Defensive fungal symbiosis on insect hindlegs

Abstract

Tympanal organs as “insect ears” have evolved repeatedly. Dinidorid stinkbugs were reported to possess a conspicuous tympanal organ on female’s hindlegs. Here we report an unexpected discovery that the stinkbug’s “tympanal organ” is actually a novel symbiotic organ. The stinkbug’s “tympanum” is not membranous but a porous cuticle, where each pore connects to glandular secretory cells. In reproductive females, the hindleg organ is covered with fungal hyphae growing out of the pores. Upon oviposition, the females skillfully transfer the fungi from the organ to the eggs. The eggs are quickly covered with hyphae and physically protected against wasp parasitism. The fungi are mostly benign Cordycipitaceae entomopathogens and show considerable diversity among insect individuals and populations, indicating environmental acquisition of specific fungal associates. These results uncover a novel external fungal symbiosis in which host’s elaborate morphological, physiological and behavioral specializations underpin the selective recruitment of benign entomopathogens for a defensive purpose.

Tympanal organs are the external auditory structures that have repeatedly evolved in diverse insects. Crickets, moths and cicadas have tympanal ears on their forelegs, thorax and abdomen, respectively, by which they perceive courtship songs and recognize approaching enemies (1,2). Not only musical cicadas but also other hemipteran insects, including diverse stinkbugs, utilize either air-borne sounds or substrate-borne vibrations in their mating behaviors (3,4). The Dinidoridae is a small stinkbug family embracing some 100 species representing 16 genera (5–7). Morphological studies reported that dinidorid stinkbugs generally develop a conspicuous tympanal organ specifically on female’s hind tibiae, which was speculated to be used for courtship communications (8). Thus far, however, no detailed studies have been conducted on this minor group of stinkbugs. How do dinidorid females listen to male’s song or dance using their hindlegs? To address this question, we investigated the Japanese dinidorid stinkbug Megymenum gracilicorne, and discovered that, unexpectedly, the stinkbug’s “tympanal organ” is not an auditory organ but a previously unknown type of symbiotic organ.

Results and Discussion

Peculiar traits of fungus-growing “tympanal organ” on hindlegs.

M. gracilicorne lives on wild cucurbitaceous plants and sometimes infests cultivated cucumbers and pumpkins (9) (Fig. 1a–c). As previously reported for over 30 dinidorid species (8), M. gracilicorne exhibits a conspicuous sexual dimorphism in the hindlegs. While the hind tibiae of adult males are normal in shape (Fig. 1d–f), the hind tibiae of adult females are widened and concave in the middle, where a flat oval area of “tympanum” develops at the center (Fig. 1g, h). However, close morphological inspection uncovered that structural features of the female-specific hindleg organ are atypical of insect’s tympanal ears in that the “tympanum” was not membranous but a rigid porous cuticle (Fig. 1h, i). We found that, in reproductively mature females, the “tympanum” was covered with wooly white materials (Fig. 1b, c, j), which looked like fungal hyphae massively growing out of the pores (Fig. 1k, l). Cultivation and microscopic observation of the white materials confirmed that they are certainly filamentous fungi, stretching hyphae and forming conidia (Fig. S1).

Fig. 1. Morphology and behavior of M. gracilicorne.

(a) An adult female in the field on Sicyos angulatus. (b) Back view of an adult female reared on cucumber, whose hindleg organs are covered with fungal hyphae (arrowheads). (c) Magnified image of the hindleg organ. (d-l) Morphology of hind tibia of adult male (d-f), immature adult female (g-i) and mature adult female (j-l). (d, g, i) Bright field images. (e, h, k) SEM images. (f, i, l) Magnified SEM images. (m-o) Peculiar behavior of egg-laying adult female. After laying each egg (m), the female rhythmically scratches the hyphae-covered hindleg organ with tarsal claws of the opposite hindleg (n) and then rubs the egg surface with the claws in a skillful manner (o). Also see Movie S1. (p-r) Egg mass covered with fungal hyphae, 0-day (p), 3-day (q) and 10-day (r) after oviposition. Though densely covered with fungal hyphae, the eggs normally exhibit high hatch rates, typically over 80%.

Female’s behavior for fungus transfer from hindleg organ to eggs.

Strikingly, we observed a peculiar behavior of the reproductive adult females during oviposition. The gravid females laid eggs in a row, and when each egg was deposited, the females rhythmically scratched the hyphae-covered hindleg organ with tarsal claws of the opposite hindleg and then rubbed the egg surface with the claws in a skillful manner, thereby smearing the fungi onto the eggs (Fig. 1m–o; Movie S1). Within a few days, fungal hyphae quickly grew and covered the entire egg mass (Fig. 1p–r). Upon hatching, the hyphae attached to the body surface of newborn nymphs (Fig. S2; Movie S2), but the fungi were subsequently lost as the nymphs molted and grew, being not maintained on the insects.

Diversity and specificity of symbiotic fungi.

In order to identify the fungi growing on the hindleg organs and the eggs, we collected reproductive adult females from different localities, each female was reared in isolation and allowed to lay two egg masses consecutively, and the hind tibiae, the first egg mass and the second egg mass were sampled and subjected to fungal cultivation (Table S1; Fig. S3). In total, over 600 fungal isolates were obtained and subjected to sequencing of ribosomal internal transcribed spacer (ITS) region (Table S1). We also performed amplicon sequencing analysis of the fungal ITS region using the DNA samples extracted from the hind tibiae and the eggs (Table S2). Molecular phylogenetic analyses based on these sequence data consistently revealed that (i) the majority of the fungi were placed within the class Sordariomycetes of the Ascomycota, (ii) the majority were classified to the family Cordycipitaceae mainly consisting of entomopathogenic fungi such as Cordyceps, Beauveria and others (10,11), and (iii) the majority represented several specific fungal lineages such as Lecanicillium, Simplicillium and Akanthomyces (12–14) (Figs. S4 and S5; Note S1).

Fungal diversity and specificity in natural host populations.

Figure 2 shows the fungal compositions on the hindlegs, the first eggs and the second eggs of field-collected reproductive females of M. gracilicorne originated from different localities and years, based on the ITS amplicon sequencing data. The following patterns were observed: (i) each female generally hosted multiple fungal species, reflecting the external open configuration of the symbiotic system vulnerable to microbial contamination; (ii) the fungal compositions on the maternal hindlegs were generally similar to the fungal compositions on the eggs, confirming vertical transmission of the fungal associates by the maternal smearing behavior; (iii) within the same collection locality, the fungal compositions were often markedly different between the insects, highlighting intra-populational diversity of the fungal associates; (iv) the predominant fungal associates may be different among the collection localities, illustrating inter-populational diversity of the fungal associates; and (v) within the same collection locality, the predominant fungal associates may be different across different years. These patterns looked quite strange – despite the fidelity of mother-offspring vertical transmission, the fungal associates exhibit remarkable diversity and variability within and between the populations of M. gracilicorne.

Fig. 2. Fungal microbiota on hindleg organs and eggs of field-collected reproductive adult females of M. gracilicorne.

(a) Nine females collected at Chikusei in 2020. (b) Eight females collected at Chikusei in 2015. (c) Four females collected at Takasaki in 2020. (d) Four females collected at Chiba in 2020. Fungal ITS amplicon-seq data from the mother’s hindlegs (Mom), the first egg mass (Egg1) and the second egg mass (Egg2) are shown by colored bar graphs. For detail, also see Fig. S3.

Vertical transmission, loss, and environmental re-acquisition of symbiotic fungi.

These results indicate that M. gracilicorne transmits the hindleg-associated fungi to eggs vertically, but the fungi are lost during the nymphal development, and the fungi are newly acquired by adult females from the environment every generation. Such a symbiont transmission trajectory is, to our knowledge, exceptional among insect-microbe and other symbiotic systems (15–17). Why the symbiotic fungi passed to offspring by mother’s specialized morphology and skillful behavior have to be lost and regained is an enigma.

Structure, cytology, and gene expression of hindleg organ.

Another enigma is how the specific fungal strains, either Lecanicillium, Simplicillium or Akanthomyces, can be selectively picked up by the female’s hindleg organ from a myriad of microbes in the environment. To address this question, we investigated the female’s hindleg organ in detail. Histologically, the “tympanum” was a thick and sclerotized porous cuticle, and the levels of thickness and sclerotization were more conspicuous in mature females than in immature females (Fig. 3a–d). Beneath the “tympanum”, no neuronal elements like chordotonal organs, the stretch receptors typically associated with auditory and vibratory sense organs of insects (3,4), were seen, but the thick cuticle was lined with a well-developed epidermal cell layer (Fig. 3a–d). Notably, the epidermal cells intruded into the cuticular layer and connected to the bottom of the cuticular pores, constituting secretory epidermal columns (Fig. 3b, d). In mature females, the epidermal columns and the cuticular pores exhibited dense signals of polysaccharide staining, indicating high secretion activities of the epidermal cells and massive fungal proliferation in the cuticular pores (Fig. 3c, d), which were also verified by TEM observations (Fig. 3e–h). Comparative transcriptomic analyses of midlegs and hindlegs of immature and mature adult insects of M. gracilicorne, from which fungus-derived and other presumably contaminating genes were removed based on sequence similarity (Tables S3 and S4), revealed that the hindlegs of mature adult females exhibit a distinct gene expression pattern (Fig. S6). Differential gene expression analyses identified 550 genes highly and preferentially expressed in the hindlegs of mature adult females (Tables S5 and S6). Functional term enrichment analysis revealed that genes with transporter activities were augmented in the hindleg organ, which was consistent with its secretory role. The top highly expressed gene encoded a Takeout family protein, which is a small secretory protein likely functioning as a carrier of hydrophobic substances (18). The other highly expressed genes included UDP-glucuronosyltransferases and glucuronate pathway genes that presumably mobilize lipophilic materials, a secretory cuticular glycoprotein possibly involved in the formation of the cuticular structure, some immune-related regulators in the Toll signaling pathway, and others (Table S6; Note S2).

Fig. 3. Histology and fine structure of female-specific hindleg organ of M. gracilicorne.

(a-d) Tissue section images of the hindleg organs subjected to PAS staining, by which sugar-rich elements such as polysaccharides, glycoproteins, mucous secretion and fungal cell wall are visualized in red. (a, b) In immature female, the “tympanal” cuticle is thinner, the epidermal cells are stained in pale pink, and the cuticular pores are empty. (c, d) In mature female, the “tympanal” cuticle is thicker and sclerotized, the epidermal cells are deeply stained in red, and the cuticular pores are full of fungal hyphae. Note that no neuronal elements like chordotonal organs are found beneath the “tympanal” cuticle. Also note that the epidermal cells form secretory columns leading to the bottom of the cuticular pores. (e) A tissue section image of the “tympanal” cuticle of mature female for guiding the histological locations corresponding to TEM images (f-h) shown by red rectangles. (f) TEM image of the bottom of a secretory column. The cytoplasm of secretory epidermal cells full of granules, vacuoles and membranous elements is seen between thick cuticles. (g) TEM image of the top of a secretory column. The cytoplasm full of secretory granules and membranous elements ends up with the narrow cuticular canals at the bottom of the cuticular pore. Arrowheads highlight secretory granules. (h) TEM image of the content of a cuticular pore. Densely packed fungal hyphae are seen.

Hindleg organ as external symbiotic organ for hosting specific fungi.

These results unequivocally rejected the conventional interpretation that the stinkbug’s female-specific hindleg organs are tympanal ears, which was based merely on superficial morphological resemblance (8). The “tympanum” is neither a vibratory membrane nor associated with sensory neurons, but is a sclerotized cuticle with numerous pores whose bottom is connected to secretory cells. The morphological and cytological features of the pores are reminiscent of the glandular mycangia, the exoskeletal cavities lined with glandular cells for fostering specific symbiotic fungi as known from ambrosia beetles (19,20). Similar exoskeletal cavities with glandular elements for hosting symbiotic actinobacteria are reported from leaf-cutter ants and digger wasps (21,22). The peculiarity in M. gracilicorne is the large number and high density of the mycangial units concentrated in the specialized region on the hindlegs. While ambrosia beetles generally have only a few mycangia (23), each hindleg organ of M. gracilicorne bears about 2,000 fungus-growing pores (mean ± SD = 2,118 ± 100; n = 11), totaling over 4,000 glandular mycangia per adult female. We conclude that the hindleg organ is a previously unknown type of external symbiotic organ consisting of highly concentrated glandular mycangia for hosting specific fungal associates.

Egg surface structure, fungal growth and proteinaceous secretion.

Following the maternal fungus-smearing behavior upon oviposition (Fig. 1m–o; Movie S1), the egg surface is soon covered with fungal hyphae (Fig. 1p–q). What support the massive and quick fungal growth on the eggs? SEM and sectioning observations revealed that the egg surface is covered with a polysaccharide-rich layer and the fungal hyphae grow presumably by consuming the secretion layer (Fig. S7a–f). The secretion layer was already observed on the surface of mature eggs in the female’s ovary (Fig. S8a). The secretion layer was solubilized when disulfide bond was reduced in the presence of detergent (Fig. S8b). From newly laid eggs, the secretion layer was extracted (Fig. S7g–k) and subjected to proteomic analysis (Table S7). An outstandingly abundant secretion protein with an odorant-binding protein (OBP)-like motif was identified. The other detected proteins were β-glucuronidase, probable antibacterial peptide, soma ferritin, and others (Table S7; Note S2). A recent work reported that, for vertical transmission upon oviposition, stinkbugs of the family Plataspidae produce symbiont-containing “capsules”, in which a predominant secretion protein with an OBP domain, called PMDP, embeds and protects the fragile symbiont cells outside the host (24). However, the abundant OBP-like motif protein of M. gracilicorne was phylogenetically not related to PMDP. On account of the quantitative predominance, it is conceivable, although speculative, that the protein might contribute to the fungal proliferation on the egg surface at least to some extent.

Low to moderate pathogenicity of hindleg-associated fungi.

The Cordycipitaceae is famous as a fungal group containing many potent entomopathogens such as Cordyceps and Beauveria (10,11). Keeping such entomopathogens on the hindlegs and eggs seems dangerous for the host insect. However, we found that the major hindleg-associated fungal strains may be not so dangerous for M. gracilicorne. Topical application of conidia, by dipping the adult insects into conidia suspension of Simplicillium Sm1, Lecanicillium Lc2 or Akanthomyces Ak1, did not kill the insects (Fig. S9a). When fungal conidia were injected into hemocoel, a representative pathogen Beauveria promptly killed all the injected insects. By contrast, the hindleg-derived Simplicillium Sm1 and Lecanicillium Lc2 showed very low mortality of the injected insects whereas Akanthomyces Ak1 exhibited some pathogenicity though lower than Beauveria (Fig. S9b). From these results, we thought that M. gracilicorne might have developed resistance to these entomopathogens during the symbiotic association. However, this idea turned out to be inappropriate when the same experiments were conducted using the pentatomid stinkbug Plautia stali with no relationship to these fungi. Similarly, conidia injection experiments revealed that Beauveria killed all the injected insects, Simplicillium Sm1 and Lecanicillium Lc2 killed few insects, and Akanthomyces Ak1 killed some insects, whereas topical application of conidia did not kill the insects (Fig. S9c, d). These results suggest that adult females of M. gracilicorne preferentially pick up relatively benign entomopathogens, Simplicillium or Lecanicillium (12–14,25,26), from environmental sources and selectively maintain them on the hindleg organ (Note S3).

Defensive symbiosis against microbial pathogens?

In such insects as leaf cutter ants (27,28), digger wasps (29,30), bark beetles (31), darkling beetles (32,33), leaf rolling weevils (34), leaf beetles (35) and others, external symbiotic actinobacteria, proteobacteria or ascomycetous fungi were reported to produce potent antimicrobial substances, thereby chemically protecting their host insects against microbial parasites and pathogens (36–38). Hence, we first suspected that the symbiotic fungi of M. gracilicorne might also play similar biological roles. Since there was no available information about microbial parasites and pathogens of M. gracilicorne, we attempted co-culture of the hindleg-derived Simplicillium, Lecanicillium and Akanthomyces with several bacteria (Escherichia coli, Pantoea dispersa, Bacillus subtilis, etc.) on the same agar plates, but conspicuous suppressive effects were not observed. The RNA sequencing data of the hindlegs of mature adult females contained fungal gene expression data, but no conspicuous expression of antibiotic-related genes like polyketide synthases was detected (Table S8; Note S4). Of course, the possibility that the hindleg-associated fungi are effective against hitherto unidentified microbial enemies of M. gracilicorne cannot be excluded. On the other hand, we discovered the relevance of the symbiotic fungi to a non-microbial natural enemy of M. gracilicorne.

Defense against wasp parasitism by egg-covering fungal hyphae.

By inspecting field-collected eggs of M. gracilicorne, we noticed that the natural eggs frequently suffer wasp parasitism. While newly laid eggs were white in color (Fig. 4a), developing fertilized eggs became reddish in color within a week (Fig. 4b) and hatched within two weeks (Fig. S2). Meanwhile, some field-collected eggs exhibited atypical black coloration (Fig. 4c), from which tiny black wasps emerged. The undescribed wasp species was named Trissolcus brevinotaulus based on the specimens we obtained (39) (Fig. 4d). When we marked and monitored 50 egg masses consisting of 223 eggs in the field, more than half of the eggs were parasitized by T. brevinotaulus (Fig. 4e), uncovering high levels of wasp parasitism in nature. In an attempt to evaluate the effect of egg-covering fungi on wasp parasitism, we generated fungus-removed egg masses by brushing and fungus-covered egg masses without brushing, and presented them to reproductive female wasps. In the experimental arena, the female wasps approached to both the fungus-removed eggs and the fungus-covered eggs, but their behaviors were markedly different. On the fungus-removed eggs, the female wasps exhibited oviposition behavior immediately after careful antennal drumming on the egg surface (Movie S3). On the fully fungus-covered eggs, by contrast, the female wasps were unable to approach to the eggs for oviposition (Movie S4). On the sparsely fungus-covered eggs, notably, the female wasps pressed down the hyphae by active antennal drumming and finally, though not always, oviposited successfully (Movie S5). Note that antennae of T. brevinotaulus are conspicuously thicker in females than in males, which are plausibly not only for sensing host eggs but also for flattening the hyphal thicket (Fig. S10a–d). Concordantly, the fungus-removed eggs suffered significantly higher wasp parasitism than the fungus-covered eggs (Fig. 4f). In addition, we experimentally generated fungus-suppressed egg masses by ablation of maternal hindlegs (Fig. S10e–h), and confirmed that the fungus-suppressed eggs suffered more oviposition trials and higher parasitism success than the fungus-covered eggs (Fig. 4g, h). It was impressive that the wasps did not avoid but actively approached and contacted the fungal hyphae, which entailed frequent self-grooming behaviors (Movies S3–S5). Even when the wasps were experimentally forced to contact with plenty of conidia of the egg-covering fungi, the survival of the fungus-treated wasps was not different from the survival of the untreated control wasps (Fig. 4i). These observations favor the idea that the nature of fungal defense is physical rather than chemical or pathogenic.

Fig. 4. Suppression of wasp parasitism by egg-covering fungal hyphae.

(a) Newly laid eggs white in color. (b) Developing fertilized eggs reddish in color. (c) Wasp-parasitized eggs black in color. (d) T. brevinotaulus ovipositing into an egg of M. gracilicorne. (e) Wasp parasitism rate in natural eggs of M. gracilicorne. In total, newly laid 50 egg masses consisting of 223 eggs were marked in the field, and inspected one week later. (f) Parasitism success rates of T. brevinotaulus on fungus-covered eggs and fungus-removed eggs by brushing. The fungus removing procedure is shown in Movie S3. (g, h) Oviposition trials per egg contact (g) and parasitism success rates (h) of T. brevinotaulus on fungus-covered eggs and fungus-suppressed eggs produced by ablation of maternal hindlegs. The fungus suppression procedure is shown in Fig. S10e–h. (i) Survival of T. brevinotaulus experimentally forced to contact with conidia of the hindleg-derived fungi. Statistical significance levels are estimated by Chi square test for (f-h) and generalized linear model test with a Poisson distribution for (i).

Discovery of defensive fungal symbiosis mediated by specialized organ on insect hindleg.

On the basis of these results, we conclude that the fungi selectively cultured on the female’s hindleg organ are transferred to eggs and play a defensive role against wasp parasitism, uncovering a previous unknown type of external symbiotic organ consisting of highly concentrated glandular mycangia for fostering the specific fungal associates. At least against the egg parasitoid T. brevinotaulus, the nature of defense is structural in that the egg-covering fungal hyphae physically distract approaching wasps and prevent parasitism. Notably, it was reported that the common cutworm moth Spodoptera litura lay eggs densely covered with scale hairs, which prevent egg parasitism by the wasp Trichogramma chilonis (40). Such structural defense against natural enemies is also known from gall-forming insects, in which thick and voluminous gall tissues, often with conspicuous projections or hairs on the surface, prevent access and oviposition of parasitoid wasps (41). In some gall midges, their symbiotic fungi contribute to the hardness of gall wall as well as serve as the nutritional source for larvae (42,43). Previous studies on defensive microbial symbioses in diverse insects have identified antibiotics (28,30–33), polyketide toxins (44,45), phage-encoded toxins (46), ribosome inactivating proteins (47) and other bioactive substances as symbiont-derived chemical factors for defense (36–38). Although not detected in this study, the possibility of chemical defense by the hindleg fungi against hitherto unidentified natural enemies of M. gracilicorne should be pursued in future studies.

Evolutionary origin of defensive fungal symbiosis on specialized hindleg organ.

The evolutionary origin of the hindleg organ and the defensive fungal symbiosis in M. gracilicorne remains to be an enigma. Enlarged hindlegs are known from several groups of heteropteran bugs (ex. Coreidae and Alydidae), but they are conspicuous in males and used for courtship fighting and territorial behaviors (48,49). Female-specific exaggerated hindlegs seem unique to the family Dinidoridae, and the conspicuous morphological trait is conserved across two major dinidorid lineages Dinidorinae (~70 species) and Megymeninae (~24 species including M. gracilicorne) (5–8). Hence, the hindleg organ is likely to have evolved in the common ancestor of the Dinidoridae. We hypothesize, although speculative, that the female-specific hindleg organ evolved originally for smearing some secretion onto eggs for chemical defense or camouflage, but environmental entomopathogenic fungi colonized the organ and exploited the secretion, and the insects established specific association with benign fungal species and co-opted them for egg defense. To test this hypothesis, the structure of the hindleg organs, associated microorganisms, and maternal behaviors upon oviposition should be comparatively investigated among diverse dinidorid stinkbugs in the world.

Relevance to evolutionary theory of symbiont transmission mode.

In M. gracilicorne, the specific fungal associates are vertically transmitted to eggs by female’s specialized structures and behaviors, but the association is cancelled during the nymphal development and restarted by newly-emerged adult females. At a glance, it was enigmatic why elaborate vertical symbiont transmission is combined with environmental symbiont acquisition in the insect life cycle. Our discovery of the symbiont’s defensive role against the egg parasitoid resolved the enigma – the maternal fungal smearing onto eggs is actually not for transmission but for defense. In principle, vertical symbiont transmission through host generations favors host-symbiont cooperation and facilitates the evolution of mutualism, whereas horizontal symbiont acquisition generally facilitates symbiont’s virulence and counters the evolution of mutualism (50–52). Meanwhile, theoretical studies predicted that mutualism without vertical transmission may evolve under the following conditions: (i) vertical transmission of the symbiont incurs some cost for the host, (ii) exploitation by the symbiont negatively affect the host, (iii) the host controls the vertical transmission process, and (iv) the host is able to discriminate benevolent symbionts from parasitic ones (53–55). Notably, all these conditions seem to apply to the fungal symbiosis in M. gracilicorne.

Impact on transmission mode of coexisting symbiont.

Here we point out that the external fungal symbiont may have evolutionarily impacted on a coexisting microbial symbiont in M. gracilicorne. In general, stinkbugs of the superfamily Pentatomoidea are dependent on specific γ-proteobacterial symbionts localized to the posterior midgut, and the beneficial/essential symbionts are vertically transmitted from mother to offspring via egg surface smearing, symbiont-containing capsules, symbiont-containing gelatinous secretion, or other means (13–15) (Fig. S11). However, our recent study uncovered that the beneficial gut symbiont of M. gracilicorne is not transmitted vertically but acquired from the environment every generation (56), which is exceptional among the Pentatomoidea (Fig. S11). It seems plausible, although speculative, that the egg-covering defensive fungi disturbed vertical transmission of the gut symbiotic bacteria and have promoted the evolution of environmental symbiont acquisition in an ancestor of the Dinidoridae.

Conclusion and perspective.

Our revisiting of the female-specific “tympanal organ”, which had been known from dinidorid stinkbugs for decades, led to the unexpected discovery of novel external fungal symbiosis for physical defense against wasp parasitism. The defensive fungal symbiosis on insect hindlegs provides an impressive case as to how evolutionarily novel traits for microbial symbiosis emerge, develop, and constitute the elaborate syndrome that integrates molecular, cellular, morphological and behavioral specializations. The bizarre coexistence of vertical symbiont transmission, symbiont loss, and environmental symbiont re-acquisition in the insect life cycle can be accounted for by benefit of the fungal egg defense and potential cost of nymphal infection, which highlights theoretical and empirical relevance of symbiont transmission mode to parasitism-mutualism evolutionary continuum. Future studies should focus on the molecular mechanisms underpinning the development of the peculiar symbiotic organ and the specificity of the fungal cultivation on the symbiotic organ, for which transcriptomic, metabolomic and RNAi screening approaches on morphogenesis and functioning of the hindleg organ will provide important clues.

Materials and Methods

Detailed Materials and Methods are available in Supplementary Information.