Skip to main content
Biology LibreTexts

12.14a: Sensory Transduction in Vision, Olfaction, and Gustation

  • Page ID
  • Genovese et al Front. Cell. Neurosci., 08 October 2021 |  Creative Commons Attribution License (CC BY)


    Sensory Transduction in Photoreceptors and Olfactory Sensory Neurons 

    Photoreceptors and olfactory sensory neurons (OSNs) have highly specialized structures that enable them to capture their respective stimuli of light and odorant ligands. Both photoreceptors and OSNs have evolved highly specific abilities to detect and discriminate light wavelengths or odors. They use intricate transduction mechanisms to convert sensory stimuli into electrical signals. Their transduction cascades not only are able to greatly amplify the signal but also to enhance the signal to noise, enabling these cells to detect and distinguish minute stimuli within very noisy backgrounds conditions. Such transduction mechanisms provide for modulation at multiple steps to adapt the sensory neurons to different background stimulation and optimize the capture of useful information about the surrounding world.

    In this review, we summarize some of the key structural and functional features of vertebrate rod and cone photoreceptors and of OSNs, and the molecular mechanisms that underlie their function. While describing features of both cell types, we emphasize the similarities and differences between photoreceptors and OSNs and the unique features of each cell type that make them perfectly suited to perform their function.


    Signal Detection in Photoreceptors and Olfactory Sensory Neurons’ Specialized Cilia


    Vertebrate rod and cone photoreceptors as well as OSNs are ciliary neurons (Figure 1) with specialized cilia where the initial detection of the sensory stimulus takes place to activate a sensory transduction cascade. Rods and cones have a single cilium that has evolved to accommodate a stack of ~1,000 membrane disks where the visual pigment is expressed at a very high 3–5 mM concentration (Figure 1APalczewski, 2006). In the case of rods, the disks are enveloped by the plasma membrane, whereas in cones the disks are formed by invaginations of the plasma membrane. As light enters the eye and reaches the retina, it travels along the length of the rod and cone outer segments. The orientation of the elongated outer segments along the light path, together with the high density of visual pigment in their disks results in ~50% probability that an incident photon is absorbed by a visual pigment molecule (Bowmaker and Dartnall, 1980). In the case of OSNs (Figure 1B), odorant ligands are detected in the ~20 cilia protruding from each dendritic knob which are immersed in the mucus layer covering the olfactory epithelium. The olfactory cilia, which are motile in amphibians but not in rodents, are only about 0.1–0.2 μm thin but can reach up to 100 μm in length depending on the species (Kleene and Gesteland, 1981Ukhanov et al., 2021). While this greatly increases the surface membrane area available to incorporate olfactory receptor (OR) proteins to detect odorants, it also greatly reduces the ciliary volume with potentially detrimental effects (see below).

    Sensory Transduction in Photoreceptors and Olfactory NeuronsFig1.svg


    Figure 1. Photoreceptors and olfactory sensory neurons (OSNs). (A) Simplified schematic representation of a rod and a cone in the retina. Photoreceptors are polarized neurons with a specialized morphology optimized to detect light stimuli. The outer segments of both rods and cones are modified sensory cilia, containing membrane disks organized in a stack. In the case of rods, the outer segment has a slim rod-like structure in which the disks are enclosed by the plasma membrane. The outer segment of the cones has a stocky conical-shaped structure, in which the disks are constituted by invaginations of the plasma membrane. The outer segment does not contain any proteins of the cell translation machinery, which are mostly localized in the inner segment, including the endoplasmic reticulum, Golgi, and mitochondria. Outer and inner segments are connected by the connecting cilium, while distal to the inner segment is the cell body containing the nucleus, followed by the axon and synaptic termini that extend into the outer plexiform layer where they synapse with the second order neurons. When the light enters the eye, after reaching the retina, it travels along the length of the rod and cone inner segment until finally reaching the outer segments. (B) Simplified schematic of an OSNs in the olfactory epithelium. OSNs are ciliated bipolar neurons, their apical dendrites extend to the surface of the epithelium terminating with a spherical structure called dendritic knob, from which the sensory cilia enter the mucus layer. The ciliary membrane contains the olfactory receptors (ORs) necessary to detect different odorants. Distal from the knob is the cell body of the OSN with its nucleus, followed by a long axon that projects to the olfactory bulb, where it synapses with the second order neurons. Images created with


    Electrophysiological Approaches to Record Light- and Odorant-Induced Responses


    The similar morphological structure of rods, cones, and OSNs, with a ciliary part able to detect the respective stimuli and an adjacent cell body, allows similar electrophysiological approaches to record stimulus-induced responses in these cell types. The cell body of a photoreceptor or an OSN can be sucked into the tip of a recording pipette by using a loose-patch (or suction pipette) recording configuration (Baylor et al., 1979Lowe and Gold, 1991). This leaves the outer segment of photoreceptors or the olfactory cilia exposed and accessible to bath solution changes, e.g., the application of pharmacological agents or to odorants, in the case of OSNs. Suction pipette recordings can be performed from isolated sensory neurons, as shown in Figures 2A,B,D (respectively, a salamander rod, salamander cone, and salamander OSN) but also from dissected retina tissue, as in the case of the outer segment of a mouse rod drawn in the recording electrode from a piece of the retina (Figure 2C). This recording configuration measures the transduction current entering the photoreceptor outer segment or olfactory cilia, and leaving via the cell body.

    Sensory Transduction in Photoreceptors and Olfactory NeuronsFig2and3.svg


    Figure 2. Suction electrode recordings of photoreceptors and OSNs. Suction electrode with a sucked salamander rod (A) and a salamander cone (B)(C) Suction electrode with a mouse rod from a fragment of the dissected retina. (D) Suction electrode with a single salamander OSN. Figures modified from Reisert and Matthews (2001)Kefalov et al. (2010), and Kefalov (2012) with permission.

    A fundamental difference between photoreceptors’ and OSNs’ responses to stimuli lies in their polarity. In the absence of light, rods and cones are kept depolarized by a standing inward current of approximately 20–40 pA for amphibian cones and rods, and 7–15 pA for mouse photoreceptors. This depolarizing current is gradually suppressed upon light stimulation until, for sufficiently high light intensities, it is reduced to zero (Figures 3A,B, mouse rod and cone responses, respectively), leading to photoreceptor hyperpolarization. Similar to rods, but differently from cones, the OSNs show comparatively little spontaneous activity in absence of stimuli (Reisert, 2010Connelly et al., 2013). Different OSNs show varying levels of spontaneous basal activity determined by the constitutive activity of their ORs (Reisert, 2010Connelly et al., 2013).

    Sensory Transduction in Photoreceptors and Olfactory NeuronsFig3.svg


    Figure 3. Light and odorant-induced responses. Typical response families from single-cell suction current recordings obtained from a mouse rod (A) and cone (B). 10 ms flashes of various intensities were delivered to each photoreceptor at time 0, with each successive flash 0.5-log times brighter than the previous one. (C) Normalized intensity-response functions showing that cones have approximately 2.5 log units lower sensitivity in comparison to rods. Suction current recordings from OSNs expressing the mOR-EG (D) or M71 (E) olfactory receptor, stimulated with increasing concentrations of the odorants eugenol and acetophenone, respectively. (F) Dose-response curves showing different sensitivity of mOR-EG and M71-expressing OSNs. Figures modified from Kefalov et al. (2010)Sakurai et al. (2011)Deng et al. (2013), and Dibattista and Reisert (2016) with permission.

    In the presence of odorants, OSNs generate an inward receptor current which leads to depolarization, and the generation of action potentials (Firestein and Werblin, 1989Kurahashi, 1989Reisert and Matthews, 1999). This receptor current is odorant concentration-depend and increases progressively with increasing stimulation until it eventually saturates at high odorant concentrations. Responses recorded from OSNs expressing different olfactory receptors can generate fairly different response amplitudes when stimulated with their respective agonists (Figures 3D,E: responses recorded from mouse OSNs that express the mOR-EG or the M71 olfactory receptor, which are activated by the ligands eugenol and acetophenone, respectively).

    The hyperpolarization and signals carried by graded potentials in photoreceptors vs. depolarization and signals carried by action potentials in OSNs represent another fundamental difference between these two types of sensory neurons. These topics and the differences in synaptic structure and transmission between photoreceptors and OSNs go beyond the focus of this review and are discussed in an excellent recent review on this topic (Lankford et al., 2020).


    Sensitivity of Photoreceptors and Olfactory Sensory Neurons


    In part due to their unique structure, photoreceptors, and, to a lesser extent, OSNs have achieved exquisite sensitivity that optimizes the detection of stimuli within the respective sensory organs. In addition, both sensory receptors use a transduction cascade to amplify the signal (see below). As a result, rod photoreceptors can reliably detect single photons (Baylor et al., 1979), enabling humans to perceive light with as few as six photons detected by adjacent rods (Hecht et al., 1942). This renders rods perfectly suited for dim light vision, with a dynamic range spanning lights from a dark cloudy night to sunrise (Fain et al., 2010). Cones are ~100-fold less sensitive than rods, making them suited for daytime light conditions. Figure 3C compares the intensity-response function of mouse rods and cones, demonstrating the much lower sensitivity of cones compared to rods.

    Most OSNs respond to odor concentrations in the low micromolar range (Bozza et al., 2002Grosmaitre et al., 2009Saito et al., 2009Lee et al., 2011Dibattista and Reisert, 2016), but they can also reach exquisite sensitivity and are capable of detecting odors at the nanomolar concentration range. Picomolar sensitivity is reached by a subset of OSNs that express receptors specialized in detecting amines, the trace-amine associated receptors (Zhang et al., 2013). In comparison to rods, OSNs do not reach such high sensitivity, and cannot be activated by a single odorant molecule but instead require around 30 odorant binding events to begin firing action potentials reliably (Bhandawat et al., 2010). The detection of odorants in the olfactory epithelium can be further enhanced by the expression of a wider number of different OR genes, more than 350 in humans and 1,000 in mice (Malnic, 2007), with overlapping response profiles to odorants. A larger number of OSNs, particularly in species relying heavily on their sense of smell, may enhance further the detection of odorants. For instance, the human olfactory epithelium covers ~3–4 cm2 and contains approximately 5–6 million OSNs while in the case of dogs, the area of the olfactory epithelium is 18–150 cm2 and contains 150–300 million OSNs (Lippi and Heaney, 2020).


    Detection of Stimuli


    In both photoreceptors and OSNs, the detection of stimuli is mediated by G protein-coupled receptors. In photoreceptors, this function is achieved by rod and cone visual pigments, which consist of a protein, opsin, covalently attached to the visual chromophore, typically 11-cis-retinal (Ebrey and Koutalos, 2001). The chromophore serves as a reverse agonist, keeping the receptor molecule in the inactive ground state (Crouch et al., 1996). Absorption of a photon by 11-cis-retinal triggers its conformational change to all-trans-retinal, which, in turn, results in rearrangement of the opsin transmembrane helices and switch of the visual pigment molecule into its active state. The activated visual pigment then binds to a G protein, transducin, activating it. The activation of transducin triggers the transduction cascade that ultimately generates the cellular response (Pugh and Lamb, 1993). Eventually, the all-trans retinal chromophore is released from opsin after the covalent Schiff base between them is hydrolyzed, leaving behind chromophore-free opsin (Saari, 2016). Notably, without chromophore, opsin has residual activity, and in sufficient quantities can produce steady activation of the photoreceptors, similar to a steady background light, thus modulating the sensitivity of photoreceptors (Fain et al., 1996). This process is known as bleaching adaptation, indicating the production of free opsin after the photoactivation of the visual pigment and dissociation of the visual chromophore.

    Unlike in photoreceptors, where the ligand, a light-sensitive reverse agonist, is covalently attached to opsin, in olfaction, the ligands are dissolved in the mucus covering the surface of the olfactory epithelium and come into direct contact with the OR proteins expressed in the OSN ciliary membrane. This results in the activation of the receptor protein that, in turn, is transduced downstream to a G protein to trigger a transduction cascade resulting in the cellular response. The binding of the ligand to the receptor protein is noncovalent and rapidly reversible. ORs, like other G protein-coupled receptors, do display antagonism, inverse, and partial agonism, leading to suppressed responses to their agonists, a reduction in basal activity in the absence of stimulation or suppression of the maximal response (Firestein et al., 1993Oka et al., 2004Reisert, 2010).


    Discrimination Between Stimuli


    The spectral sensitivity of individual rod and cone photoreceptors is dictated by the absorption properties of their visual pigments. Typically, each photoreceptor type expresses only one type of opsin; in the case of the human retina, rods express rod opsin, whereas cones express long wavelength (LW, red), middle wavelength (MW, green), or short wavelength (SW, blue) opsin (Nathans, 1987). When bound to the chromophore, the amino acid structure of each opsin determines the optical properties of the resulting visual pigment and the spectral sensitivity of the photoreceptors expressing it. As a result, species existing in environments with characteristic light distribution, such as deep-sea fish, have visual pigments that have evolved to optimize their spectral sensitivity (Hope et al., 1997). A second factor controlling the optical properties of the visual pigment is the structure of the visual chromophore. Most species, including mice and humans, use 11-cis-retinal, a derivative of Vitamin A, also known as A1. However, some amphibians and fish also use 3,4-dehydro 11-cis retinal, also known as A2. This chromophore has an extra conjugated double bond in its structure, which shifts the absorption spectrum of A2 visual pigments to longer wavelengths compared to A1 visual pigment embedded in the same opsin molecule (Corbo, 2021). Some aquatic and amphibian species use the A1/A2 chromophores to shift their spectral sensitivity from murky waters dominated by longer wavelengths of light to seawater and air, dominated by shorter wavelength lights (Bridges, 1964). One notable example includes the toad, where the retina is populated by A1 visual pigment in its ventral section, receiving light from above the surface of the water, and by A2 visual pigment in its dorsal section, receiving light from below the surface of the water (Reuter et al., 1971). A shift in the chromophore can also occur during the lifetime of the animal as its environment changes, such as the A2 to A1 shift in salamanders as they metamorphose from larval (aquatic) to the adult (terrestrial) stage (Ala-Laurila et al., 2007), or the A2 to A1 shift in Atlantic salmon during migration from sea to freshwater (Beatty, 1966).

    Similar to photoreceptors, the ligand specificity of the OSNs is also dictated by the expression of OR genes in their cilia. As photoreceptors, each OSN expresses generally only one receptor gene so that its ligand specificity is determined by the structure of the OR expressed in that particular cell. However, photoreceptors typically use no more than five opsin genes to cover the visible spectrum, while OSNs can use hundreds, in the case of humans, to thousand and more, for rodents and dogs, OR genes to cover the odor space (Malnic, 2007). The same OR can be activated by multiple odorants with different sensitivities, and a given odorant can activate different ORs with different half-maximal concentrations (Buck, 1996Ache, 2020). This generates a complex mosaic of ORs and odorants response pairs. Figure 3F compares the dose responses of OSNs expressing either the mOR-EG or the M71 OR to eugenol and acetophenone, respectively. In this case, mOR-EG OSNs display higher sensitivity to its agonist compared to M71 OSNs. However, this does not preclude the possibility that the M71 OR is more sensitive to another ligand resulting in a more left-shifted dose response relation than the one seen with acetophenone. Conversely, the dose response relation of M71 OSNs to benzaldehyde is approximately 10-fold right-shifted compared to acetophenone (Bozza et al., 2002).

    Determining the ligand specificity of ORs is an ongoing endeavor (Abaffy et al., 2006Saito et al., 2009Kurian et al., 2021). Due to the large number and diversity of OR genes, as well as the near endless number of odorant molecules, understanding the overall mechanisms that control their ligand binding affinity and specificity remains a challenge. Receptor modeling approaches to understand and predict OR–odorant molecule interactions can provide valuable insights but are somewhat hampered by the lack of a crystal structure of any vertebrate OR. The rhodopsin structure is often used as a guide and homology model to predict the structure of ORs (Katada et al., 2005Bavan et al., 2014).


    Sensory Transduction Activation


    In both photoreceptors and OSNs, the detection of stimuli by their respective G protein-coupled receptors is converted into electrical signals via the activation of a G protein coupled to a second messenger transduction cascade. The two pathways, though clearly distinct, share an amazing level of similarity (Figure 4). Thus, in both cases the second messenger is a cyclic nucleotide, cGMP in photoreceptors (Pugh and Lamb, 1990) and cAMP in OSNs (Sklar et al., 1987Bakalyar and Reed, 1990). As a result, the activation of both transduction cascades results in a rapid shift in the equilibrium between synthesis and hydrolysis of the respective cyclic nucleotide, which is then sensed by the cyclic nucleotide-gated (CNG) transduction channels in the plasma membrane of the photoreceptor outer segment or olfactory cilium.

    Sensory Transduction in Photoreceptors and Olfactory NeuronsfFig4.svg


    Figure 4. Activation of the transduction cascade in rod photoreceptors and OSNs. (A) Schematic representation of phototransduction cascade in rods. Abbreviations: rhodopsin (Rh), Tα, β, and γ subunits of the retinal G protein, transducin (T), guanosine-5′-triphosphate (GTP), guanosine-5′-diphosphate (GDP), phosphodiesterase (PDE), guanosine monophosphate (GMP), and cyclic guanosine monophosphate (cGMP), and cyclic nucleotide-gated (CNG) channel. (B) Schematic representation of the olfactory transduction cascade in OSNs. Abbreviations: Olfactory receptor (OR), guanosine-5′-triphosphate (GTP), guanosine-5′-diphosphate (GDP), Gαolfβ, and γ, subunits of the olfactory G protein; adenylyl cyclase 3 (AC3), adenosine-5′-triphosphate (ATP), cyclic adenosine monophosphate (cAMP), cyclic nucleotide-gated (CNG) channel; Ca2+-activated Cl channel anoctamin 2 (ANO2). Images created with

    In the case of photoreceptors, the photoactivated visual pigment binds to and activates the trimeric G protein transducin (T) (Figure 4A), causing the exchange of GDP for GTP on its α-subunit, which is part of the Gαt protein family. Following the subsequent dissociation of the α-subunit (Tα) from its β/γ complex (Tβγ), Tα then binds to the cGMP phosphodiesterase (PDE) complex, relieving the inhibition of its catalytic α- and β-subunits by its inhibitory γ-subunits (Ebrey and Koutalos, 2001Burns and Arshavsky, 2005). All these transduction proteins are embedded in or tethered to the disc membranes inside rods or are contained in the cell membrane of cones. As a result of their activation, the hydrolysis of free cGMP in the outer segment is upregulated, causing its rapid decline, and partial or complete closure of the cGMP-gated channels expressed in the rod and cone cell membrane (Luo et al., 2008). The closure of the CNG channels leads to the reduction of the inward transduction current, followed by the hyperpolarization of the cells, and a reduction of neurotransmitter release to second order neurons within the retina. Inversely, in the absence of light, the opening of CNG channels and the resulting inward transduction current is sustained by the continuous cGMP production by guanylyl cyclase (GC).

    Similarly, in OSNs (Figure 4B), the ligand-activated OR proteins bind to the G protein Golf, causing its dissociation into active Gαolf and olfactory β- and γ-subunit, Gβγolf. In contrast to transducin, however, Gαolf is part of the Gαs protein family and binds to adenylyl cyclase 3 (AC3), activating it. As a result, the synthesis of cAMP in the olfactory cilia is upregulated, causing its rapid increase and the opening of cAMP-gated channels (Kleene, 2008Su et al., 2009Boccaccio et al., 2021).

    While both photoreceptors and OSNs use CNG transduction channels, their respective channels have different subunit compositions (Bradley et al., 2005). Rods and cones express heterotetramers consisting of the main A1 and A3 and the modulatory B1a and B3 subunits in 3:1 and 2:2 stoichiometries respectively. The olfactory CNG channel is a heterotetramer consisting of two units of the main A2 subunits and one each of the modulatory A4 and B1b subunits. Interestingly, the rod and the olfactory CNG channels express different splice variants of the same B1 subunit. In OSNs, the initial inward Na+ and Ca2+ current generated by the opening of the CNG channel raises ciliary Ca2+ and opens a secondary ion channel, the Ca2+-activated Cl channel Anoctamin 2. A high intraciliary Cl maintained by the Na+/K+/2Cl cotransporter 1 ensures a Cl efflux which further depolarizes the OSNs (Dibattista et al., 2017Boccaccio et al., 2021). This depolarization triggers the generation of action potentials which further propagate along the axons, inducing glutamate release at synapses with the second order neurons in the olfactory bulb (Murphy et al., 2004). In photoreceptors, the transduction cascade upon stimulation does not ultimately generate action potentials in the receptor cell, but only a graded receptor potential that directly causes a change in neurotransmitter release.




    As for any other sensory modality, proper amplification of the signal is required for the detection of small stimuli and the resulting high sensory sensitivity is critical for the survival and propagation of the species. Nature has reached the highest physically possible sensitivity in the case of rod photoreceptors that can produce a detectable electrical response to the absorption of a single photon. This impressive feat is achieved by employing a transduction cascade that allows tremendous amplification of the signal. During the ~50 ms active lifetime, a single photoactivated rhodopsin molecule activates ~20 transducins, producing an immediate 20-fold amplification (Burns and Pugh, 2010). The following activation of PDE by transducin does not directly produce amplification as each transducin has to bind to a PDE molecule to activate it. However, once activated, each PDE enzyme can hydrolyze thousands of cGMP molecules. Lastly, as the binding of cGMP to the CNG transduction channels is cooperative, a slight change in cGMP levels can reduce the number of cGMP molecules bound to the channel from 3 to 2. This results in channel closure and a sharp reduction in the transduction current, further enhancing the detection of photostimulation. Despite the similarities in the transduction cascades of rods and cones, the amplification in cone photoreceptors is substantially lower as a result of fine-tuning at several of the phototransduction steps (Yau, 1994Kawamura and Tachibanaki, 2008). Interestingly, even though rod and cone visual pigments activate transducin with similar efficiencies, the lower thermal stability of the cone visual pigment results in higher intrinsic activity in cones compared to rods in darkness (Kefalov et al., 2003), effectively desensitizing the cones and shifting their function towards brighter daytime light conditions (see Figure 3C).

    Curiously, the activation of Golf by the OR molecule does not result in amplification. Indeed, the dwell time of the odorant ligand on the OR appears to be very short and on a millisecond timescale (Bhandawat et al., 2005). As a consequence, on average, this results in the activation of less than one G protein per activated receptor. As such, in contrast to phototransduction, where the lifetime of the activated rhodopsin greatly influences the response size and kinetics, in OSNs the response depends more prominently on the coupling efficacy of downstream transduction components while the odorant presence keeps the OR activated. To compensate for the lack of initial amplification at the G protein level, OSNs employ a secondary amplification step on top of the cAMP transduction cascade. The activation of AC3 by Golf results in the synthesis of most likely hundreds of cAMP molecules, the opening of the CNG channels which is followed by a unique secondary amplification based on excitatory Ca2+-activated Cl channels in the cilia (Figure 4B). The Cl current carries up to 80% of the overall transduction current (Dibattista et al., 2017). Physiological experiments with pharmacological and genetic modulation of the Cl conductance indicate that the Cl channels serve to set the length of the action potential train generated in response to odorant stimulation (Pietra et al., 2016) and to promote recognition of novel odorants (Pietra et al., 2016Neureither et al., 2017).

    A puzzling aspect of the secondary amplification step is why Cl is the charge-carrying ion and not Na+, which could be achieved easily by increasing the expression level and/or the ion permeation and conductance of the olfactory CNG channel. Recent theoretical work hinted at two main advantages of Cl, instead of Na+, as the charge carrier. As the external environment of cilia is the nasal mucus, currents will depend on the ion concentration in the mucus, which can be unstable. A current that depends on the intracellular ion concentration, as is the case for Cl but not for Na+, is much less dependent on the mucosal ion concentration. For instance, this could become an issue in the case of a cold with a runny nose or during swimming, when the mucus becomes diluted. The second advantage results from the “compromise” to increase the ciliary surface area, at the expense of having a very small ciliary volume, in the femtoliter range. In such small volumes, even small ionic currents can lead to large changes in ion concentration and osmotic pressures. If the main charge carrier was Na+ this would lead to a large increase (tens of mM). This would cause a large increase in osmotic pressure and also would prevent Ca2+ clearance via the olfactory Na+/Ca2+, K+ exchanger (see below) with greatly deleterious effects. In contrast, high intracellular Cl is maintained throughout the OSN so that its local depletion in the cilia upon ligand activation is rapidly reversed by diffusion from the cell soma. Both these issues do not exist for photoreceptors as they are embedded in the interstitial fluid of the eye and photoreceptors are sufficiently large and their transduction currents are sufficiently small that ion concentration changes due to changes in transduction current are relatively small (Reisert and Reingruber, 2019). Nevertheless, rod photoreceptors undergo osmotically-driven length changes upon light activation, an effect that is mitigated by the translocation of G protein subunits into the cytosol (Zhang et al., 2017).


    Receptor and G Protein Inactivation


    Timely and effective transduction inactivation is critical for allowing sensory neurons to continue to detect stimuli with high temporal resolution. Equally important is to extract behaviorally relevant information from the presented stimuli. In both photoreceptors and OSNs, all active transduction components need to be turned off and the level of cyclic nucleotides within the cells needs to be restored to the rest level before the sensory cell can be reset to the inactive state and become ready for subsequent activation (Figure 5). In the case of photoreceptors, the identity of the step determining the overall kinetics of the photoresponse inactivation was the subject of intense research and debate over several decades. As the visual chromophore ligand is covalently attached to opsin, inactivation of the visual pigment could potentially be extremely slow. Indeed, if left on its own, the active state of rhodopsin decays with a time constant of ~50 s (Imai et al., 2007). Its inactivation in photoreceptors is a two-step process, involving phosphorylation of the rhodopsin C-terminus by rhodopsin kinase (GRK1) which partially quenches its activity, followed by the binding of arrestin1, which completely inactivates the visual pigment (Figure 5A). Though the decay of the active state of cone pigment is significantly faster at ~2 s (Fu et al., 2008), this is still clearly too slow to enable the timely termination and reset of phototransduction. Thus, in both rods and cones, the visual pigments are inactivated by phosphorylation by rhodopsin kinase and the subsequent binding of arrestin long before they would decay spontaneously (Makino et al., 2003). The effective time constant of rod visual pigment inactivation is ~50 ms (Krispel et al., 2006). The slowest step in the inactivation of rod phototransduction turned out to be the hydrolysis of GTP which shuts off Tα, a reaction driven by the transducin GTPase activity and enhanced by a GTPase (GTP-ase activating protein, GAP) complex consisting of Gβ5 and the membrane anchoring protein R9AP (Arshavsky and Wensel, 2013). Inactivation of transducin results in its release from PDE, allowing the two PDE γ inhibitory subunits to resume their inhibition on the two catalytic subunits (α and β) of this enzyme. The kinetics of this reaction determines the overall kinetics of response inactivation in rod photoreceptors. In contrast, work from amphibian cones suggests that in cones the photoresponse duration is Ca2+-dependent and involves the quenching of the cone visual pigment (Matthews and Sampath, 2010).

    Sensory Transduction in Photoreceptors and Olfactory NeuronFig5.svg


    Figure 5. Termination of transduction cascade in rod photoreceptors and OSNs. (A) Schematic representation of the termination of phototransduction in rods. Abbreviations: Phosphorylated light-activated rhodopsin (Rh*-P), arrestin (ARR), G protein-coupled receptor kinase 1 (GRK1), Tα, β, and γ subunits of the retinal G protein, transducin (T), guanosine-5′-triphosphate (GTP), guanosine-5′-diphosphate (GDP), phosphodiesterase (PDE), guanosine monophosphate (GMP) and cyclic guanosine monophosphate (cGMP), and cyclic nucleotide-gated (CNG) channel, guanylate cyclase (GC), guanylate cyclase activating protein (GCAP), cyclic nucleotide-gated (CNG) channels, K+ dependent Na+/Ca2+ exchanger 1, 2 and 4 (NCKX1, NCKX2, NCKX4). (B) Schematic representation of the termination of the olfactory transduction cascade. Abbreviations: Olfactory receptor (OR), arrestin (ARR), G protein-coupled receptor kinase 3 (GRK3) Gαolfβ, and γ, subunits of the olfactory G protein, guanosine-5′-triphosphate (GTP), guanosine-5′-diphosphate (GDP), adenylyl cyclase 3 (AC3), activated phosphodiesterase 1C (PDE1C), cyclic adenosine monophosphate (cAMP), adenosine monophosphate (AMP), Ca2+/calmodulin-dependent protein kinase II (CaMKII), cyclic nucleotide-gated channel (CNG); Ca2+-activated Clchannel anoctamin 2 (ANO2), K+-dependent Na+/Ca2+ exchanger 4 (NCKX4). Images created with

    In OSNs, the inactivation by phosphorylation and arrestin are potentially not needed for the timely shutoff of the olfactory transduction cascade, due to the extremely short lifetime of the active ligand-bound receptor molecule. Early biochemical experiments suggested that OR phosphorylation does control cAMP kinetics (Dawson et al., 1993Schleicher et al., 1993Peppel et al., 1997), but it seems to play little, if any, role in the control of odorant-response kinetics for one particular OR, mOR-EG (Kato et al., 2014). It still remains to be established whether this applies to all ORs, or whether a subset of ORs is subject to phosphorylation and inactivation. β-arrestin interacts with ORs, mediating internalization during prolonged stimulation and altering adaptation to repetitive odor stimuli (Mashukova et al., 2006). Experiments on isolated human and rat OSNs suggested a role for protein kinases A (PKA) and C (PKC) in the termination of the olfactory response. Ca2+ imaging showed that the inhibition of PKA and PKC increases intracellular Ca2+ responses in the presence of odorant mixtures, and blocks their termination after odorant stimulation ceases. While the inhibition of both PKA and PKC modulated the odor-induced intracellular Ca2+ increase in the human OSNs, only PKC and not PKA affected the Ca2+ response to odorants in rat OSNs, suggesting differences among species in the termination of the olfactory response (Gomez et al., 2000).

    The control of the lifetime of the olfactory G protein seems to be more complex and less well understood compared to phototransduction. Ric-8B (resistant to inhibitors of cholinesterase-8B) has been identified as a GTP exchange factor (GEF) expressed in OSNs, which facilitates the exchange of GDP for GTP on Gαolf and its activation. Unusually, Ric-8B not only interacts with the G protein α-subunit, but also with γ13, the olfactory γ-subunit. In a heterologous system, Ric-8B co-expression with olfactory transduction components can greatly increase cAMP production, suggesting that it could indeed modulate olfactory transduction (Von Dannecker et al., 2005Kerr et al., 2008). A knockout mouse for Ric-8B displays impaired olfactory behavior, and, surprisingly, greatly reduced odorant responses. Ric-8B is localized primarily in the cell body and the dendritic knob of OSNs. Ric-8B knockout OSNs are devoid of Gαolf (Machado et al., 2017), suggesting that this gene is needed for the stable expression of Gαolf, and excludes addressing its potential role as a GEF in the odorant response. The Ric-8B knockout mice also display higher OSN cell death. Regulators of G protein signaling (RGS) are GAPs that modulate the lifetime of an activated G protein as described above. RGS2, instead of functioning as a GAP, directly inhibits AC3 to control the size of the odorant response (Sinnarajah et al., 2001). However, inconsistent and contradictory data on RGS2 and RGS3 expression and their roles in OSNs suggest that more research is needed (Norlin and Berghard, 2001Kanageswaran et al., 2015Saraiva et al., 2015).




    Adaptation plays a critical role in the capacity of our sensory neurons to remain able to detect stimuli above the background in a complex and rapidly changing environment. For instance, in constant light conditions, the dynamic range for both rods and cones is only 100-fold, spanning a range from threshold stimulation to saturation (Figure 3C). However, as a result of light adaptation, photoreceptors can shift their functional range over a very wide span of light conditions, ranging from cloudy night to sunrise for rods, and starry night to bright sunny day for cones (Weale, 1961). Thus, using the adaptation of individual photoreceptors, the visual system is able to remain responsive to stimuli over a wide range of light conditions. In contrast, the ability of OSNs to adapt is rather limited even at modest levels of background odorant (Reisert and Matthews, 1999). Nevertheless, increasing concentrations of the same odorant are able to recruit less sensitive ORs, and therefore less sensitive OSNs, preserving its perception at higher concentrations and ensuring to report the presence of that odorant to the brain.

    In both types of sensory neurons, adaptation is mediated by a change in Ca2+ upon stimulation. This change is sensed by several Ca2+-binding proteins that trigger a negative feedback on the vision and olfaction transduction cascades by modulating several of their steps. In the outer segments of rods and cones and in olfactory cilia, Ca2+ levels are controlled by the balance of influx via the CNG channels, whose current is carried in part by Ca2+, and efflux via Na+/Ca2+, K+ exchangers (NCKXs) that use the electrochemical gradient for Na+ and K+ to extrude Ca2+ (Figure 5Yau and Nakatani, 1984). In rods (Figure 5A), this task is accomplished by the rod-specific NCKX1, whereas cones employ two separate exchangers, NCKX2 and NCKX4 (Vinberg et al., 2017). At rest, both in darkness and in steady state light, the influx of Ca2+ is matched with its extrusion and, as a result, the level of free Ca2+ in the outer segments is maintained constant. Upon photostimulation, the transduction cascade is activated, resulting in depletion of cGMP, closure of CNG channels, and reduction in the influx of Ca2+ into the outer segments. However, Ca2+ extrusion by the Na+/Ca2+, K+ exchangers carry on for at least a while and, as a result, the level of Ca2+ in the outer segments declines. Direct Ca2+ measurements in amphibian photoreceptors indicate a dynamic range from 670 to 30 nM in rods (Sampath et al., 1998) and 400–5 nM in cones (Sampath et al., 1999), in darkness and bright light, respectively.

    The light-driven decline in Ca2+ causes its release from several Ca2+-binding proteins. The dominant Ca2+-dependent feedback mechanism in both rods and cones controls the synthesis of cGMP by membrane-bound GC via a pair of GC activating proteins (GCAPs)—GCAP1 and GCAP2. When Ca2+ in the outer segments is high, Ca2+-bound GCAPs bind to and partially inhibit the activity of GC. Upon photoactivation and the decline in Ca2+, GCAPs become Ca2+-free and release from GC, resulting in upregulation of cGMP synthesis which restores the dark current after photostimulation and modulates the activation of the transduction cascade in the presence of background light (Dizhoor, 2000Sakurai et al., 2011). Another mechanism by which Ca2+ modulates phototransduction involves the Ca2+-binding protein recoverin. As GCAPs, recoverin is a member of the EF-hands protein family, and when bound to Ca2+ in darkness, it inhibits rhodopsin kinase, thus slowing down the inactivation of the visual pigment (Makino et al., 2004Sakurai et al., 2015). When the photoreceptors are activated and Ca2+ declines, it is released from recoverin, which in turn dissociates from rhodopsin kinase and relieves its inhibition. This enhances the phosphorylation of visual pigments and accelerates their inactivation, effectively reducing the activation of the transduction cascade by the background light. Finally, direct modulation of the CNG channels has also been suggested. However, in the case of rods, such modulation appears to play a marginal, at best, role (Koutalos and Yau, 1996) and is not mediated by the Ca2+-binding protein calmodulin (Chen et al., 2010). In zebrafish cones, the modulation of the CNG channels appears to play a more substantial role and is mediated by the Ca2+-binding protein CNG-modulin (Korenbrot et al., 2013). It is still unclear whether the mammalian homolog of CNG modulin, EML1 plays a similar role in mammalian cones.

    Adaptation in OSNs is less well understood compared to phototransduction. Early data, mostly of biochemical nature or obtained from heterologously-expressed proteins of interest, suggested three main molecular targets for adaptation. All three of them are mediated by the Ca2+ influx during the odorant response: Ca2+/calmodulin-mediated desensitization of the olfactory CNG channel to close the channel even in the presence of high cAMP (Chen and Yau, 1994); phosphorylation via CaM-kinase 2 of AC3 to reduce the rate of cAMP production (Wei et al., 19961998); and Ca2+-mediated upregulation of phosphodiesterase 1C, which is expressed in olfactory cilia, and is assumed to degrade cAMP to AMP to terminate the response (Borisy et al., 1992). Follow-up experiments using recordings from OSNs all seem to indicate that none of these mechanisms plays as prominently or as originally thought of role in transduction (Reisert and Zhao, 2011). A mouse with a mutation in the CNGB1b channel subunit that entirely prevents desensitization by Ca2+ surprisingly displays normal olfactory adaptation but instead shows a delayed response termination, suggesting that Ca2+/calmodulin-mediated desensitization of the CNG channel speeds up response termination (Song et al., 2008). A mouse model that carries a mutation in AC3 that prevents phosphorylation does not show a discernable phenotype of the olfactory response (Cygnar et al., 2012), although it might be possible that other, unknown phosphorylation sites in AC3 might be important. Finally, a knockout mouse for PCE1C has no deficits in response termination but instead shows much reduced response amplitudes for unclear reasons (Cygnar and Zhao, 2009). This begs the obvious question as to what the role of PDE1C might be and what might actually happen to cAMP that is generated during the odorant response. For the latter, an interesting option is that cAMP diffuses out of the cilia into the cell body as a means to reduce ciliary cAMP, allowing OSNs to recover from stimulation (Cygnar and Zhao, 2009). One aspect that is reasonably understood is NCKX4, the Ca2+ exchanger in OSNs that is required to lower intraciliary Ca2+ during and after odorant stimulation, allowing the transduction cascade to recover from adaptation (Reisert and Matthews, 1998Stephan et al., 2012).




    Disorders affecting photoreceptors are among the leading causes of blindness in the human population. One of the prevalent visual disorders, called retinitis pigmentosa, is a complex disease caused by a wide range of mutations in photoreceptors. Many of these mutations affect the expression, structure, and function of the rod visual pigment (Athanasiou et al., 2018). Because of the very high expression of opsin in the outer segments of rods, this protein plays not only a functional role, but is also critical for the proper formation of the outer segment itself. As a result, mutations affecting the expression, folding, or targeting of opsin to the rod outer segments, cause gradual degeneration of the rods. Other genes implicated in rod dysfunction and degeneration include these for phosphodiesterase (e.g., rd1, rd10; Chang et al., 2002), the CNG channels A and B subunits (channelopathies; Michalakis et al., 2018), GC, and GCAPs (Olshevskaya et al., 2002). Another diverse set of visual disorders is caused by abnormal chromophore production or supply to photoreceptors, which limits the ability to detect light and can also lead to degeneration (Ku and Pennesi, 2020). Notably, the efficiency of the visual system to produce chromophore seems to decline with age, which may result in poor rod function in dim light even in normally aging adults. It is also an early indicator for age-related macular degeneration, a devastating blinding disorder that affects the function of cones in the central retina responsible for acute vision and color discrimination (Jackson et al., 2002). Interestingly, rods and cones seem to coexist synergistically in the retina, and diseases caused by rod-specific mutations that result in rod degeneration, eventually lead to the loss of cones and central vision as well. Thus, considerable efforts are currently focused on developing methods for preserving rods even when they are not functional, as a way of protecting daytime cone-driven vision. Because the eye is a relatively accessible organ, novel therapeutic approaches for vision protection and restoration have led the field, with successful examples of gene therapy and stem cell therapy in experimental and clinical trial phases (Ovando-Roche et al., 2017).

    Compared to vision, in olfactory transduction, very few mutations in transduction components are known that lead to deleterious effects. Several aspects might account for this. Mutations causing a partial reduction of olfaction might go unnoticed in the human population as very little systematic olfactory testing is done. OSNs regenerate throughout life and only have a lifespan of a few weeks. Hence any slow degeneration as those seen in photoreceptors might not manifest in that time window. In a screen of families with congenital anosmia, no potentially causative mutations were found in three main transduction proteins (Gαolf, CNGA2, AC3), with these genes also being under purifying selection (Feldmesser et al., 2007). An interesting exception are patients suffering from retinitis pigmentosa, which is caused by mutations in the gene encoding the CNGB1 subunit expressed in both rods and OSNs. Those patients, identified because of their visual function decline, were found to be hyposmic or anosmic when tested for their olfactory ability (Charbel Issa et al., 2018). If congenital anosmia is considered to be a relatively rare and little understood condition, more known and frequently detected are specific anosmias, which manifest in the inability to detect certain odorants (Keller et al., 2007Trimmer et al., 2019). Broadly speaking, this is the olfactory equivalent of color blindness, and is caused by known OR mutations.

    Arguably, the most common causes of smell loss are events that lead to the destruction of the olfactory epithelium and/or the olfactory nerves connecting it to the central nervous system (CNS). These events include head or face trauma, inhalation of toxic chemicals, or viral infection (as SARS-CoV2), and, neurodegenerative diseases such as Alzheimer’s and Parkinson’s disease (Attems et al., 2015Cooper et al., 2020). In the former, the origin of the smell disorder can be tracked down to the periphery, the olfactory epithelium. In the case of neurodegenerative diseases, it has been thought that olfactory dysfunction originates centrally in the CNS, but it is becoming clearer now that peripheral olfaction can be affected in these cases as well, although the respective mechanisms have not been fully elucidated.



    G Protein-Coupled Receptors in Taste Physiology and Pharmacology


    Ahmad and Dalziel.  Front. Pharmacol., 30 November 2020 |  Creative Commons Attribution License (CC BY).

    • Food Nutrition and Health Team, Food and Bio-based Products Group, AgResearch, Palmerston North, New Zealand

    Heterotrimeric G protein-coupled receptors (GPCRs) comprise the largest receptor family in mammals and are responsible for the regulation of most physiological functions. Besides mediating the sensory modalities of olfaction and vision, GPCRs also transduce signals for three basic taste qualities of sweet, umami (savory taste), and bitter, as well as the flavor sensation kokumi. Taste GPCRs reside in specialised taste receptor cells (TRCs) within taste buds. Type I taste GPCRs (TAS1R) form heterodimeric complexes that function as sweet (TAS1R2/TAS1R3) or umami (TAS1R1/TAS1R3) taste receptors, whereas Type II are monomeric bitter taste receptors or kokumi/calcium-sensing receptors. Sweet, umami and kokumi receptors share structural similarities in containing multiple agonist binding sites with pronounced selectivity while most bitter receptors contain a single binding site that is broadly tuned to a diverse array of bitter ligands in a non-selective manner. Tastant binding to the receptor activates downstream secondary messenger pathways leading to depolarization and increased intracellular calcium in TRCs, that in turn innervate the gustatory cortex in the brain. Despite recent advances in our understanding of the relationship between agonist binding and the conformational changes required for receptor activation, several major challenges and questions remain in taste GPCR biology that are discussed in the present review. In recent years, intensive integrative approaches combining heterologous expression, mutagenesis and homology modeling have together provided insight regarding agonist binding site locations and molecular mechanisms of orthosteric and allosteric modulation. In addition, studies based on transgenic mice, utilizing either global or conditional knock out strategies have provided insights to taste receptor signal transduction mechanisms and their roles in physiology. However, the need for more functional studies in a physiological context is apparent and would be enhanced by a crystallized structure of taste receptors for a more complete picture of their pharmacological mechanisms.




    G protein-coupled receptors (GPCRs) are the largest and the most diverse group of membrane receptors in eukaryotes. They are activated by a wide variety of ligands in the form of light energy, lipids, sugars, peptides and proteins (Billington and Penn, 2003Schoneberg et al., 2004Lundstrom, 2009) which convey information from the outside environment into the cell to mediate their corresponding functional responses. The conformational changes of GPCRs upon ligand binding initiate a series of biochemical reactions within the cell. These intracellular reactions regulate sensory functions of smell, taste, and vision, and a wide variety of physiological processes such as secretion, neurotransmission, metabolism, cellular differentiation, inflammation and immune responses (Lagerström and Schiöth, 2008Rosenbaum et al., 2009Venkatakrishnan et al., 2013Ahmad et al., 2015). Taste is one of the most important sensations for human life, enabling us to perceive different tastes from the diverse range of food available in nature and is a major determinant of our ingestion decisions.

    The anatomical units of taste detection are taste receptor cells (TRCs) that are assembled into taste buds distributed across different papillae of the tongue and palate epithelium. Taste processing is first achieved at the level of TRCs that are activated by specific tastants. They transmit information via sensory afferent fibers to the gustatory cortex in the brain for taste perception (Figure 1). Three different morphologic subtypes of TRCs in taste buds sense the different tastes we perceive. Type I glial-like cells detect salty taste while type II cells expressing GPCRs detect sweet, umami, and bitter tastes. Type III cells sense sour stimuli (Janssen and Depoortere, 2013).



    FIGURE 1. Schematic diagram shows taste signal transmission between tongue and brain. Taste buds present in different papillae in tongue and palate contain taste receptor cells (TRC) which contain taste G protein-coupled receptors (GPCRs). Left side shows how afferent nerves transmit a signal to the gustatory cortex in brain via cranial/glossopharyngeal nerves. Right side shows taste bud with taste TRCs and simplified signal transduction pathway of taste receptor where taste GPCRs are activated by a tastant that in turn recruits a specific G protein that further induces intracellular calcium release (created with

    Sweet and umami stimuli are transduced by Type 1 taste GPCRs while bitter taste is sensed by Type 2 taste GPCRs (Figure 2Table 1). The more recently described kokumi sensation is mediated by another GPCR, the calcium-sensing receptor (CaSR) (Figure 2Table 1). Taste GPCRs are activated by specific taste ligands present in foods and recruit G proteins to activate downstream signaling effectors (Figure 3).



    FIGURE 2. Schematic representation of different types of taste receptor cells (TRCs) in taste bud with their attributed taste modalities and signal transduction. Type I TRCs exhibit a support function similar to glial cells and express enzymes and transporters that remove extracellular neurotransmitters (Lawton et al., 2000Bartel et al., 2006Vandenbeuch et al., 2013), and ion channels linked with the redistribution and spatial buffering of K+ (Dvoryanchikov et al., 2009). A subpopulation of type I cells are thought to be involved in low salt taste perception (Vandenbeuch et al., 2008) but this remains to be confirmed. Type II TRCs are receptor cells and express G protein-coupled receptors (GPCRs) on their surface that respond to sweet, umami and bitter tasting stimuli. The type II TRCs are fine-tuned and express either type 1 (TAS1R2/TAS1R3: sweet and TAS1R1/TAS1R3: umami) or type 2 taste (TAS2Rs; bitter) GPCRs and correspondingly respond to sweet/umami or bitter stimuli (Matsunami et al., 2000DeFazio et al., 2006Yoshida et al., 2009) (see also Table 2 for classification). Moreover, three isoforms of type 1 taste GPCRs (TAS1R1, TAS1R2 and TAS1R3) are often co-expressed and responses to both sweet and umami stimuli can be detected in the same cell (Kusuhara et al., 2013). Interestingly, recent studies reported a novel subpopulation of cells with type II TRCs that transduce a signal in response to high salt concentrations (>150 mM) (AI) (Roebber et al., 2019). Type III TRCs are the least abundant and sense sour stimuli through the proton selective channel, otopterin 1 (Tu et al., 2018Zhang et al., 2019). As a consequence of expressing several synaptic proteins, they are termed presynaptic cells (DeFazio et al.,2006). Although both Type II and Type III TRCs require action potentials for transmitter release, their working mechanisms are quite different. Whereas, type III TRCs use a conventional synapse and SNARE mechanism like that in neurons to affect the release of synaptic vesicles, type II TRCs rely on action potentials to trigger the release of ATP through voltage gated channels (DeFazio et al., 2006Vandenbeuch et al., 2013) (see also Figures 13) (created with





    TABLE 1. Taste GPCRs classification and their downstream signaling regulators.



    FIGURE 3. Schematic representation of signal transduction pathway of sweet, umami, bitter and kokumi-calcium sensing receptors (CaSR) in taste receptor cells on the tongue. Ligand-induced stimulation of the sweet (TAS1R2/TAS1R3), umami (TAS1R1/TAS1R3), bitter receptors (TAS2Rs) and kokumi sensation expressed in type II taste cells within taste bud activates a trimeric G protein composed of α-gustducin (Gα-gust) in sweet, umami, bitter and Gα-q/11 in kokumi-receptor and a complex consisting of Gβγ proteins. The released Gβγ-complex activates phospholipase C isoform β2 (PLCβ2) which then induces production of inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (DAG); the second messenger IP3, in turn activates the IP3 receptor (IP3R), an intracellular ion channel that allows Ca2+ release from the intracellular endoplasmic reticulum (ER store). Increase in intracellular Ca2+ then activates the complex of transient receptor potential cation channel subfamily M member 4 and 5 (TRPM4/5) that are plasma membrane localized sodium-selective channels which leads to depolarization and subsequent activation of voltage-gated sodium channels (VGSC). The combined action of increased Ca2+ and membrane depolarization activates the complex of calcium homeostasis modulator 1 and 3(CALHM1/3) channel and pannexin1 channels, thus resulting in the release of the neurotransmitter ATP. Increased ATP, in turn activates P2X ionotropic purinergic receptors 2 and 3 (P2X2/P2X3) on afferent cranial nerve generating an action potential that subsequently signals to the gustatory cortex for sensory perception. Besides well-known taste GPCR pathways, connecting proteins semaphorin 7A (Sem 7A) and 3A (Sem 3A) are depicted in close contact with sweet and bitter receptors as they provide instructive signals that fine tune to sweet or bitter ganglion neurons, respectively. VFT, venus flytrap domain; CRD, cystine rich domain; ECD, extracellular domain. (created with

    In this review, we will first explore the basic architecture of the gustatory sensory system and its peripheral signal transmission. Then we will discuss taste GPCR signal transduction mechanisms for the different taste modalities, their molecular structure, and the conformational changes that occur following orthosteric/allosteric binding of endogenous and food-derived ligands.


    Taste Buds and Neural Transmission


    In mammals, taste buds on the tongue comprise 50–100 elongated epithelial cells and a small number of proliferative basal cells (Sullivan et al., 2010). Ultrastructural studies and patterns of gene expression with cell function reveal three distinct anatomical types of TRCs within each taste bud: Type I, Type II and Type III (Murray, 1986) (refer to Figure 2Table 2).


    TABLE 2. Summary of taste receptor cell characteristics.

    Type II TRCs express either sweet, umami, or bitter taste receptors at their cell surface. These receptors share some commonality to their signal transduction mechanisms that are intrinsic to TRCs. Taste GPCRs (sweet, umami and bitter) couple to heterotrimeric G proteins that include Gα-gustducin, Gβ3, and Gγ13 (Huang et al., 1999) and initiate a series of signal transduction cascades involving activation of phospholipase C-β2 (PLCB2), production of inositol-1,4,5-triophosphate (IP3), and IP3-dependent Ca2+ release from the endoplasmic reticulum (ER) via the IP3 receptor (IP3R). The increased intracellular [Ca2+]i then activates the transient receptor potential cation channel subfamily M member 4 and 5 (TRPM4/5) in the basolateral plasma membrane, leading to membrane depolarization that triggers Na+ action potential firing, and depolarization-induced release of ATP. In turn, ATP acts as the primary neurotransmitter stimulating purinergic receptors 2 and 3 (P2X2 and P2X3) on afferent cranial nerves whose activation triggers an action potential which subsequently activates the gustatory cortex in the brain (McLaughlin et al., 1992Wong et al., 1996Margolskee, 2002). α-gustducin is a distinct G protein selectively expressed in ∼30% of type II TRCs and shares 80% identity with retinal protein α-transducin (McLaughlin et al., 1992) and is a key contributor to signal transduction for sweet and bitter taste receptors (McLaughlin et al., 1992Wong et al., 1996Margolskee, 2002).

    An important aspect of taste transduction is how ATP signaling is conducted. Recent studies have discovered that calcium homeostasis modulators 1 and 3 (CALHM1/3) are enriched in type II TRCs where they interact and form a functional complex. Their genetic deletion abolishes responses to sweet, bitter and umami tastes, supporting the requirement of the CALHM1/3 complex as an ATP release channel for the GPCRs mediated tastes (Taruno et al., 2013Ma et al., 2018).

    New information has provided insight into how specific taste qualities are fine-tuned in order to recognize their partner ganglionic neurons in the brain. Lee et al. (2017) discovered semaphorin proteins, 7A and 3A as the physical links between sweet and bitter TRCs, respectively, and their partner ganglion neurons in the brain. It remains to be determined what physical links exist between umami TRCs and their corresponding neurons in the brain. Delineating the underlying molecular basis for this interaction would provide further understanding of purinergic transmission in the taste system. In addition, whether these mechanisms are relevant for kokumi sensation has not yet been investigated, despite CaSR having distinct expression in TRCs and significant functional synergy with other prominent taste qualities (sweet, umami and salty). Moreover, there is still debate regarding the recognition of kokumi as a sixth taste entity, consequently the calcium sensing receptor (CaSR) is not yet included in the nomenclature for any subtypes of taste GPCRs, although it would best fit with Type 1 taste receptors.


    Type 1 Taste G Protein-Coupled Receptors (Sweet and Umami)


    The type 1 taste receptors (TAS1Rs) belong to the class C GPCRs, which possess a large N-terminal extracellular domain (ECD) fused to the heptahelical seven transmembrane domain (TMD). The ECD is further divided into two ligand-binding domains (LBD1 and LBD2), having a bi-lobed structure called a Venus flytrap domain (VFT) due to its resemblance to this shape (Hoon et al., 1999). With the exception of GABAB receptors, a cysteine-rich domain (CRD) connects the VFT to the TMD (Leach and Gregory, 2017).

    In contrast to other receptors from this class C of GPCRs, such as the metabotropic glutamate receptor (mGluR) or γ-aminobutyric acid type B receptors (GABABRs) which function as homo- or heterodimers, respectively (Jones et al., 1998Kaupmann et al., 1998White et al., 1998Kunishima et al., 2000), the TAS1Rs function as obligatory heterodimers. The distinct expression pattern of TAS1R1 and TAS1R2 in different subsets of murine cells led to the idea that they could detect two different taste profiles. However, following the discovery of the TAS1R3 subtype, it was clear that when TAS1R1 heterodimerizes with TAS1R2, the receptor detects sweet taste substances (Nelson et al., 2001Ohkuri et al., 2009Kim et al., 2017). On the other hand, if heterodimerized with TAS1R3 (TAS1R1/TAS1R3), it is responsible for umami or amino acid taste detection (Li et al., 2002Nelson et al., 2002). Please refer to figure 4A for the basic structure of sweet and umami receptors.


    FIGURE 4. Figure depicting basic structural features of sweet/umami/kokumi dimeric receptor (A) and monomeric bitter receptor (B) (created with


    Sweet Taste Signal Transduction Mechanisms


    The TAS1R2/TAS1R3 receptor recognizes a wide variety of sweet substances including natural sugars, artificial sweeteners, amino acids and proteins (Li et al., 2002Xu et al., 2004Jiang et al., 2005aJiang et al., 2005c) (Table 3). This was demonstrated in studies using heterologous expression systems as well as knockout mice for TAS1R2 and/or TAS1R3 subtypes that showed a blunted response to sugars, sweeteners, and D-amino acids, confirming the TAS1R2/TAS1R3 heterodimer as the main sweet taste receptor in vivo (Li et al., 2002Zhao et al., 2003Xu et al., 2004).


    TABLE 3

    TABLE 3 | Agonists of sweet taste receptor along with their EC50 values.




    Binding pocket

    EC50 (mM)



    Natural carbohydrate

    VFT (TAS1R2 and TAS1R3)


    (Li et al., 2002; Servant et al., 2010; Zhang et al., 2010; Zhang





    et al., 2003)



    VFT (TAS1R2)


    (Li et al., 2002; Liu et al., 2011; Masuda et al., 2012)



    VFT (TAS1R2)


    (Li et al., 2002; Masuda et al., 2012)



    TMD (TAS1R3)


    (Xu et al., 2004; Jiang et al., 2005c)



    CRD (TAS1R3)


    (Li et al., 2002; Jiang et al., 2004; Ide, et al., 2009; Masuda et al.,








    CRD (TAS1R3)


    Masuda et al., 2012; Jiang et al., 2004



    VFT (TAS1R3), VFT (TAS1R2)


    Koizumi et al., 2007; Jiang et al., 2004



    VFT (TAS1R2)


    (Jiang et al., 2004; Koizumi et al., 2007)


    N sulfonyl amide

    VFT (TAS1R2)


    (Li et al., 2002; Masuda et al., 2012; DuBois, 2016)

    Suosan, cyanosuasan


    VFT (TAS1R2)


    (Tinti and Nofre, 1991; Du Bois, 2016)


    Guanidinoacetic acid

    VFT (TAS1R2)


    (DuBois, 1995; Sanematsu et al., 2014)


    Halogenated carbohydrate

    VFT (TAS1R2 and TAS1R3)


    (Li et al., 2002; Masuda et al., 2012)

    Acesulfame K

    Sulfamate ester

    VFT (TAS1R2)


    (Li et al., 2002; Masuda et al., 2012)


    Oxime, ethoxyphenyl urea, alkoxyaryl urea,

    TMD (TAS1R2)


    (Li et al., 2002; Servant et al., 2010)


    Ethoxyphenyl urea

    TMD (TAS1R2)


    (Servant et al., 2010)


    Alkoxyaryl urea

    TMD (TAS1R2)


    (Zhang et al., 2008)


    Amino acid

    VFT (TAS1R2)


    (Li et al., 2002; Masuda et al., 2012)

    Xyletol, sorbitol


    VFT (TAS1R2)


    (Mahalapbutr et al., 2019)

    Maltotriose, acarbose

    Oligosaccharide, pseudotetrasaccharide



    (Pullicin et al., 2017; Pullicin et al., 2019)

    Where VFT, venus ytrap domain; TMD, transmembrane domain; ND, not determined.


    TABLE 3. Agonists of sweet taste receptor along with their EC50 values.

    The sweet receptor couples to heterotrimeric Gα-gustducin which includes Gβ3 and Gγ13, as mice lacking Gα-gustducin showed a reduced response to sweet substances either natural or artificial (McLaughlin et al., 1992Wong et al., 1996Margolskee, 2002). Moreover, a point mutation in the C-terminal region of gustducin (G352P) (critical for its receptor interaction) results in loss of its ability to activate taste GPCRs while keeping other functions intact. Further, G352P acts as a dominant negative to block heterotrimeric G protein interaction with taste receptors and disrupts the responses to sweet and bitter compounds in both wild type (WT) and null mice (Ruiz-Avila et al., 2001). In addition, the G352 mutant further reduces any residual sweet/bitter taste responses of the null mice by acting as a “βγ sink” to bind all unbound βγ-subunits and remove them from the viable pool of G protein heterotrimers available to the receptor (Ruiz-Avila et al., 2001). These observations confirm the essential requirement of Gα-gustducin in sweet and bitter taste transduction.

    In addition to the Gα-gustducin pathway, sweet taste transduction occurs via two additional signaling pathways involving different secondary messengers. The first one involves cAMP and the second one involves IP3. Normally, sugars elevate the level of cAMP, while sweeteners stimulate IP3 production (Tonosaki and Funakoshi, 1988Uchida and Sato, 1997). Sucrose or other sugars bind to either TAS1R2 or TAS1R3 and recruit Gαs protein that leads to increased cAMP levels which initiates the influx of cations through ion channels. Alternatively, cAMP activates protein kinase A that leads to TRC cell depolarization resulting in an influx of calcium ions and neurotransmitter release (Avenet et al., 1988Tonosaki and Funakoshi, 1988Margolskee, 2002). Sweetener binding to the TAS1R2/TAS1R3 heterodimer recruits Gα-gustducin proteins that stimulate PLCβ2 which in turn hydrolyzes phosphatidylinositol 4,5-bisphosphate (PIP2) to diacylglycerol (DAG) and IP3 (Margolskee, 2002Chandrashekar et al., 2006). IP3R3 (Hisatsune et al., 2007) induced Ca2+ release from ER stores (Figure 3) activates TRPM5 (Zhao et al., 2003Hisatsune et al., 2007Dutta Banik et al., 2018) that leads to an action potential (Yoshida et al., 2005Yoshida et al., 2006) and subsequent release of neurotransmitters.

    Interestingly, Dutta Banik et al. (2018) confirmed that TRPM4 also mediates taste signaling independent of TRPM5, and knocking out both channel proteins (TRPM4/5) abolishes the sweet, umami and bitter taste response completely. This revealed another layer of complexity to sweet signal transmission. This in-depth mechanistic research has increased our understanding of sweet and bitter receptors and presents a challenge to dissect the taste signal transmission pathways for umami and kokumi as well.


    Structural, Molecular and Conformational Changes of Sweet Receptor


    Since, the sweet taste receptor has not yet been crystallized, determining the structure of the sweetener binding site and mechanism of activation has been a challenge. Based on homology with other class C GPCRs (mGluRs and GABABRs), multiple studies propose similar activation mechanisms for the sweet receptor (Kunishima et al., 2000Tsuchiya et al., 2002Jingami et al., 2003Muto et al., 2007Perez-Aguilar et al., 2019). The many different sweet agonists and their diverse binding sites across receptor domains (VFT, TMD and CRD) (Table 3) may explain its complex yet broadly tuned nature. For example, a single residue in VFT (I60) of TAS1R3 of the TAS1R2/TAS1R3 heteromer is required for a saccharin preference in in-bred mouse strains (Max et al., 2001Reed et al., 2004).

    Several studies utilizing homology and computational modeling based on the crystal structure of mGluR and GABABRs have predicted structural and functional aspects of orthosteric and allosteric binding sites for the sweet receptor (Kim et al., 2017Cheron et al., 2019Park et al., 2019). They reported that both VFT regions undergo ligand dependent conformational changes and intersubunit interactions between ECDs that further stabilize heterodimer formation for subsequent downstream signaling (Perez-Aguilar et al., 2019). Binding of orthosteric agonists to VFT of TAS1R2 leads to major conformational changes that form a TMD6/TMD6 interface between TMDs of TAS1R2 and TAS1R3, which is consistent with the activation process observed biophysically on the mGluR2 homodimer. The initial role of the bound agonist is to pull the bottom part of VFT3 (VFT of TAS1R3) toward the bottom part of VFT2 (VFT of TAS1R2) in order to transmit this movement from VFT2 (where agonists bind) through the VFT3 and the CRD3 (VFT and CRD of TAS1R3) to the TMD3 (TMD of TAS1R3). This facilitates G protein coupling and downstream signaling. The CRDs are crucial in this streamlined relay of structural changes where disulfide bonds provide rigidity to the CRD and amplify the mechanical constraints that help in attaining an active conformation (Cheron et al., 2019). This is empirically supported by a study in which a single mutation (A537P) in the CRD of TAS1R3 abolished the response to all sweeteners, indicating that the CRD3 must couple ligand binding in VFT2 to the conformational changes required in TMD3 for receptor activation.

    Trafficking and cell surface expression are also crucial factors for sweet taste transduction. Molecular modeling with mutagenesis scanning revealed specific regions consisting of hydrophobic residues in ECD (site II; at the tip of CRD) and TMD regions (site IV; includes TMD6 and the cytoplasmic base of TMD5) of the TAS1R2 subunit to be important for dimerization with TAS1R3. Moreover, the CRD region and ECL2 domain of the transmembrane region seem to be important for surface co-expression of the TAS1R2/TAS1R3 dimer. In particular, the cytosolic C-terminus portion of the CRD region of TAS1R2 needs to be properly folded for coexpression and trafficking (Park et al., 2019). This reflects the difficulty in expressing these receptors at consistent levels in mammalian cell lines (Li et al., 2002Shimizu et al., 2014).


    Positive Allosteric Modulation of Sweet Receptor


    Class C GPCRs pose an ideal target for allosteric modulation either positive (PAM) or negative (NAM). PAMs show little or no agonist activity on their own but significantly enhance agonist activity. Sweet taste is a major target of the food industry globally and non-caloric sweeteners are highly sought to exploit a huge commercial market. In a first comprehensive high throughput screen by Servant et al. (2010), novel PAMs (SE1, SE2, SE3; Table 4) for the sweet heteromer were reported that were not sweet on their own but significantly enhanced the sweetness of sucralose or sucrose. Agonist binding to the VFT region of TAS1R2, facilitates a closed conformation which constitutes an active state of the sweet receptor, while its open conformation represents an inactive state. Molecular modeling and mutagenesis studies revealed that these PAMs follow a similar mode of binding as that reported for umami PAMs (IMP and GMP). They bind near the opening of the binding pocket of the VFT region adjacent to their agonists, through Van der Waals and hydrogen bonding interactions, and utilize several critical residues for their activity. Although these residues are not in direct contact with any receptor bound sweetener, mutation of some of them (K65, Y103, L279, D307, and R383) diminishes the response to sweeteners suggesting that these residues normally stabilize the closed conformation. Initial closing of the VFT region by agonist binding and further stabilization of the closed conformation by subsequent binding of SE modulators occurs in two steps. First, by interacting with the ECD region of TAS1R2, and second, by strengthening the hydrophobic interactions between the two lobes of ECD and lowering the free energy needed for their closure (Zhang et al., 2010).


    TABLE 4 | Sweet taste receptor’s positive allosteric regulators with concentration (used in cell based assays in studies) and negative allosteric modulators with their IC50 values.

    Positive allosteric modulators (PAMs)


    Binding pocket

    Conc. (mM)


    SE1, SE2, SE3


    VFT (TAS1R2)


    (Servant et al., 2010; Zhang et al., 2010)

    Neohesperidin dihydrochalcone (NHDC)


    TMD (TAS1R3)


    (Jiang et al., 2005c; Winnig et al., 2007)

    Unnatural tripeptides (several)

    Biaryl derivative tripeptides


    2 – 20

    Yamada et al., 2019

    Sodium, cholestrol

    Cation, lipid

    TMD (TAS1R2)


    Perez-Aguilar et al., 2019


    IC50 (mM)


    Carboxylic acid salt

    TMD (TAS1R3)


    (Jiang et al., 2005c)

    (2-(2,4-dichlorophenoxy)propionic acid)

    Carboxylic acid salt

    TMD (TAS1R3)


    (Nakagita et al., 2019)

    Gymnemic acid

    Triterpenoid glycoside

    TMD (TAS1R3)


    (Sanematsu et al., 2014)

    Clofibric acid


    TMD (TAS1R3)


    (Maillet et al., 2009; Kochem and Breslin, 2017)



    TMD (TAS1R2)


    (Imada et al., 2010; Zhao et al., 2018)

    Umami compounds: MSG, Glu-Glu, Glu-Asp


    VFT (TAS1R2)


    (Shim et al., 2015)

    Where VFT, Venus ytrap domain; TMD, transmembrane domain; ND, not determined.

    TABLE 4. Sweet taste receptor’s positive allosteric regulators with concentration (used in cell based assays in studies) and negative allosteric modulators with their IC50 values.

    Using a high throughput chemical screening approach and heterologous expression of the TAS1R2/TAS1R3 heteromer, several unnatural tripeptides with a novel core biaryl structure were found as potential sweet enhancers (Yamada et al., 2019). This study divided the potential molecule into three parts namely, “head and linker” which together are essential for its sweet enhancer activity, while the “tail” determines the level of activity. This approach provided some useful inputs toward synthesis of potent PAMs. Firstly, an amine incorporated at the α-position of carbonyl moiety in the tail structure interacts with the TAS1R2 subunit thereby increasing allosteric activity. Secondly, additional hydrophobic substitutions in the tail structure provided increased allosteric activity to the molecule. Lastly, distance between the head and linker and insertion of an amide bond is crucial for its synthesis. Although, their binding characteristics and allosteric mechanisms are not yet known, these observations provide a starting point to identify and synthesize new sweet PAMs in the future.

    Small molecule PAMs can also bind to the transmembrane domain in class C GPCRs, in contrast to agonist which binds to the extracellular domain (Urwyler, 2011). For example, the flavonoid sweetener, neohesperidin dihydrochalcone (NHDC) binds to TMD regions to enhance the agonist induced sweet response. It interacts with a receptor binding pocket in the TMD of TAS1R3 and requires seventeen critical residues in TMDs and extracellular loop 2 for its allosteric activity (Winnig et al., 2007). These residues also contribute to cyclamate and lactisole binding sites. Among seventeen residues, eight alter receptor activation by NHDC (Q6373.29, S6403.32, H6413.33, Y6994.60, W7756.48, F7786.51, L7826.55, and C8017.39) and influence lactisole mediated inhibition. Similarly, nine of the seventeen residues (Q6373.29, H6413.33, H721ex2, S7265.39, F7305.43, W7756.48, F7786.51, L7826.55, and C8017.39) mediate activation by cyclamate, while six (Q6373.29, H6413.33, W7756.48, F7786.51, L7826.55, and C8017.39) influence receptor inhibition by lactisole as well as receptor activation by cyclamate [superscript refers to the nomenclature suggested for class C GPCRs by Pin et al. (2003) where first number denotes TMD region and the second number denotes residue position from the most conserved residue].

    Notably, three critical residues in TMD6 (W7756.48, F7786.51, L7826.55) and one in TMD7 (C8017.39) of TAS1R3 were found crucial for allosteric binding, as their mutation to alanine altered the receptor's sensitivity to NHDC and cyclamate, as well as to the inhibitor lactisole (Winnig et al., 2005). Therefore, TMD6 and TMD7α helices of TAS1R3 are integral to allosteric modulation of the sweet receptor, implicating them in TAS1R2 and TAS1R3 subunit interactions and indicating an important role for this structural region in the conformational changes involved in receptor activation. Furthermore, these residues are conserved across mammalian species (Cheron et al., 2019).


    Negative Allosteric Modulation of Sweet Receptor


    Like PAMs, negative allosteric modulators (NAM) such as lactisole and gymnemic acid bind to the TMD region of TAS1R3 and inhibit sweet substance induced responses. Lactisole, an aralkyl carboxylic acid not only inhibits sweet but also the umami receptor response in humans and presents a rare opportunity to study the structural cross talk between these two taste qualities. Using heterologous expression and mutagenesis, Jiang et al. (2005b) reported that lactisole's sweet inhibition might be mediated by its binding to TMD3, TMD5, and TMD6 of TAS1R3 and induce a conformation change which restricts the movement required to stabilize the active state. Residues A7335.46 in TMD5, L7987.36 in TMD7, and R790ex3 in extracellular loop 3 were found to be crucially important for sensitivity to lactisole in humans (Jiang et al., 2005b). These observations were confirmed in a recent study where 2-(2,4-dichlorophenoxy)propionic acid (2,4-DP) was found to be a more potent antagonist and utilise the same residues as well as four additional ones (H6413.37, H7345.43, F7786.53 and Q7947.32) in binding to TAS1R3. Moreover, the (S)- isomer of both compounds was found to be more strongly bound to the TMD of TAS1R3 and be a more effective inhibitor [lactisole; (S)-lactisole IC50, 20 µM while (R)- lactisole exerted no inhibition at this conc.; 2,4-DP: (S)-isomer was 10 fold more effective than (R)-2,4DP]. The (S)- lactisole isomer interacts with the TMD via its carboxyl group and stabilizes in only one orientation in the binding pocket that does not allow for very strong binding. In contrast, (S)-2,4- DP binds through two moieties simultaneously, a carboxyl group and an aromatic ring with two Cl groups and stabilizes in several different orientations through hydrophobic interactions that allow stronger binding, resulting in stronger negative allosteric modulation (Nakagita et al., 2019).

    These observations provide information about the relevance of structural modification in NAM compounds that could affect their interaction with the receptor. Although TMDs of TAS1R3 are the most likely regions responsible for allosteric modulation, TMDs and VFT regions of TAS1R2 cannot be ruled out completely. For example, the diuretic amiloride binds to TAS1R2 (TMD3, TMD5, TMD7) and inhibits the sweet response in a species dependent manner (Zhao et al., 2018). Further, the umami compound [monosodium glutamate (MSG)] and peptides (Glu-Asp, Glu-Glu) bind to the VFT region of TAS1R2 and inhibit the sweet induced response (Shim et al., 2015). These observations suggest that both subunits are important for allosteric activity of TAS1R2/TAS1R3 and further structural studies are required to design novel sweet allosteric modulators.


    Umami Taste Signal Transduction Mechanisms


    In contrast to four well-known basic human tastes (sweet, bitter, salty and sour), umami or ‘savoury taste’ is relatively recent and was introduced in early 2000 by Kikuna Ikeda (Ikeda, 2002) as a new seasoning element in food. The main stimulus for umami taste is the amino acid, L-glutamate present in the diet mainly in the form of MSG (Roper, 2007). Glutamate was first extracted from konbu/kombu (dried kelp of Fucus vesiculosus) and described as a “unique taste” and “very different from other tastes”. The terminology “umami” comes from the Japanese word “umai” meaning “delicious.” Moreover, the taste of umami is also produced by food such as mushrooms and soy sauce that contain amino acids (L-aspartate), peptides and synthetic ingredients similar to glutamate and some organic acids (Roper, 2007Kinnamon, 2009) (Table 5).


    TABLE 5

    TABLE 5 | Umami receptor agonists with their EC50 values and other pharmacological properties.




    EC50 (mM)

    Binding pocket


    L-amino acids (glutamate, aspartate,

    alanine, serine, asparagine, arginine, histidine, threonine, glutamine)


    Amino acids



    Amino acid (plant origin)

    3 (glutamate), ND for others



    VFT (TAS1R1)



    VFT (TAS1R1)

    (Li et al., 2002; Nelson et al., 2002; Zhang et al., 2008; Toda et al., 2013)


    (Narukawa et al., 2014)

    VFT, venus ytrap domain; ND, not determined.





    TABLE 5. Umami receptor agonists with their EC50 values and other pharmacological properties.

    The umami receptor (TAS1R1/TAS1R3) is a heteromeric member of the class C GPCRs, whereas most other receptors of this class exist as homodimers (Nelson et al., 2002Temussi, 2009Leach and Gregory, 2017). TAS1R1/TAS1R3 is the predominant umami taste receptor (Zhao et al., 2003Behrens and Meyerhof, 2011) and the TAS1R1 subtype is critical for sensing umami taste as its deletion abolished the response to umami taste stimuli (Mouritsen and Khandelia, 2012). However, TAS1R1/TAS1R3 is not the only receptor capable of detecting umami ligands (Chaudhari et al., 2000Kunishima et al., 2000Li et al., 2002Nelson et al., 2002). Studies using heterologous expression, afferent nerve recordings, and behavioral experiments have confirmed that metabotropic glutamate receptor 1, and 4 (taste-mGluR1 and taste-mGluR4) also sense umami stimuli (Chaudhari et al., 2000Kunishima et al., 2000Li et al., 2002Nelson et al., 2002). Notably, TAS1R3 knock out mice show a strongly diminished response to glutamate and sweet stimuli (Damak et al., 2003) and taste cells isolated from these mice respond to IMP and glutamate which is abolished in presence of mGluR antagonists (Pal Chaudhry et al., 2016). TAS1R1/TAS1R3 is not only activated by glutamate, but this activation is strongly enhanced in the presence of 5′-ribonucleotides, (inosine 5′ monophosphate; IMP) a response that is a hallmark of umami taste (Rifkin and Bartoshuk, 1980).

    The main transduction components following the activation of TAS1R1/TAS1R3 are similar to those for sweet taste (Zhang et al., 2003), i.e., α-gustducin (and γ13/β1 or β3), PLCβ2, IP3R and TRPM4/5. Cyclic nucleotides may also contribute to transduction of umami taste in TRCs. When taste tissue is stimulated with umami, its cyclic AMP level is decreased (Abaffy et al., 2003). However, the consequence of decreased cAMP in TRCs has not yet been fully elucidated. Both, α-transducin and α-gustducin are involved in umami taste signal transduction, as mice lacking the gene for one of these proteins showed a reduced response to this taste (He et al., 2004Leach and Gregory, 2017). In the taste palate fungiform papillae, α-gustducin and α-transducin activate PDE that reduces cAMP levels. Ligand binding to the TAS1R1/TAS1R3 heterodimer, releases Gβγ subunits to stimulate PLCβ2, which hydrolyzes PIP2 to DAG and IP3 (Kinnamon, 2009). IP3 then activates IP3R3 which results in release of calcium ions from intracellular compartments (Clapp et al., 2001Leach and Gregory, 2017) (Figure 3). Calcium ions activate TRPM5 and TRPM4 channels, leading to an influx of sodium ions, subsequent cell membrane depolarization, and finally release of ATP, which activates ionotropic purinergic receptors located in sensory fibers (Perez et al., 2002Sugita, 2006). This pathway was confirmed when mice devoid of TRPM5, TRPM4, PLCβ2, and IP3R3 showed a reduced response to umami taste perception following glutamate stimuli (Damak et al., 2006Kinnamon, 2009Eddy et al., 2012).


    Structural, Molecular and Conformational Changes of Umami Receptor


    In the last decade, several in depth modeling and mutagenesis approaches have our improved structural and molecular understanding of the umami receptor. The VFT regions of both subunits of TAS1R1/TAS1R3 comprise orthosteric and allosteric ligand binding sites for umami stimuli.

    Mutagenesis and molecular modeling studies reveal that the cognate agonist glutamate binds in the VFT region of the TAS1R1 subunit of TAS1R1/TAS1R3 and stabilizes the closed active receptor conformation. Moreover, four residues in the TAS1R1 VFT region (S172, D192, Y220 and E301) showed no detectable response to glutamate when they were mutated to alanine suggesting that they are critical for glutamate binding. The glutamate binding and stabilization of the closed conformation of TAS1R1, activates the downstream signaling pathway, while TAS1R3 remains in an open (inactive) conformation. Therefore, closure of the VFT is the key event that sensitizes umami taste receptor signal transduction (Lopez Cascales et al., 2010). Apart from glutamate, other L amino acids were also found to elicit functional responses by binding to the corresponding VFT region of TAS1R1. Six residues that contributed to the acidic amino acid agonist (L-glutamate and L-alanine) responses have been identified (S148, R151, A170, E174, A302, and D435).


    Allosteric Modulation of Umami Receptor


    Because of significant advancement in understanding and food industry application of umami taste, its allosteric modulators are sought after. Several allosteric umami ligands have been discovered with varying potency, only a few of which have been characterized at the molecular level. The best characterized umami PAMs, the 5′-ribonucleotides: inosine 5′-monophosphate (IMP) and guanosine 5′-monophosphate (GMP), interact with the VFT region of the TAS1R1 subunit to enhance the glutamate induced response that is hallmark of umami taste (Table 6). IMP and GMP binding sites in the VFT are adjacent to that for glutamate binding. Mutation of four residues (H71, R277, S306, and H308) abolished the IMP/GMP induced glutamate response suggesting their involvement in allosteric binding of these nucleotides. Structurally, IMP and GMP stabilize the closed form of the TAS1R1 VFT region through electrostatic interactions and coordinate the positively charged residues that act as pincers. The ability of IMP and GMP to interact with the VFT region (as opposed to the TMD region) represents a unique mechanism of positive allosteric regulation within class C GPCRs (Urwyler, 2011).


    TABLE 6 | Umami receptor allosteric modulators with conc. used in cell based assays and other pharmacological properties.


    Allosteric modulators


    Conc. (mM)

    Binding pocket





    VFT (TAS1R1)

    (Li et al., 2002; Nelson et al., 2002; Zhang et al.,









    TMD (TAS1R3)

    (Xu et al., 2004)






    Methional (3-methylsulfanylpropanal)


    TMD (TAS1R3)

    (Toda et al., 2018)

    Lactisole (2-4-methoxyphenoxy propionic acid)

    Carboyxlic acid salt


    TMD (TAS1R3)

    (Xu et al., 2004)

    Clofibric acid (4- chlorophenoxy)-2-methylpropanoic

    Herbicide acid


    TMD (TAS1R3)

    (Maillet et al., 2009; Kochem and Breslin, 2017)






    Where VFT, Venus ytrap domain; TMD, transmembrane domain.

    TABLE 6. Umami receptor allosteric modulators with conc. used in cell based assays and other pharmacological properties.

    In contrast to IMP and GMP that bind to the TAS1R1 extracellular domain, the well-known flavor compound methional and its analogs bind to the TMD region and allosterically regulate the umami receptor in a species dependent manner (Toda et al., 2018). Importantly, methional utilizes several distinct residues in different TAS1R1 transmembrane domains (TMD2-7) to act as a PAM in the human umami receptor, yet it behaves as a NAM in the mouse counterpart. This unusual phenomenon provided an opportunity to study the mechanisms of both positive and negative modulation in TAS1R1 simultaneously (Toda et al., 2018).

    Construction of chimeric receptors between human (h) and mouse (m) and their functional analysis demonstrated that the TMD of TAS1R1 is the key domain for switching the PAM/NAM activities of methional. Point mutation substitutions between these species identified four residues (h/m; F768/L769, N769/H770, S799/T800, and S802/G803) that are collectively required to switch PAM/NAM activities. A similar mode of allosteric regulation and PAM/NAM mode switching has been reported for mGluR5 (Gregory et al., 2013) suggesting this as an unusual and distinct phenomenon of the class C GPCRs. Further, alanine scanning mutagenesis in TAS1R1 of the corresponding residues vital for the activity of other taste inhibitors (sweetener inhibitors; NHDC and cyclamate; sweet and umami taste inhibitor; lactisole) revealed three residues required for PAM (W6974.50 F7285.40 and F7325.44) and a single residue (F6423.40) for NAM. These results suggest that both the PAM and NAM activities of methional are conferred by residues that are distinct from those required for the PAM/NAM switch. Knowing that methional is an important part of food seasoning globally, these observations could help in maximizing its use in enhancing flavors along with amino acids and nucleotides.

    Despite PAMs being a central focus for umami allosteric modulation, there has also been considerable research on negative allosteric modulation where lactisole emerged as a prominent NAM of the umami receptor, TAS1R2/TAS1R3. Because umami and sweet receptors share the TAS1R3 subunit, findings from studies on sweet receptor lactisole binding are relevant. A comprehensive study on the sweet receptor identified critical residues within the TMD regions (S6403.32, H6413.33 in TMD3 and F7786.51, L7826.55 in TMD6) of TAS1R3 required for lactisole binding pocket and showed a large effect on sensitivity to lactisole (Xu et al., 2004Jiang et al., 2005b). Because lactisole shares structural similarity with two other classes of compound: fibrates and phenoxy-herbicides, researchers studied them to search for novel sweet/umami inhibitors (Maillet et al., 2009). The lipid lowering drug, clofibric acid inhibits the TAS1R3 umami receptor mediated response both in vitro and in vivo (Table 6). Like lactisole, clofibrate inhibits the umami taste from glutamate by binding with a similar affinity to TAS1R1/TAS1R3. However, its specificity against the umami receptor still needs to be validated alongside other umami taste receptors (mGluR1, mGluR4, or NMDA).



    Type 2 taste GPCRs are represented by bitter taste receptors that have a distinct subset of bitter sensing cells in type II TRCs and notably 25 bitter taste receptors (TAS2Rs) are reported to be expressed in humans (Chandrashekar et al., 2000Devillier et al., 2015Behrens and Meyerhof, 2018). A significant amount of work has been done to explore the diversity among TAS2Rs and their agonists in taste biology (Adler et al., 2000Behrens and Meyerhof, 2009Behrens and Meyerhof, 2018). Some TAS2Rs (TAS2R3, TAS2R5, TAS2R13, TAS2R50) are narrowly tuned to structurally similar bitter compounds, whereas others are broadly tuned (TAS2R10, TAS214, TAS2R46), responding to several bitter compounds. Initially it was believed that each bitter-sensitive type II TRC expressed every TAS2R isoform (Adler et al., 2000) but other studies suggest that TAS2Rs can be expressed differentially, allowing for possible discrimination among bitter compounds (Caicedo and Roper, 2001Behrens and Meyerhof, 2009Behrens et al., 2009). Please refer to figure 4B for the basic structure of the bitter receptor.


    Bitter Taste Signal Transduction Mechanisms


    Bitter taste is the most complex of all the five basic tastes and provides protection against ingestion of toxic substances by eliciting an innate aversive response across species (Chandrashekar et al., 2006Behrens and Meyerhof, 2018). The TAS2Rs that mediate bitter taste perception are among ∼50 TAS2Rs identified in mammals, and 25 known to be expressed in humans (Adler et al., 2000Devillier et al., 2015Yoshida et al., 2018). TAS2R family is the most diverse and binds to a wide range of agonists compared with the other taste GPCRs (Jaggupilli et al., 2016) (Supplementary Table 1).

    TAS2Rs are distinctive among class A GPCRs in that many of them bind agonist with low apparent affinity in the micromolar range, rather than the nanomolar range (Di Pizio et al., 2016). The activation of TAS2Rs by harmless, minute amounts of bitter compounds such as those contained in most vegetables would limit the availability of food resources appearing safe for consumption and therefore could negatively affect survival. Hence, the concentration ranges at which bitter taste receptors are activated are well-balanced to allow species to maintain a healthy diet yet avoid ingestion of spoiled food containing strongly bitter ligand.

    Hundreds of bitter compounds have been reported to evoke bitterness and activate human bitter receptors in different cell based assays. These bitter agonists include plant-derived and synthetic compounds such as peptides, alkaloids and many other substances (Supplementary Table 1). (Pronin et al., 2004Meyerhof et al., 2010Iwata et al., 2014). Some TAS2Rs are activated by a wide range of compounds, whereas others show strict specificity for a single bitter compound (Behrens et al., 2009Sakurai et al., 2010aBorn et al., 2013). Interestingly, TAS2R31, TAS2R43, and TAS2R46 have around 85% sequence homology, but they bind to different agonists (Brockhoff et al., 2010Jaggupilli et al., 2016), reinforcing the idea that each TAS2R might have a unique ligand-binding pocket.

    The canonical TAS2R signal transduction cascade signaling molecules shared among bitter sweet and umami receptors (Wong et al., 1996Huang et al., 1999Mueller et al., 2005), include the heterotrimeric G protein subunits (Gα-gustducin, Gβ3, and Gγ13), (Ishimaru, 2009Shi and Zhang, 2009), a phospholipase C (PLCβ2), an inositol trisphosphate receptor (InsP3R), and the TRPM5 ion channel. Upon receptor activation by bitter ligands the G protein α-gustducin dissociates from its βγ subunits. The latter activates PLCβ2, leading to a release of Ca2+ from IP3-sensitive Ca2+ stores, resulting in Na+ influx through TRPM5 channels. This Na+ influx depolarizes the cells and causes the release of neurotransmitter ATP through gap junction hemichannels or CALHM1 ion channels (Finger et al., 2005Chaudhari and Roper, 2010Taruno et al., 2013) (Figure 3).


    Structural, Molecular and Conformational Changes of Bitter Receptors


    Classification of TAS2Rs has always been ambiguous because they were originally considered to be a distinct family (Horn et al., 2003) or grouped with the frizzled receptors (Fredriksson et al., 2003Jaggupilli et al., 2016), but most recent analyses (Di Pizio et al., 2016) support their classification with Class A GPCRs. The ability of bitter taste receptors to interact with numerous structurally diverse substances compared to other GPCRs is remarkable and includes a wide range of drugs/antibiotics, polyphenols, bacterial metabolites, salts and metal ions (Supplementary Table 1). Therefore, exploring the criteria for identification of highly heterogeneous bitter compounds with pronounced selectivity has become a major research area. Some of these studies rely solely on in silico homology/computational modeling (Dai et al., 2011Tan et al., 2012Di Pizio et al., 2020Dunkel et al., 2020) and others on in vitro genetic modification and functional assay systems (Pronin et al., 2004Nowak et al., 2018Jaggupilli et al., 2019).

    As a group of over ∼50 receptor subtypes, TAS2Rs recognize structurally diverse agonists where some are broadly tuned (TAS2R46, TAS2R14, TAS2R10, and TAS2R43) recognizing diverse agonists, while others (TAS2R1, TAS2R4, TAS2R7) show strong selectivity and narrow tuning (Liu et al., 2018Wang et al., 2019). The agonist binding cavity in most bitter GPCRs is located deep within their transmembrane domain (TMD), with the exception of TAS2R7 in which it resides on the extracellular surface (Liu et al., 2018). TAS2Rs are also distinct in containing highly conserved TMD regions, with thirteen key residues and two motifs (LXXXR in TMD2 and LXXSL in TMD5) that are absent in class A GPCRs, and may reflect their different activation mechanisms (Singh et al., 2011). LXXSL plays a structural role by stabilizing the helical conformation of TMD5 at the cytoplasmic end and a functional role by interacting with residues in intracellular loop 3 (ICL3) which is important for proper receptor folding and function (Singh et al., 2011). Moreover, mutation of the conserved residues in LXXSL and LXXXR motifs results in protein misfolding and poor surface expression (Singh et al., 2011Pydi et al., 2014a).

    The initial study highlighting the structure–activity relationship of bitter taste receptors was performed with receptors belonging to a subfamily of closely related TAS2Rs (Pronin et al., 2004). By physically swapping the extracellular loop 1 (ECL1) between TAS2R43 and TAS2R31, chimeric TAS2R31/TAS2R43 (ECL) gained responsiveness to the compound n-isopropyl-2methyl-5-nitrobenzenesulfonamide (IMNB), whereas the reverse chimera TAS2R31 (ECL)/TAS2R43 lost responsiveness for IMNB. Although this report supports an important contribution of residues located within the transmembrane region of the investigated receptors, the extracellular loops appear to be of importance for agonist selectivity. This empirical finding contrasts with earlier computational studies which predicted the agonist binding site to lie within the helical bundle of TAS2Rs without particular contacts between extracellular loops and docked agonists (Floriano et al., 2006Miguet et al., 2006).


    Bitter Receptor Ligand Binding Pocket


    The emergence of TAS2Rs as the most broadly tuned taste receptors might give the impression that their specific interaction with numerous agonists is because of several binding pockets that accommodate subgroups of bitter compounds. However, structure–function analysis of TAS2Rs (except for TAS2R7) has demonstrated the presence of only a single agonist binding pocket comprising the upper parts of TMD2, TMD3, TMD5, TMD6 and TMD7. The reason for their broad tuning and recognition of such a broad spectrum of agonist might most likely be attributed to the presence of an additional extracellular binding site called a “vestibular site,” in addition to the orthosteric selecting as reported for TAS2R46 (Sandal et al., 2015). This two site architecture offers more ligand recognition points than a single one, and thus might help in selecting the appropriate agonists. Moreover, the presence of the vestibular site may also help to discriminate among the wide spectrum of bitter ligands.

    Although broadly tuned receptors (TAS2R46, TAS2R31 and TAS2R43) have high homology in amino acid sequence, their agonist profiles only slightly overlap (Kuhn et al., 2004Brockhoff et al., 2007Di Pizio and Niv, 2015) which suggests the involvement of key residues at different positions in agonist specificity. Consequently, when strychinine interacting positions in TAS2R46 (residues differ at this position in TAS2R31, TAS2R43) were exchanged between these two receptors not only was the strychnine responsiveness transferred to the recipient receptor (TAS2R31, TAS2R43), but also sensitivity to additional TAS2R46 agonists (absinthin and dentaonium). Sensitivity to activation by aristolochic acid was lost in the mutant receptors (Brockhoff et al., 2010). This experimental evidence supports the presence of a common agonist binding pocket and agrees with other studies on TAS2R16, TAS2R14 and TAS2R7 receptors (Sakurai et al., 2010aSakurai et al., 2010bThomas et al., 2017Liu et al., 2018Nowak et al., 2018).

    Recent studies used homology modeling and mutagenesis to elucidate the nature of the ligand binding pocket in TAS2R7, TAS2R14 and TAS2R16 receptors (Thomas et al., 2017Liu et al., 2018Nowak et al., 2018). They reported that the binding pocket is flexible and wide open to accommodate molecules of diverse size and shape, and thus permits chemical modifications among agonists as well (Thomas et al., 2017Liu et al., 2018Nowak et al., 2018). Although the molecular basis for the promiscuity of bitter receptors is attributed to their apparent flexible spacious binding site, future work elucidating the contact points between TAS2Rs binding site residues and its agonists in terms of additional binding locations is required.


    Bitter Receptors Ligand Binding Domain and Amino Acid Residues


    A majority of the TAS2R studies based on molecular modeling, mutagenesis and heterologous expression systems (Biarnes et al., 2010Brockhoff et al., 2010Tan et al., 2012Nowak et al., 2018Shaik et al., 2019) suggest that the ligand binding pocket is formed by several key residues in most TMDs (TMD1, TMD2, TMD3, TMD5, TMD6 and TMD7), with the exception of TMD4.

    Studies show similarities as well as differences regarding residues and positions involved in agonist-receptor interactions. However, most of them agree that besides position N3.36 in TMD3 (superscript as per Ballestros-Weinstein nomenclature for class A GPCRs) (Ballesteros and Weinstein, 1995) and other residues (L3.32, L3.33, and E3.37) in its close proximity, play a role in agonist activation of several broadly tuned TAS2Rs (TAS2R1, TAS2R16, TAS2R30, TAS2R38, TAS2R46) (Pronin et al., 2004Biarnes et al., 2010Brockhoff et al., 2010Sakurai et al., 2010bDai et al., 2011). In contrast, for the narrowly tuned TAS2R7, one position in TMD3 (H943.37) and another in TMD7 (E2647.32) were found crucial for metal ion binding (Wang et al., 2019). Mutagenesis and molecular modeling revealed that these two residues contribute to the metal ion binding pocket in TAS2R7. Moreover, metal ions bind distinctively to residues lining the binding pocket and interestingly, the presence of calcium in the assay solution appears to affect the TAS2R7 response to metal ions. It is not clear how calcium affects metal ion binding to TAS2R7, but it might work cooperatively with certain ions and not others. Future studies focusing on structural interactions between the receptor and metal ions will provide further insights into how they activate the receptor.

    In TMD2, two studies suggest that position N2.61 is critical for binding in TAS2R1 (Singh et al., 2011) and TAS2R46 (Brockhoff et al., 2010). Likewise, in TMD7, position 2657.39 is implicated in binding to TAS2R46 (E265) and TAS2R1 (I263) (Dai et al., 2011). In TMD5, position H5.43 is implicated in binding in TAS2R16 and E5.46 in TAS2R1 (Dai et al., 2011) while, in TMD7, position E7.32 was crucial for metal ion binding (Wang et al., 2019). These residues represent putative contact points for agonist interaction and form a pattern in being spaced one helical turn from each other.

    Recent mutagenesis studies (Nowak et al., 2018Di Pizio et al., 2020) performed in broadly tuned TAS2R14 with agonists (aristolochic acid, picrotoxinin, thujone) found several residues in TMDs to be involved in agonist binding. However, in contrast to TAS2R10 (Born et al., 2013) and TAS2R46 (Brockhoff et al., 2007), mutation of TAS2R14 did not result in complete loss of function for all agonists but a varied reduction in responsiveness or selectivity toward agonists. Among several mutants, only mutation of W89A resulted in complete loss of responsiveness against picrotoxinin while others showed more subtle agonist selective changes. This indicates that TAS2R14 is not streamlined for the most sensitive detection of selected agonists, but rather tailored to detect numerous diverse agonists, with comparatively lower apparent affinity.

    The binding characteristics of bacterial acyl homoserine lactones (AHLs) on TAS2Rs (TAS2R4, TAS2R14 and TAS2R20) suggest the presence of a single orthosteric site situated close to the extracellular surface and reinforce the significant role of the extracellular loop structure (ECL2) in TAS2R ligand binding and activation (Jaggupilli et al., 2018). The crucial AHL binding residues in TAS2R4 and TAS2R14 are predominantly located in the ECL2, while in TAS2R20 they are present in TMD3 and TMD7 helices. The ECL2 residues, N165 in TAS2R4, and R160 and K163 in TAS2R14 were found crucial for lactone binding. In contrast, TAS2R20 residues W88 (TMD3) and Q265 (TMD7) are essential for agonist binding (Pydi et al., 2014cZhang et al., 2017Jaggupilli et al., 2018). In addition, the hydrophobic amino acids in the three TAS2Rs are considered important in directing the orientation of the hydrophobic acyl chains of lactones that facilitate receptor activation.

    The transmembrane domain in GPCRs is composed mainly of hydrophobic amino acids accommodated in the plasma membrane. Therefore, hydrophobic properties of the receptor binding pocket are important for any membrane accessible agonist. Hydrophobic residues in TMD3 and TMD7 of TAS2R16 are important in forming a wide ligand-binding pocket (Thomas et al., 2017) that accommodates larger ligands like the β-glycosides. By using salicin analogs as TAS2R16 novel agonists (differ structurally to salicin in β-glucoside core constituents), several critical residues were identified that are required for signaling. Interestingly, these were identical to the residues critical for salicin signaling, except for W261, which was not required for activation by the analog 4-NP-β-mannoside. Importantly, all these residues are in the TMD helices or intracellular face of the receptor, consistent with classical GPCR signal transduction. These results suggest that larger ligands bind to the wide binding pocket of TAS2R16 on the extracellular side, and then their signal is transduced via conserved residues on the intracellular side. This can account for the broad spectrum of ligand recognition conferred by TAS2R16.

    Unlike broadly tuned receptors, narrowly tuned ones like TAS2R7 show two different types of critical residue in ligand binding. The first type includes D86, W170 and S181 that are agonist independent and their mutation significantly reduces the ability of TAS2R7 to bind agonist, while a second group consisting of D65 and W89 are selective for quinine and enhance binding to a specific category of ligand (Liu et al., 2018).

    Despite the variation in the amino acid type and location important for agonist binding among receptors of the bitter family, for the most part, ligand binding pockets are present on the extracellular surface of TMDs or on ECL2. The function of the residues at these binding pockets is dictated by multiple factors that include the type of ligand, the movements in TMDs, and the associated movement of ECL2 to accommodate the ligand. Structure–function studies have identified a conserved KLK/R motif in the intracellular carboxyl terminal domain of 19 TAS2Rs that is critical for cell surface expression, trafficking and receptor activation (Upadhyaya et al., 2015Jaggupilli et al., 2016).


    Agonist, Antagonist Binding and Modulation of Bitter Receptors


    In simple pharmacological terms an antagonist is a ligand that inhibits the biological response induced by an agonist and does not induce any response of its own, while a ligand that reduces the constitutive/basal activity of a GPCR is considered an inverse agonist. An antagonist acts as a competitive inhibitor to block receptor activity. Large numbers of agonists have been identified for bitter receptors, but few antagonists have been found so far (Table 7). Finding an antagonist/inhibitor for bitter taste would not only help in understanding the TAS2R mechanism of signal transduction but have potential use in foods to overcome unwanted bitterness in consumer products. Such bitter blockers have been proposed to increase the palatability of bitter tasting food and beverages, increase the compliance in taking bitter tasting drugs, especially children’s formulations and reduce or prevent off-target drug effects in extra-oral tissues (Clark et al., 2012).



    TABLE 7 | Bitter taste receptor inhibitors with their IC50 values and other pharmacological properties.


    Mode of action

    Bitter receptors

    Tested agonists

    IC50 (µM)


    GIV3727or 4-(2,2,3-trimethylcyclopentyl) butanoic acid

    Competitive orthosteric inhibitor




    (Slack et al., 2010)




    Aristolochic acid















    Gamma-aminobutyric acid (GABA)

    Orthosteric inhibitor




    (Pydi et al., 2014b)

    3β-hydroxydihydrocostunolide (3HDC)





    (Slack et al., 2010; Brockhoff et al., 2011)






























    (Brockhoff et al., 2011)


























    Allosteric inhibitor




    (Greene et al., 2011)






    (Fletcher et al., 2011)






    (Fletcher et al., 2011)






    (Fletcher et al., 2011)




    Epicatechin gallate (ECG)


    (Roland et al., 2014)










    Epicatechin gallate (ECG)


    (Roland et al., 2014






    (Pydi et al., 2014b)

    (±) abscisic acid (ABA)





    (Pydi et al., 2015)

    ND, not determined.







    TABLE 7. Bitter taste receptor inhibitors with their IC50 values and other pharmacological properties.

    To date ∼12 bitter inhibitors have been reported to interact with only 10 TAS2Rs subtypes (Table 5) by binding to transmembrane domains in a similar manner to agonist. GIV3727 (4-(2,2,3-trimethylcyclopentyl) butanoic acid) was the first TAS2R antagonist discovered and to be well characterized structurally (Slack et al., 2010) that acts as an orthosteric competitive antagonist for TAS2R31. It competes with the acesulfame K agonist both in vitro and in vivo. GIV3727 is moderately selective because it inhibits multiple bitter receptors including, TAS2R4, TAS2R40 and TAS2R43. Homology modeling revealed that the -COOH group in GIV3727 is important for ligand-receptor interactions as its replacement with an ester or the corresponding alcohol abolished its antagonist activity. Moreover, a mutagenesis study in TAS2R31 and TAS2R43 revealed residues K2657.39 and R2687.39 in TMD7 to be crucial for its antagonistic activity (Slack et al., 2010). Similarly, another non-selective inhibitor, probenecid (p-(dipropylsulfamoyl) benzoic acid) was found to act as NAM of TAS2R16 activity and inhibits TAS2R38 and TAS2R43 as well (Greene et al., 2011). Two point mutations, P44T and N96T in TMD3 of hTAS2R16 were found to significantly suppress the ability of probenecid to inhibit salicin activity. Hydrophobicity seems important for their pharmacological activity as observed for both probenecid and GIV3727. The sesquiterpene lactone, 3β-hydroxydihydrocostunolide (3HDC) is an interesting bitter blocker as it acts as a competitive antagonist of TAS2R46, TAS2R30, TAS2R40, yet activates TAS2R4, TAS2R10, TAS2R14 and TAS2R31 as an agonist (Brockhoff et al., 2011).

    Similarly, various flavonones were also noted as antagonists for TAS2R31, TAS2R39 with varying efficacy. Taken together most of the currently known antagonists are non-selective and there is an urgent need for studies that focus on selective antagonists of major broadly tuned TAS2Rs (such as TAS2R10, TAS2R14, TAS2R16 and TAS2R46). In order to target bitterness in terms of food industry needs, potential peptide inhibitors from different protein sources such as hen protein hydrolysates (inhibits TAS2R4, TAS2R7, TAS2R14) and beef proteins (inhibits TAS2R4) (Zhang et al., 2018Xu et al., 2019) are reported to be effective. Several umami glutamyl peptides isolated from soyabean have been found to act as non-competitive allosteric inhibitors of TAS2R16 against the salicin induced response (Kim et al., 2015).


    Constitutive Activity of Bitter Receptors


    A phenomenon in GPCR activity is that of constitutive activity, essentially an active state occurring in the absence of agonist which has been demonstrated in more than 60 GPCRs (Seifert and Wenzel-Seifert, 2002). It is the production of a second messenger or downstream signaling by a receptor in ligand independent manner. Constitutive activity provides another possibility for taste inhibitor discovery using inverse agonists. Inverse agonists can inhibit both agonist-dependent and agonist-independent activity, while antagonists can inhibit only agonist-dependent activity (Chalmers and Behan, 2002). Interestingly, some mutations in GPCRs can lead to constitutive activity and receptors with this characteristic (including constitutively active mutants or CAM) are important tools to investigate new bitter inhibitors. Although constitutive activity has not been observed naturally in TAS2Rs, when induced by mutation these receptors provide a useful means to investigate the relationship between an active receptor conformation and inverse agonist pharmacology.

    Molecular modeling and functional assays report five CAMs critical residues for TAS2Rs, one in TMD7 (S2857.47) and four others in intracellular loop 3 (H214A, Q216A, V234A, and M237A) (Pydi et al., 2014aPydi et al., 2014b). Of the five CAMs, only the TAS2R4 with H214A mutation shows a 10 fold increase in constitutive activity. This histidine residue is highly conserved in most TAS2Rs. Mutation of H214 (H214A) helped in finding two new inverse agonists (GABA and ABA; Table 7) (Pydi et al., 2015). Similar pharmacological approaches can be used to generate mutants of all TAS2Rs to screen for their inverse agonist/bitter taste blockers. However, for better characterization and interpretation of TAS2Rs, future in vivo studies should be performed to understand the functional relevance of these CAMs. At the same time, it is worth noting that the potential presence of endogenous agonists makes it difficult to determine the true constitutive activity of GPCRs including TAS2Rs (Devillier et al., 2015).


    Kokumi Sensation Signal Transduction


    In addition to the five basic tastes, sensations beyond these add another dimension to taste perception. One such example is “kokumi” that is distinct from the other five tastes in that it does not have a taste as such but rather induces a sensation of “mouthfulness,” depth, thickness and aftertaste in the flavors. Although, this flavor has been used historically and is well recognized in Japanese cuisine, it was first characterized by Ueda et al. (1990) who isolated a kokumi taste substance from water extracts of garlic and onion and identified, γ-glutamylcysteinylglycine or glutathione (GSH) as the main active ingredient of kokumi flavor (Ueda et al., 1990Ueda et al., 1997Dunkel et al., 2007). GSH is abundantly present in food-grade yeast extract and has been used to make foods more flavoursome.

    Kokumi signal transduction was unknown until CaSR expression was reported in a subpopulation of taste cells in mice and rats that suggested it could function as a taste receptor for calcium and amino acids (San Gabriel et al., 2009Bystrova et al., 2010). However, its apparent role in kokumi stimuli detection was not confirmed. Ohsu et al. (2010) for the first time reported that kokumi peptides (GSH, γ-Glu-Val-Gly and various γ-glutamyl peptides; Table 8) signal through CaSR and can synergise with sweet, salty, and umami taste qualities to impart an augmented kokumi sensation, i.e., increased depth of flavor which was further complemented by later studies (Maruyama et al., 2012Kuroda and Miyamura, 2015). By using heterologous expression systems and human sensory analysis these studies demonstrated that kokumi peptides impart kokumi sensation to sweet, salty and umami taste via CaSR as the kokumi component was specifically suppressed in the presence of the CaSR-specific NAM NPS-2143. To further validate this idea, Maruyama et al. (2012) identified a distinct population of taste cells expressing CaSR in mouse lingual tissue which did not express either sweet or umami receptors. Notably, these cells are specifically responsive to kokumi substances and elicit a Ca2+ response to focally applied kokumi stimuli in mouse lingual slices. Moreover, this response was inhibited in the presence of NPS-2143. These findings support the idea that CaSR mediates kokumi sensation effects in TRCs.


    TABLE 8 | Kokumi sensation receptor agonists, allosteric modulators with concentrations used in cell based assays.

    Ca2+  Orthosteric agonist/cation 1a VFT
    Mg2+  Orthosteric agonist/cation 10a VFT
    Gd2+  Orthosteric agonist/cation 0.02a VFT
    Al2+  Orthosteric agonist/cation 0.5a VFT
    Sr2+  Orthosteric agonist/cation 0.5a VFT
    Mn2+  Orthosteric agonist/cation 0.5a VFT
     Ni2+  Orthosteric agonist/cation 0.5a VFT
    Ba2+  Orthosteric agonist/cation 0.2a VFT
    Ca2+  Orthosteric agonist/cation 1a VFT Ca2+ 


    Orthosteric agonist/polyamine



    (Nemeth et al., 2018)


    Orthosteric agonist/aminoglycoside antibiotic



    (Katz et al., 1992)


    Orthosteric agonist/aminoglycoside antibiotic



    Katz et al., 1992)


    Orthosteric agonist/aminoglycoside antibiotic



    (Katz et al., 1992)

    Amyloid β-peptides

    Orthosteric agonist/Peptide


    (Ye et al., 1997)


    Orthosteric agonist/peptide

    0.03 µMa


    (Brown et al., 1991; Nemeth et al., 2018)

    Poly L-arginine

    Orthosteric agonist/peptide

    0.004 µMa


    Brown et al., 1991; Nemeth et al., 2018)





    (Yamamoto et al., 2020)





    (Yamamoto et al., 2020)

    Aromatic L-amino acids (Trp, Phe, His, Ala, Ser)




    (Conigrave et al., 2000; Mun et al., 2004; Geng et al., 2016)

    Anions (SO42-)




    (Geng et al., 2016)



    0.051 µMa


    (Miedlich et al., 2002; Petrel et al., 2004; Nemeth et al., 2004)



    0.31 µMa


    Miedlich et al., 2002; Petrel et al., 2004)

    NPS R-568


    0.5 µMa


    (Miedlich et al., 2002; Petrel et al., 2004)

    NPS R-467




    (Miedlich et al., 2002; Petrel et al., 2004)



    0.041 µMa

    (Ohsu et al., 2010)

    γ-Glu-Cys-Gly (Glutathione)


    76.5 µMa


    (Ohsu et al., 2010; Wang et al., 2006



    3.65 µMa


    (Wang et al., 2006; Ohsu et al., 2010)

    γ -Glu-Val


    1.34 µMa


    (Wang et al., 2006; Ohsu et al., 2010)

    γ -Glu-Cys


    0.45 µMa


    (Ohsu et al., 2010; Wang et al., 2006)

    γ -Glu-α-aminobutyryl-Gly (Opthalmic acid)


    0.018 µMa


    (Ohsu et al., 2010)



    0.0003 (IC50)


    (Gowen et al., 2000; Petrel et al., 2004)

    Calhex 231

    Mixed PAM/NAM

    0.1–1 µM (PAM); 3–10 µM (NAM)


    (Petrel et al., 2003; Petrel et al., 2004; Gregory et al., 2018)

    Where VFT, venus flytrap domain; TMD, transmembrane domain; ND, not determined. a shows EC50 value.

    TABLE 8. Kokumi sensation receptor agonists, allosteric modulators with concentrations used in cell based assays.

    More recently, kokumi peptides have been found to have an extraoral physiological role in the gastrointestinal tract where they stimulate secretion of hormones (cholecystokinin and glucagon-like peptide1 by activating CaSR (Depoortere, 2014Yang et al., 2019). However, future studies with tissue specific deletion of CaSR in taste buds would be helpful in delineating its role in taste physiology.

    CaSR involvement in taste is a relatively recent discovery, but its central role in extracellular calcium homeostasis in mammals is well recognised (Brown et al., 1993Brown, 2013). Diverse ligands activate CaSR, including cations (Ca2+ and Gd3+), peptides, polyamines (Brown and MacLeod, 2001) and amino acids (Conigrave et al., 2000Conigrave and Hampson, 2006) (Table 8). Unlike other taste modalities (sweet, bitter and umami), CaSR–ligand binding and recruitment of G protein results in the activation of an intricate, amplifying signaling network which initiates numerous intracellular functions. The functional diversity of CaSR results from its ability to activate multiple Gα proteins (Gq/11, Gi/o, G12/13 and Gs) (Magno et al., 2011Conigrave and Ward, 2013) which subsequently affect multiple signaling pathways related to the pathophysiology of parathyroid hormone secretion, cancer and metastasis (Kelly et al., 2007Wettschureck et al., 2007Mamillapalli et al., 2008).

    Kokumi substrates activate CaSR and transmit their signal through Gαq/11 proteins which further activate PLCβ that results in release of intracellular Ca2+ store through activation of IP3 receptor channels in the ER. Whether the kokumi pathway strictly relies on Gαq/11 protein or can also use Gα-gustducin, like other taste modalities for downstream signaling, is still unknown (Figure 3). The growing number of reports on kokumi flavor signal transduction are shedding light on its potential use as a flavor enhancer.


    Structural, Molecular and Conformational Changes of Kokumi Receptor


    CaSR belongs to the class C GPCR. Within this class, CaSR and metabotropic glutamate receptors (mGluRs) are known to function as disulfide-linked homodimers (Bai et al., 1998Ward et al., 1998Pidasheva et al., 2006) (Figure 4A). Structurally, the human CaSR is similar to sweet and umami taste receptors but differs in being a homodimer instead of a heterodimer (Hendy et al., 2013). The ECD of CaSR not only senses nutrients (Ca2+, L-Phe and polypeptides; Table 8) and allows ligand to modulate CaSR cooperatively, but is also required for its dimerization (Ray et al., 1999Zhang et al., 2014). Binding of Ca2+ and other ligands to the ECD changes the conformation of the seven transmembrane domains, causing alterations in the intracellular loops and the intracellular domain (ICD), which further trigger downstream signaling pathways (Brown et al., 1975). The ICD is relatively diverse among species and participates in controlling CaSR signaling in multiple ways by modulating receptor expression, trafficking and desensitization (Gama and Breitwieser, 1998Ward, 2004Huang et al., 2006).

    Homology modeling, mutagenesis and heterologous expression revealed distinct and closely located binding sites for Ca2+ and aromatic L-amino acids, in VFT and the cleft of the VFT, respectively (Silve et al., Conigrave et al.,2000Huang et al., 2009). Notably, four putative Ca2+ binding sites of varying affinity have been predicted in the VFT of the CaSR and in which the interaction between site 1 and the other three sites plays a central role in positive cooperativity in sensing Ca2+ (Zhang et al., 2014). Besides Ca2+, aromatic L amino acids (L-Trp, L-Phe) also activate the CaSR by binding adjacent to the VFT region through three serine and one threonine residue (S169/S170/S171/T145). Interestingly, the double mutation T145/S170 was found to selectively impair L amino acid (Phe, Trp, His) sensing of CaSR, while Ca2+ sensing remained intact (Mun et al., 2004Mun et al., 2005).

    The recent crystal structure of the entire extracellular domain of CaSR (Geng et al., 2016) identified four novel Ca2+ binding sites in each protomer of the homodimer including one at the homodimer interface which does not correspond to any of the sites reported previously by Huang et al., (2007). It is unclear why these additional calcium binding sites were not found in earlier studies. This might be due to the different expression systems used, crytallization conditions and methods of analysis. The conditions of the more recent studies may have stabilised an active conformational state in which these calcium sites become available (Geng et al., 2016). Among these four Ca2+-binding sites, site 4 seems most relevant to receptor activation as it directly participates in the active CaSR conformation. Moreover, a previously reported natural mutation G557E (Hendy et al., 2009) reduced the potency of Ca2+ possibly by affecting backbone conformation, thereby weakening the affinity of Ca2+ for this site. This confirms that a Ca2+ ion at site 4 stabilizes the active conformation of the receptor by facilitating homodimer interactions between the membrane proximal LBD2 region and CRD of CaSR.

    The most interesting aspect of Ca2+ and L-amino acid interplay was reported by Zhang et al. (2014) who studied L-Phe binding characteristics by monitoring intracellular [Ca2+]i oscillations in living cells and performing molecular dynamic simulations. Their findings supported a previous observation that the L-Phe binding pocket is adjacent to the Ca2+ binding site 1. Importantly, by binding to this site, L-Phe influences all Ca2+ binding sites in the VFT region and enhances CaSR functional cooperativity through positive heterotropic cooperativity to Ca2+. Moreover, the dynamic communication of L-Phe at its predicted binding site in the hinge region with the Ca2+ binding sites not only influences the adjacent Ca2+ binding site 1, but also globally enhances cooperative activation of the receptor in response to alterations in extracellular Ca2+.

    The crystal structures (Geng et al., 2016) of the entire ECD region of CaSR in the resting and active conformations have provided additional information about the dynamics between calcium and L-amino acid binding (Geng et al., 2016). Most importantly, by using L-Trp, the study provided direct evidence that L-amino acids are CaSR co-agonists, and they act concertedly with Ca2+ to achieve full receptor activation. Several lines of evidence support this contention: 1) L-Trp binds at the interdomain cleft of the VFT, which is a canonical agonist-binding site for class C GPCRs (Kunishima et al., 2000Muto et al., 2007Geng et al., 2016) and shares a common receptor-binding mode with the endogenous agonists (amino acids or their analogs) of mGluR and GABAB receptors, (Kunishima et al., 2000Tsuchiya et al., 2002Muto et al., 2007Geng et al., 2016). 2) L-Trp interacts with both LBD1 and LBD2 in ECD to facilitate its closure, a crucial first step during CaSR activation. In contrast, no Ca2+ ion is found at the putative orthosteric agonist-binding site to induce domain closure. 3) Mutations of L-Trp-binding residues (S147A, S170A, Y218A, and E297K) severely reduced Ca2+ induced IP accumulation and intracellular Ca2+ mobilization (Zhang et al., 2002Silve et al., 2005), indicating that L-Trp is required for a Ca2+ induced receptor response. Notably, the presence of extracellular Ca2+ above a threshold level is required for amino-acid-mediated CaSR activation, amino acids increase the sensitivity of the receptor toward Ca2+. Taken together, amino acids and Ca2+ ions act jointly to trigger CaSR activation.

    Knowing that aromatic L-amino acids (Trp, Phe, His) are important tastants in kokumi flavor, CaSR becomes more relevant for taste biology. Moreover, the kokumi tripeptide, glutathione (GSH) and glutamyl peptide are suggested to bind allosterically to CaSR at the same site as L-amino acids (Wang et al., 2006Broadhead et al., 2011) and enhance its activity in the presence of 0.5–1 mM free calcium, thereby acting as a positive allosteric modulator. In addition, an ECD crystal structure might help to explain structural and molecular details of the GSH binding pocket such as the nature of critical residues and their binding characteristics. In view of recent reports of calcium emerging as taste modifier, it would be worth investigating how GSH and Ca2+ operate in kokumi human perception.


    Allosteric Modulation of Calcium-Sensing Receptor


    Classically CaSR is known to be involved in pathophysiology of parathyroid and renal related diseases by sensing calcium ions in extracellular fluid (Brown, 2007). Research on related therapeutic applications has identified several classes of PAMs and NAMs that modulate CaSR agonist sensitivity. More recently this has been applied to kokumi taste signal transduction.


    Endogenous Modulators (L-amino Acids, Anions and Glutathione Analogs)


    Several studies based on molecular modeling and mutagenesis report L-amino acids (L-Phe, L-Tyr, L-His and L -Trp) as PAMs because they enhance the Ca2+ induced response of CaSR. Aromatic L-amino acids bind in the VFT domain (Mun et al., 2004) and require a highly conserved five residue binding motif (S147, S170, D190, Y218 and E297) (Conigrave and Hampson, 2006Geng et al., 2016). Among these residues, E297 was identified through the natural mutation E297K as essential for structural and functional activity (Table 8) (Pollak et al., 1993Bai et al., 1998Conigrave et al., 2000Zhang et al., 2002Mun et al., 2004).

    As recently identified NAMs, anions SO42 and PO43 are important modulators of the Ca2+ induced response. They bind in the VFT region and act as moderate NAMs for CaSR activity (Geng et al., 2016Centeno et al., 2019). Based on anomalous difference maps, four anion-binding sites were identified in the inactive and active CaSR ECD structures. Sites 1 and 3 are located above the interdomain cleft in LBD1, while site 4 lies in the LBD2 region. Sites 1 and 3 appear to stabilize the inactive conformation while site 2, which is present in both active and inactive conformations appears important for receptor function as mutation in its residues (R66H, R69E, and S417L) abolished the Ca2+ induced response. In addition, each protomer structure contains one Ca2+ ion and three SO42 ions which together contribute to the structural integrity of the receptor (Geng et al., 2016). Taken together, anions along with Ca2+ and amino acids are involved in an intricate interplay for CaSR activation to maintain conformational equilibrium between inactive and active states.

    As positive allosteric modulators, γ glutamyl peptides including glutathione (γGlu-Cys-Gly) and its analogs (Table 8) are predicted to have overlapping binding sites with L-amino acids in the VFT region (Wang et al., 2006Ohsu et al., 2010Broadhead et al., 2011). Kokumi peptides that activate CaSR resemble amino acids in having free α-amino and free α-carboxylate groups because they contain both amide bond formation between the γ-carboxylate group of L-glutamate and the α-amino group of its neighboring Cys residue. However, compared to amino acids, glutathione analogs have much larger side chains and are more potent activators of CaSR (Wang et al., 2006). Nonetheless, the free sulfhydryl is not required for CaSR activation (Ohsu et al., 2010Maruyama et al., 2012).

    The crystal structure of ECD enables mapping of the GSH binding site and investigation into how GSH binding works in synergy with Ca2+ to modulate the kokumi sensation. NPS2143, the sole kokumi NAM identified to date has been reported to inhibit kokumi taste sensation to GSH and its analogs which provides an opportunity to screen for novel kokumi enhancing molecules in a cell-based assay.


    Synthetic Drugs as Allosteric Ligands of Calcium-Sensing Receptor


    Because of its pathophysiological importance, various synthetic PAMs and NAMs of CaSR have been identified and are in clinical use. The allosteric modulation of CaSR by synthetic drugs has been recently reviewed (Hannan et al., 2016Chavez-Abiega et al., 2020Leach et al., 2020). Since the 1990’s the term calcimimetics and calcilytics, have been used for drugs that mimic or antagonize the effect of extracellular Ca2+ on CaSR activity, respectively. Pharmacologically, a calcimimetic activates the CaSR and includes agonists (type I) and allosteric ligands (type II). Most type I calcimimetics are either inorganic or organic polycations (e.g., Mg2+, Gd3+, neomycin), whereas type II calcimimetics are small naturally occurring molecules (aromatic amino acids or GSH) or synthetic drugs and peptides (NPS R-568, cinacalcet). Type II calcimimetics (like aromatic amino acids) bind in the ECD while others (e.g., NPS R-568, NPS R-467) bind in the TMD of the CaSR. Calcilytics are thus small organic molecules that appear to act as NAMs and bind in the TMD of the receptor (Widler, 2011Nemeth, 2013).

    Homology modeling and mutational studies show that both PAMs and NAMs have overlapping but non-identical binding sites in TMD and can partially allosterically modulate CaSR activity in the complete absence of the ECD, but their potencies vary among structurally different compounds (Collins et al., 1998Ma et al., 2011) (Table 8). Several residues reportedly critical for allosteric modulation, W8186.48, F8216.51 (TMD6) and E8377.39, I8417.43 (TMD7), R6803.28, F6843.32, F6883.36 (TMD3) impair calcimimetic and calcilytic induced CaSR signaling (Miedlich et al., 2004Petrel et al., 2004Leach et al., 2016). Nevertheless, subtle differences in ligand–receptor interactions drive negative vs. positive modulation of CaSR signaling, by NPS2143 or cinacalcet and NPSR-568, respectively (Miedlich et al., 2004Leach et al., 2016Keller et al., 2018). The details of CaSR allosteric modulation by synthetic drugs is out of the scope of the current review, for a comprehensive explanation refer to these studies (Chaves-López et al., 2014Hannan et al., 2016Leach et al., 2020).




    Taste GPCR research has advanced rapidly over the past two decades providing a more thorough understanding of receptor molecular pharmacology and signal transduction pathways. With the exception of the kokumi receptor ECD, high-resolution crystal structures for any taste receptor would be a major step toward designing novel and potent surrogate taste receptor ligands and selective antagonists. This has been a challenge due to low taste GPCR functional heterologous expression, appropriate post-translational modifications, high conformational flexibility, and low detergent stability. However, significant advancements in structural biology technologies of serial femtosecond crystallography using X-ray free-electron lasers and high-resolution cryo-electron microscopy provide promising tools for understanding conformational dynamics and visualizing the process of receptor activation with high spatial and temporal resolution. The physiological relevance of taste GPCRs will be further advanced through in vivo studies to help provide information on potential synergies in taste signal transduction mechanisms particularly among bitter, umami, sweet and kokumi receptors.

    • Was this article helpful?