Introduction

Chemical communication is one of the most widespread forms of communication occurring among bacteria, fungi, plants, and animals (Wyatt 2003). Social insects have evolved sophisticated societies, characterized by efficient communication systems based on chemical signals (Wilson 1971; Vander Meer et al. 1998). The capacity to synthesize chemical compounds is biosynthetically finite and costly, potentially leading to strong evolutionary pressure to use single products parsimoniously for multiple purposes (Blum 1996; Steiger et al. 2011). Because various substances are emitted for non-communicative purposes, these chemicals provide multiple starting points for the evolution of communication. For example, cuticular anti-desiccation compounds are also used for recognition in the social Hymenoptera (Howard 1993; Greene and Gordon 2003; Steinmetz et al. 2003; Howard and Blomquist 2005; Nehring et al. 2011). Most organisms emit a multitude of defensive chemicals to protect themselves against enemies, such as antimicrobial compounds against pathogens, and repellents and venoms against predators. The secondary use of defensive compounds, which are primarily emitted for non-communicative purposes, as communication signals for sex, aggregation, alarm, or trail following occurs in a variety of insects (Blum and Brand 1972; Blum 996; Nojima et al. 2005; de Brujin et al. 2006; Geiselhardt et al. 2009).

Recently, a queen-produced volatile pheromone consisting of n-butyl-n-butyrate (nBnB) and 2-methyl-1-butanol (2M1B) was identified in the subterranean termite Reticulitermes speratus Kolbe (Matsuura et al. 2010; Yamamoto et al. 2012). This queen pheromone plays a variety of roles in reproductive regulation by indicating queen fertility, including inhibition of neotenic queen differentiation (Matsuura et al. 2010) and regulation of colony-level egg production (Yamamoto and Matsuura 2011). Importantly, the two volatiles in the queen pheromone are also emitted by eggs (Matsuura et al. 2010). These compounds indicate egg presence and function as an orientation pheromone guiding workers to care for eggs. The dual production of an inhibitory pheromone by reproductive females and eggs may provide a mechanism ensuring honest signaling of reproductive status, with a close relationship between fertility and inhibitory power (Matsuura 2012).

In social insect colonies, individual members cooperate to ensure colony growth, survival and reproduction; only a few individuals, the queens and their mates, produce offspring, whereas most individuals perform tasks such as foraging, nest construction, and offspring care. The dependence of the colony on a small number of reproductive individuals means that the fitness of all members of the society is jeopardized when the queen succumbs to a pathogen infection. Therefore, we would expect that the queen in a social insect colony should be subject to special protection, similar to germ lines that are subject to immune privilege (Cremer et al. 2007).

Brood survivorship is another determining factor of colony productivity. Brood protection is especially important in Reticulitermes termites, since eggs cannot survive in their microorganism-rich habitat without being groomed by workers (Matsuura et al. 2000). Soon after being laid by queens, eggs are carried into nursery chambers and frequently groomed by workers, whereby they are coated with saliva rich in antibiotic substances that protect them from desiccation and pathogenic infection. Brown fungal balls, called termite balls, are often found in egg piles of various termite species (Matsuura et al. 2000; Matsuura 2006) (Fig. 1). Termite balls tended by Reticulitermes termites are the sclerotia of an athelioid fungus (Basidiomycota, Agaricomycotina) of the genus Fibularhizoctonia (Matsuura et al. 2000). To date, this egg-mimicking fungus has been identified from seven Reticulitermes species in Japan and the United States (Matsuura 2005; Yashiro and Matsuura 2007), as well as from Coptotermes formosanus Shiraki (Matsuura and Yashiro 2010). By mimicking termite eggs chemically (Matsuura et al. 2009) and morphologically (Matsuura 2006), the termite ball fungus inhabits a nearly competitor-free habitat inside termite nests. This fungus relies on termites for defense against desiccation and other microorganisms, and the frequent grooming by workers keeps the survival rate near 100 %. In practice, most termite balls are inhibited from germination, and the fungus rarely consumes the eggs (Matsuura 2006). The evidence obtained to date indicates that the interaction is parasitic, in that it is beneficial for the fungus but costly for the host termites, at least in the short term. In this sense, the termite ball fungus is a sort of fungal cuckoo, taking advantage of its host’s brood care through egg mimicry.

Fig. 1
figure 1

Termite balls in egg piles of the termite Reticulitermes speratus. a An egg pile in a nest, b Close view of the egg pile. Termite balls are light brown and spherical, whereas eggs are transparent and ovoid, c Mycelial growth of the termite ball fungi in the egg pile, where eggs are consumed by the fungus

Full size image

Considering the privileged status of queens and brood in defense against parasites, together with the fact that queens and eggs emit the same volatile compounds in R. speratus (Matsuura et al. 2010), we hypothesized that queen pheromone compounds were originally antimicrobial defense chemicals that were later exapted to function as queen fertility signals or neotenic queen differentiation inhibitors. In this study, we examined the antifungal activity of the R. speratus queen pheromone, which contains nBnB and 2M1B, as a first step in testing the antimicrobial-origin hypothesis of queen pheromone evolution. We investigated the compounds’ influence on the growth of the egg-mimicking parasitic fungus Fibularhizoctonia sp., the closely related sclerotium-forming fungus Athelia rolfsii (Curzi) Tu & Kimbrough, and the distantly related sclerotium-forming fungus Sclerotium tuliparum Klebahn. We also tested the entomopathogenic fungi Beauveria bassiana (Balsamo-Crivelli) Vuillemin, Metarhizium anisopliae (Metschnikoff) Sorokin, and Isaria farinosa (Holmskjold) Fries. Pathogenicity of these entomopathogenic fungi to R. speratus has been demonstrated by laboratory bioassays (Shimizu and Yamaji 2002). Additionally, we compared the inhibitory effect of the queen pheromone compounds on the germination of termite balls harvested from different strains.

Materials and methods

Bioassays for queen pheromone compound effects on termite ball germination

We collected termite nests from rotten pinewood, Pinus densiflora Siebold & Zuccarini, because both the host termite R. speratus and the termite ball fungi Fibularhizoctonia sp. are most commonly found there (Matsuura et al. 2000; Matsuura 2005). When termites were found, the nest wood was dismantled completely to locate the reproductive center cells, which harbored reproductive individuals, eggs, and larvae. Termite balls used in this experiment were obtained from the egg piles of three termite colonies (SE110810A, KA110727A, KA110627A) collected in Okayama, western Japan.

The termite balls collected from each termite colony were placed in a 2 mL test tube, and washed three times in sterilized distilled water (DW) using a vortex mixer. Fifty termite balls were randomly chosen from each colony and arranged on a moist unwoven cloth (25 × 25 mm) (REED Healthy-Cooking Paper, Lion Corp., Tokyo) placed on one side of a Petri dish (90 × 15 mm) as shown in Fig. 2a. We placed a filter paper square (10 × 10 mm) on a glass plate (20 × 25 mm) in the other side of the Petri dish. We applied 1 μL of a test chemical onto the filter paper as follows: (1) nBnB; (2) 2M1B; (3) 2:1 nBnB:2M1B mixture; (4) 1:2 nBnB:2M1B mixture; (5) 1:1 nBnB:2M1B mixture; and (6) DW as a control. The Petri dishes were wrapped with two layers of Parafilm and incubated at 27 °C for 1 week. Germination rates were determined by observing the termite balls under a stereomicroscope (SZX7; Olympus Corp., Tokyo) every 24 h. Germination rates were analyzed using the Kaplan–Meier method followed by a log-rank test with Bonferroni correction. We used JMP 8.0.2 (SAS Institute, Cary, NC, USA) for this analysis.

Fig. 2
figure 2

Experimental set-up for the bioassay. a Germination test. Fifty termite balls were arranged on moist unwoven cloth. We added 1 µL of a test chemical to the filter paper, b Mycelial growth test. A 5 mm diameter plug of growing mycelium was placed onto potato–dextrose agar (PDA). We added 5 µL of a test chemical to the filter paper

Full size image

Bioassays for queen pheromone compound effects on fungal growth

The termite ball strains (TMB strains I, II, III, and IV) used in this bioassay were isolated from the termite colonies HA100725A, TU100904A, SE110810A, and KA110727A, respectively. Colonies HA100725A, SE110810A, and KA110727A were collected in Okayama, and TU100904A was collected in Tsukuba, Japan. To isolate termite ball fungi, we extracted 5–12 termite balls from each nest. These were rinsed with sterilized DW, and then arranged on an agar plate containing 200 ppm tetracycline soon after field collection. After germination, each termite ball was inoculated on a potato–dextrose agar (PDA; BD Difco, Franklin Lakes, NJ, USA) plate and incubated at 27 °C for 3 weeks. One newly developed sclerotium was re-isolated from each plate and cultured on a new PDA plate. The strains of the sclerotium-forming fungi A. rolfsii (NBRC4476) and Sclerotium tuliparum (NBRC6168) and of the entomopathogenic fungi B. bassiana (NBRC5838), I. farinosa (NBRC8296), and M. anisopliae (NBRC31961), were provided by the Biological Resource Center (NBRC), National Institute of Technology and Evaluation, Tokyo, Japan. These fungal strains were cultured on PDA plates at 27 °C, as were the strains of termite ball fungi.

The assay was performed by placing a 5 mm diameter plug of growing mycelia from a 3-week-old culture onto a Petri dish with each test chemical (Fig. 2a). We placed a glass plate (20 × 25 mm) on PDA in the Petri dish (Fig. 2b). A filter paper square (10 × 10 mm) was placed at the center of the glass plate. We applied 5 μL of a test chemical onto the filter paper as follows: (1) nBnB; (2) 2M1B; (3) 2:1 nBnB:2M1B mixture; (4) 1:2 nBnB:2M1B mixture; (5) 1:1 nBnB:2M1B mixture; and (6) DW as a control. The Petri dishes were wrapped with two layers of Parafilm and incubated at 27 °C for 1 week. Five replicates were done for each treatment of each fungal strain. The radial growth of mycelia (colony diameter in mm) in all plates was measured every 24 h after inoculation. The means of two measurements of each growing colony were used for analyses. We conducted measurements for a maximum of 11 days, terminating observations when mycelial growth reached the edge of the Petri dish or glass plate. Data were analyzed by repeated measures ANOVA followed by Tukey’s HSD test. We used STATISTICA 06 J (Stat Soft, Tulsa, OK, USA) for this analysis.

Results

Inhibitory effects on termite ball germination

The germination rates of the termite balls were significantly different among colonies (p < 0.0001, Fisher’s exact probability test; Fig. 3). In the control treatment (DW), termite balls obtained from colonies SE110810A, KA110727A, and KA110627A showed germination rates of 0.88 (44/50), 0.86 (43/50), and 0.32 (16/50), respectively. Termite balls from the colony SE110810A were significantly inhibited by queen pheromone substances (df = 5, overall x 2 = 96.21, p < 0.0001, log-rank test) in the order 2M1B = 2TO1 = 1TO2 = 1TO1 > nBnB > DW (p < 0.05, log-rank test with Bonferroni correction; Fig. 3a). There was a significant difference among termite balls from KA110727A with respect to chemical treatment (CT) (df = 5, overall x 2 = 65.66, p < 0.0001, log-rank test), wherein the order of inhibition was nBnB = 2TO1 > 2M1B = 1TO1 = 1TO2 > DW (Fig. 3b). Termite balls from KA110627A had low germination rates in all treatments, and none of the queen pheromone substances showed significant inhibition (df = 5, overall x 2 = 4.56, p = 0.47, log-rank test; Fig. 3c).

Fig. 3
figure 3

Effects of queen pheromone compounds on the germination of termite balls. a Termite balls from colony SE110810A, b Termite balls from colony KA110727A, c Termite balls from colony KA110627A. Different letters indicate significant differences (p < 0.05, log-rank test). nBnB n-butyl-n-butyrate, 2M1B 2-methyl-1-butanol

Full size image

Inhibitory effects on mycelial growth

The volatile compounds of termite queen pheromone showed significant inhibitory effects on mycelial growth of the termite ball fungus (CT: df = 5, F = 189.94, p < 0.0001, repeated measures ANOVA; Fig. 4). The interaction effect between CT and the TMB strain was significant (Strain × CT: df = 15, F = 6.86, p < 0.0001, repeated measures ANOVA), where nBnB showed greater inhibition than 2M1B in strains I, II, and III, but not in strain IV. In TMB strain I, the order of inhibition was nBnB = 2TO1 = 1TO1 > 1TO2 > 2M1B > DW (p < 0.05, Tukey’s HSD test; Fig. 4a). In TMB strain II, the order was nBnB > 2TO1 = 1TO1 > 1TO2 > 2M1B > DW (Fig. 4b). In TMB strain III, the order was nBnB = 2TO1 = 1TO1 = 1TO2 > 2M1B > DW (Fig. 4c). However, in TMB strain IV, the order was nBnB = 2TO1 = 1TO1 = 1TO2 = 2M1B > DW (Fig. 4d).

Fig. 4
figure 4

Effects of queen pheromone compounds on mycelial growth of the termite ball fungus. The fungal strains I (a), II (b), III (c), and IV (d) were isolated from termite colonies HA100725A, TU100904A, SE110810A, and KA110727A, respectively. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test). Error bars denote SEM. nBnB n-butyl-n-butyrate, 2M1B 2-methyl-1-butanol

Full size image

Queen pheromone substances significantly inhibited the mycelial growth of the other sclerotium-forming fungi, S. tuliparum (CT: df = 5, F = 26.668, p < 0.0001; Fig. 5a) and A. rolfsii (CT: df = 5, F = 555.75, p < 0.0001, repeated measures ANOVA; Fig. 5b). The order of inhibition was nBnB = 2TO1 = 1TO1 = 1TO2 > 2M1B, DW, in both S. tuliparum and A. rolfsii (p < 0.05, Tukey’s HSD test).

Fig. 5
figure 5

Effects of queen pheromone compounds on the mycelial growth of sclerotiun-forming fungi Sclerotium tuliparum (a) and Athelia rolfsii (b). Different letters indicate significant differences (p < 0.05, Tukey’s HSD test). Error bars denote SEM. nBnB n-butyl-n-butyrate, 2M1B 2-methyl-1-butanol

Full size image

Queen pheromone significantly inhibited the entomopathogenic fungi M. anisopliae (CT: df = 5, F = 147.06, p < 0.0001; Fig. 6a) and B. bassiana (CT: df = 5, F = 51.42, p < 0.0001, repeated measures ANOVA; Fig. 6b). In M. anisopliae, the order of inhibition was nBnB > 2TO1 = 1TO1 > 1TO2 > 2M1B > DW (Fig. 6a). In B. bassiana, nBnB showed significant inhibition (p < 0.001), whereas 2M1B showed no significant difference from DW (Fig. 6b). In contrast, no CT significantly inhibited the entomopathogenic fungus I. farinose relative to DW (Fig. 6c).

Fig. 6
figure 6

Effects of queen pheromone compounds on the mycelial growth of entomopathogenic fungi Metarhizium anisopliae (a), Beauveria bassiana (b), and Isaria farinose (c). Different letters indicate significant differences (p < 0.05, Tukey’s HSD test). Error bars denote SEM. nBnB n-butyl-n-butyrate, 2M1B 2-methyl-1-butanol

Full size image

Discussion

Insects and fungi have a long history of association in common habitats, where they share similar environmental conditions (Vega and Blackwell 2005). The remarkable diversity and ecological success of the social insects has been attributed to their ability to cope with the infectious microbial community inhabiting their nests and feeding sites (Wilson 1971; Hölldobler and Wilson 1990; Traniello et al. 2002). In addition to behavioral adaptations to lower disease risk (Rosengaus et al. 1999; Matsuura et al. 2002; Yanagawa et al. 2012), social insects also rely on biochemical secretions (Brough 1983; Beattie et al. 1986; Rosengaus et al. 2000; Poulsen et al. 2002; Bulmer and Crozier 2004; Rosengaus et al. 2004; de Lima Mendonça et al. 2009). In this study, the volatile compounds of a termite queen pheromone nBnB and 2M1B showed antifungal activities against the entomopathogenic termite ball fungi M. anisopliae and B. bassiana. Considering the parsimonious evolution of pheromones (Blum 1996), it is reasonable that pheromone compounds should have antimicrobial activity in various organisms (Cole et al. 1975; Matsumoto et al. 1979; Kuwahara et al. 1989; Vander Meer and Morel 1995; Ruther et al. 2001; de Brujin et al. 2006; Matsuura et al. 2007). Our results imply that the compounds nBnB and 2M1B might have originally evolved as antimicrobial defense, and that their function as termite queen pheromones might be secondary.

Germination of termite balls in egg piles imposes enormous costs on the colony because the fungus consumes surrounding eggs. The queen pheromone compounds nBnB and 2M1B are known to also be emitted by eggs (Matsuura et al. 2010). This study demonstrates that the inhibitory effects of the compounds on termite ball germination differ among colonies. Interestingly, 2M1B showed stronger inhibition than nBnB on termite balls collected from colony SE110810A (Fig. 3a), whereas nBnB showed higher suppression than 2M1B on balls from KA110727A (Fig. 3b). On average, a 2:1 ratio of nBnB and 2M1B, which matches that of queen pheromone, yielded the most stable inhibition of termite ball germination. Mycelial growth also indicated that the inhibitory effects of nBnB and 2M1B were significantly different among strains. This might explain why the termite has evolved two compounds for queen pheromones and egg volatiles.

In nature, most termite balls are inhibited from germination, and the fungus rarely consumes the eggs (Matsuura 2006). Old termite balls become shrunken and deformed. Such old, deformed, termite balls are removed from the egg piles by workers and left in a corner of the nest as garbage. The dumped termite balls germinate and grow in the nest, and newly formed termite balls are carried into egg piles. Variation in suppression level among the strains of termite ball fungus suggests that the fungus may be able to develop resistance to antifungal compounds. A single termite colony can harbor multiple strains of termite balls in its egg piles (Yashiro et al. 2011). Therefore, there might be competition among termite ball strains within a colony. If a strain were resistant to the antifungal compounds, it would be able to germinate in the egg pile and consume surrounding eggs, thereby propagating more than other strains. However, such a resistant strain would prevent reproduction by the host termite and eventually exterminate the colony. Because this fungus relies on the host for defense against desiccation and other microorganisms, severely harming the host termite would be disadvantageous for the fungus itself. This type of multi-level selection might have determined the chemical interactions between termite ball fungi and their host termites.