mT1R3-GFP, tWGA-DsRed, α-gustducin, PGP-9.5, and 5-HT in taste buds of mT1R3-WGA mice. Location of mT1R3-GFP and tWGA-DsRed was clarified by detecting GFP and DsRed fluorescence. Locations of α-gustducin, PGP-9.5, and 5-HT were detected using the primary antibodies against those proteins and Alexa-633-conjugated secondary antibodies. However, Alexa-633 fluorescence was replaced by the pseudocolor blue and overlaid. White arrows indicate the mT1R3-GFP- and tWGA-DsRed-expressing cells without immunoreactivity for α-gustducin, PGP-9.5, or 5-HT. Red arrows indicate the mT1R3-GFP- and tWGA-DsRed-expressing cells with immunoreactivity for α-gustducin, PGP-9.5, or 5-HT.

Then, we focused on the spatial distribution of tWGA-DsRed-labeled neurons in the brain. To characterize the precise distribution of tWGA-DsRed-labeled neurons in mT1R3-WGA mice in comparison with mT2R5-WGA mice and to permanently and repetitively observe the labeled neuron in the brain, we performed immunohistochemical detection of WGA in the brain. First, in geniculate ganglion neurons, which provide innervation to taste buds via the corda tympani nerve branch (11), the clusters of tWGA-DsRed-labeled neurons were dispersed in serial sections of ganglions, isolated from both the mT1R3-WGA and mT2R5-WGA mice (Fig. 3A). WGA immunoreactivity was also detected in subsets of nerve fibers (Fig. 3A). However, we could not deduce the differences in the spatial distributions of tWGA-DsRed-labeled geniculate ganglion neurons between mT1R3-WGA and mT2R5-WGA mice. In higher brain centers, segregation of inputs from mT1R3 and mT2R5 was revealed. tWGA-DsRed-labeled neurons are located in the coronal sections of the transgenic mouse brains (Fig. 3, B and C). In the mT2R5-WGA mouse brain, WGA immunoreactivity was detected in the broad but specific regions including Sol, the medial PB (MPB), Gus, and DI. tWGA-DsRed-labeled neurons were also observed in the amygdala (ACo, BLA), the olfactory cortex (Pir), and the primary somatosensory cortex (S1ULp) (Fig. 3B). Similarly, in the mT1R3-WGA mouse brain, tWGA-DsRed-labeled neurons were widely but restrictedly distributed to Sol, PB (MPB, LPBD, LPBC), Gus, DI, amygdala (APir, PMCo, BLA, MeAV), the olfactory cortex (LEnt, Pir), and the primary somatosensory cortex (S1J) (Fig. 3C). Differences in the tWGA-DsRed-labeled patterns of the two strains are detailed in online supporting text and fig. S3.

Fig. 3.

Spatial distribution of tWGA-DsRed in the mT2R5-WGA and mT1R3-WGA mouse brains, revealed by immunohistochemical detection of WGA. (A) Visualizing the spatial distribution of tWGA-DsRed-labeled neurons in geniculate ganglions of mT2R5-WGA mice (left) and mT1R3-WGA mice (right). Arrows indicate the tWGA-DsRed-labeled nerve fibers. (B) Locations of tWGA-DsRed-labeled neurons in the coronal sections of the mT2R5-WGA mouse brain. Arrows in the seventh panel indicate the tWGA-DsRed-labeled neurons in the amygdala. (C) Locations of tWGA-DsRed-labeled neurons in the coronal sections of the mT1R3-WGA mouse brain. The distance to pfr was calculated and denoted in each section.

It is noteworthy that the positions of labeled neuronal clusters in Sol, PB, Gus, DI, and the amygdala revealed similar patterns in different individuals in each strain (n = 5 and 7 for mT1R3-WGA mice and mT2R5-WGA mice, respectively), which show that the gustatory neurons dispersed in those regions were organized with sweet inputs rostral and with bitter inputs caudal, except for bitter inputs into ELPB and EMPB and the complex inputs into the amygdala (Fig. 4 and fig. S3). Although densely labeled neurons were scattered in the broad areas, most of the clusters appeared to be bilaterally asymmetrical in the left and right hemispheres of both the mT1R3-WGA and mT2R5-WGA mouse brains (fig. S3, A, E, and F). Judging from the distributional patterns of densely labeled neurons, inputs from mT2R5 and mT1R3 may be concentrated on the small size of the gustatory relays in the brain stem, the thalamus, and the cortex. However, it is possible that efficiency of tWGA-DsRed transfer might not parallel the strength of synaptic transmission, and that the neurons that contained only a small amount of tracer and were not detected might also relay taste information.

Fig. 4.

Schematic representation of the spatial distribution of tWGA-DsRed-labeled cells in mT2R5 and mT1R3 mice. Locations of tWGA-DsRed-labeled cells in the mouth, the solitary tract nucleus, the pontine parabrachial nucleus, the thalamic gustatory area, the amygdala, and the cortical gustatory area are plotted using green circles for mT2R5-WGA mice and red triangles for mT1R3-WGA mice.

Thus, the two strains of mice, expressing a tracer transgene in specific taste cells, enabled us to map connections formed by small subsets of neurons, which process and integrate the information of bitter taste, separated from sweet-umami taste. The mapping may be influenced by a certain equilibrium among the efficiencies of biogenesis, transport, and degradation of the tracer, which may vary depending on the developmental stages of the mouse brain and peripheral taste systems. Nevertheless, these transgenic mice can reveal the molecular aspects underlying the construction and refinement of taste neuronal circuitries, especially in combination with the gene-targeted mutant mice for key molecules.

Supporting Online Material

www.sciencemag.org/cgi/content/full/309/5735/781/DC1

Materials and Methods

SOM Text

Figs. S1 to S3

References and Notes

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article