Identification of a queen primer pheromone in higher termites

Abstract

It is long established that queens of social insects, including termites, maintain their reproductive dominance with queen primer pheromones (QPPs). Yet, the QPP chemistry has only been elucidated in a single species of lower termites. By contrast, the most diversified termite family Termitidae (higher termites), comprising over 70% of termite species, has so far resisted all attempts at QPP identification. Here, we show that the queen- and egg-specific sesquiterpene (3R,6E)-nerolidol acts as the QPP in the higher termite Embiratermes neotenicus. This species has a polygynous breeding system, in which the primary queen is replaced by multiple neotenic queens of parthenogenetic origin. We demonstrate that (3R,6E)-nerolidol suppresses the development of these parthenogenetic queens and thus mimics the presence of mature queen(s). It acts as an airborne signal and may be used to optimize the number of queens, thus being the key regulatory element in the special breeding system of E. neotenicus.

Introduction

How the division of labor between reproducing and helping castes of eusocial insects has evolved and how it is established and maintained in their colonies are the central questions of social insect biology. Current evidence suggests that the reproductive division of labor in socially advanced eusocial insects (most eusocial wasps, ants, termites, and eusocial bees) is controlled via queen-emitted chemical signals, generally referred to as queen pheromones^1,2,3. The definition of queen pheromones encompasses a broad array of effects of these signals on the receivers², allowing the queens to advertise their presence, to be physically identified by their nestmates, and ultimately to monopolize colony reproduction.

Queen pheromones can act as releaser pheromones by directly affecting the behavior of the nestmates towards the queen upon her physical recognition (tending, feeding, defensive and retinue behaviors). At the same time, they may alter the physiology and development of the receivers and make them forgo or postpone their own reproduction in favor of the reigning queen(s), acting thus as primer pheromones^1,2. Because queen primer pheromones (QPPs) inhibit the reproductive potential of other females in the colony, they have traditionally been considered tools of manipulation or control over the subordinate colony members. From the modern perspective, QPPs are viewed as honest signals of the queen's presence and quality, allowing the nestmates to choose between individual reproduction and indirect gains from helping^4,5,6. QPPs evolved independently in individual eusocial lineages to supersede the ancestral situation characterized by physical conflicts among nestmates over the queen status, encountered in some socially primitive taxa^3,6,7.

In spite of the vital importance of QPPs for social homeostasis and prevention of conflict over reproduction in social insect colonies, their chemistry has long remained unknown. For nearly half a century, the honeybee Queen Mandibular Pheromone, in part described in the 1960s, remained the only chemically characterized QPP⁸. Unambiguous chemical identifications in other social insects only took place in 2010 and later, when the role of QPPs was ascribed to cuticular hydrocarbons (CHCs) on the queen's body surface (or also on queen-laid eggs) in several species of ants and wasps and a bumblebee^{9,10,11,12,13}, suggesting that multiple independent origins of eusociality within Hymenoptera were accompanied by co-options of a conserved ancestral CHC signaling system for the new role in queen fertility signaling^2,14. By coincidence, in 2010 the first (and so far, the only) QPP was characterized in the lower termite Reticulitermes speratus¹⁵, indicating that termite QPPs may be chemically very different from CHCs (see below).

Most termite colonies are established by a pair of winged primary reproductives (primary king and queen), who may remain the only reproducing individuals in the colony throughout its life cycle. However, in many species an alternative reproductive phenotype exists, i.e. neotenic reproductives. These can develop from male and female immatures (larvae, workers, nymphs) as replacement kings or queens after colony orphaning, or as supplementary kings or queens in the presence of other reproductives¹⁶. A special context of neotenic formation has been described in three lower termites of the genus Reticulitermes (Rhinotermitidae)^17,18,19 and several higher termites (Termitidae) from subfamilies Termitinae^20,21 and Syntermitinae^22,23. In these species, the founding primary queen prematurely dies, but before doing so, she produces a generation of female nymphs via thelytokous parthenogenesis. These nymphs develop into neotenic queens (NQs), replace the foundress, and reproduce in a polygynous breeding system with the founding primary king. At the same time, they also give rise asexually to subsequent generations of NQs, which supplement the pool of queens. Hence, the systematic production of NQs is the keystone of this unique reproductive strategy dubbed asexual queen succession (AQS).

Potential access of multiple colony members to neotenic reproductive options requires an effective mechanism by which the reigning queen(s) and king(s) advertise their presence and control the formation of neotenics. It has long been known that queens and kings do so via chemical cues which comply with the definition of primer pheromones^24,25,26. Observations of king- and/or queen-specific profiles of CHCs in several termite species indicated that convergently with eusocial Hymenoptera, these non-volatile cues may be involved in fertility signaling^27,28. CHCs were recently shown to act as releaser pheromones responsible for king and queen physical recognition^29,30, and pieces of indirect evidence suggest that they may also act as QPPs, inhibiting neotenic formation³¹. Nevertheless, the only termite QPP was identified in 2010 in Reticulitermes speratus (Rhinotermitidae) as a volatile blend of n-butyl-n-butyrate and 2-methylbutan-1-ol¹⁵. The pheromone is secreted by NQs and inhibits the formation of further NQs from workers. The pheromone is also emitted by eggs, thus providing complementary information on the presence and quality of the queen(s). It also has a releaser function and serves as an attractant of the nestmates, regulates the oviposition rate in other queens³², and promotes the production of salivary lysozyme used by workers for egg protection³³. Last but not least, it has antifungal properties against a specialized parasitic fungus³⁴. This multifunctionality is a neat example of chemical parsimony, suggesting that antifungal defense was the primary role of the volatile blend, which only secondarily became the "central signal" ensuring colony homeostasis³⁵.

Discovery of the QPP in R. speratus prompts questions about termite queen pheromones and queen pheromone communication of social insects in general, especially with respect to two conspicuous aspects: (i) the first identified termite QPP uses very different chemistry from that of QPPs in the honeybee and other Hymenoptera, and (ii) the first identified termite QPP is composed of highly volatile components, unlike the contact CHC cues in most eusocial Hymenoptera, seemingly to facilitate the spread of the signal across the large termite nests. Both these characteristics call for a comparison of different termite species to detect commonalities and idiosyncrasies in the QPP communication system of different social insect taxa. This perspective is particularly appealing for higher termites, the most diversified termite lineage (comprising 70% of living termite species³⁶), whose colonies may contain up to millions of inhabitants. Himuro et al.³⁷ reported phenylethanol as a queen-specific volatile of Nasutitermes takasagoensis (Nasutitermitinae), and we later identified the macrolide (5Z,9S)-tetradec-5-en-9-olide in queens and eggs of Silvestritermes minutus (Syntermitinae)³⁸ and the sesquiterpene alcohol (3R,6E)-nerolidol in queens and eggs of four other Syntermitinae³⁹. Yet, the biological role of these queen pheromone candidates could not be rigorously studied due to difficulties in maintaining colonies of higher termites in captivity and in experimental groups.

In the present study, we search for the QPP in the higher termite Embiratermes neotenicus (Syntermitinae). This species is among the rare termites having the AQS breeding system; the founding primary queen is replaced early in the colony cycle by tens to hundreds of NQs, which in turn produce further generations of parthenogenic queens²³. The polygynous AQS breeding system calls for the presence of a control mechanism that would ensure the correct timing of development of NQs from parthenogenetic nymphs upon the death (or senescence) of the foundress and later regulate the numbers of supplementary NQs according to the colony population size and the fitness of the reigning queen(s). Our observations on E. neotenicus and the related AQS species Silvestritermes minutus revealed that throughout the year, a pool of units to tens (S. minutus) or tens to hundreds (E. neotenicus) of parthenogenetic female nymphs of the fourth developmental stage (NY4♀) is present in most colonies. When isolated from queens, many of these NY4♀ rapidly, within a few days, molt into NQs^22,23. Therefore, NY4♀ can be considered arrested in their developmental trajectory, awaiting the appropriate moment to develop into NQs, and may thus be a potential target of the QPP control. We previously reported that primary queens and NQs of E. neotenicus emit significant quantities of enantiomerically pure (3R,6E)-nerolidol (hereafter referred to as RNERO) to the headspace³⁹ and that the compound is also present in washes of eggs while being absent in all other castes and stages in the colony, including NY4♀ and young subfertile NQs. Because of its queen- and egg-specific distribution, RNERO is a suitable candidate for a QPP function. Here, we test its potential role as QPP in a series of short-term experiments performed with freshly collected E. neotenicus colonies.

A remarkable characteristic of the AQS breeding system is that the NQs predominantly develop from NY4♀ of parthenogenetic origin while the sexually produced female nymphs occurring in colonies prior to dispersal flights mostly develop into winged dispersers. Even though the proximate mechanisms of these developmental priorities remain unknown and are broadly discussed^40,41,42,43, the differential developmental readiness of female nymphs according to their genetic origin is vital for the function of AQS systems. Here, we address this question by including in our experiments both colonies having relatively small numbers of mostly female NY4 presumed to be parthenogenetic as well as colonies nearing dispersal and containing hundreds to thousands of male and female nymphs of sexual origin. We test the readiness of female nymphs from both colony types to become NQs in the presence or absence of NQs and RNERO. In queen inhibition bioassays, we assess the genetic origin of nymphs and newly developed NQs by genotyping them at previously developed microsatellite markers.

We show that RNERO inhibits the formation of new NQs from NY4♀ of parthenogenetic origin and thus mimics the effect of live NQs in experimental groups. By contrast, in queenless conditions or in absence of RNERO, the parthenogenetic NY4♀ respond by rapid differentiation into new NQs unlike the sexually produced NY4♀ from colonies nearing dispersal. We prove that RNERO acts as an airborne signal, and shows selective olfactory preference for RNERO by NY4♀, compared to its opposite (S) enantiomer and related sesquiterpenoids, putting forward the role of RNERO as a signal mediated via olfaction. These results confirm RNERO as the second QPP identified in termites and the first to be described in higher termites. At the same time, they reveal that the two termite QPPs identified so far present fundamentally different chemistries, suggesting that the array of termite QPPs can be unexpectedly diverse.

Results

Readiness of nymphs to develop into neotenic queens

In the initial experiment I, we tested the readiness of NY4♀ to develop into NQs and possible differences between the readiness of NY4♀ from non-dispersing colonies (expected to be predominantly parthenogens) and NY4♀ from large dispersing colonies (dominantly nymphs of sexual origin). To do so, we isolated groups of 17 NY4♀ from six different colonies (A–F), three of them representing typical non-dispersing colonies (low hundreds of NY4♀, male nymphs being rare), while the other three were nearing dispersal and contained high hundreds to thousands of NY4 with a substantial proportion of males (41–42%) (Fig. 1). NY4♀ from each of the three non-dispersing colonies started to differentiate into NQs soon after their isolation and reached the final counts of four to five NQs per one group within 72 h, i.e., 27% of all nymphs from non-dispersing colonies. By contrast, we did not observe any of the 51 nymphs from the dispersing colonies transform into NQ. The survival curves constructed from total cumulative counts of new NQs in dispersing vs. non-dispersing colonies were significantly different (Mantel–Cox test, χ² = 16.13, df = 1, p < 10⁻⁴). The transformation of NY4♀ into NQ and subsequent NQ maturation are documented in Fig. 2.

**Fig. 1: Readiness of female nymphs to develop into neotenic queens.**

**Fig. 2: Development and maturation of neotenic queens in *E. neotenicus*.**

Presence of RNERO in queens and eggs

Experiments II and III, testing the effect of NQs and RNERO on the development of new NQs, were performed on three colonies (G–I) having relatively small numbers of NY4 (<300), since NY4♀ in such colonies responded to orphaning by differentiation into NQs unlike the nymphs from large dispersing colonies. During the establishment of queen inhibition bioassays, we sampled mature NQs, NY4♀, workers, soldiers, eggs, and primary king (when found) for chemical analyses to verify the production of RNERO by NQs and eggs. The sample set was later supplemented by young NQs which were obtained from the bioassays and subsequently kept in a colony fragment for 9 h, 10 or 14 days.

Two-dimensional gas chromatography–mass spectrometry (GC×GC-MS) analyses confirmed that the extracts of mature NQs and eggs from the three colonies contained RNERO, as evidenced by characteristic retention parameters and MS fragmentation pattern (Fig. 3). By contrast, the peak of RNERO was absent in extracts of workers, soldiers, primary king (colony I) and NY4♀. We did not detect RNERO in freshly differentiated, non-physogastric and non-pigmented NQs 9 h after the molt from NY4♀ (colony I), while pigmented but non-physogastric NQs 10 and 14 days after the molt already produced RNERO in significant quantities (colonies G and H). Full chromatograms and comparisons of MS spectra are given in Supplementary Figs. 1–4.

**Fig. 3: Mature NQs and eggs from the three colonies used in queen inhibition bioassays produce RNERO.**

NQs and RNERO inhibit differentiation of new NQs

In experiment II, we exposed groups of 9 NY4♀ from colony G, accompanied by workers and soldiers, to four treatments: solvent control, 10 live mature NQs (10NQ), 1 and 10 queen headspace equivalents of RNERO (1QE and 10QE) for 72 h, with five replicate groups per treatment and a design allowing direct contact of the termites with NQs and the treated filter paper (Fig. 4a).

**Fig. 4: Presence of mature queens or RNERO inhibits the development of new queens from female nymphs.**

During those 72 h, 13 out of the 45 NY4♀ in the solvent control groups differentiated into NQs (28.8%), 6 new queens were observed in 1QE treatment (13.3%) and 2 in the 10NQ and 10QE treatments (4.4%). Survival curves generated from cumulative counts of new NQs in individual treatments (Fig. 4b) were significantly different (Mantel–Cox test, χ² = 17.24, df = 3, p = 6 × 10⁻⁴), especially due to differences between the control and 10NQ and 10QE treatments (p = 1.5 × 10⁻³ and 1.7 × 10⁻³, respectively, corrected α_(0.05) = 8.5 × 10⁻³).

Likewise, significant differences were observed when the variability among the five replicates was considered (K–W rank-sum test, H = 10.24, groups = 4, n = 20, p = 1.67 × 10⁻²), resulting especially from significant differences between the control (median = 2) and 10NQ and 10QE treatments (median = 0, Dunn's multiple comparison test, adjusted p = 0.033 for both) (Fig. 4c). In brief, both stronger treatments (10NQ, 10QE) significantly reduced the numbers of new NQs, while the weaker treatment (1QE) revealed an intermediate effect, which did not differ significantly from other treatments and control.

RNERO inhibits differentiation of new NQs as an airborne factor

Based on the above observations we designed Experiment III with colonies H and I. The general design was identical to Experiment II (Fig. 5a). Termite groups were exposed for 60 h to three treatments, i.e., solvent control, 10 live mature NQs (10NQ) and 10 queen headspace equivalents of RNERO loaded on a piece of filter paper surrounded with a fine double mesh, preventing, in this case, the direct contact of the termites with the paper (10QE#).

**Fig. 5: RNERO acts as an airborne signal and inhibits the development of new queens from female nymphs without physical contact with its source.**

While in solvent controls 8 (17.7%) and 9 (20%) new NQs developed during 60 h in colony H and I respectively, only a single NQ was retrieved in the 10NQ and 10QE# treatments (2.2%) in each colony (Fig. 5b, e). Survival curves of cumulative counts of new NQs were significantly different in both colonies (colony H: χ² = 10.96, df = 2, p = 4.2 × 10⁻³, Fig. 5b; colony I: χ² = 13.16, df = 2, p = 1.4 × 10⁻³, Fig. 5e) due to the significant differences between the control and both treatments (colony H: p = 0.0142 for control vs. 10NQ and 0.0133 for control vs. 10QE#; colony I: p = 7.1 × 10⁻³ for control vs. 10NQ and 6.9 × 10⁻³ for control vs. 10QE#; corr. α = 0.0169). When the variability among replicates was compared (Fig. 5c, f), significant overall statistics were also obtained (colony H: H = 9.618, groups = 3, n = 15, p = 6 × 10⁻³; colony I: H = 9.556, groups = 3, n = 15, p = 6 × 10⁻³), driven by the differences between the controls and the 10NQ and 10QE# treatments (adjusted p = 0.022 for both comparisons in both colonies).

Genetic origin of NY4♀ and NQs in colonies G–I

To establish the genetic background of female nymphs in the colonies G–I used for queen inhibition bioassays, and to address the potential differential readiness of parthenogenetic vs. sexually produced NY4♀ to develop into NQs, we genotyped subsets of workers (12–14 per colony), soldiers (9–12), NY4♀ (38–65), and NQs newly molted in the experiments (7–23). We inferred the parental genotypes for each colony and deduced the genetic origin of NY4♀ and NQs from the presence/absence of exclusive parental allele(s) in their genotypes. Genotypes recorded at the nine studied microsatellite markers in the 259 individuals are listed in Supplementary Table 1.

In agreement with our expectations based on the small number of NY4 (269) and low proportion of NY4 males (1.5%), colony G revealed to be a non-dispersing colony with only one out of 61 genotyped females (38 NY4♀ + 23 NQs) being sexually produced (Fig. 4d). Likewise, all but one among 54 genotyped females (44 NY4♀ + 10 NQs) were of parthenogenetic origin in colony H, having 249 NY4 and 1% of males (Fig. 5d). Out of the 72 successfully genotyped females (65 NY4♀ + 7 NQs) from colony I, 37 were parthenogenetic and the remaining 35 sexually produced, suggesting a roughly balanced initial ratio of sexually vs. parthenogenetically produced NY4♀ randomly selected for the inhibition bioassay (Fig. 5g). This indicates, together with the presence of 32% of males among NY4, that in spite of the relatively low total number of NY4 (261), typical of non-dispersing nests, colony I was in fact nearing a small dispersal event and the pool of parthenogens was in part diluted by sexually produced NY4♀. By contrast, among the 7 successfully genotyped NQs only one was of sexual origin.

In total, among all NY4♀ randomly selected from their natal colonies, randomly attributed to one of the treatments in the queen inhibition bioassay and randomly selected for genotyping, only two sexually produced individuals developed into NQs out of 37 (5.4%). This contrasts with 60 new NQs that differentiated from 150 parthenogenetic NY4♀ (25%) (Figs. 4d, 5d, g, Supplementary Table 1).

Olfactory preference for RNERO by NY4♀

The results of queen inhibition bioassays suggested that RNERO acts as an airborne signal detected by olfaction. Therefore, we decided to test electroantennographic responses of NY4♀ to RNERO and three structurally related sesquiterpenoids, i.e. (E,E)-α-farnesene, (E,E)-farnesol and (3S,6E)-nerolidol in order to address the potential olfactory preference for RNERO over related structures. Electroantennographic measurements revealed significant differences in antennal responses of NY4♀ to the four sesquiterpenoids and the solvent control at both tested doses (50 ng: ANOVA, F_4,70 = 43.46, p < 10⁻⁴; 500 ng: F_4,70 = 62.18, p < 10⁻⁴) (Fig. 6a–c). At both doses, similar ranking of the mean responses was observed: control < (E,E)-α-farnesene < (E,E)-farnesol < (3S,6E)-nerolidol < RNERO, with most pairwise comparisons being significant at α = 0.05 [except for farnesol vs. (3S,6E)-nerolidol at both doses and (E,E)-α-farnesene vs. farnesol at 50 ng]. Detailed test statistics of all pairwise comparisons are provided in the Supplementary Data 1 file. The difference between the mean response to RNERO and (3S,6E)-nerolidol was strikingly larger than among other pairs of treatments, suggesting a strong stereoselective olfactory preference for RNERO over all other compounds including its opposite (S) enantiomer.

**Fig. 6: Olfactory responses of female nymphs show a significant preference for RNERO over its opposite enantiomer and related sesquiterpenoids.**

Discussion

In this study, we show that RNERO suppresses the differentiation of new NQs from NY4♀ of parthenogenetic origin and thus mimics the presence of mature NQs in experimental groups. By contrast, when liberated from exposure to RNERO or mature NQs, these parthenogenetic NY4♀ rapidly respond by differentiation into new NQs, with up to almost 30% of NY4♀ molting into NQs within 3 days. RNERO also effectively suppresses the development of new NQs in experiments preventing direct contact of the termites with its source, indicating its function as an airborne signal acting via olfaction. In line with these observations, electrophysiological experiments revealed a strong antennal response of NY4♀ to RNERO and significant enantioselective preference for RNERO over its opposite (S) enantiomer and other structurally related acyclic sesquiterpenoids. Our results also showed that sexually produced NY4♀, present in colonies nearing dispersal, only rarely respond to orphaning or RNERO absence by differentiation into NQs.

Our observations lead us to conclude that RNERO acts as the queen pheromone by which the mature queen(s) inhibits the formation of new neotenic queens from parthenogenetic female nymphs. We showed here and previously³⁹ that both the mature founding primary queen and mature NQs emit RNERO on the body surface and to the headspace. By doing so, the primary queen may delay the onset of the polygynous breeding system from à priori-produced parthenogens as long as she remains fertile and fit. Likewise, after the replacement of the foundress by multiple parthenogens, the queen pheromone secreted by NQs may delay the differentiation of new supplementary NQs. The same function can be ascribed to RNERO emitted by queen-laid eggs, which represent a direct indicator of queen fecundity. RNERO is the second functionally confirmed termite QPP, and the first to be characterized in higher termites. It may serve for the correct timing of primary queen replacement and optimization of the number of NQs according to the colony size and fitness of the reigning queen(s). Pheromone control over the development of new parthenogenetic queens is a key regulatory element in the AQS breeding system and a negative feedback complement of the genetic control via the production of sexual vs. parthenogenetic female nymphs (Fig. 7).

**Fig. 7: Proposed scheme of genetic and pheromonal regulation of the AQS breeding system in *E. neotenicus*.**

In E. neotenicus, we did not previously detect any sexually produced NQs across tens of colonies^23,44. Indeed, our present experiments show that also under experimental conditions the NY4♀ of sexual origin only rarely differentiate into NQs, irrespective of the presence or absence of the queen(s) or QPP in the experimental groups. Therefore, whatever the mechanism determining the developmental priorities of female nymphs, our experiments underline the adaptive significance of the combined (epi)genetic and pheromonal control over queen development in AQS breeding systems: while the sexually produced female nymphs are almost unreservedly channeled into winged dispersers to ensure the genetic diversity of future dispersing queens and prevent inbred mating with their father, the developmentally arrested parthenogenetic NY♀ respond to the absence of QPP to allow an optimized number of non-dispersing parthenogens of the foundress for outbred mating with the primary king (Fig. 7).

The search for volatile queen pheromones in higher termites was motivated by the pioneering description of the first queen pheromone in the lower termite R. speratus¹⁵. Yet, while the chemical identifications of pheromone candidates were rather straightforward^37,38,39, functional studies were hampered by the difficulties in performing bioassays with higher termites and keeping their colonies in captivity^38,39. Our previous attempts to study the role of RNERO with E. neotenicus laboratory colonies were unsuccessful. Therefore, we developed a simple bioassay adjusted for soil-feeding termites, which we applied to freshly collected field colonies. Female nymphs from non-dispersing colonies proved to be very responsive to orphaning, which allowed successful testing of RNERO effects. This experimental design should facilitate functional tests of the queen- and egg-specific volatiles identified in other South American Syntermitinae, i.e., (6E)-nerolidol in Silvestritermes heyeri, Cyrilliotermes angulariceps and Labiotermes labralis, and (5Z,9S)-tetradec-5-en-9-olide in S. minutus, the former three having colonies almost exclusively headed by primaries, and the last being an emblematic case of AQS²². The potential function of these volatiles is even more interesting when considering the different chemical nature and presumed biosynthesis of the sesquiterpene nerolidol and the macrolide tetradecenolide, likely arising through cyclization of hydroxy fatty acid. Yet a different biosynthetic origin can be expected for n-butyl-n-butyrate and 2-methylbutan-1-ol in R. speratus, suggesting altogether multiple independent co-options of fundamentally different secondary metabolites for queen pheromone function in different termite taxa. This chemical diversity of termite QPPs (or presumed QPPs) is unexpected given their vital importance for the social homeostasis in termite colonies and contrasts with the conservative situation in eusocial Hymenoptera (except for the honeybee), in which structurally related CHCs have been recruited for the QPP function across the multiple independently arising eusocial lineages.

In contrast to R. speratus, in which the volatile QPP also has a releaser function by attracting workers¹⁵, we did not observe any influence of RNERO on worker or soldier behavior, even though the compound has electroantennographic activity also in workers³⁹. Hence, it is likely that other compounds may be responsible for physical queen recognition. Recent observations in the genus Reticulitermes have shown that recognition of reproductives, accompanied by appropriate behavior, is mediated by king/queen-specific CHCs^29,30,45. Therefore, in line with some predictions (e.g. ref. ⁴⁶) and reports of differential CHC profiles in reproductive and sterile colony members^27,28, termite fertility signaling may consist of multiple components, such as volatile queen (king) primer pheromones suppressing the reproductive potential of nestmates and contact releaser pheromone cues contained in the CHC profiles.

The apparent correlation between RNERO production and queen fertility, as well as RNERO presence on eggs, complies with the definition of an honest signal, truly reflecting queen fertility^4,6,14,47. It would be of interest to study whether RNERO production is subject to the same endocrine regulation as fertility itself, i.e., intuitively the juvenile hormone pathway, as evidenced for sex/aggregation volatile pheromones in solitary insects (e.g. refs. ^48,49,50), but also for fertility-linked CHC profiles of termites⁵¹ and eusocial Hymenoptera, including those confirmed as queen pheromones^14,52. Enantioselective biosynthesis of RNERO, presumably from farnesyl diphosphate precursor, would be another appealing topic of investigation, given the recent advances in functional characterizations of terpene synthases in insects, which revealed to be unrelated to the terpene synthases known in plants and appear to have evolved independently in individual insect lineages from isoprenyl diphosphate synthases^53,54,55,56. Along with the biosynthesis, the production organ of the queen pheromone should also be located, because it has not been identified in either E. neotenicus or R. speratus^15,39. Last but not least, the role of RNERO should be addressed from the perspective of its possible ancestral function, preceding its co-option for the queen pheromone signal. Such an evolutionary pattern has been proposed for the queen pheromone of R. speratus and may also be realistic for RNERO, given the wide range of reported biological functions of different nerolidol isomers, including antimicrobial properties⁵⁷.

Methods

Origin of termites, sampling

Ten colonies (A–J) of Embiratermes neotenicus (Holmgren 1906) were used in this study. The colonies were collected in French Guiana in February 2019, March 2020, and February 2022 at three forest localities along the Road to Petit Saut (N5°04.250′ W52°58.770′–N5°04.650′ W53°01.360′) and one locality 13 km South-West from Kourou (N5°06.440′ W52°45.521′).

E. neotenicus colonies live in soil-made nests situated at the ground level of the forest, whose volume may reach up to multiple dozen liters. We collected large nests, having more than 50 l in volume. In such colonies, the founding queen has already been replaced by multiple ...

Raccoon Diseases