- Email: [email protected]
Cuckoo Fungus Mimics Termite Eggs by Producing the Cellulose-Digesting Enzyme β-Glucosidase
Current Biology 19, 30–36, January 13, 2009 ª2009 Elsevier Ltd All rights reserved
DOI 10.1016/j.cub.2008.11.030
Report Cuckoo Fungus Mimics Termite Eggs by Producing the Cellulose-Digesting Enzyme b-Glucosidase Kenji Matsuura,1,2,* Toshihisa Yashiro,1 Ken Shimizu,1 Shingo Tatsumi,1 and Takashi Tamura3 1Laboratory of Insect Ecology Graduate School of Environmental Science Okayama University Okayama 700-8530 Japan 2Department of Organismic and Evolutionary Biology Harvard University Cambridge, MA 02138 USA 3Graduate School of Natural Science and Technology Okayama University Okayama 700-8530 Japan
Summary Insects and fungi share a long history of association in various habitats, including the wood-decomposition niche [1]. Fungal mimicry of termite eggs is one of the most striking evolutionary consequences of insect–fungus association [2, 3]. Termites of the genus Reticulitermes often harbor fungal sclerotia, called ‘‘termite balls,’’ along with eggs in nursery chambers, whereby the fungus gains a competitor-free habitat in termite nests. Sophisticated morphological and chemical camouflage are needed for the fungus to mimic termite eggs [3]. However, the mechanism of chemical egg mimicry by the fungus is unknown. Here, we show that the fungus mimics termite eggs chemically by producing the cellulose-digesting enzyme b-glucosidase. We found that the termite egg-recognition pheromone consists of b-glucosidase and lysozyme. Both enzymes are major salivary compounds in termites and are also produced in termite eggs. Termite balls were tended by termites only when the fungus produced b-glucosidase. Our results demonstrated that the overlap of the cellulose digestion niche between termites and the fungus sharing the same chemicals provided the opportunity for the origin of termite egg mimicry by the fungus. This suggests that pheromone compounds might have originally evolved within other life history contexts, only later gaining function in chemical communication. Results and Discussion Cellulose in the form of dead plants is a superabundant potential resource, although digesting it is not an easy task for most animals. Termites have developed cellulose digestion capabilities with both symbiotic [4] and endogenous [5] cellulases that allow them to use dead plants as food resources. This wood-decomposition niche was occupied by various microorganisms, including bacteria, fungi, and protozoa, before the ancestors of termites began exploiting it. In this sense, termites live in the world of microorganisms, and how they coexist with
*Correspondence: [email protected]
various microorganisms is one of the most important selection pressures on termites [6–10]. The interactions between termites and microorganisms vary from total dependence of the termite on the microorganism for food at one end of the spectrum to total dependence of the microorganism on the termite at the other, with many degrees of mutualism in between. Termite egg mimicry by parasitic fungi is a phenomenon representative of the great diversity of termite-fungal interactions [2, 3]. Mechanism of recognition is essential to the evolution of parasitic interactions between species. Chemical mimicry is a wellknown strategy for various social parasites because social insects rely heavily on chemical communication for many aspects of their social interactions, including recognition of colony members [11, 12]. Here we report the chemical control of insect social behavior by a fungus. In their microorganism-rich habitat, termite eggs cannot survive without protection by workers [2]. Soon after being laid by queens, eggs are carried into nursery chambers and frequently groomed by workers, whereby they are coated with saliva, which contains antibiotic substances that protect the eggs from desiccation and pathogenic infection. Brown fungal balls, called ‘‘termite balls,’’ are often found in egg piles of various termite species (Figure 1A). Termite balls tended by Reticulitermes termites are sclerotia of an athelioid fungus (Basidiomycota, Agaricomycotina) of the genus Fibularhizoctonia [2]. To date, this egg-mimicking fungus has been found in four Reticulitermes termites in Japan (R. speratus, R. kanmonensis, R. amamianus, and R. miyatakei) and four species in the United States (R. flavipes, R. virginicus, R. hageni, and R. malletei) [13, 14]. The number of termite balls sometimes exceeds the number of true eggs in egg piles. Most of termite balls are inhibited from germination in egg piles, whereas the fungus rarely consumes the eggs in natural colonies [3]. Old termite balls become shrunken and deformed. Such old, deformed termite balls are removed from egg piles and thrown in the corner of the nest as garbage. The dumped termite balls germinate and grow in the nest (Figure 1B), and newly formed termite balls are carried into egg piles. Thus, the termite ball fungus is able to live in the termite nest, where the other naturally occurring fungi hardly access due to the antifungal defense of termites (e.g., [9, 10]). Egg mimicry provides a nearly competitor-free habitat for the fungus. Termites exclusively tend termite balls both with diameters that exactly match their egg size and with smooth spherical surface [3]. Unlike other sclerotiumforming fungi, termite balls have an extremely stable size corresponding to the size of the termite egg, and they have lost the outer shell-like layer and thus have acquired a smooth surface as a morphological adaptation for egg mimicry. However, they are easily killed by desiccation as they have lost desiccation resistance as a cost of morphological mimicry [3]. This fungus relies on termites for defense against desiccation and other microorganisms, where frequent grooming by workers keeps the survival rate of the termite balls at almost 100% [2]. 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 a short term. In this sense, the termite ball fungus is a fungal cuckoo, so to speak, which takes advantage of brood care of host species by mimicking eggs.
Chemical Mimicry of Termite Eggs by Fungus 31
A Lysozyme activity (U)
0.4 0.3 b 0.2 0.1
0
B
a
c
c
c
NM- NM+ TMB
Termite WFC Salivary gland
Pheromone activity
1.0 a 0.8 0.6 bc
0.4
b c
0.2 0
L
C E-vp39 promoter
Lt
EE
Control
XhoI Native signal XbaI
pV23-Rs-Lys2
Figure 1. Scheme to Determine Chemical Mimicry by the Termite EggMimicking Fungus (A) Termite balls in a termite egg pile. Termite eggs are transparent and oval, whereas termite balls are brown and spherical. Workers frequently groom the eggs, whereby they are coated with saliva to protect against desiccation and pathogen infection. Neither eggs nor termite balls survive without protection by workers in the nursery chambers. (B) Termite ball fungus growing on the termite nest material in the corner of the nest. (C) Comparison of mycelial growth of 2-week-old fungal colonies cultured on minimal media (1.34% yeast nitrogen base, 4 3 10%–5% biotin, 4 3 10%–3% histidine) either with 1% cellobiose or 1% glucose, or without sugar as a negative control. This shows that the fungus produces b-glucosidase, which is the enzyme to digest cellobiose into glucose. (D) Scheme of candidate protein screening based on the concept of chemical overlap in the wood-decomposition niche. The observation that the saliva of the wood-feeding cockroach has high termite-egg-recognition pheromone activity suggests that the target protein is a fundamental enzyme of wood feeders. Lysozyme is located in area W (aXbXgXdc), where it is produced by termites and wood-feeding cockroaches but cannot be produced by the fungus. The cellulose-digesting enzyme b-glucosidase remains in area X (aXbXgXd).
In addition to morphological mimicry, chemical camouflage is also needed for the fungus to be recognized as eggs by termites [3]. The antibacterial protein lysozyme (EC 3.2.1.17) is the egg-recognition pheromone in termites [15]. However, lysozyme cannot explain chemical mimicry by the fungus because lysozyme production by the fungus was not observed under the culture conditions employed (Figure 2A). In addition, lysozyme alone cannot fully explain the high pheromone activity of termite saliva and termite egg extracts (Figure 2B). Therefore, we postulated that another major component of
Control
FLAG
S kDa 220 120 80 60 50 40
Baculovirus homologous region
30
Amp r
20
S
2.5 0.25
FT W E1 E2 2.5 2.5 2.5 2.5
Figure 2. Lytic Activity and Pheromone Activity of Lysozyme (A) Comparison of lysozyme activity. Lysozyme activity of extracts of termite balls (TMB) grown on potato dextrose agar with (NM+) or without (NM2) nest material was measured by the lysis of Micrococcus lysodeikticus. Extracts from 25 mg termite balls were used in each assay. Extracts of salivary glands of termite workers (5 individual equivalent) and wood-feeding cockroach (WFC) larvae (1/3 individual equivalent) were used as positive controls. Error bars represent standard deviation of three replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F4,10 = 361.78, p < 0.0001, one-way ANOVA). Lysozyme production by the termite ball fungus was not observed under the culture conditions employed. (B) Egg-recognition pheromone activity of lysozyme. Dummy eggs were coated with each enzyme at a final concentration of 1 mg/bead. L, lysozyme standard from hen egg; Lt, termite lysozyme obtained in the baculovirussilkworm expression system; EE, crude termite egg extract as a positive control. The bioassay was replicated three times for each of the three colonies. Error bars represent standard error of the mean of nine replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F3, 30 = 20.65, p < 0.0001, two-way ANOVA). (C) Structure of Rs-Lys2 cDNA and a transfer vector (left). The vector, pV23, contains the baculovirus homologous sequence including E-vp39 gene promoter fused to Rs-Lys2 cDNA. Ampr and FLAG represent ampicillinresistance gene and FLAG tag, respectively. Western blot (right) for the purified termite lysozyme. S, hemolymph sample from larvae infected with recombinant pV23-RsLys2; FT, flow-through fraction; W, wash fraction; E1, the first elution fraction; E2, the second elution fraction. The amount of sample loaded (ml) was shown on each lane. Lysozyme alone cannot fully explain termite egg recognition pheromone and egg mimicry by the fungus.
the egg-recognition pheromone remains to be identified and might explain chemical mimicry by the fungus. To determine whether lysozyme is the single chemical component of the termite egg-recognition pheromone, we first obtained pure
A
1.0
Pheromone activity
Current Biology Vol 19 No 1 32
0.8
a ab
ab
0.6
b 0.4 0.2 0
c SG
Egg
Termite B
Larva
Adult
Control
WF cockroach
β-glucosidase activity (U/mg)
0.14 a 0.13 b
0.03 0.02
c
0.01
d e
0
SG
Egg
Termite
Larva
Adult
Control
WF cockroach
1.0
C
Pheromone activity
a
a
0.8 a 0.6 0.4 b
0.2 0 β
Lβ
EE
Control
Figure 3. Evidence that b-Glucosidase Is a Component of Termite Egg-Recognition Pheromone (A) Piling rates of dummy eggs made of glass beads coated with extracts of salivary glands of termite workers (SG), eggs dissected from termite queens, salivary glands of wood-feeding cockroach larvae, adult woodfeeding cockroach salivary glands, or control without chemical coating. Error bars represent standard error of the mean of eight replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F4, 30 = 20.09, p < 0.0001, two-way ANOVA). (B) b-glucosidase activity of the extracts used in (A). Error bars represent standard deviation of three independent replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F4, 10 = 6886.83, p < 0.0001, one-way ANOVA). (C) Egg-recognition pheromone activity of enzyme standards. Dummy eggs were coated with each enzyme at a final concentration of 1 mg/bead. b,
termite lysozyme by using the baculovirus-silkworm larvae Bombyx mori protein expression system (Figure 2C). A sufficient amount of termite lysozyme was expressed in silkworm larvae and then purified with anti-FLAG affinity gel (SigmaAldrich; Figure 2C). The termite lysozyme produced in the silkworm expression system showed significant termite egg-recognition pheromone activity (Figure 2B). However, its pheromone activity was not significantly different from that of hen egg lysozyme. Regardless of its origin, lysozyme alone cannot achieve egg-recognition activity as high as that of crude pheromone extract from termites. These observations raise questions regarding the nature of the other component(s) of the egg-recognition pheromone. The target chemical must fulfill the following requirements. (1) It must be a component of termite saliva and eggs because their extracts show high pheromone activity at the same level (Figure 3A). (2) The target chemical is a protein because its pheromone activity disappears completely after proteinase digestion [15]. (3) The target protein should also be produced by the fungus to mimic termite eggs. To screen for candidates from salivary proteins, we first examined cross-species activity of egg-recognition pheromone obtained from the salivary glands of five different lower termite species. The eggrecognition bioassay revealed that all salivary gland extracts from R. miyatakei, R. kanmonensis, Coptotermes formosanus (Rhinotermitidae), Cryptotermes brevis (Kalotermitidae), and Hodotermopsis sjostedti (Termopsidae) showed egg-recognition activity for R. speratus as high as the saliva of the parent species (see Figure S1 available online). When considering the evolution of termites, it is important to note that termites are a eusocial form of cockroach [16, 17]. Modern species of termites and wood-feeding cockroaches should therefore share some ancestral traits. We found that even the salivary gland extracts of the wood-feeding cockroach Panesthia angustipennis spadica (Blattaria: Blaberidae) showed strong termite egg-recognition pheromone activity (Figure 3A). This observation suggests that their common ancestors already had the target protein in their saliva for some life history function other than as an egg-recognition pheromone before the origin of eusociality. When considering the evolution of fungi, it is also important to note that the phylogenetic position of the termite ball fungus is located in the decay fungus genus Athelia [2]. Indeed, termite ball fungus grows on the termite nest material, indicating that the fungus can obtain energy and nutrition from wood (Figures 1B and 1C). Our candidate protein screen based on evolutionary reasonability and pheromone bioassays limited the target protein to cellulose-digesting enzymes (Figure 1D). Both termites and wood-feeding cockroaches produce the endogenous cellulose-digesting enzymes endo-b-1,4-glucanase (EC 3.2.1.4) and b-glucosidase (EC 3.2.1.21) in their salivary glands [5, 18, 19], although cellulose digestion was previously thought to rely solely on microbial gut symbionts. These cellulose-digesting enzymes appear not to be a common necessity of insect eggs. However, the target protein must also be produced in termite eggs to act as an egg-recognition pheromone. Reverse transcriptase polymerase chain reaction (RT-PCR) with
b-glucosidase standard from almond; Lb, mixture of lysozyme and b-glucosidase standards; EE, crude termite egg extract as a positive control. The bioassay was replicated three times for each of the three colonies. Error bars represent standard error of the mean of nine replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F3, 30 = 23.81, p < 0.0001, two-way ANOVA).
Chemical Mimicry of Termite Eggs by Fungus 33
A
egg
Micrograph
salivary gland
worker abdomen
termite ball
SR SD HG
SG
Fluorescent micrograph
SR SR SD SG
HG
0.6
a
8.0 6.0 4.0 2.0 b
c
Pheromone activity
B
β-glucosidase activity (mU)
SR
a 0.4
0.2
b
b
Figure 4. Detection of b-Glucosidase with a Fluorescent Probe (A) Micrograph (upper panels) or fluorescent micrographs (lower panels) of a termite salivary gland, termite eggs dissected from the ovary of a queen, cross section of the abdomen of a termite worker, and termite balls harvested from a potato dextrose agar (PDA) plate containing termite nest material. SG, salivary gland; SR, salivary reservoir; SD, salivary duct; HG, hind gut. Scale bars represent 0.1 mm in panels of SG and HG and represent 0.5 mm in other panels. Termite eggs and termite balls share b-glucosidase, whereas no b-glucosidase was present in the Argentine ant egg (in the square). (B) Conditional production of b-glucosidase and chemical mimicry by termite ball fungus. b-glucosidase activity (left) of extracts of termite balls grown on PDA with (NM+) or without (NM2) nest material. Error bars represent standard deviation of three replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F2, 6 = 22284.43, p < 0.0001, one-way ANOVA). Piling rates of dummy eggs (right) coated with extracts of termite balls (TMB) grown on PDA with or without nest material, or control without chemical coating. Error bars represent standard error of the mean of four independent experiments. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F2, 9 = 9.91, p < 0.001, one-way ANOVA). Termites recognized termite balls only when they produced b-glucosidase.
0
0 NM-
NM+
TMB
Control
NM-
NM+
Control
TMB
primers specific for the endo-b-1,4-glucanase of Reticulitermes [19] showed that endo-b-1,4-glucanase genes are not expressed in eggs, whereas their expression was observed in worker salivary glands (Figure S2). Indeed, endo-b-1,4glucanase standard alone showed no pheromone activity (p > 0.05, Tukey’s HSD test). By contrast, enzymatic activity of b-glucosidase was detected in both the salivary glands and eggs (Figure 3B). TokyoGreen-bGlu (Sekisui Medical) is an excellent novel probe for b-glucosidase, which has a number of advantages over previously used probes, especially in terms of sensitivity and for real-time imaging of living cells [20]. Experiments using this fluorescent probe confirmed the existence of b-glucosidase in the salivary glands, eggs, and hindgut of termites (Figure 4A). The primary function of b-glucosidase as a cellulose-digesting enzyme can explain its presence in salivary glands and the hindgut but cannot explain its localization in eggs. We did not detect b-glucosidase in the eggs of Argentine ants (Linepithema humile) used as a negative control (Figure 4A), indicating that it is not a common developmental necessity for insect eggs. Thus, the production of b-glucosidase in eggs in termite queen ovaries strongly suggests its secondary function as an egg-recognition pheromone. As predicted, b-glucosidase standard showed strong eggrecognition pheromone activity (Figure 3C). Perfect pheromone activity was achieved with a mixture of b-glucosidase and lysozyme standards, whereas even b-glucosidase alone showed relatively high pheromone activity. No other common insect salivary enzyme [21], including endo-b-1,3-glucanase (EC 3.2.1.6), a-amylase (EC 3.2.1.1), a-glucosidase (EC 3.2.1.20), and a-galactosidase (EC 3.2.1.22), showed significant termite
egg-recognition pheromone activity compared with the negative control (p > 0.05, Tukey’s HSD test). We found that the wood-feeding cockroach had both b-glucosidase (Figure 3B) and lysozyme (Figure 2A) in its salivary glands, which explains why the egg-recognition pheromone activity of wood-feeding cockroach saliva was as high as that of termite saliva (Figure 3A). In addition, saliva of cockroach larvae showed higher b-glucosidase activity than that of adult cockroaches (Figure 3B), which was consistent with the observation that larval saliva had higher pheromone activity than adult saliva (Figure 3A). We conclude that the termite egg-recognition pheromone consists of b-glucosidase and lysozyme, although we do not completely deny the possibility of other minor chemical compounds involved in termite egg recognition. To determine whether the termite ball fungi mimic termite eggs by producing b-glucosidase, we examined b-glucosidase production by the fungus and tested its function in chemical mimicry. Production of b-glucosidase, which is a common fungal enzyme, has been reported in various fungi including the sclerotium-forming fungus Sclerotium rolfsii [22], which is closely related to the termite ball fungi. We found that the termite ball fungus produces b-glucosidase conditionally; the fungus produced only a marginally detectable amount of b-glucosidase when cultured on plain potato dextrose agar (PDA), whereas addition of termite nest material to PDA markedly increased b-glucosidase production by the fungus (Figure 4B). The fluorescent probe confirmed the existence of b-glucosidase on the surface of termite balls produced on PDA with nest material (Figure 4A). There were no morphological differences between termite balls produced on PDA with
Current Biology Vol 19 No 1 34
and without nest material. This conditional enzyme production seems likely because b-glucosidase production is inhibited in the absence of the substrate of the enzyme but with sufficient amounts of other nutritional resources. In fact, addition of 1% cellobiose, the substrate of b-glucosidase, also promoted b-glucosidase production. Conditional production of b-glucosidase provided the ideal opportunity to examine whether this enzyme can explain chemical egg mimicry by the fungus. The extracts of termite balls showed egg recognition activity only when they contained b-glucosidase (Figure 4B). The conditional acceptance of termite balls by termites in accordance with b-glucosidase production by the fungus is compelling evidence that the termite ball fungi mimic termite eggs by producing this enzyme. Our results were parsimonious with the evolution of chemical signals involved in within-colony social communications and in interspecific associations. The primary functions of the pheromone components, lysozyme and b-glucosidase, are as an antibacterial defense agent and digestion enzyme, respectively. Termites have evolved to use the pre-existing chemicals on the egg surface as an egg-recognition signal, without the evolution of any additional specific chemical for this purpose. b-glucosidase in termite eggs might still have any practical function (e.g., egg shell formation) because b-glucosidase is an essential enzyme for oothecal formation in other dictyopteran insects including mantis [23]. Secondary use of chemical compounds that have evolved for other primary functions can be seen in various social insects [24, 25]. It is notable that both lysozyme and b-glucosidase are enzymes that hydrolyze b-1,4 glycosidic linkages. In addition, termites cannot distinguish the enzymes from different origins based on differences in amino acid sequence. For example, termite lysozyme shows approximately 40% amino acid similarity with hen egg lysozyme [15]. Even a mixture of hen egg lysozyme and almond b-glucosidase showed perfect pheromone activity. We confirmed that heat inactivation of enzymes did not significantly affect their pheromone activity, suggesting that the pheromones are not the products of digestion but the enzymes themselves (Figure S3). However, our results do not rule out the possibility that other supporting compounds are involved in termite egg recognition. In the bioassay, dummy eggs were coated with the b-glucosidase standard at the concentration of 1 mg/bead (7.6 mU/bead), whereas the amount of b-glucosidase detected in each termite egg was 0.013 mU/egg. Future quantitative approaches are needed to understand the relative function and the recognition specificity of each compound. When considering the evolutionary process of chemical mimicry, it is important to note that the fungus originally had the potential to produce the pheromone substance for use as an enzyme for cellulose digestion. This would have served as an important preadaptation for the evolution of termite-egg mimicry by the fungus. Social insect colonies have evolved diverse defensive strategies to prevent parasite invasion [25]. However, once a way to overcome these defenses has been established, it can be an ideal environment for parasites providing a stable competitor-free habitat. Recently, another egg-mimicking fungus Trechispora sp. (Basidiomycota, Agaricomycotina) was found in the higher termite Nasutitermes takasagoensis, indicating that egg mimicry has evolved at least twice independently (K.M. and T.Y., unpublished data). Egg mimicry, by which the fungus can easily gain access to the center of the nest, seems to be an evolutionary loophole around antiparasite defense in termites. Striking similarities can be
seen in the parasitic interaction between Maculinea butterfly and Myrmica ant, where chemical larval mimicry plays a role as a signal to be picked up outside the nest and actively brought into the colony by ants [26]. Recent cost-benefit analyses of aphid-ant [27] and lycaenid-ant [28] interactions have demonstrated that even apparently peaceful mutualisms involve many potential conflicts. In the Lasius ant-Stomaphis aphid mutualism, ant workers bring aphid eggs into the nest and groom the eggs to protect them against pathogenic fungi [29]. Even in such a mutualistic interaction, egg tending by ants may be the result of the aphids manipulating the ants. To understand the evolution of chemical mimicry for social parasitism, we should carefully consider the net outcome of the interaction not only in the short term but also in the long term, frequency of the interaction in the host populations, and preadaptation of parasites, as well as the mechanism of recognition. Experimental Procedures Egg-Recognition Bioassay The workers, queens, and eggs used here were collected from R. speratus colonies in Okayama, Japan. To examine egg-recognition pheromone activity of extracted chemicals, we used dummy eggs made of glass beads coated with each test chemical as described previously [15]. Extraction of crude egg-recognition pheromone from the salivary glands of workers and eggs were carried out by a previously published method [15]. Dummy eggs consisting of 0.5 mm glass beads coated with 1.0 mg/bead of each test chemical were used in egg recognition bioassays. We randomly arranged ten eggs and 20 dummy eggs on moist unwoven cloth in a Petri dish, and ten workers were released in the dish. Acceptance rates were determined by counting the number of dummy eggs carried into egg piles. Eggs used for chemical extraction and enzyme detection were obtained from the ovaries of queens to avoid contamination by enzymes from worker saliva. The egg-recognition bioassay was replicated four times for each of the two R. speratus colonies unless otherwise stated. Piling rates were arcsine-square-root-transformed and analyzed by two-way ANOVA followed by Tukey’s HSD test for multiple comparisons if ANOVA indicated a significant difference. Termite species for the study of cross-species pheromone activities were collected in Okayama (C. formosanus and C. brevis), Yamaguchi (R. kanmonensis), and Amami Island (R. miyatakei and H. sjostedti) in Japan. Colonies of the wood-feeding cockroach Panesthia angustipennis spadica were collected in Takahashi, Okayama, Japan. We also examined pheromone activity of following enzyme standards: lysozyme (EC 3.2.1.17, from hen egg, #L7651, Sigma-Aldrich), b-glucosidase (EC 3.2.1.21, from almond, #G0395, Sigma-Aldrich), endo-b-1,4-glucanase (EC 3.2.1.4, from Trichoderma viride, #C1794, Sigma-Aldrich), endo-b-1,3-glucanase (EC 3.2.1.6, from Aspergillus niger, #49101, BioChemica), a-amylase (EC 3.2.1.1, from Bacillus sp. #A6380, Sigma-Aldrich), a-glucosidase (EC 3.2.1.20, from Bacillus stearothermophilus, #G3651, Sigma-Aldrich), and a-galactosidase (EC 3.2.1.22, from Mortierella vinacea, #100560, Hokkaido Sugar Company). Purified termite lysozyme expressed in the B. mori-baculovirus system and enzyme standards were dialyzed (MWCO 3500) against distilled water to remove buffer salts before use to avoid interference with egg recognition [15]. Lysozyme Assay Termite balls harvested from potato dextrose agar plates were homogenized by grinding with a mortar and pestle in 0.1 M phosphate buffer (pH 6.0) on ice and then centrifuged at 20,000 3 g for 10 min at 4 C. The supernatant was used as the enzyme extract. The lytic activity of enzyme extract was measured against Micrococcus lysodeikticus (Sigma-Aldrich) suspended in the phosphate buffer. Lysis of M. lysodeikticus was measured at 450 nm with a spectrophotometer (SmartSpec Plus, BioRad). One unit of the activity was defined as the amount of enzyme decreasing the absorbance at a rate of O.D. 0.01/min. Phosphate buffer (pH 3.6) was used for the salivary gland extracts of the wood-feeding cockroach instead of pH 6.0 because the optimum pH of wood-feeding cockroach lysozyme was 3.6. Extracts of salivary glands of termite workers and wood-feeding cockroach larvae were used as positive controls. Five individual equivalent and 1/3 individual equivalent of salivary gland extract were used in each assay for termites and wood-feeding cockroach, respectively. Only extraction
Chemical Mimicry of Termite Eggs by Fungus 35
buffer was added to the reaction mixture in the negative control. Assay was replicated three times for each treatment. Expression of Termite Lysozyme in Silkworm-Baculovirus System To obtain sufficient amounts of termite lysozyme for the egg-recognition bioassay, we used the Bombyx mori-baculovirus expression system. Termite lysozyme cDNA (Rs-Lys2) [30] was cloned and ligated into the baculovirus transfer vector pV23. Total RNA was extracted from the salivary glands of 50 worker termites with NucleoSpin RNA II (Macherey-Nagel), and one-step RT-PCR was performed with a kit (QIAGEN) and the primers 50 -CTCGAGATGGACGTGAGAAATTCTCCGAT-30 and 50 -TCTAGAACGAAA ACTGGAATTACGACCACA-30 according to the manufacturer’s protocol. The amplified PCR product encoding Rs-Lys2 with the putative native signal peptide was subcloned with a TOPO TA cloning kit (Invitrogen). The plasmid containing the Rs-Lys2 nucleotide sequence was propagated with a Quantum Prep Plasmid Midiprep kit (BioRad). The Rs-Lys2 fragment was then cut out with Xhol and XbaI (Takara Bio) and ligated into the same sites of the transfer vector (pV23, Katakura Industries), yielding pV23-Rs-Lys2 (Figure 2C). Rs-Lys2 cDNA was fused in-frame to the C-terminal FLAG tag sequence (Sigma-Aldrich). Recombinant Rs-Lys2 was produced by KaikoExpress (Katakura Industries, http://www.katakura.co.jp/research/ ke-top.htm) and purified with Anti-FLAG M2 Agarose Affinity Gel column (Sigma-Aldrich; see Supplemental Experimental Procedures). RT-PCR for Endo-b-1,4-glucanase Gene Expression Analysis Total RNA was extracted from eggs dissected from two queens, legs of the queens, hind gut of a queen, salivary glands of 100 workers, and 20 eggs collected from the nest, with the NucleoSpin RNA II (Macherey-Nagel). We performed one-step RT-PCR to detect expression of an endogenous cellulase in termite, using a kit from QIAGEN according to the manufacturer’s protocol, with the following primers [19]: Cell-1 ORF-F 50 -GCCATGAAGGTC TTCGTTTG-30 and Cell-1 ORF-R 50 -GTTACACGCCAGCCTTGA-30 . To verify the integrity of the mRNA used in this experiment, samples were amplified by PCR with the housekeeping gene b-actin primers; forward 50 -AGAGGGA AATCGTGCGTGAC-30 ; reverse 50 -CAATAGTGATGACCTGGCCGT-30 . Detection of b-Glucosidase b-Glucosidase activity was assayed with p-nitrophenyl-b-D-glucoside (pNPbG) by standard methods [31]. Samples of salivary glands, eggs, and termite balls were homogenized in sodium acetate buffer (200 ml) on ice by grinding with a mortar and pestle and then centrifuged at 20,000 3 g for 10 min at 4 C. The supernatant was used as the enzyme extract. b-Glucosidase was assayed with a mixture (1 ml) containing 4 mM pNPbG, 50 mM sodium acetate buffer (pH 5.0), and 50 ml of enzyme solution. After incubation at 50 C for 30 min, the reaction was stopped by adding 1 ml of ice-cold 0.5 M Na2CO3, and the absorbance at 405 nm was recorded with a spectrophotometer (SmartSpec Plus, BioRad). One unit of b-glucosidase was defined as the release of 1 mmol of p-nitrophenol min21 in the reaction mixture. Data were standardized as units per milligram (lyophilized dry weight) of each extract. To determine the localization of b-glucosidase, we also used the enzyme activity-based fluorescent probe TokyoGreen-bGlu [20] (Sekisui Medical). b-Glucosidase hydrolyses nonfluorescent TokyoGreen-bGlu into highly fluorescent TokyoGreen (Ex./Em. = 490/510 nm). For fluorescence imaging, TokyoGreen-bGlu was used at a concentration of 10 mM in 10 mM phosphate buffer (pH 7.0). Salivary glands and eggs were dissected from termites in phosphate buffer and then soaked in TokyoGreen-bGlu mixture for 30 min at 27 C. We added 0.4% Triton X-100 in buffer for insect eggs. Eggs of the Argentine ant were tested as a negative control. Fluorescence images were captured with a fluorescence microscope (BZ8000, Keyence). Fungal Culture The three strains of termite ball fungus used in this experiment were isolated from the egg piles of R. speratus in Okayama, Japan. Termite balls extracted from egg piles were inoculated onto PDA in Petri dishes 90 mm in diameter and incubated at 27 C for 20 days. One newly developed sclerotium was reisolated from each termite ball and cultured on a new PDA plate. Two mycelial plugs 5 mm in diameter were taken from the 1-month-old colony of each isolate and placed on either a new plate of plain PDA or on a plate of PDA containing termite nest material and incubated at 27 C for 20 days. The nest materials consisting of nest structure and nest wood were obtained from a nest of R. speratus in red pine wood, dried at room temperature, milled (MX-X41; National), and added to PDA at a concentration of 15% (w/v). Termite balls were harvested from each plate and used in enzyme assays and pheromone assays.
Supplemental Data Supplemental Data include Supplemental Experimental Procedures, three figures and can be found with this article online at http://www. current-biology.com/supplemental/S0960-9822(08)01546-7. Acknowledgments We thank N.E. Pierce, S.-P. Quek (Harvard University), T. Nishida, C. Tanaka, K. Fujisaki (Kyoto University, Japan), T. Miyatake, Y. Murata, and F. Nakasuji (Okayama University, Japan) for their helpful advice and discussions. This work was funded by the Japan Society for the Promotion of Science and the Program for Promotion of Basic Research Activities for Innovative Biosciences (PROBRAIN). Received: September 13, 2008 Revised: November 5, 2008 Accepted: November 6, 2008 Published online: December 24, 2008 References 1. Vega, F.E., and Blackwell, M. (2005). Insect-Fungal Associations: Ecology and Evolution (New York: Oxford University Press). 2. Matsuura, K., Tanaka, C., and Nishida, T. (2000). Symbiosis of a termite and a sclerotium-forming fungus: Sclerotia mimic termite eggs. Ecol. Res. 15, 405–414. 3. Matsuura, K. (2006). Termite-egg mimicry by a sclerotium-forming fungus. Proc. R. Soc. Lond. B. Biol. Sci. 273, 1203–1209. 4. Cleveland, L.R. (1923). Symbiosis between termites and their intestinal protozoa. Proc. Natl. Acad. Sci. USA 9, 424–428. 5. Watanabe, H., Noda, H., Tokuda, G., and Lo, N. (1998). A cellulase gene of termite origin. Nature 394, 330–331. 6. Matsuura, K. (2003). Symbionts affecting termite behavior. In Insect Symbiosis, K. Bourtzis and T.A. Miller, eds. (Boca Raton: CRC Press), pp. 131–143. 7. Matsuura, K. (2001). Nestmate recognition mediated by intestinal bacteria in a termite, Reticulitermes speratus. Oikos 92, 20–26. 8. Matsuura, K., Fujimoto, M., Goka, K., and Nishida, T. (2002). Cooperative colony foundation by termite female pairs: Altruism for survivorship in incipient colonies. Anim. Behav. 64, 167–173. 9. Rosengaus, R.B., Guldin, M.R., and Traniello, J.F.A. (1998). Inhibitory effect of termite fecal pellets on fungal spore germination. J. Chem. Ecol. 24, 1697–1706. 10. Rosengaus, R.B., Jordan, C., Lefebvre, M.L., and Traniello, J.F.A. (1999). Pathogen alarm behavior in a termite: A new form of communication in social insects. Naturwissenschaften 86, 544–548. 11. Howard, R.W., McDaniel, C.A., and Blomquist, G.J. (1980). Chemical mimicry as an integrating mechanism: Cuticular hydrocarbons of a termitophile and its host. Science 210, 431–433. 12. Meer, R.K., and Wojcik, D.P. (1982). Chemical mimicry in the myrmecophilous beetle Myrmecaphodius excavaticollis. Science 218, 806–808. 13. Matsuura, K. (2005). Distribution of termite egg-mimicking fungi (‘‘termite balls’’) in Reticulitermes spp. (Isoptera: Rhinotermitidae) nests in Japan and the United States. Appl. Entomol. Zool. (Jpn.) 40, 53–61. 14. Yashiro, T., and Matsuura, K. (2007). Distribution and phylogenetic analysis of termite egg-mimicking fungi ‘‘termite balls’’ in Reticulitermes termites. Ann. Entomol. Soc. Am. 100, 532–538. 15. Matsuura, K., Tamura, T., Kobayashi, N., Yashiro, T., and Tatsumi, S. (2007). The antibacterial protein lysozyme identified as the termite egg recognition pheromone. PLoS ONE 2, e813. 10.1371/journal. pone.0000813. 16. Kambhampati, S. (1995). A phylogeny of cockroaches and related insects based on DNA sequence of mitochondrial ribosomal RNA genes. Proc. Natl. Acad. Sci. USA 92, 2017–2020. 17. Lo, N., Tokuda, G., Watanabe, H., Rose, H., Slaytor, M., Maekawa, K., Bandi, C., and Noda, H. (2000). Evidence from multiple gene sequences indicates that termites evolved from wood-feeding cockroaches. Curr. Biol. 10, 801–804. 18. Scrivener, A.M., and Slaytor, M. (1994). Properties of the endogenous cellulase from Panesthia cribrata Saussure and purification of major endo-beta-1,4-glucanase components. Insect Biochem. Mol. Biol. 24, 223–231.
Current Biology Vol 19 No 1 36
19. Zhou, X., Smith, J.A., Koehler, P.G., Oi, F.M., Bennett, G.W., and Scharf, M.E. (2007). Correlation of cellulase gene expression and cellulolytic activity throughout the gut of the termite Reticulitermes flavipes. Gene 395, 29–39. 20. Urano, Y., Kamiya, M., Kanda, K., Ueno, T., Hirose, K., and Nagano, T. (2005). Evolution of fluorescein as a platform for finely tunable fluorescence probes. J. Am. Chem. Soc. 127, 4888–4894. 21. House, C.R., and Ginsborg, B.L. (1985). Salivary gland. In Comprehensive Insect Physiology, Biochemistry and Pharmacology, Vol. 11, G.A. Kerkut and L.I. Gilbert, eds. (Oxford: Pergamon Press), pp. 195–224. 22. Chinnathambi, S., and Lachke, A. (1997). Activities of glycosidases from Sclerotium rolfsii in non-aqueous media. Biotechnol. Tech. 11, 589–591. 23. Yago, M., Sato, H., Oshima, S., and Kawasaki, H. (1990). Enzymatic activities involved in the oothecal sclerotization of the praying mantid, Tenodera aridifolia sinensis Saussure. Insect Biochem. 20, 745–750. 24. Turillazzi, S., Dapporto, L., Pansolli, C., Boulay, R., Dani, F.R., Moneti, G., and Pieraccini, G. (2006). Habitually used hibernation sites of paper wasps are marked with venom and cuticular peptides. Curr. Biol. 16, R530–R531. 25. Cremer, S., Armitage, S.A.O., and Schmid-Hempel, P. (2007). Social immunity. Curr. Biol. 17, R693–R702. 26. Nash, D.R., Als, T.D., Maile, R., Jones, G.R., and Boomsma, J.J. (2008). A mosaic of chemical coevolution in a large blue butterfly. Science 319, 88–90. 27. Stadler, B., Kindlmann, P., Smilauer, P., and Fiedler, K. (2003). A comparative analysis of morphological and ecological characters of European aphids and lycaenids in relation to ant attendance. Oecologia 135, 422–430. 28. Pierce, N.E., Braby, M.F., Heath, A., Lohman, D.J., Mathew, J., Rand, D.B., and Travassos, M.A. (2002). The ecology and evolution of ant association in the Lycaenidae (Lepidoptera). Annu. Rev. Entomol. 47, 733–771. 29. Matsuura, K., and Yashiro, T. (2006). Aphid-egg protection by ants: a novel aspect of the mutualism between the tree-feeding Stomaphis hirukawai and its attendant ant Lasius productus. Naturwissenschaften 93, 506–510. 30. Fujita, A., Minamoto, T., Shimizu, I., and Abe, T. (2002). Molecular cloning of lysozyme-encoding cDNAs expressed in the salivary gland of a wood-feeding termite, Reticulitermes speratus. Insect Biochem. Mol. Biol. 32, 1615–1624. 31. Saha, B.C., and Bothast, R.J. (1996). Production, purification, and characterization of a highly glucose-tolerant novel beta-glucosidase from Candida peltata. Appl. Environ. Microbiol. 62, 3165–3170.
DOI 10.1016/j.cub.2008.11.030
Report Cuckoo Fungus Mimics Termite Eggs by Producing the Cellulose-Digesting Enzyme b-Glucosidase Kenji Matsuura,1,2,* Toshihisa Yashiro,1 Ken Shimizu,1 Shingo Tatsumi,1 and Takashi Tamura3 1Laboratory of Insect Ecology Graduate School of Environmental Science Okayama University Okayama 700-8530 Japan 2Department of Organismic and Evolutionary Biology Harvard University Cambridge, MA 02138 USA 3Graduate School of Natural Science and Technology Okayama University Okayama 700-8530 Japan
Summary Insects and fungi share a long history of association in various habitats, including the wood-decomposition niche [1]. Fungal mimicry of termite eggs is one of the most striking evolutionary consequences of insect–fungus association [2, 3]. Termites of the genus Reticulitermes often harbor fungal sclerotia, called ‘‘termite balls,’’ along with eggs in nursery chambers, whereby the fungus gains a competitor-free habitat in termite nests. Sophisticated morphological and chemical camouflage are needed for the fungus to mimic termite eggs [3]. However, the mechanism of chemical egg mimicry by the fungus is unknown. Here, we show that the fungus mimics termite eggs chemically by producing the cellulose-digesting enzyme b-glucosidase. We found that the termite egg-recognition pheromone consists of b-glucosidase and lysozyme. Both enzymes are major salivary compounds in termites and are also produced in termite eggs. Termite balls were tended by termites only when the fungus produced b-glucosidase. Our results demonstrated that the overlap of the cellulose digestion niche between termites and the fungus sharing the same chemicals provided the opportunity for the origin of termite egg mimicry by the fungus. This suggests that pheromone compounds might have originally evolved within other life history contexts, only later gaining function in chemical communication. Results and Discussion Cellulose in the form of dead plants is a superabundant potential resource, although digesting it is not an easy task for most animals. Termites have developed cellulose digestion capabilities with both symbiotic [4] and endogenous [5] cellulases that allow them to use dead plants as food resources. This wood-decomposition niche was occupied by various microorganisms, including bacteria, fungi, and protozoa, before the ancestors of termites began exploiting it. In this sense, termites live in the world of microorganisms, and how they coexist with
*Correspondence: [email protected]
various microorganisms is one of the most important selection pressures on termites [6–10]. The interactions between termites and microorganisms vary from total dependence of the termite on the microorganism for food at one end of the spectrum to total dependence of the microorganism on the termite at the other, with many degrees of mutualism in between. Termite egg mimicry by parasitic fungi is a phenomenon representative of the great diversity of termite-fungal interactions [2, 3]. Mechanism of recognition is essential to the evolution of parasitic interactions between species. Chemical mimicry is a wellknown strategy for various social parasites because social insects rely heavily on chemical communication for many aspects of their social interactions, including recognition of colony members [11, 12]. Here we report the chemical control of insect social behavior by a fungus. In their microorganism-rich habitat, termite eggs cannot survive without protection by workers [2]. Soon after being laid by queens, eggs are carried into nursery chambers and frequently groomed by workers, whereby they are coated with saliva, which contains antibiotic substances that protect the eggs from desiccation and pathogenic infection. Brown fungal balls, called ‘‘termite balls,’’ are often found in egg piles of various termite species (Figure 1A). Termite balls tended by Reticulitermes termites are sclerotia of an athelioid fungus (Basidiomycota, Agaricomycotina) of the genus Fibularhizoctonia [2]. To date, this egg-mimicking fungus has been found in four Reticulitermes termites in Japan (R. speratus, R. kanmonensis, R. amamianus, and R. miyatakei) and four species in the United States (R. flavipes, R. virginicus, R. hageni, and R. malletei) [13, 14]. The number of termite balls sometimes exceeds the number of true eggs in egg piles. Most of termite balls are inhibited from germination in egg piles, whereas the fungus rarely consumes the eggs in natural colonies [3]. Old termite balls become shrunken and deformed. Such old, deformed termite balls are removed from egg piles and thrown in the corner of the nest as garbage. The dumped termite balls germinate and grow in the nest (Figure 1B), and newly formed termite balls are carried into egg piles. Thus, the termite ball fungus is able to live in the termite nest, where the other naturally occurring fungi hardly access due to the antifungal defense of termites (e.g., [9, 10]). Egg mimicry provides a nearly competitor-free habitat for the fungus. Termites exclusively tend termite balls both with diameters that exactly match their egg size and with smooth spherical surface [3]. Unlike other sclerotiumforming fungi, termite balls have an extremely stable size corresponding to the size of the termite egg, and they have lost the outer shell-like layer and thus have acquired a smooth surface as a morphological adaptation for egg mimicry. However, they are easily killed by desiccation as they have lost desiccation resistance as a cost of morphological mimicry [3]. This fungus relies on termites for defense against desiccation and other microorganisms, where frequent grooming by workers keeps the survival rate of the termite balls at almost 100% [2]. 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 a short term. In this sense, the termite ball fungus is a fungal cuckoo, so to speak, which takes advantage of brood care of host species by mimicking eggs.
Chemical Mimicry of Termite Eggs by Fungus 31
A Lysozyme activity (U)
0.4 0.3 b 0.2 0.1
0
B
a
c
c
c
NM- NM+ TMB
Termite WFC Salivary gland
Pheromone activity
1.0 a 0.8 0.6 bc
0.4
b c
0.2 0
L
C E-vp39 promoter
Lt
EE
Control
XhoI Native signal XbaI
pV23-Rs-Lys2
Figure 1. Scheme to Determine Chemical Mimicry by the Termite EggMimicking Fungus (A) Termite balls in a termite egg pile. Termite eggs are transparent and oval, whereas termite balls are brown and spherical. Workers frequently groom the eggs, whereby they are coated with saliva to protect against desiccation and pathogen infection. Neither eggs nor termite balls survive without protection by workers in the nursery chambers. (B) Termite ball fungus growing on the termite nest material in the corner of the nest. (C) Comparison of mycelial growth of 2-week-old fungal colonies cultured on minimal media (1.34% yeast nitrogen base, 4 3 10%–5% biotin, 4 3 10%–3% histidine) either with 1% cellobiose or 1% glucose, or without sugar as a negative control. This shows that the fungus produces b-glucosidase, which is the enzyme to digest cellobiose into glucose. (D) Scheme of candidate protein screening based on the concept of chemical overlap in the wood-decomposition niche. The observation that the saliva of the wood-feeding cockroach has high termite-egg-recognition pheromone activity suggests that the target protein is a fundamental enzyme of wood feeders. Lysozyme is located in area W (aXbXgXdc), where it is produced by termites and wood-feeding cockroaches but cannot be produced by the fungus. The cellulose-digesting enzyme b-glucosidase remains in area X (aXbXgXd).
In addition to morphological mimicry, chemical camouflage is also needed for the fungus to be recognized as eggs by termites [3]. The antibacterial protein lysozyme (EC 3.2.1.17) is the egg-recognition pheromone in termites [15]. However, lysozyme cannot explain chemical mimicry by the fungus because lysozyme production by the fungus was not observed under the culture conditions employed (Figure 2A). In addition, lysozyme alone cannot fully explain the high pheromone activity of termite saliva and termite egg extracts (Figure 2B). Therefore, we postulated that another major component of
Control
FLAG
S kDa 220 120 80 60 50 40
Baculovirus homologous region
30
Amp r
20
S
2.5 0.25
FT W E1 E2 2.5 2.5 2.5 2.5
Figure 2. Lytic Activity and Pheromone Activity of Lysozyme (A) Comparison of lysozyme activity. Lysozyme activity of extracts of termite balls (TMB) grown on potato dextrose agar with (NM+) or without (NM2) nest material was measured by the lysis of Micrococcus lysodeikticus. Extracts from 25 mg termite balls were used in each assay. Extracts of salivary glands of termite workers (5 individual equivalent) and wood-feeding cockroach (WFC) larvae (1/3 individual equivalent) were used as positive controls. Error bars represent standard deviation of three replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F4,10 = 361.78, p < 0.0001, one-way ANOVA). Lysozyme production by the termite ball fungus was not observed under the culture conditions employed. (B) Egg-recognition pheromone activity of lysozyme. Dummy eggs were coated with each enzyme at a final concentration of 1 mg/bead. L, lysozyme standard from hen egg; Lt, termite lysozyme obtained in the baculovirussilkworm expression system; EE, crude termite egg extract as a positive control. The bioassay was replicated three times for each of the three colonies. Error bars represent standard error of the mean of nine replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F3, 30 = 20.65, p < 0.0001, two-way ANOVA). (C) Structure of Rs-Lys2 cDNA and a transfer vector (left). The vector, pV23, contains the baculovirus homologous sequence including E-vp39 gene promoter fused to Rs-Lys2 cDNA. Ampr and FLAG represent ampicillinresistance gene and FLAG tag, respectively. Western blot (right) for the purified termite lysozyme. S, hemolymph sample from larvae infected with recombinant pV23-RsLys2; FT, flow-through fraction; W, wash fraction; E1, the first elution fraction; E2, the second elution fraction. The amount of sample loaded (ml) was shown on each lane. Lysozyme alone cannot fully explain termite egg recognition pheromone and egg mimicry by the fungus.
the egg-recognition pheromone remains to be identified and might explain chemical mimicry by the fungus. To determine whether lysozyme is the single chemical component of the termite egg-recognition pheromone, we first obtained pure
A
1.0
Pheromone activity
Current Biology Vol 19 No 1 32
0.8
a ab
ab
0.6
b 0.4 0.2 0
c SG
Egg
Termite B
Larva
Adult
Control
WF cockroach
β-glucosidase activity (U/mg)
0.14 a 0.13 b
0.03 0.02
c
0.01
d e
0
SG
Egg
Termite
Larva
Adult
Control
WF cockroach
1.0
C
Pheromone activity
a
a
0.8 a 0.6 0.4 b
0.2 0 β
Lβ
EE
Control
Figure 3. Evidence that b-Glucosidase Is a Component of Termite Egg-Recognition Pheromone (A) Piling rates of dummy eggs made of glass beads coated with extracts of salivary glands of termite workers (SG), eggs dissected from termite queens, salivary glands of wood-feeding cockroach larvae, adult woodfeeding cockroach salivary glands, or control without chemical coating. Error bars represent standard error of the mean of eight replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F4, 30 = 20.09, p < 0.0001, two-way ANOVA). (B) b-glucosidase activity of the extracts used in (A). Error bars represent standard deviation of three independent replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F4, 10 = 6886.83, p < 0.0001, one-way ANOVA). (C) Egg-recognition pheromone activity of enzyme standards. Dummy eggs were coated with each enzyme at a final concentration of 1 mg/bead. b,
termite lysozyme by using the baculovirus-silkworm larvae Bombyx mori protein expression system (Figure 2C). A sufficient amount of termite lysozyme was expressed in silkworm larvae and then purified with anti-FLAG affinity gel (SigmaAldrich; Figure 2C). The termite lysozyme produced in the silkworm expression system showed significant termite egg-recognition pheromone activity (Figure 2B). However, its pheromone activity was not significantly different from that of hen egg lysozyme. Regardless of its origin, lysozyme alone cannot achieve egg-recognition activity as high as that of crude pheromone extract from termites. These observations raise questions regarding the nature of the other component(s) of the egg-recognition pheromone. The target chemical must fulfill the following requirements. (1) It must be a component of termite saliva and eggs because their extracts show high pheromone activity at the same level (Figure 3A). (2) The target chemical is a protein because its pheromone activity disappears completely after proteinase digestion [15]. (3) The target protein should also be produced by the fungus to mimic termite eggs. To screen for candidates from salivary proteins, we first examined cross-species activity of egg-recognition pheromone obtained from the salivary glands of five different lower termite species. The eggrecognition bioassay revealed that all salivary gland extracts from R. miyatakei, R. kanmonensis, Coptotermes formosanus (Rhinotermitidae), Cryptotermes brevis (Kalotermitidae), and Hodotermopsis sjostedti (Termopsidae) showed egg-recognition activity for R. speratus as high as the saliva of the parent species (see Figure S1 available online). When considering the evolution of termites, it is important to note that termites are a eusocial form of cockroach [16, 17]. Modern species of termites and wood-feeding cockroaches should therefore share some ancestral traits. We found that even the salivary gland extracts of the wood-feeding cockroach Panesthia angustipennis spadica (Blattaria: Blaberidae) showed strong termite egg-recognition pheromone activity (Figure 3A). This observation suggests that their common ancestors already had the target protein in their saliva for some life history function other than as an egg-recognition pheromone before the origin of eusociality. When considering the evolution of fungi, it is also important to note that the phylogenetic position of the termite ball fungus is located in the decay fungus genus Athelia [2]. Indeed, termite ball fungus grows on the termite nest material, indicating that the fungus can obtain energy and nutrition from wood (Figures 1B and 1C). Our candidate protein screen based on evolutionary reasonability and pheromone bioassays limited the target protein to cellulose-digesting enzymes (Figure 1D). Both termites and wood-feeding cockroaches produce the endogenous cellulose-digesting enzymes endo-b-1,4-glucanase (EC 3.2.1.4) and b-glucosidase (EC 3.2.1.21) in their salivary glands [5, 18, 19], although cellulose digestion was previously thought to rely solely on microbial gut symbionts. These cellulose-digesting enzymes appear not to be a common necessity of insect eggs. However, the target protein must also be produced in termite eggs to act as an egg-recognition pheromone. Reverse transcriptase polymerase chain reaction (RT-PCR) with
b-glucosidase standard from almond; Lb, mixture of lysozyme and b-glucosidase standards; EE, crude termite egg extract as a positive control. The bioassay was replicated three times for each of the three colonies. Error bars represent standard error of the mean of nine replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F3, 30 = 23.81, p < 0.0001, two-way ANOVA).
Chemical Mimicry of Termite Eggs by Fungus 33
A
egg
Micrograph
salivary gland
worker abdomen
termite ball
SR SD HG
SG
Fluorescent micrograph
SR SR SD SG
HG
0.6
a
8.0 6.0 4.0 2.0 b
c
Pheromone activity
B
β-glucosidase activity (mU)
SR
a 0.4
0.2
b
b
Figure 4. Detection of b-Glucosidase with a Fluorescent Probe (A) Micrograph (upper panels) or fluorescent micrographs (lower panels) of a termite salivary gland, termite eggs dissected from the ovary of a queen, cross section of the abdomen of a termite worker, and termite balls harvested from a potato dextrose agar (PDA) plate containing termite nest material. SG, salivary gland; SR, salivary reservoir; SD, salivary duct; HG, hind gut. Scale bars represent 0.1 mm in panels of SG and HG and represent 0.5 mm in other panels. Termite eggs and termite balls share b-glucosidase, whereas no b-glucosidase was present in the Argentine ant egg (in the square). (B) Conditional production of b-glucosidase and chemical mimicry by termite ball fungus. b-glucosidase activity (left) of extracts of termite balls grown on PDA with (NM+) or without (NM2) nest material. Error bars represent standard deviation of three replicates. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F2, 6 = 22284.43, p < 0.0001, one-way ANOVA). Piling rates of dummy eggs (right) coated with extracts of termite balls (TMB) grown on PDA with or without nest material, or control without chemical coating. Error bars represent standard error of the mean of four independent experiments. Different letters indicate significant differences (p < 0.05, Tukey’s HSD test; F2, 9 = 9.91, p < 0.001, one-way ANOVA). Termites recognized termite balls only when they produced b-glucosidase.
0
0 NM-
NM+
TMB
Control
NM-
NM+
Control
TMB
primers specific for the endo-b-1,4-glucanase of Reticulitermes [19] showed that endo-b-1,4-glucanase genes are not expressed in eggs, whereas their expression was observed in worker salivary glands (Figure S2). Indeed, endo-b-1,4glucanase standard alone showed no pheromone activity (p > 0.05, Tukey’s HSD test). By contrast, enzymatic activity of b-glucosidase was detected in both the salivary glands and eggs (Figure 3B). TokyoGreen-bGlu (Sekisui Medical) is an excellent novel probe for b-glucosidase, which has a number of advantages over previously used probes, especially in terms of sensitivity and for real-time imaging of living cells [20]. Experiments using this fluorescent probe confirmed the existence of b-glucosidase in the salivary glands, eggs, and hindgut of termites (Figure 4A). The primary function of b-glucosidase as a cellulose-digesting enzyme can explain its presence in salivary glands and the hindgut but cannot explain its localization in eggs. We did not detect b-glucosidase in the eggs of Argentine ants (Linepithema humile) used as a negative control (Figure 4A), indicating that it is not a common developmental necessity for insect eggs. Thus, the production of b-glucosidase in eggs in termite queen ovaries strongly suggests its secondary function as an egg-recognition pheromone. As predicted, b-glucosidase standard showed strong eggrecognition pheromone activity (Figure 3C). Perfect pheromone activity was achieved with a mixture of b-glucosidase and lysozyme standards, whereas even b-glucosidase alone showed relatively high pheromone activity. No other common insect salivary enzyme [21], including endo-b-1,3-glucanase (EC 3.2.1.6), a-amylase (EC 3.2.1.1), a-glucosidase (EC 3.2.1.20), and a-galactosidase (EC 3.2.1.22), showed significant termite
egg-recognition pheromone activity compared with the negative control (p > 0.05, Tukey’s HSD test). We found that the wood-feeding cockroach had both b-glucosidase (Figure 3B) and lysozyme (Figure 2A) in its salivary glands, which explains why the egg-recognition pheromone activity of wood-feeding cockroach saliva was as high as that of termite saliva (Figure 3A). In addition, saliva of cockroach larvae showed higher b-glucosidase activity than that of adult cockroaches (Figure 3B), which was consistent with the observation that larval saliva had higher pheromone activity than adult saliva (Figure 3A). We conclude that the termite egg-recognition pheromone consists of b-glucosidase and lysozyme, although we do not completely deny the possibility of other minor chemical compounds involved in termite egg recognition. To determine whether the termite ball fungi mimic termite eggs by producing b-glucosidase, we examined b-glucosidase production by the fungus and tested its function in chemical mimicry. Production of b-glucosidase, which is a common fungal enzyme, has been reported in various fungi including the sclerotium-forming fungus Sclerotium rolfsii [22], which is closely related to the termite ball fungi. We found that the termite ball fungus produces b-glucosidase conditionally; the fungus produced only a marginally detectable amount of b-glucosidase when cultured on plain potato dextrose agar (PDA), whereas addition of termite nest material to PDA markedly increased b-glucosidase production by the fungus (Figure 4B). The fluorescent probe confirmed the existence of b-glucosidase on the surface of termite balls produced on PDA with nest material (Figure 4A). There were no morphological differences between termite balls produced on PDA with
Current Biology Vol 19 No 1 34
and without nest material. This conditional enzyme production seems likely because b-glucosidase production is inhibited in the absence of the substrate of the enzyme but with sufficient amounts of other nutritional resources. In fact, addition of 1% cellobiose, the substrate of b-glucosidase, also promoted b-glucosidase production. Conditional production of b-glucosidase provided the ideal opportunity to examine whether this enzyme can explain chemical egg mimicry by the fungus. The extracts of termite balls showed egg recognition activity only when they contained b-glucosidase (Figure 4B). The conditional acceptance of termite balls by termites in accordance with b-glucosidase production by the fungus is compelling evidence that the termite ball fungi mimic termite eggs by producing this enzyme. Our results were parsimonious with the evolution of chemical signals involved in within-colony social communications and in interspecific associations. The primary functions of the pheromone components, lysozyme and b-glucosidase, are as an antibacterial defense agent and digestion enzyme, respectively. Termites have evolved to use the pre-existing chemicals on the egg surface as an egg-recognition signal, without the evolution of any additional specific chemical for this purpose. b-glucosidase in termite eggs might still have any practical function (e.g., egg shell formation) because b-glucosidase is an essential enzyme for oothecal formation in other dictyopteran insects including mantis [23]. Secondary use of chemical compounds that have evolved for other primary functions can be seen in various social insects [24, 25]. It is notable that both lysozyme and b-glucosidase are enzymes that hydrolyze b-1,4 glycosidic linkages. In addition, termites cannot distinguish the enzymes from different origins based on differences in amino acid sequence. For example, termite lysozyme shows approximately 40% amino acid similarity with hen egg lysozyme [15]. Even a mixture of hen egg lysozyme and almond b-glucosidase showed perfect pheromone activity. We confirmed that heat inactivation of enzymes did not significantly affect their pheromone activity, suggesting that the pheromones are not the products of digestion but the enzymes themselves (Figure S3). However, our results do not rule out the possibility that other supporting compounds are involved in termite egg recognition. In the bioassay, dummy eggs were coated with the b-glucosidase standard at the concentration of 1 mg/bead (7.6 mU/bead), whereas the amount of b-glucosidase detected in each termite egg was 0.013 mU/egg. Future quantitative approaches are needed to understand the relative function and the recognition specificity of each compound. When considering the evolutionary process of chemical mimicry, it is important to note that the fungus originally had the potential to produce the pheromone substance for use as an enzyme for cellulose digestion. This would have served as an important preadaptation for the evolution of termite-egg mimicry by the fungus. Social insect colonies have evolved diverse defensive strategies to prevent parasite invasion [25]. However, once a way to overcome these defenses has been established, it can be an ideal environment for parasites providing a stable competitor-free habitat. Recently, another egg-mimicking fungus Trechispora sp. (Basidiomycota, Agaricomycotina) was found in the higher termite Nasutitermes takasagoensis, indicating that egg mimicry has evolved at least twice independently (K.M. and T.Y., unpublished data). Egg mimicry, by which the fungus can easily gain access to the center of the nest, seems to be an evolutionary loophole around antiparasite defense in termites. Striking similarities can be
seen in the parasitic interaction between Maculinea butterfly and Myrmica ant, where chemical larval mimicry plays a role as a signal to be picked up outside the nest and actively brought into the colony by ants [26]. Recent cost-benefit analyses of aphid-ant [27] and lycaenid-ant [28] interactions have demonstrated that even apparently peaceful mutualisms involve many potential conflicts. In the Lasius ant-Stomaphis aphid mutualism, ant workers bring aphid eggs into the nest and groom the eggs to protect them against pathogenic fungi [29]. Even in such a mutualistic interaction, egg tending by ants may be the result of the aphids manipulating the ants. To understand the evolution of chemical mimicry for social parasitism, we should carefully consider the net outcome of the interaction not only in the short term but also in the long term, frequency of the interaction in the host populations, and preadaptation of parasites, as well as the mechanism of recognition. Experimental Procedures Egg-Recognition Bioassay The workers, queens, and eggs used here were collected from R. speratus colonies in Okayama, Japan. To examine egg-recognition pheromone activity of extracted chemicals, we used dummy eggs made of glass beads coated with each test chemical as described previously [15]. Extraction of crude egg-recognition pheromone from the salivary glands of workers and eggs were carried out by a previously published method [15]. Dummy eggs consisting of 0.5 mm glass beads coated with 1.0 mg/bead of each test chemical were used in egg recognition bioassays. We randomly arranged ten eggs and 20 dummy eggs on moist unwoven cloth in a Petri dish, and ten workers were released in the dish. Acceptance rates were determined by counting the number of dummy eggs carried into egg piles. Eggs used for chemical extraction and enzyme detection were obtained from the ovaries of queens to avoid contamination by enzymes from worker saliva. The egg-recognition bioassay was replicated four times for each of the two R. speratus colonies unless otherwise stated. Piling rates were arcsine-square-root-transformed and analyzed by two-way ANOVA followed by Tukey’s HSD test for multiple comparisons if ANOVA indicated a significant difference. Termite species for the study of cross-species pheromone activities were collected in Okayama (C. formosanus and C. brevis), Yamaguchi (R. kanmonensis), and Amami Island (R. miyatakei and H. sjostedti) in Japan. Colonies of the wood-feeding cockroach Panesthia angustipennis spadica were collected in Takahashi, Okayama, Japan. We also examined pheromone activity of following enzyme standards: lysozyme (EC 3.2.1.17, from hen egg, #L7651, Sigma-Aldrich), b-glucosidase (EC 3.2.1.21, from almond, #G0395, Sigma-Aldrich), endo-b-1,4-glucanase (EC 3.2.1.4, from Trichoderma viride, #C1794, Sigma-Aldrich), endo-b-1,3-glucanase (EC 3.2.1.6, from Aspergillus niger, #49101, BioChemica), a-amylase (EC 3.2.1.1, from Bacillus sp. #A6380, Sigma-Aldrich), a-glucosidase (EC 3.2.1.20, from Bacillus stearothermophilus, #G3651, Sigma-Aldrich), and a-galactosidase (EC 3.2.1.22, from Mortierella vinacea, #100560, Hokkaido Sugar Company). Purified termite lysozyme expressed in the B. mori-baculovirus system and enzyme standards were dialyzed (MWCO 3500) against distilled water to remove buffer salts before use to avoid interference with egg recognition [15]. Lysozyme Assay Termite balls harvested from potato dextrose agar plates were homogenized by grinding with a mortar and pestle in 0.1 M phosphate buffer (pH 6.0) on ice and then centrifuged at 20,000 3 g for 10 min at 4 C. The supernatant was used as the enzyme extract. The lytic activity of enzyme extract was measured against Micrococcus lysodeikticus (Sigma-Aldrich) suspended in the phosphate buffer. Lysis of M. lysodeikticus was measured at 450 nm with a spectrophotometer (SmartSpec Plus, BioRad). One unit of the activity was defined as the amount of enzyme decreasing the absorbance at a rate of O.D. 0.01/min. Phosphate buffer (pH 3.6) was used for the salivary gland extracts of the wood-feeding cockroach instead of pH 6.0 because the optimum pH of wood-feeding cockroach lysozyme was 3.6. Extracts of salivary glands of termite workers and wood-feeding cockroach larvae were used as positive controls. Five individual equivalent and 1/3 individual equivalent of salivary gland extract were used in each assay for termites and wood-feeding cockroach, respectively. Only extraction
Chemical Mimicry of Termite Eggs by Fungus 35
buffer was added to the reaction mixture in the negative control. Assay was replicated three times for each treatment. Expression of Termite Lysozyme in Silkworm-Baculovirus System To obtain sufficient amounts of termite lysozyme for the egg-recognition bioassay, we used the Bombyx mori-baculovirus expression system. Termite lysozyme cDNA (Rs-Lys2) [30] was cloned and ligated into the baculovirus transfer vector pV23. Total RNA was extracted from the salivary glands of 50 worker termites with NucleoSpin RNA II (Macherey-Nagel), and one-step RT-PCR was performed with a kit (QIAGEN) and the primers 50 -CTCGAGATGGACGTGAGAAATTCTCCGAT-30 and 50 -TCTAGAACGAAA ACTGGAATTACGACCACA-30 according to the manufacturer’s protocol. The amplified PCR product encoding Rs-Lys2 with the putative native signal peptide was subcloned with a TOPO TA cloning kit (Invitrogen). The plasmid containing the Rs-Lys2 nucleotide sequence was propagated with a Quantum Prep Plasmid Midiprep kit (BioRad). The Rs-Lys2 fragment was then cut out with Xhol and XbaI (Takara Bio) and ligated into the same sites of the transfer vector (pV23, Katakura Industries), yielding pV23-Rs-Lys2 (Figure 2C). Rs-Lys2 cDNA was fused in-frame to the C-terminal FLAG tag sequence (Sigma-Aldrich). Recombinant Rs-Lys2 was produced by KaikoExpress (Katakura Industries, http://www.katakura.co.jp/research/ ke-top.htm) and purified with Anti-FLAG M2 Agarose Affinity Gel column (Sigma-Aldrich; see Supplemental Experimental Procedures). RT-PCR for Endo-b-1,4-glucanase Gene Expression Analysis Total RNA was extracted from eggs dissected from two queens, legs of the queens, hind gut of a queen, salivary glands of 100 workers, and 20 eggs collected from the nest, with the NucleoSpin RNA II (Macherey-Nagel). We performed one-step RT-PCR to detect expression of an endogenous cellulase in termite, using a kit from QIAGEN according to the manufacturer’s protocol, with the following primers [19]: Cell-1 ORF-F 50 -GCCATGAAGGTC TTCGTTTG-30 and Cell-1 ORF-R 50 -GTTACACGCCAGCCTTGA-30 . To verify the integrity of the mRNA used in this experiment, samples were amplified by PCR with the housekeeping gene b-actin primers; forward 50 -AGAGGGA AATCGTGCGTGAC-30 ; reverse 50 -CAATAGTGATGACCTGGCCGT-30 . Detection of b-Glucosidase b-Glucosidase activity was assayed with p-nitrophenyl-b-D-glucoside (pNPbG) by standard methods [31]. Samples of salivary glands, eggs, and termite balls were homogenized in sodium acetate buffer (200 ml) on ice by grinding with a mortar and pestle and then centrifuged at 20,000 3 g for 10 min at 4 C. The supernatant was used as the enzyme extract. b-Glucosidase was assayed with a mixture (1 ml) containing 4 mM pNPbG, 50 mM sodium acetate buffer (pH 5.0), and 50 ml of enzyme solution. After incubation at 50 C for 30 min, the reaction was stopped by adding 1 ml of ice-cold 0.5 M Na2CO3, and the absorbance at 405 nm was recorded with a spectrophotometer (SmartSpec Plus, BioRad). One unit of b-glucosidase was defined as the release of 1 mmol of p-nitrophenol min21 in the reaction mixture. Data were standardized as units per milligram (lyophilized dry weight) of each extract. To determine the localization of b-glucosidase, we also used the enzyme activity-based fluorescent probe TokyoGreen-bGlu [20] (Sekisui Medical). b-Glucosidase hydrolyses nonfluorescent TokyoGreen-bGlu into highly fluorescent TokyoGreen (Ex./Em. = 490/510 nm). For fluorescence imaging, TokyoGreen-bGlu was used at a concentration of 10 mM in 10 mM phosphate buffer (pH 7.0). Salivary glands and eggs were dissected from termites in phosphate buffer and then soaked in TokyoGreen-bGlu mixture for 30 min at 27 C. We added 0.4% Triton X-100 in buffer for insect eggs. Eggs of the Argentine ant were tested as a negative control. Fluorescence images were captured with a fluorescence microscope (BZ8000, Keyence). Fungal Culture The three strains of termite ball fungus used in this experiment were isolated from the egg piles of R. speratus in Okayama, Japan. Termite balls extracted from egg piles were inoculated onto PDA in Petri dishes 90 mm in diameter and incubated at 27 C for 20 days. One newly developed sclerotium was reisolated from each termite ball and cultured on a new PDA plate. Two mycelial plugs 5 mm in diameter were taken from the 1-month-old colony of each isolate and placed on either a new plate of plain PDA or on a plate of PDA containing termite nest material and incubated at 27 C for 20 days. The nest materials consisting of nest structure and nest wood were obtained from a nest of R. speratus in red pine wood, dried at room temperature, milled (MX-X41; National), and added to PDA at a concentration of 15% (w/v). Termite balls were harvested from each plate and used in enzyme assays and pheromone assays.
Supplemental Data Supplemental Data include Supplemental Experimental Procedures, three figures and can be found with this article online at http://www. current-biology.com/supplemental/S0960-9822(08)01546-7. Acknowledgments We thank N.E. Pierce, S.-P. Quek (Harvard University), T. Nishida, C. Tanaka, K. Fujisaki (Kyoto University, Japan), T. Miyatake, Y. Murata, and F. Nakasuji (Okayama University, Japan) for their helpful advice and discussions. This work was funded by the Japan Society for the Promotion of Science and the Program for Promotion of Basic Research Activities for Innovative Biosciences (PROBRAIN). Received: September 13, 2008 Revised: November 5, 2008 Accepted: November 6, 2008 Published online: December 24, 2008 References 1. Vega, F.E., and Blackwell, M. (2005). Insect-Fungal Associations: Ecology and Evolution (New York: Oxford University Press). 2. Matsuura, K., Tanaka, C., and Nishida, T. (2000). Symbiosis of a termite and a sclerotium-forming fungus: Sclerotia mimic termite eggs. Ecol. Res. 15, 405–414. 3. Matsuura, K. (2006). Termite-egg mimicry by a sclerotium-forming fungus. Proc. R. Soc. Lond. B. Biol. Sci. 273, 1203–1209. 4. Cleveland, L.R. (1923). Symbiosis between termites and their intestinal protozoa. Proc. Natl. Acad. Sci. USA 9, 424–428. 5. Watanabe, H., Noda, H., Tokuda, G., and Lo, N. (1998). A cellulase gene of termite origin. Nature 394, 330–331. 6. Matsuura, K. (2003). Symbionts affecting termite behavior. In Insect Symbiosis, K. Bourtzis and T.A. Miller, eds. (Boca Raton: CRC Press), pp. 131–143. 7. Matsuura, K. (2001). Nestmate recognition mediated by intestinal bacteria in a termite, Reticulitermes speratus. Oikos 92, 20–26. 8. Matsuura, K., Fujimoto, M., Goka, K., and Nishida, T. (2002). Cooperative colony foundation by termite female pairs: Altruism for survivorship in incipient colonies. Anim. Behav. 64, 167–173. 9. Rosengaus, R.B., Guldin, M.R., and Traniello, J.F.A. (1998). Inhibitory effect of termite fecal pellets on fungal spore germination. J. Chem. Ecol. 24, 1697–1706. 10. Rosengaus, R.B., Jordan, C., Lefebvre, M.L., and Traniello, J.F.A. (1999). Pathogen alarm behavior in a termite: A new form of communication in social insects. Naturwissenschaften 86, 544–548. 11. Howard, R.W., McDaniel, C.A., and Blomquist, G.J. (1980). Chemical mimicry as an integrating mechanism: Cuticular hydrocarbons of a termitophile and its host. Science 210, 431–433. 12. Meer, R.K., and Wojcik, D.P. (1982). Chemical mimicry in the myrmecophilous beetle Myrmecaphodius excavaticollis. Science 218, 806–808. 13. Matsuura, K. (2005). Distribution of termite egg-mimicking fungi (‘‘termite balls’’) in Reticulitermes spp. (Isoptera: Rhinotermitidae) nests in Japan and the United States. Appl. Entomol. Zool. (Jpn.) 40, 53–61. 14. Yashiro, T., and Matsuura, K. (2007). Distribution and phylogenetic analysis of termite egg-mimicking fungi ‘‘termite balls’’ in Reticulitermes termites. Ann. Entomol. Soc. Am. 100, 532–538. 15. Matsuura, K., Tamura, T., Kobayashi, N., Yashiro, T., and Tatsumi, S. (2007). The antibacterial protein lysozyme identified as the termite egg recognition pheromone. PLoS ONE 2, e813. 10.1371/journal. pone.0000813. 16. Kambhampati, S. (1995). A phylogeny of cockroaches and related insects based on DNA sequence of mitochondrial ribosomal RNA genes. Proc. Natl. Acad. Sci. USA 92, 2017–2020. 17. Lo, N., Tokuda, G., Watanabe, H., Rose, H., Slaytor, M., Maekawa, K., Bandi, C., and Noda, H. (2000). Evidence from multiple gene sequences indicates that termites evolved from wood-feeding cockroaches. Curr. Biol. 10, 801–804. 18. Scrivener, A.M., and Slaytor, M. (1994). Properties of the endogenous cellulase from Panesthia cribrata Saussure and purification of major endo-beta-1,4-glucanase components. Insect Biochem. Mol. Biol. 24, 223–231.
Current Biology Vol 19 No 1 36
19. Zhou, X., Smith, J.A., Koehler, P.G., Oi, F.M., Bennett, G.W., and Scharf, M.E. (2007). Correlation of cellulase gene expression and cellulolytic activity throughout the gut of the termite Reticulitermes flavipes. Gene 395, 29–39. 20. Urano, Y., Kamiya, M., Kanda, K., Ueno, T., Hirose, K., and Nagano, T. (2005). Evolution of fluorescein as a platform for finely tunable fluorescence probes. J. Am. Chem. Soc. 127, 4888–4894. 21. House, C.R., and Ginsborg, B.L. (1985). Salivary gland. In Comprehensive Insect Physiology, Biochemistry and Pharmacology, Vol. 11, G.A. Kerkut and L.I. Gilbert, eds. (Oxford: Pergamon Press), pp. 195–224. 22. Chinnathambi, S., and Lachke, A. (1997). Activities of glycosidases from Sclerotium rolfsii in non-aqueous media. Biotechnol. Tech. 11, 589–591. 23. Yago, M., Sato, H., Oshima, S., and Kawasaki, H. (1990). Enzymatic activities involved in the oothecal sclerotization of the praying mantid, Tenodera aridifolia sinensis Saussure. Insect Biochem. 20, 745–750. 24. Turillazzi, S., Dapporto, L., Pansolli, C., Boulay, R., Dani, F.R., Moneti, G., and Pieraccini, G. (2006). Habitually used hibernation sites of paper wasps are marked with venom and cuticular peptides. Curr. Biol. 16, R530–R531. 25. Cremer, S., Armitage, S.A.O., and Schmid-Hempel, P. (2007). Social immunity. Curr. Biol. 17, R693–R702. 26. Nash, D.R., Als, T.D., Maile, R., Jones, G.R., and Boomsma, J.J. (2008). A mosaic of chemical coevolution in a large blue butterfly. Science 319, 88–90. 27. Stadler, B., Kindlmann, P., Smilauer, P., and Fiedler, K. (2003). A comparative analysis of morphological and ecological characters of European aphids and lycaenids in relation to ant attendance. Oecologia 135, 422–430. 28. Pierce, N.E., Braby, M.F., Heath, A., Lohman, D.J., Mathew, J., Rand, D.B., and Travassos, M.A. (2002). The ecology and evolution of ant association in the Lycaenidae (Lepidoptera). Annu. Rev. Entomol. 47, 733–771. 29. Matsuura, K., and Yashiro, T. (2006). Aphid-egg protection by ants: a novel aspect of the mutualism between the tree-feeding Stomaphis hirukawai and its attendant ant Lasius productus. Naturwissenschaften 93, 506–510. 30. Fujita, A., Minamoto, T., Shimizu, I., and Abe, T. (2002). Molecular cloning of lysozyme-encoding cDNAs expressed in the salivary gland of a wood-feeding termite, Reticulitermes speratus. Insect Biochem. Mol. Biol. 32, 1615–1624. 31. Saha, B.C., and Bothast, R.J. (1996). Production, purification, and characterization of a highly glucose-tolerant novel beta-glucosidase from Candida peltata. Appl. Environ. Microbiol. 62, 3165–3170.