A structural perspective on insect gustatory receptors

Abstract

Across the animal kingdom, a remarkable diversity of chemoreceptors has evolved, reflecting the ecological and evolutionary pressures that shape species-specific sensory demands. Insects, the most biodiverse class of animals, play crucial roles in ecosystems and have an extensive chemosensory repertoire. At the heart of insect gustation lie the gustatory receptors (GRs), a large and remarkably diverse family of proteins characterized by their seven-transmembrane domain structure and tetrameric stoichiometry. These receptors are phylogenetically distinct from the taste receptors present in most other animal groups, including mammals. Functionally, GRs operate as ligand-gated cation channels. Upon binding to specific chemical compounds (tastants), these receptors undergo conformational changes that lead to the opening of ion-conducting channels in the neuronal membrane, ultimately triggering neuronal activation and initiating the perception of taste. Recent advancements in structural biology, particularly the use of cryo-electron microscopy, have enabled the visualization of the three-dimensional structure of several insect GRs that detect sugars. These structures, in unbound and ligand-bound states, have begun to reveal the principles of sugar recognition and discrimination. Here, we highlight recent advances in our understanding of insect GRs.

1. Introduction

The sense of taste, or gustation, plays a pivotal role in the lives of animals, orchestrating a diverse array of essential behaviors that are critical for their survival and reproduction. These behaviors encompass fundamental actions such as feeding, where taste enables animals to distinguish between nutritious food sources and potentially harmful substances, thereby influencing their food selection and overall fitness. Beyond the realm of nutrition, gustation is also integral to more complex interactions with the environment, including the selection of suitable mates and the identification of appropriate sites for laying eggs (Montell 2009; Baik and Carlson 2020). The diverse ways in which animals utilize their sense of taste underscore its significance in their ecology and evolutionary adaptations. Understanding the molecular mechanisms that underpin gustation, particularly the receptors involved and the chemical compounds they detect, can therefore provide profound insights into a wide spectrum of animal behaviors and their adaptations to various ecological niches.

Given the nearly infinite diversity of chemical compounds, recognizing a specific arrangement of atoms within a complex milieu is a remarkable trait. Chemical senses are traditionally—and contentiously—classified into olfaction (smell) and gustation (taste), following an anthropomorphic categorization based on human neuroanatomy. This classification is limiting as it does not fully encompass the diverse mechanisms by which many animals recognize molecules, such as the detection of volatile compounds independently of nasal structures or the existence of extraoral gustatory receptors that sense substances beyond tastants (Behrens et al. 2021; Derby and Caprio 2024). Nevertheless, this classification will be maintained throughout this review for consistency and clarity, with a particular focus on the chemoreception of external stimuli by insects.

Prokaryotes and early algae exhibit basal chemotactic systems that drive diverse behaviors (Shimizu et al. 2010; Choi et al. 2016; Stirling et al. 2024; Brunet et al. 2025). As organisms adapted to distinct ecological niches, evolution gave rise to several specialized organs across Animalia to track down or avoid substances (Fig. 1a, b). In addition, animals lost and acquired chemosensory capabilities based on contextual environmental pressures, such as changes in nutritional needs or food source availability (Ferrero et al. 2012 ; Jiang et al. 2012; Baldwin et al. 2014). Hummingbirds, which evolved from insectivorous ancestors, repurposed their umami taste receptors to detect carbohydrates, facilitating a shift in their dietary preference toward flower nectar (Baldwin et al. 2014). Likewise, loss-of-charge mutations tuned large primates' umami receptors to sense proteins in plants rather than insects (Janiak et al. 2018; Toda et al. 2021). Conversely, some chemoreceptors are preserved across hundreds of millions of years, as evidenced by the conserved tuning of bitter receptors between coelacanth and zebrafish (Behrens et al. 2021).

Fig. 1.

Diversity of chemical recognition in the animal kingdom. a) Phylogenetic tree of selected Metazoan species. The estimated timescale is shown at the bottom. b) Chemosensory organs of various species. c) Examples of cryo-EM structures of taste receptors bound to ligands: CRT1 and diosgenin (PDB 8EIS); BmGr9 and D-fructose (PDB 8UVU); T2R46 and strychnine (PDB 7XP6); T1R2/T1R3 and sucralose (PDBs 9NOT and 9NOV). Ligand locations and chemical structures are highlighted.

Insects have wandered this planet for about 450 million years and are indispensable architects of nearly every terrestrial ecosystem (Misof et al. 2014). They serve as primary decomposers, recycling nutrients, and as vital pollinators, ensuring plant reproduction. They form critical links in food webs, supporting numerous bird, mammal, and amphibian populations as a primary food source (Baik and Carlson 2020; Stephens et al. 2023). Unlike mammals, insects detect chemicals in their environment using sensory neurons located throughout their bodies, especially in their mouthparts, antennae, legs, and wings (Vosshall and Stocker 2007; Baik and Carlson 2020). These distributed taste organs express gustatory receptors (GRs) that detect the presence of sugars, an indication of nutritive value, and bitter compounds, an indication of toxicity, and many other chemical signals. This review will explore recent advances in our understanding of the GR family, the biophysical mechanisms that underlie chemical recognition, and the tuning mechanisms that link ligand binding to receptor activation.

2. Diversity of animal chemosensory receptors

Animals around the world share the same Earth and experience similar physical and chemical environments. It is therefore not surprising that many distantly related species have evolved to rely on similar receptor types to detect these universal phenomena. For instance, opsins, the primary photoreceptors involved in both visual and nonvisual stimuli, are ubiquitous across the animal kingdom, with their evolutionary origin traceable to ctenophores (Shichida and Matsuyama 2009; Schnitzler et al. 2012; Cronin and Johnsen 2016). Similarly, Piezo and Transient Receptor Potential (TRP) ion channels are essential for somatosensation and mechanotransduction in both vertebrates and invertebrates (Sidi et al. 2003; Cheng et al. 2010; Moroni et al. 2018; Moon et al. 2021; Millet et al. 2022; Tadge et al. 2024). By contrast, chemoreception relies on a diverse array of independently evolved chemoreceptors, likely reflecting the differing ecological and evolutionary pressures that shape species-specific sensory demands (Fig. 1b). For instance, humans detect sugars, amino acids, and bitter compounds using several classes of G-protein coupled receptors (GPCRs) found in their tongues (Nelson et al. 2001; Breslin and Spector 2008). Arthropods, however, detect these molecules utilizing structurally unrelated ionotropic GRs that are expressed in a wide variety of sensory organs (Dahanukar et al. 2007; Vosshall and Stocker 2007; Sato et al. 2011). C. elegans uses GPCRs expressed in polymodal nociceptive neurons to avoid bitter molecules, but cephalopods rely on pentameric ionotropic channels similar to nicotinic acetylcholine receptors (AChRs) found in their suckers (Hilliard et al. 2004; Allard et al. 2023).

While much has been learned about animal chemoreceptors through genetic, behavioral, and biochemical studies, the molecular mechanisms of chemoreception have remained elusive for many systems. More recently, advances in single-particle cryogenic electron microscopy (cryo-EM) methodology have enabled a more detailed understanding of this process. Several structures of animal chemoreceptors have been elucidated in recent years (Fig. 1c). For example, the structure of chemotactile receptor for terpenes 1 (CRT1) of an octopus showed a hydrophobic pocket that accommodates hydrophobic molecules, unlike the vertebrate ortholog's hydrophilic pocket that binds acetylcholine (Allard et al. 2023). The first structure of a human GPCR taste receptor revealed how strychnine, a bitter tastant, activates the type 2 taste receptor 46 (TAS2R46) through movement of an aromatic residue (Xu et al. 2022). An analogous mechanism was observed in the narrowly tuned Bombyx mori gustatory receptor 9 (BmGr9), where D-fructose selectively engages aromatic residues that sit between the pocket and ion-conducting pore (Gomes et al. 2024). Although representing only a tiny subset of the extensive repertoire of animal chemoreceptors, these structural investigations have significantly enhanced our understanding of chemosensory biology and how chemical recognition drives animal behavior.

3. The superfamily of insect chemoreceptors

GRs, together with the related olfactory receptors (ORs), comprise a major class of insect chemoreceptors. Our understanding of GRs (and ORs) has advanced significantly over the past 2 decades. GRs were first identified in the fruit fly Drosophila melanogaster through molecular and genetic studies (Clyne et al. 2000). Initial sequence analysis suggested that these receptors likely contain 7 transmembrane helices. Therefore, it was originally hypothesized that these insect chemoreceptors were GPCRs like the mammalian olfactory receptors (Clyne et al. 2000). However, subsequent experiments, including antibody-tagging studies, revealed that the C-terminus of GRs, like ORs (Benton et al. 2006), is oriented extracellularly (Zhang et al. 2011), opposite to the topology of canonical GPCRs. Further research demonstrated that GRs and ORs operate independently of classical GPCR signaling pathways and instead function as ligand-gated ion channels (Sato et al. 2008, 2011). Structural modeling based on evolutionary covariation supported the idea that GRs and ORs adopt a transmembrane architecture distinct from GPCRs (Hopf et al. 2015). More recently, some of the most compelling evidence came from structural studies showing that GRs adopt a novel tetrameric fold with a quadrivial ion-conducting pore (Chen et al. 2024; Frank et al. 2024; Gomes et al. 2024; Ma et al. 2024), similar to the architecture of ORs (Butterwick et al. 2018; del Marmol et al. 2021). Collectively, these findings established that insect GRs are structurally and functionally distinct from mammalian chemoreceptors.

Originally, scientists believed that GRs were unique to arthropods. However, recent work highlighted that these receptors are likely part of a larger superfamily of seven-transmembrane ion channels (7TMICs) (Fig. 2a). Sequence and structure-based screening identified homologs in many other species, first in yeast and plants (Benton 2015; Robertson 2015) and later in humans and unicellular eukaryotes (Benton et al. 2020; Benton and Himmel 2023). The most recent work, which used 3D structure-based screening, ab initio protein folding predictions, phylogenetics, and expression analyses to identify potential homologs, uncovered similarities between 7TMICs and the regrettably named Putative Homeodomain Transcription Factor (PHTF) family (Benton and Himmel 2023). Human homologs have been identified within the PHTF family and are expressed in testis, cerebellum, and muscle, providing evidence that this protein family, originally thought to only exist in invertebrates, may have a common eukaryotic ancestor. However, whether these proteins function as chemoreceptors remains unknown.

Fig. 2.

Superfamily of 7TMICs and insect gustatory reception specialization. a) Unrooted phylogenetic tree of 7TMIC superfamily generated using data from Benton and Himmel (2023), color-coded according to different classes. Select species from each class of 7TMIC, color-coded accordingly. b) Phylogenetic analysis of D. melanogaster (Dm) and B. mori (Bm) gustatory receptors, color-coded according to ligand sensitivity. Gustatory receptors that have been referenced in the text are in bold font.

Despite orthologous receptors being found across the animal and plant kingdoms, the use of GRs for gustation seems to be unique to arthropods (Gardiner et al. 2009; Robertson 2019). Among insects, GR gene numbers vary widely, which may reflect differences in ecological niche and feeding behavior. For instance, the generalist herbivore Helicoverpa armigera possesses a highly expanded GR repertoire, comprising 197 genes (Xu et al. 2016), whereas the honey bee Apis mellifera, which primarily relies on floral resources and social communication for nutritional intake, expresses only 12 (Robertson and Wanner 2006). The model organism D. melanogaster expresses 68 GRs (Dunipace et al. 2001; Scott et al. 2001; Robertson et al. 2003), whereas mosquito species show intermediate numbers, with Aedes aegypti expressing 95 and Anopheles gambiae 76 (Hill et al. 2002; Matthews et al. 2018).

GRs are broadly categorized into 4 functional groups: bitter receptors (including a subset specialized for pheromone detection), sugar receptors (divided into 2 clades, one specific to D-fructose and the other responsive to a range of sugars), carbon dioxide receptors, and a smaller group responsive to heat (Fig. 2b) (Robertson and Kent 2009; Montell 2009, 2013, 2021; Agnihotri et al. 2016). Notably, bitter receptor clades often show species-specific segregation, likely reflecting evolutionary adaptations to distinct ecological niches (Robertson 2019). While most individual GRs fall within these broad functional categories, this classification is not absolute. For instance, D. melanogaster Gr32a (DmGr32a) has been implicated in both bitter taste perception (Weiss et al. 2011; Dweck and Carlson 2020) and pheromone detection (Moon et al. 2009). Similarly, the D. melanogaster Gr28 family (discussed in more detail later) includes receptors that are not only heat-sensitive but also responsive to appetitive RNA and ribonucleosides (Mishra et al. 2018; Fujii et al. 2023) as well as bitter compounds (Ahn and Amrein 2023). These examples underscore the functional diversity and complexity within the GR family. Moreover, many GRs remain orphan receptors with no identified ligands; in such cases, putative functions are typically inferred from sequence similarity to characterized receptors (Robertson 2015; Agnihotri et al. 2016).

Identifying specific ligands for many GRs is challenging, a difficulty compounded by the fact that multiple GRs are often co-expressed within the same neuron and many appear to function as heteromeric channels. Fundamental in vivo studies have shown that members of the sugar GR family operate through a combinatorial mechanism (Dahanukar et al. 2007; Jiao et al. 2008; Yavuz et al. 2014; Fujii et al. 2015), and similar work has explored how numerous bitter receptor GRs are required for ligand detection (Weiss et al. 2011; Kim et al. 2016; Sung et al. 2017). To better understand their function, many GRs—both those that form homomeric and heteromeric channels—are now being expressed and studied in controlled heterologous systems (Table 1). Xenopus laevis oocytes excel in producing high expression levels suitable for robust electrophysiological characterization due to their large size and low intrinsic noise, while cultured cells offer higher throughput potential and environments ranging from mammalian (e.g. HEK293) to insect (e.g. Sf9, S2). Indeed, subtle variations in activity have been observed for chemoreceptors expressed in different systems, suggesting cellular context affects receptor function (e.g. Hou et al. 2020).

Table 1.

Heterologously expressed GRs.

Class	GR	Activating ligand(s)	Species	# subunits	Expression system	References
Sugar	DmGr5a	Trehalose	D. melanogaster	1	S2 cells	Chyb et al. (2003)
	DmGr64a	D-Fructose, Trehalose, Sucrose, Maltose	D. melanogaster	1	HEK293 cells	Chen et al. (2024)
	BtGr1	Sucrose	B. tabaci	1	Xenopus oocytes	Aidlin et al. (2023)
	AmGr1	Sucrose, Glucose	A. mellifera	1	Xenopus oocytes	Değirmenci et al. (2023)
	AmGr2	Sucrose, Glucose	A. mellifera	1 to 2a	Xenopus oocytes	Değirmenci et al. (2023)
	BmGr10	Inositol	B. mori	1	Xenopus oocytes, HEK293 cells	Kikuta et al. (2016)
	TcGR64f1	Sucrose	T. chilonis	2	Xenopus oocytes	Liu et al. (2020a)
	TcGR64f2	Sucrose	T. chilonis	2	Xenopus oocytes	Liu et al. (2020a)
	NlGr10a	D-Fructose, Cellobiose	N. lugens	1	Sf9 cells	Chen et al. (2019)
	NlGr10b	Arabinose	N. lugens	1	Sf9 cells	Chen et al. (2019)
Fructose	DmGr43a	D-Fructose	D. melanogaster	1	Xenopus oocytes, HEK293 cells	Sato et al. (2011) and Chen et al. (2024)
	BmGr9	D-Fructose	B. mori	1	Xenopus oocytes, HEK293 cells	Sato et al. (2011) and Gomes et al. (2024)
	HaGr4	D-Fructose	H. armigera	1	Xenopus oocytes	Jiang et al. (2015)
	TcGr43a	D-Fructose	T. chilonis	1	Xenopus oocytes	Liu et al. (2019a)
	SlGr8	D-Fructose	S. litura	1	Xenopus oocytes	Liu et al. (2019b)
	PxGr43a-1	D-Fructose	P. xylostella	1	Xenopus oocytes	Liu et al. (2020b)
	PxGr43a-2	D-Fructose	P. xylostella	1	Xenopus oocytes	Liu et al. (2020b)
	AmGr3	D-Fructose	A. mellifera	1	Xenopus oocytes	Değirmenci et al. (2023)
	PxGr1	Synephrine	P. xuthus	1	Sf9 cells	Ozaki et al (2011)
Bitter	DmGr8a/DmGr66a/DmGr98b	L-Canavanine	D. melanogaster	3	S2 cells	Shim et al. (2015)
	PrGr28	Sinigrin	P. rapae	1	Xenopus oocytes	Yang et al. (2021)
	HaGr180	Coumarin	H. armigera	1	Xenopus oocytes	Chen et al. (2022)
CO2	DmGr21a/DmGr63a	Sodium Bicarbonate	D. melanogaster	1 to 2	Xenopus oocytes	Ziemba et al. (2023)
	HaGr1/HaGr2/HaGr3	Sodium Bicarbonate	H. armigera	2 to 3	Xenopus oocytes	Ning et al. (2016)
	CqGr1/CqGr2/CqGr3	Carbon Dioxide	C. quinquefasciatus	2 to 3	Xenopus oocytes	Xu et al. (2020)
Temperature	DmGr28b.d	Heat	D. melanogaster	1	HEK293 cells	Capek et al. (2025)

^aAuthors noted slight activation with AmGr2 when individually expressed, albeit very low levels of current were detected.

Moving forward, it will be critical to integrate both in vitro and in vivo approaches to validate heteromeric states and clarify the functional properties of GRs. Numerous examples in the field highlight how different expression systems can yield conflicting results. For instance, the Drosophila carbon dioxide receptor requires both DmGr21a and DmGr63a for proper function in vivo (Suh et al. 2004; Jones et al. 2007; Kwon et al. 2007), yet some studies have reported CO₂ responses from individual subunits expressed in oocytes (Ziemba et al. 2023). Similarly, DmGr5a—originally shown to respond robustly to trehalose when expressed alone in Drosophila S2 cells (Chyb et al. 2003)—was later found to be insufficient on its own in vivo (Jiao et al. 2008). Genetic deletion of the DmGr64a–f cluster significantly reduced trehalose-driven feeding behavior, suggesting that Gr5a requires co-expression with other sugar GRs to function effectively (Jiao et al. 2007, 2008; Slone et al. 2007). Although heterologous expression has been critiqued for producing inconsistent results (Amrein 2014), in vivo data also underscore the complexity of interpreting receptor function, especially where expression levels may modulate receptor tuning. For example, DmGr5a may be capable of forming a functional homomeric complex, but if expressed at lower levels than its heteromeric species, its contribution may be obscured in physiological contexts. Behavioral readouts and action potential measurements may fail to detect such subtle contributions.

To address these challenges, it is essential to apply both approaches with careful attention to expression dynamics. Overexpression in heterologous systems can produce artifacts, while low levels of endogenous expression may mask receptor function altogether. Thus, interpreting GR activity requires a nuanced understanding of expression context, stoichiometry, and cellular environment. Our review primarily focuses on recent advancements in heterologous systems as in vivo work has been extensively reviewed elsewhere (Hallem et al. 2006; Montell 2009; Agnihotri et al. 2016; Benton 2017; Xu et al. 2020; Arntsen et al. 2024). Notably, expression in heterologous systems have enabled major advances in GR structural biology, providing unprecedented insight into receptor architecture. These tools are instrumental in uncovering binding sites, subunit organization, and activation mechanisms—features that cannot be resolved through sequence analysis or heterologous expression alone.

4. Major families of insect GRs

The insect GR family is highly diverse, with receptors capable of detecting a wide range of chemical and nonchemical stimuli. While recent structural breakthroughs have significantly advanced our understanding of some insect taste receptors, particularly sugar-sensing GRs, several crucial families remain structurally uncharacterized. We anticipate that many of the structural features revealed by sugar receptors will be conserved across other GR families. Future structural studies will be invaluable for identifying conserved motifs and mechanisms across the broader GR family, particularly regarding whether the location of the ligand-binding pocket is consistent and if diverse ligands employ a similar S5-bridge mechanism to couple ligand binding to pore opening. Structural studies will also be important for elucidating the different roles of each subunit in heteromeric GR complexes.

4.1. Sugar receptors

Sugars are the most abundant biomolecules on the planet and are a primary energy source for nearly all living organisms. Insect gustatory receptor neurons express many GRs dedicated to recognizing sugars. Drosophila, for example, have at least 8 discrete sets of sweet-sensing neurons that express different combinations of the 9 sugar GRs (DmGr5a, DmGr43a, DmGr61a, and DmGr64a-f) (Fujii et al. 2015). While it is thought that sugar GRs may form heteromeric complexes in vivo (Dahanukar et al. 2007; Wanner and Robertson 2008; Değirmenci et al. 2023), several have been shown to produce functional receptors when expressed in isolation. The first such receptor was DmGr5a, whose transient expression in Drosophila S2 cells yielded robust responses to the disaccharide trehalose (Chyb et al. 2003). Recently, 4 groups independently determined the structures of 2 types of receptors within the sugar-sensing GR family: a set of homologous fructose-selective receptors and a disaccharide-sensing receptor (Chen et al. 2024; Frank et al. 2024; Gomes et al. 2024; Ma et al. 2024). These molecular descriptions of sugar GRs highlight the similarities and differences by which insects can detect and discriminate among sweet compounds.

Several structures from the clade of fructose-selective receptors were determined: Gr9 from B. mori (Frank et al. 2024; Gomes et al. 2024) and Gr43a from D. melanogaster or D. mojavensis (Chen et al. 2024; Ma et al. 2024). Receptors from this clade are highly conserved among insect species and are remarkably selective for the monosaccharide D-fructose (Wanner and Robertson 2008; Sato et al. 2011; Miyamoto et al. 2012; Miyamoto and Amrein 2014). Unlike other sugar receptors, fructose receptors are expressed independently in several neuronal populations in the gut and brain and are thought to be involved in regulating feeding and satiety (Sato et al. 2011; Miyamoto et al. 2012; Xu et al. 2012; Miyamoto and Amrein 2014). After a sugar-rich meal, metabolizable sugars are rapidly converted to D-fructose in the insect's gut, potentially via the polyol pathway (Miyamoto et al. 2012; Miyamoto and Amrein 2014), resulting in an increase in D-fructose concentration in the hemolymph. By contrast, the concentrations of D-glucose and trehalose, the most common sugars found in the hemolymph, remain stable, making fructose an optimal signal for modulating satiation (Miyamoto et al. 2012; Miyamoto and Amrein 2014).

Although Gr43a-like receptors are generally described as fructose-specific, in vivo work showed that peripheral tarsal neurons of mutant flies expressing DmGr43a and not any other sugar GR responded to both fructose and sucrose. However, central brain neurons remained strictly fructose-tuned, hinting at an unknown regulatory mechanism that constrains ligand selectivity according to neuronal context (Miyamoto et al. 2012). Moreover, a Gr43a homolog in Bactrocera dorsalis was recently shown to sense sucrose rather than D-fructose, raising the possibility that ligand preference in this receptor family may have diverged across Dipterans (Dong et al. 2024).

4.2. Bitter receptors

Bitter GRs form the largest subfamily in many insect species (Fig. 2b). For example, the agricultural pest Helicoverpa armigera genome contains 197 GRs, of which 180 comprise a bitter-sensing family (Xu et al. 2016). This diversity has been suggested to reflect the structural diversity of bitter compounds, especially those from plants that are detected by herbivorous insects, in comparison to sugars (Montell 2021). Significant effort has been directed toward mapping the expression patterns of bitter receptors and identifying which neurons respond to bitter compounds (Thorne et al. 2004; Weiss et al. 2011; Delventhal et al. 2017; Sung et al. 2017). In Drosophila, for example, there are 33 bitter GRs expressed within the labellum. DmGr32a, DmGr33a, DmGr39a, DmGr66a, and DmGr93a are often considered the “core” bitter receptors because they are expressed in all bitter-sensing neurons, and their removal abolishes responses to several compounds (Lee et al. 2009; Weiss et al. 2011; Dweck and Carlson 2020).

A key unanswered question that remains is how many and which subunits are required to form a functional receptor. For example, DmGr8a, DmGr66a, and DmGr98b, together, can sense L-canavanine, a plant-derived insecticide (Shim et al. 2015). Coumarin, another plant-produced toxic compound, is avoided by Drosophila, especially during oviposition. While DmGr33a, DmGr66a, and DmGr93a contribute to coumarin sensing, only DmGr33a is necessary for coumarin avoidance in an egg-laying behavioral assay (Poudel and Lee 2016). However, another study shows that coumarin detection, along with theophylline, theobromine, caffeine, and umbelliferone, requires 4 subunits for an electrophysiological response in sugar neurons (Dweck and Carlson 2020). Strychnine detection may require 5 GRs (Poudel et al. 2017). Given that GRs are thought to form tetramers, this raises the possibility of multiple functional channels acting synergistically.

Similar to some olfactory neurons, Drosophila labellar bitter neurons generate both ON (during stimulation) and distinct OFF responses (upon bitter stimulus removal), a mechanism mediated by a particular set of bitter GRs. Researchers hypothesize that bitter OFF responses could generate a lingering signal—or “aftertaste” —that reinforces aversive cognition and dictates a behavioral output (Devineni et al. 2021; Dweck and Carlson 2023). How might chemoreceptors generate OFF responses? One possibility is that bitter GRs cycle through an inhibited conformational state such that ligand unbinding triggers rebound activation, leading to an action potential. Structural studies and detailed electrophysiological analyses of bitter GRs will be critical to test this idea and to clarify subunit stoichiometry for channel function, ligand specificity of subunits, the presence of potential allosteric sites, and the mechanisms of inter-subunit communication.

4.3. Pheromone receptors

Within the bitter family lies a subclass of GRs that respond to pheromones. Pheromones are generally long-chain hydrocarbons whose chemical characteristics differ dramatically from sugars and bitter compounds (Yew and Chung 2015). Pheromone detection plays an important role in insect courtship, mating, and post-mating behaviors. Disrupting pheromone detection is a viable strategy for trapping disease-carrying insects. For example, fatty acid methyl esters are detected by male tsetse flies, the carriers of African sleeping sickness, and promote mating behaviors (Ebrahim et al. 2023). When another tsetse species was treated with the pheromone methyl palmitoleate, Glossina morsitans males mounted Glossina fuscipes females, illustrating the potential of pheromones as attractants. While olfactory receptors detect volatile pheromones, contact pheromones are detected via ionotropic receptors (IRs), pickpocket (Ppk) channels, and GRs. OR, IR, and Ppk detection of pheromones are described elsewhere (see Montell 2021).

In Drosophila, several GRs have been identified as potentially involved in contact pheromone detection. DmGr32a, for example, is expressed in the Drosophila labellum, pharynx, and tarsal leg segments. Gr32a male mutants show high courtship towards other males and previously mated females, indicating a role for this receptor in mate detection (Miyamoto and Amrein 2008). More specifically, this receptor is thought to detect 7-tricosene (7-T), a male pheromone involved in female receptivity and mate-guarding. Male Gr32a-positive neurons sense 7-T and suppress courtship with other males, non-virgin females, and females of other Drosophila species. However, DmGr32a is widely expressed, suggesting that other co-expressed subunits likely contribute to selective 7-T detection (Moon et al. 2009). Other potential pheromone receptors that have been reported in Drosophila include DmGr68a, DmGr39a, and DmGr33a (Bray and Amrein 2003; Ejima and Griffith 2008; Watanabe et al. 2011; Hu et al. 2015). There remains a notable absence of biochemical and structural data on pheromone receptors, which would elucidate which pheromones activate them and how their unique tuning is achieved. Structures of pheromone receptors will likely reveal highly unique binding pockets compared to those of sugar GRs, given the hydrophobic nature of pheromone ligands (Fig. 2c).

4.4. Carbon dioxide receptors

Many insect species, including flies, ants, termites, mosquitoes, and honeybees, can detect gaseous carbon dioxide (CO₂) (Jones 2013), leading to a variety of behavioral responses (Torr 1990; Healy and Copland 1995; Suh et al. 2004; Dekker et al. 2005; Buehlmann et al. 2012; Wasserman et al. 2013). CO₂-sensitive neurons have been characterized on the antennae, maxillary palps, and labial appendages of many insect species (Stange and Stowe 1999; Lu et al. 2007), and the genes responsible for CO₂ detection have been identified. In D. melanogaster, chemoreceptors DmGr21a and DmGr63a are expressed in the ab1C neuron and are both required to mediate CO₂ sensitivity (Suh et al. 2004; Jones et al. 2007; Kwon et al. 2007, but see Ziemba et al. 2023). Orthologs in mosquito species have been identified in specialized cpA neurons as Gr1 and Gr3 (Robertson and Kent 2009). Additionally, in mosquitoes and many other non-Drosophilid species, a second Gr1 paralog receptor has been identified and called Gr2 (McMeniman et al. 2014). In these species, studies have suggested that only 2 of the 3 CO₂ receptor genes are required to detect CO₂, and the third may serve to modulate responses to CO₂ and other ligands (Ning et al. 2016; Kumar et al. 2020; Xu et al. 2020).

A challenging question that structural data could resolve is whether the endogenous ligand is CO₂ or sodium bicarbonate. While studies using Drosophila and mosquito CO₂ receptors in the empty neuron system have indicated CO₂ as the endogenous ligand (Jones et al. 2007; Kwon et al. 2007), work using Xenopus oocytes suggested that sodium bicarbonate can also activate the channel (Ziemba et al. 2023; but see Xu et al. 2020). The identity of the endogenous ligand thus remains an open question. Other research has identified species-specific agonists and antagonists. For example, in vivo screening was used to identify CO₂ receptor agonists and antagonists specific to mosquito and Drosophilid species (Turner and Ray 2009; Turner et al. 2011). Structural studies could illuminate whether these receptors bind ligands in the canonical ligand-binding pocket, shed light on ligand specificity across different species, and define the contributions of individual subunits in ligand gating.

4.5. Gr28 family

While most insect GRs are specialized for detecting chemical compounds, the highly conserved Gr28 family has evolved to sense a variety of non-canonical ligands (Fujii et al. 2023). In Drosophila, the Gr28 family comprises six GRs: DmGr28a and 5 isoforms—DmGr28b.a through DmGr28b.e—each of which expresses a unique exon from distinct transcriptional start sites (Thorne and Amrein 2008; Ni et al. 2013; Montell 2013). In vivo studies have revealed intriguing roles for these receptors in Drosophila larvae, including mediating the appetitive taste of RNA and ribonucleosides (Mishra et al. 2018 ; Fujii et al. 2023) as well as detecting bitter compounds (Ahn and Amrein 2023). However, there remain many open questions about the roles of other members of this family.

One particularly interesting function involves light avoidance. It has been proposed that at least one DmGr28b isoform enables Drosophila larvae to avoid light stimuli (Xiang et al. 2010). This hypothesis is supported by the fact that DmGr28b shares distant homology with the C. elegans LITE protein, a GR-like receptor activated by ultraviolet light (Edwards et al. 2008; Ward et al. 2008; Liu et al. 2010). Xiang et al. (2010) identified specific larval neurons responsive to ultraviolet, violet, and blue light, and found that DmGr28b mutations significantly reduced these light responses. However, the precise DmGr28b isoform responsible for this behavior remains unidentified.

Recent studies have clarified the function of the specific DmGr28b.d. isoform. Previously, it was unclear whether temperature preference in Drosophila was mediated by internal or external sensors (Hamada et al. 2008; Gallio et al. 2011). Subsequent research showed that both TRPA1 and DmGr28b.d contribute to thermal sensing (Ni et al. 2013). TRPA1 functions as an internal sensor for subtle temperature changes, whereas DmGr28b.d is expressed in “hot cells” located in the arista—a feathery structure on the antenna—where it detects larger temperature shifts. Remarkably, ectopic expression of DmGr28b.d in other cell types conferred heat sensitivity, providing the first evidence that a gustatory receptor in insects can function as a thermosensor. Further analysis of other DmGr28b isoforms revealed that only DmGr28b.e could restore temperature-sensitive behavior in mutant flies. However, because DmGr28b.e is not naturally expressed in the arista, it may serve a temperature-sensing function elsewhere in the body (Thorne and Amrein 2008; Montell 2013; Ni et al. 2013). More recently, Capek et al. (2025) investigated DmGr28b.d thermal activation across Drosophila species from diverse climates and found that the activation threshold of DmGr28b.d correlates with species-specific behavioral heat avoidance thresholds. GRs that mediate light avoidance and temperature sensing are structurally intriguing, as they do not interact with traditional chemical ligands. This raises compelling questions about whether the structural features observed in sugar-sensing GRs are conserved in these non-canonical, non-chemical-sensing GRs.

5. Fundamental structural features of GRs

Significant advances in structural biology have begun to illuminate the overall structural features of insect GRs and the molecular mechanisms by which they bind to their specific chemical ligands. A pivotal contribution in this area has been the recent determination of the three-dimensional structures of sugar-sensing GRs from the silk moth B. mori (BmGr9), D. melanogaster (DmGr43a and DmGr64a), or D. mojavensis (DmojGr43a) using cryo-electron microscopy (Chen et al. 2024; Frank et al. 2024; Gomes et al. 2024; Ma et al. 2024). These structures, determined in the absence or presence of sugar molecules, provide atomic-level views of how these receptors interact with their chemical cues, revealing the precise binding pockets and conformational changes that lead to receptor activation. This unprecedented structural insight moves beyond inferential studies, enabling a mechanistic understanding of insect taste receptors.

Like insect ORs (Butterwick et al. 2018; Del Marmol et al. 2021; Wang et al. 2024; Zhao et al. 2024), GRs are tetramers with 4 protomers symmetrically arranged around a quadrivial aqueous pore, which consists of a central extracellular opening connected to 4 lateral conduits leading to the cytosol (Fig. 3a). Each protomer has seven-transmembrane helices (S1 to S7), with S7 subdivided into S7a and S7b, separated by a β-hairpin loop. S7a and the intracellular portions of S4, S5, and S6 constitute the anchor domain, the cytosolic region where most intermolecular interactions between protomers occur to stabilize the quaternary structure. S7b from each of the 4 protomers lines the central ion-conducting pore and contains the only GR signature motif, TYhhhhhQF, with “h” being any hydrophobic residue (Robertson 2015). While the conserved threonine (T) and tyrosine (Y) residues do not seem to play an obvious role in channel gating, the glutamine (Q) and phenylalanine (F) alternate their orientations in open versus closed channel conformations (Frank et al. 2024; Gomes et al. 2024; Ma et al. 2024). In the absence of an agonist, the extracellular ends of S7b interact with one another, and the conserved phenylalanine side chain from each helix projects into the pore, creating a hydrophobic gate that constricts the central pore at the extracellular membrane surface, providing a barrier to prevent ion conduction (Fig. 3b). Other nearby hydrophobic side chains that line the pore are also involved in constricting the channel. When bound to a ligand, the ends of S7b tilt away from the pore axis, and the hydrophobic groups blocking the pore swing away to be replaced by the neighboring glutamine residue. This structural rearrangement both widens the extracellular gate and creates a polar environment that facilitates the passage of hydrated cations. This hydrophobic-to-hydrophilic “wetting transition” seems to be a general mechanism of pore opening among insect chemoreceptors, as a similar rearrangement was observed in Or5 from the bristletail Machilis hrabei (MhOr5), which has a valine (Val468) instead of a phenylalanine lining the pore in the closed state (Del Marmol et al. 2021).

Fig. 3.

The general activation mechanism of insect GRs. a) Side and b) top views of unbound (left) and ligand-bound GR models (right). Helices are numbered, and important regions are highlighted. The pore helix (S7b) is highlighted, the anchor domain (S7a and cytosolic regions of S4, S5, and S6) is depicted by dashed lines, and the rest of the protein (transmembrane and extracellular segments of S1–S6) is light blue. Cations are depicted as small circles with a positive sign, and their path through the central pore (between S7b's) and lateral conduits (beneath the β-hairpin loops) is shown. Ligands are represented as a circle, which binds to an extracellular-facing pocket within the transmembrane domains of S1–S6. Vertical dashed lines represent the intra- and intermolecular interactions between S5 and S7b, and S5 and the ligand, respectively.

Thus far, all open structures of GRs were determined with saturating amounts of ligand, such that all 4 subunits appear to be bound to a ligand. It remains unclear how many subunits are required to bind the receptor to open the pore and if there is a concerted movement of all 4 protomers to facilitate the passage of ions, or if the movement of the pore helix from a single bound subunit would be enough to asymmetrically activate the channel, as suggested by recent work in heteromeric ORs (Wang et al. 2024; Zhao et al. 2024).

6. Ligand binding in GRs: a case study of sugar recognition

BmGr9 and DmGr43a present almost identical structures with the same set of conserved residues contacting D-fructose (Fig. 4a–f). A tryptophan sits at one side of the pocket (Trp354 in BmGr9, Trp333 in DmGr43a), which interacts with the sugar through CH–π interactions. Notably, an analysis of protein–carbohydrate complexes deposited in the Protein Data Bank revealed that aromatic amino acids, especially tryptophans, are frequently found in proximity to sugars (Hudson et al. 2015). Thus, this sugar-aromatic interaction likely helps to orient the D-fructose molecule so that its hydroxyl groups can all make polar interactions with residues in the binding pocket. D-Fructose interacts with all helices that line the binding pocket (S2–S6) and significantly reduces the pocket size upon binding. The binding of D-fructose also appears to stabilize the S1–S2 extracellular loop: density for this loop is much stronger in the presence of D-fructose, likely due to an interaction between D-fructose and a conserved arginine in the loop (Arg86 in BmGr9; Arg70 in DmGr43a). Generally, mutations in the pocket substantially impair channel function, indicating that pocket chemistry must be preserved to retain D-fructose activity (Frank et al. 2024; Gomes et al. 2024; Ma et al. 2024).

Fig. 4.

Structures of insect sweet-sensing GRs. a) BmGr9 bound to D-fructose (highlighted) EM density (EMD-42629). b) Top and c) side close-up views of BmGr9 interactions with D-fructose in β-pyranose form (PDB 8UVU). d) DmGr43a bound to D-fructose (highlighted) EM density (EMD-36411). e) Top and f) side close-up views of DmGr43a interactions with D-fructose in β-furanose form (PDB 8JMA). g) DmGr64a bound to sucrose (highlighted) EM density (EMD-36421) h) Top and i) side close-up views of the DmGr64a interactions with sucrose (PDB 8JMH). Hydrogen bonds are depicted by dashed lines. Helices and residues are labeled accordingly.

Like many sugars, D-fructose exists in a rapidly interconverting set of ring configurations. The five-membered ring β-D-fructofuranose and the six-membered ring β-D-fructopyranose are the most abundant forms, at 23% and 68%, respectively, under physiological conditions (Shi et al. 2018). Interestingly, the different groups modeled different ring forms in the structures, with DmGr43a bound to pyranose (Ma et al. 2024) and BmGr9 bound to a mixture of both forms (Gomes et al. 2024). Although the pyranose form is more abundant in solution, most structures with D-fructose deposited into the Protein Data Bank include the furanose form. The density for the bound fructose is not sufficiently detailed to discriminate between these 2 forms in any of the structures. Although the ligand density is consistent with both forms binding to the receptors, their relative affinities and whether they both activate the receptor remain unknown.

Several structures of DmGr64a have also been reported (Ma et al. 2024). DmGr64a has a significantly larger pocket than either BmGr9 or DmGr43a, which can easily accommodate larger disaccharide molecules (Fig. 4g–i). Sucrose and maltose bind in the same spot with similar residues interacting with both molecules, although mutating His197 or Thr257 affects activation by sucrose but not maltose (Ma et al. 2024). Like in BmGr9 and DmGr43a, a tryptophan sits at the base of the pocket (Trp260), and most hydroxyls make hydrogen bonds. Several distinct features can be observed in DmGr64a, which are not previously seen in other insect GRs or ORs. For instance, there is a 30° bend in S4 (at Trp260), which widens the ligand binding cavity, and Met115 is (thus far) the only residue in S1 that interacts with a ligand (not illustrated in Fig. 4 for clarity) in insect chemoreceptors. Unlike the binding of D-fructose to BmGr9 or DmGr43a, the binding of neither sucrose nor maltose induced prominent conformational changes in DmGr64a, either in the binding pocket or the pore. Thus, it is unclear what state these structures represent. The concentrations of sugars may have been too low to open the pore, or DmGr64a might reflect a constitutively open state under these conditions, and this protein may need to form an obligate heteromer with other sugar GRs in vivo, as previously suggested (Dahanukar et al. 2007).

7. Chemical tuning of sweet GRs

A notable feature of the homomeric sugar-sensing GRs studied in heterologous systems is their remarkably narrow chemical tuning profiles. BmGr9 and DmGr43a represent an extreme example of this, being activated only by a single type of sugar, D-fructose. The size and chemical properties of the binding pocket are important determinants for ligand recognition and constrain the range of potential ligands in insect chemoreceptors. Sugar GRs present multiple polar side chains to interact with their hydrophilic ligands, which are conserved among receptors with similar binding profiles, like BmGr9 and DmGr43a, but not between different classes of sugar receptors, such as DmGr43a versus DmGr64a. Concordantly, ORs have significantly more hydrophobic pockets to accommodate volatile lipophilic molecules (Butterwick et al. 2018; Del Marmol et al. 2021; Wang et al. 2024; Zhao et al. 2024).

Despite the critical role of the binding pocket, studies show that its characteristics alone do not fully account for receptor tuning. For example, L-sorbose, an epimer of D-fructose, can bind to BmGr9 without activating it, instead inhibiting the D-fructose response (Sato et al. 2011; Gomes et al. 2024). Structural analysis of L-sorbose-bound BmGr9 reveals that while both sugars make nearly identical interactions in the binding cavity, L-sorbose's specific hydroxyl orientation at C5 creates steric hindrance, blocking the movement of an “aromatic bridge” (Tyr332 and Phe333 in S5) (Gomes et al. 2024) (Fig. 5a, b). Conversely, D-fructose's orientation allows this bridge to engage, thereby disrupting inter-subunit interactions in the pore and leading to its opening. This suggests that the precise spatial configuration of the sugar is critical for triggering the allosteric activation pathway in fructose receptors. While the S5-bridge motif is conserved in related receptors, whether this aromatic residue movement is a universal activation mechanism for all GRs remains an open question. However, S5 clearly serves as a crucial link connecting ligand binding to channel activation. The emerging model for sugar GR tuning, particularly in fructose receptors, emphasizes the coupling of pocket chemistry to an allosteric activation pathway selectively triggered by the sugar's spatial arrangement (Fig. 5c).

Fig. 5.

Chemical tuning of insect fructose receptors. a) Top views of apo BmGr9 (white, PDB 8UVT) and D-fructose-bound BmGr9 (blue, PDB 8UVU) aligned, showing conformational changes induced by the sugar. b) Top views of apo BmGr9 (white, PDB 8UVT) and L-sorbose-bound BmGr9 (pink, PDB 8VV3) aligned. c) Fischer projections of D-fructose and L-sorbose, two epimers that change hydroxyl orientation at a single chiral center (C5, asterisk). Molecules with similar size and chemistry can bind to the receptor (represented by the bigger circle), although the receptor is narrowly tuned (only activated by) D-fructose (smaller circle).

8. Convergent evolution of sweetness reception

For millennia, humans have wondered what makes a substance taste sweet. Mammals detect a wide range of chemically distinct sweet-tasting molecules, including natural sugars, artificial sweeteners, amino acids, and sweet proteins (Fig. 6a), using only a single sweet receptor complex composed of 2 class C GPCRs, T1R2, and T1R3 (Zhang et al. 2010; Zheng et al. 2024). Early molecular models focused on the matching of hydrogen bonding partners between the sugars and the receptor (Shallenberger and Acree 1967). This was refined as more chemical structures of sweet and nonsweet molecules were characterized, leading to the development of a multipoint attachment model of interaction (Nofre and Tinti 1996) (Fig. 6b). This model posits that multiple polar and steric (hydrophobic) interactions are involved between sugar and its receptor, including an ionic triad formed by one basic, originally proposed to be a lysine, and 2 bordering acidic amino acids, either aspartate or glutamate. Recently, structures of the human T1R2/T1R3 complex were determined by cryo-EM, showing the sugar-like sweetener sucralose bound to the Venus Flytrap Domain (VFD) of T1R2, interacting with the ionic triad of Lys65, Asp142, and Asp278 (Fig. 6c) (Juen et al. 2025).

Fig. 6.

Sugar recognition mode across species. a) 2D chemical structures of selected sweet molecules (to humans). b) Simplified multipoint attachment theory model of sweet compounds interacting with a sweet taste receptor (adapted from Nofre and Tinti 1996). “Φ” (Phi) represents a hydrophobic residue, “H” represents residues forming hydrogen bonds, while “−” and “+” represent negatively (red) and positively (blue) charged residues, respectively. Vertical dashed lines depict hydrophobic interactions, and dots represent polar interactions. Protein models of c) hT1R2/R3 bound to sucralose (only hT1R2 is shown), d) BmGr9 bound to D-fructose, e) E. coli 2GBP bound to D-galactose. f,g,h) Close-up views of the respective interactions between the sweet compound and receptor. Residues that form the ionic triad are color-coded and oriented as seen in (b).

Remarkably, in the structures of BmGr9 and DmGr43a bound to D-fructose, a T1R2-like ionic triad in the fructose-binding pocket was observed, despite human and insect taste receptors having no shared evolutionary history (Kent and Robertson 2009; Fig. 6d). Notably, although sugars are neutral molecules, the hydroxyl groups of anomeric carbons are relatively more acidic (Feng et al. 2013). Indeed, in BmGr9 and DmGr43a, anomeric hydroxyl groups interact with arginine, a basic residue (Arg86 and Arg70, respectively) (Gomes et al. 2024; Ma et al. 2024). This interaction appears to be crucial, as mutating this arginine to alanine abolishes sugar binding in BmGr9. In contrast, the conservative substitution to lysine retains activity comparable to that of wild-type receptors (Gomes et al. 2024).

The striking similarity of sugar–receptor interactions among unrelated receptors suggests convergent evolution of sugar-binding pockets, indicating that there may be preferred ways in which taste receptors interact with natural sugars. Indeed, similar ionic triads can be found in other sugar-sensing receptors, such as the Glucose Galactose Binding Protein (2GBP) of the prokaryote Escherichia coli bound to D-galactose (Vyas et al. 1988) (Fig. 6e). This chemoreceptor is involved in strong chemotactic behavior, helping E. coli “taste” sugars (Vyas et al. 1988). Here, an arginine (Arg158) is surrounded by aspartic acids (Asp154 and Asp236), while a tryptophan (Trp183) also helps position the galactose. This mechanism may be unique to gustatory detection, as no similar triad is present in the sugar transporters GLUT1 and GLUT5, which transport D-glucose and D-fructose across cell membranes, respectively (Deng et al. 2014; Nomura et al. 2015).

9. Conclusion

The burgeoning structural insights into insect GRs mark a pivotal moment in chemosensory biology, transitioning from inferential models to atomic-level understanding. While sugar-sensing GRs have provided the initial blueprint for their unique architecture and ligand-gating mechanisms, the vast landscape of bitter, carbon dioxide, pheromone, and even thermo/light-sensing GRs remains largely unexplored structurally. Unraveling the three-dimensional structures of these diverse GR families will be crucial for deciphering the full spectrum of insect chemical recognition, resolving lingering questions about ligand identity and subunit stoichiometry, and ultimately accelerating the development of highly specific tools for pest control and disease vector management. The continuing revolution in structural biology promises to unlock the remaining mysteries of these essential insect senses.

Funding

This research was supported by grants from the National Institutes of Health: 1F31AI194815 (to R.A.R.), and 1RM1GM149406 and 1R01GM149636 (to J.A.B).

Data availability

No new data were generated or analyzed in support of this research.