Termites (Blattodea: Isoptera) are the oldest group of eusocial insects; they evolved 140 Ma from wood-dwelling cockroaches within the Blattoidea lineage (Inward et al., 2007; Bucek et al., 2019). Termites display the most pronounced division of labor and caste polyphenism among social insects, manifested by the presence of multiple well-defined caste phenotypes, such as primary and secondary reproductives (kings, queens) and up to several types of workers and soldiers (Roisin and Korb, 2011).

Like other social insects, termites extensively use chemical communication to coordinate the tasks in their colonies, mediate the division of labor, and orient in the environment. Termite chemical communication includes a variety of releaser pheromones directly influencing the behavior of the receivers. These pheromones are used, among others, for marking of foraging trails, mate search, alarm signaling, nestmate recognition, and marking of food sources (Bagnères and Hanus, 2015; Mitaka and Akino, 2021). In addition, termites use primer pheromones having impact on the physiology, reproduction and development of the receivers (Matsuura et al., 2010; Mitaka et al., 2017; Dolejšová et al., 2022).

Trail-following pheromones (TFPs) and sex-pairing pheromones (SPPs) belong among the best studied termite chemical signals. The first TFP has been identified more than five decades ago in the Eastern subterranean termite Reticulitermes virginicus (Matsumura et al., 1968). Since then, TFPs have been characterized in 68 species, and SPPs in 17 species of termites, along with a range of other pheromones (primer pheromones, alarm pheromones, etc.; Bordereau and Pasteels, 2010; Bagnères and Hanus, 2015; Mitaka and Akino, 2021). TFPs and SPPs are noteworthy for their conservation and parsimony with respect to chemical diversity. Only 8 different structures occur as TFP across the studied species, some of which are also used as SPPs, or are closely related to the SPPs. At the same time, similarities in chemistry and glandular origin of TFPs and SPPs pinpoint the shared evolutionary origin of trail-following communication and sex-pairing communication (Bordereau and Pasteels, 2010). Experimentally confirmed or expected chemical identities of TFPs and SSPs in the five species studied here – are provided in Table 1.

Like in other insects, antennae are key chemosensory organs in termites (Saran et al., 2007; Du et al., 2019). The antennal flagellum of moniliform antennae is covered by antennal sensilla of various morphological types (Castillo et al., 2021). Irrespective of their different morphologies, all olfactory sensilla house dendrites of olfactory sensory neurons (OSNs), expressing chemosensory receptor proteins that detect the odorants (Hansson and Stensmyr, 2011).

Odorant Receptors (ORs) and Ionotropic Receptors (IRs) are the two arguably most important protein families involved in the detection of volatile cues in insects (Clyne et al., 1999; Gao and Chess, 1999; Vosshall et al., 1999; Benton et al., 2009; Gomez-Diaz et al., 2018). The ORs and IRs have different structures and are expressed in OSNs associated with distinct olfactory sensillum types (Benton et al., 2009; Scalzotto et al., 2022). Insect ORs are derived from insect Gustatory Receptors (GRs); ORs and GRs form a superfamily with a common phylogenetic origin (Robertson et al., 2003; Thoma et al., 2019). ORs are seven-transmembrane receptors that exhibit unusual topology compared to classical GPCRs, with an intracellular N-terminus. OR proteins function as heteromultimeric channels formed by a ligand-specificity defining OR protein, and the ubiquitous OR-Coreceptor (ORCo; Vosshall et al., 1999; Hansson and Stensmyr, 2011; Butterwick et al., 2018). Silencing ORCo in two termite species (Reticulitermes chinensis and Odontotermes formosanus) impaired their ability to perceive TFPs and perform oriented locomotory behavior, indicating a role of ORs in termite pheromone detection (Gao et al., 2020). ORs exhibit extreme sequence variability, with rapid birth-and-death evolution leading to gene diversifications (McBride and Arguello, 2007; Sánchez-Gracia et al., 2009; Ramdya and Benton, 2010).

ORs can detect a wide range of environmental odors. The remarkable diversity of potential chemical cues, whose importance changes in conjunction with lifestyle changes across insects, presumably drives the rapid birth-and-death evolution seen in ORs (McBride and Arguello, 2007; Sánchez-Gracia et al., 2009; Ramdya and Benton, 2010). Accordingly, OR repertoire sizes differ significantly among insect species, with for example 62 ORs reported in Drosophila melanogaster (Clyne et al., 1999; Gao and Chess, 1999; Vosshall et al., 1999), 131 in mosquitoes (Bohbot et al., 2007), 77 in bark beetles and as many as 375 ORs in the ant Acromyrmex echinatior (Zhou et al., 2015). The evolution of this gene family is clearly driven by the diversity of chemical recognition needs, and the evolution of eusocial organization in social Hymenoptera (bees, ants, wasps) shows correlation with the diversification of OR repertoire (Yan et al., 2020).

Gustatory receptors (GRs) are seven-transmembrane domain proteins that are first identified in D. melanogaster, involved in the detection of tastants: sugar, bitter and CO₂ (Clyne et al., 2000; Scott et al., 2001). Together with ORs they form a superfamily of proteins. They share the same inverted topology when compared to classical G-protein coupled receptors (Scott et al., 2001; Zhang et al., 2011). Molecular evolutionary analysis has proposed GRs as ancestral sequences to ORs (Robertson et al., 2003). A number of GRs have been functionally characterized in D. melanogaster, DmelGR32a and DmelGr68a are involved in pheromone detection (Montell, 2009), and DmelGr21a and DmelGr63a, which function together as CO₂ receptor (Jones et al., 2007). Various other fly receptors are involved in the detection of sugars and bitter tastants (Dahanukar et al., 2001, 2007; Vosshall and Stocker, 2007; Montell, 2009; Isono and Morita, 2010; Freeman et al., 2014; Delventhal and Carlson, 2016).

IRs are transmembrane proteins distantly related to a variant of ionotropic glutamate receptor receptors, iGluRs, expressed with co-receptors IR8a or IR25a (Benton et al., 2009; Croset et al., 2010). In general, iGluRs are classified into subfamilies defined by their main agonist: α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), kainate, and N-methyl-D-aspartate (NMDA; Croset et al., 2010). AMPA and kainate receptors were further verified by the presence of threonine (T) in the first half of the S2 domain; iGluRs that lack this T residue were classified as NR1 or delta variants (Benton et al., 2009). IRs lack the conserved aspartate (D) or glutamate (E) that interacts with the α-amino group of the glutamate ligand, that non-IR iGluRs possess (Benton et al., 2009). Like other iGluRs, IRs function as heterotetrameric channels; they participate in various sensory modalities, including olfaction. Antennal IRs are considerably more conserved than ORs, with a similarly conserved ligand profile (Benton et al., 2009; Abuin et al., 2019).

Additionally, other non-receptor proteins like odorant binding proteins (OBPs), chemosensory proteins (CSPs), and sensory neuron membrane proteins (SNMPs; Hansson and Stensmyr, 2011; Leal, 2011) are involved in olfaction (Zhou et al., 2006). OBPs and CSPs are both small, globular proteins that are present in high concentrations within the sensillum lymph (Venthur and Zhou, 2018; Pelosi et al., 2018b) and both bind odorant molecules (Pelosi et al., 2005). Their exact function is so far unclear, but there is evidence that they might be involved in the transport of hydrophobic odorants through the lymph space to the dendrite membranes, or that they might protect odorants from enzymatic degradation (Pelosi et al., 2018a). SNMPs belong to the CD36 family of transmembrane proteins; their exact function in olfaction is unclear, but functional knock-down in D. melanogaster indicated that they are involved in pheromone detection (Benton et al., 2007; Pregitzer et al., 2014).

Chemosensory genes have been identified in a wide range of insects using genomic and transcriptomic approaches (Clyne et al., 1999; Gao and Chess, 1999; Vosshall et al., 1999) which to a large extent enabled us to understand their evolutionary and behavioral adaptations in different contexts of biology including eusociality (Robertson and Wanner, 2006; Zhou et al., 2012, 2015; Engsontia et al., 2014; Terrapon et al., 2014; De Fouchier et al., 2017; Pask et al., 2017; Auer et al., 2020; Obiero et al., 2021; Keesey et al., 2022). Termites, despite being eusocial insects with well-studied chemical ecology and pheromone biology, have not been examined in detail in this regard until recently, with chemosensory genes identified only in three species: Zootermopsis nevadensis (Archotermopsidae) with 85 ORs in the genome (Terrapon et al., 2014), Cryptotermes secundus (Kalotermitidae) with 42 ORs in the genome (Harrison et al., 2018) and Reticulitermes speratus (Rhinotermitidae) with 22 ORs in the whole-body transcriptome (Mitaka et al., 2016). The relatively modest repertoire of termite ORs contrasts with the situation in eusocial Hymenoptera and the assumptions made on the impact of social evolution on OR expansion. At the same time, it highlights the independent origin of eusociality in Hymenoptera and Isoptera, and may potentially reflect the chemical parsimony observed in termite semiochemicals. By contrast, the repertoire of termite IRs is substantially expanded, with 141 and 135 IRs reported, respectively, for Z. nevadensis (Terrapon et al., 2014) and C. secundus (Harrison et al., 2018), which may intuitively be interpreted as a correlate of the social lifestyle (Harrison et al., 2018).

Here, we present the results of antennal transcriptome analysis of termite workers for the identification of the main olfactory sensory genes of three termite species, Neotermes cubanus (Kalotermitidae), Prorhinotermes simplex (Rhinotermitidae), and Inquilinitermes inquilinus (Termitidae). The three species were selected to cover the phylogenetic diversity of Isoptera, from the relatively basal Kalotermitidae to the modern lineage of Termitidae, with Rhinotermitidae being situated on the mid-way between the two. At the same time, the life histories of the three species represent different levels of social complexities encountered in termites, from the socially primitive N. cubanus devoid of true worker caste through P. simplex situated at the boundary of the emergence of true foraging workers, to the socially advanced higher termite I. inquilinus having true worker caste (Rupf and Roisin, 2008; Roisin and Korb, 2011). Our findings should serve as a basis in elucidating the evolution of pheromone detection in termites.

Here, we present the analysis of the main chemosensory families (ORs, GRs, IRs, SNMPs, OBPs, and CSPs) in three species of termites from different phylogenetic lineages, encompassing basal and advanced clades. The number of ORs in the transcriptomes of N. cubanus (30), P. simplex (50), and I. inquilinus (28) was roughly similar to that reported from the genomes of Z. nevadensis (69) and C. secundus (42), and higher than R. speratus (22; Terrapon et al., 2014; Mitaka et al., 2016; Harrison et al., 2018). However, the highest number of ORs among Blattodea was found in B. germanica (134 ORs), which may be partially explained by its large genome size, chromosomal translocations, and a higher rate of gene family expansions (Harrison et al., 2018). Ants, belonging among eusocial Hymenoptera, also possess massive OR expansions, leading to ~350 ORs in H. saltator and Camponotus floridanus (Zhou et al., 2012). It has been hypothesized that these expansions are connected to their eusocial behavior (Zhou et al., 2012). However, high number of ORs has been reported also in non-eusocial Hymenoptera, such as Nasonia vitripennis (301, Robertson et al., 2010), suggesting that the expansion of ORs is an ancestral trait shared by Hymenoptera, which might potentially had facilitated the multiple independent evolutions of eusociality in hymenopteran insects. By contrast, the high OR repertoire reported recently in basal solitary apoid wasps phylogenetically positioned between ants and bees indicates that the OR repertoire in fact reduced during the evolution of eusocial apoids (Obiero et al., 2021).

Termites, despite being eusocial insects, exhibit numbers of OR genes comparable to non-eusocial insects (Mitaka and Akino, 2021). If an expansion of ORs preceded the emergence of eusociality in Hymenoptera, the same is clearly not true for Isoptera. The phylogenetic analysis revealed the highly conserved ORCo lineage and multiple Isoptera-specific OR expansions, which were analogous to the recent report in C. secundus based on the gene tree analysis (Harrison et al., 2018). Within these Isoptera-specific expansions, we found 1:1 orthologous relationship between the ORs of distinct termite species. This is rather unusual for the highly divergent OR family, indicating a high degree of OR conservation across termites.

The transcriptome screening performed in the three species of termites yielded 20, 25 and 26 GRs, respectively, from N. cubanus, P. simplex, and I. inquilinus. The phylogeny (Figure 2) also reveals the isopteran specific expansion of GRs in all three major sub clades: sugar, bitter and CO₂ receptors. The dendrogram clearly separates different GR sub-classes as taste and CO₂ and pheromone receptors. The basal clade includes D. melanogaster GR5a and Gr64a, which are tuned towards trehalose and sucrose, respectively, and who exhibit complementary functional profiles in D. melanogaster (Jiao et al., 2008). We have identified 6, 3 and 4 putative sugar receptors each, respectively, from N. cubanus, P. simplex, and I. inquilinus. The CO₂ receptor-containing clades was also largely expanded in termites. There were no orthologs found for Drosophila pheromone receptors GR32a and Gr68a. The fructose receptor (Gr43a) sub clade was located within the large bitter receptor clade, as previously reported in the B. germanica GR expansions (Robertson et al., 2018), indicating a conserved phylogenetic pattern across insect orders.

IRs, a subfamily of iGluRs, were found to be involved in detecting environmental as well as intracellular chemical signals (Benton et al., 2009; Croset et al., 2010; Ai et al., 2013). They were first identified in D. melanogaster and are well described in terms of the functional and evolutionary origins (Croset et al., 2010; Rytz et al., 2013). In contrast to other insect orders, IRs are numerous in Isoptera; in fact, the IR expansion in termites is considered to be analogous to OR expansions in Hymenoptera, signifying the importance of this protein family (Harrison et al., 2018; Robertson et al., 2018). A recent genome-based annotation in the cockroach B. germanica recovered 455 IRs, the highest number reported in insects. Nevertheless, nearly half of them were pseudogenes (Harrison et al., 2018; Robertson et al., 2018). While our findings of 98, 95 and 77 IRs from N. cubanus, P. simplex, and I. inquilinus, respectively, exceed the numbers identified in most other insect species, they fall in the range of the numbers reported from other termites (Z. nevadensis: 141; C. secundus: 135; Terrapon et al., 2014, Harrison et al., 2018). Although the olfactory perception of social signals in ants (Slone et al., 2017; Trible et al., 2017) and TFPs in termites (Gao et al., 2020) have been demonstrated as OR/ORCo dependent, it was also proposed that a parallel ionotropic receptor gene family expansion has favoured the evolution of colony communication in termites (Harrison et al., 2018). It seems likely that the total count of IR coding genes will be higher within the full genomes, but not substantially so. In contrast, only 12 IRs reported from R. speratus could be explained by the limited coverage of chemosensory genes in the whole-body transcriptome (Mitaka et al., 2016). It should be noted that genes with expression limited to one or a few tissues, like antennal IRs, will be underrepresented in a whole-body RNA pool, a reason we used antennal transcriptomes in our study. The expansion and positive selection in IRs have been reported recently in Z. nevadensis and B. germanica (Harrison et al., 2018). Rapid expansions in chemosensory receptor gene families provide functional divergence, crucial for adaption to different niches (Arguello et al., 2016). The caste and sex-biased expression of IRs reported in Z. nevadensis and C. secundus indicates the possible role of these genes in the pheromone communication (Harrison et al., 2018). The different subsets of iGluRs including IRs were added to the phylogenetic analysis. iGluRs exist across kingdoms, including plants, animals and prokaryotes (Croset et al., 2010; Rytz et al., 2013). We found orthologs of all major iGluR subfamilies in all three transcriptomes; our analysis revealed both antennal and divergent IRs. Similar to ORs, Isoptera-specific expansions were previously observed in termite IRs (Harrison et al., 2018). The antennal IRs are considered to be involved in olfaction, divergent IRs in gustation (Benton et al., 2009; Croset et al., 2010; Abuin et al., 2011; Prieto-Godino et al., 2017). As per the functional studies in Drosophila, the IR20a clade includes both taste and pheromone receptors (Koh et al., 2014). In our analysis this clade grouped with the divergent IRs. We identified five candidates in this clade, two each from N. cubanus, P. simplex, and one from I. inquilinus, and further research is required to confirm the role of these receptors.

SNMPs are broadly conserved CD36 (cluster of differentiation 36) family of transmembrane proteins in animals and are reported to be involved in the detection of lipid-derived pheromones in insects (Benton et al., 2007; Pregitzer et al., 2014). Among the two SNMP types reported in insects, SNMP1 was found to be expressed in both sensory neurons and supporting cells of insect pheromone-sensitive sensilla, whereas SNMP2 was found only in the sensory supporting cells as reported in the moths Heliothis virescens and Antheraea polyphemus (Forstner et al., 2008). The recent structural studies indicate that SNMP1 might function as a co-receptor or acts as a tunnel to pass the signal molecules to the pheromone receptor (Gomez-Diaz et al., 2016). The number of SNMP1 proteins identified from our antennal transcriptomes was similar to the number reported from C. secundus (5) and higher than the one reported from Z. nevadensis (Terrapon et al., 2014; Harrison et al., 2018). Like termite ORs and IRs, SNMP1 showed 1:1 orthologous pattern among the five termite species compared. The higher number of SNMP1 proteins in termites could be correlated with the pheromone diversity in termites (Mitaka and Akino, 2021). In SNMP2 proteins, we found a single orthologous transcript in all five termite species compared. SNMP2, proteins are mainly found in the sensory neuron supporting cells and are proposed to be involved in pheromone clearance processes (Forstner et al., 2008).

OBPs and CSPs expressed in antennae and pheromone glands, respectively, are involved in both the reception and broadcast of the chemical message (Pelosi et al., 2018a). OBPs are highly abundant in the insect sensillar lymph and thus found abundantly in antennal transcriptomes (Venthur and Zhou, 2018). In Isoptera, OBPs and CSPs have been found to be differentially expressed among castes (Mitaka et al., 2016). The number of OBPs identified, i.e., 37, 35, 28 from N. cubanus, P. simplex, and I. inquilinus, respectively, are higher than the OBPs reported from other termites (Z. nevadensis: 19; C. secundus: 19; R. speratus: 9; Mitaka et al., 2016). Since OBPs are highly divergent in amino acid composition, using a basal hexapod L. y-signata (Missbach et al., 2015) as an outgroup helped in understanding OBP evolutionary pattern (Pelosi et al., 2005). All four major OBP sub-groups (classic, Minus-C, Plus-C and ABP-II types) have been identified based on the structure-based annotations reported earlier (Venthur et al., 2014). Unlike in Lepidoptera and other insect orders, Isopteran OBPs are understudied. However, we found two transcripts each in our transcriptomes with similarity to the well-studied protein BmorPBP from the moth B. mori (Lautenschlager et al., 2007). The 1:1 orthologous pattern observed in the other termite chemosensory genes continued in the case of OBPs. The number of CSPs identified was also higher in our transcriptomes as these were not annotated from the genomes of the other two termite species Z. nevadensis and C. secundus (Harrison et al., 2018).

Our research provides candidate genes of the major insect chemosensory gene families from three termite species belonging to three families of Isoptera of different phylogenetic positions, life histories and social complexities. We found comparatively large repertoires of chemosensory genes in all studied gene families as in other analysed termite species. The evolutionary analysis of termite chemosensory proteins revealed Isoptera-specific expansions with 1:1 orthologous pattern, indicating the existence of conserved olfactory functions. Our findings on basal eusocial insects will further enhance our understanding of the molecular underpinnings of eusociality.

The datasets generated for this study can be found in the NCBI repository: https://www.ncbi.nlm.nih.gov/bioproject/PRJNA885453/; accession numbers SRR21753787, SRR21753788, SRR21753789.

Ethical review and approval was not required for the study of animals in accordance with the local legislation and institutional requirements.

OL and RH performed initial sequencing. Bioinformatic analysis was performed mainly by JJ, with assistance by SD, OL, MS, and EG-W. The study was conceived by RH and EG-W. Data interpretation and manuscript drafting was shared by all authors. All authors contributed to the article and approved the submitted version.

This study was supported by the grant No. 20-17194S from the Czech Science Foundation and by Institute of Organic Chemistry and Biochemistry, CAS (RVO: 61388963). EG-W was further supported by the EXTEMIT-K project financed by Operational Programme Research, Development and Education at the Czech University of Life Sciences, Prague, Czech Republic (CZ.02.1.01/0.0/0.0/15–003/0000433). RH was also supported by OP RDE as part of project grant no. CZ.02.1.01/0.0/0.0/16_019/0000803, “Advanced research supporting the forestry and wood-processing sector’s adaptation to global change and the 4th industrial revolution.”

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.