A Note: Gut bacteria produce components of a locust cohesion pheromone

Authors


R.J. Dillon Department of Biology and Biochemistry, University of Bath, Bath, BA2 7AY, UK (e-mail: r.j.dillon@bath.ac.uk).

Abstract

Aims: Faecal pellets from germ-free locusts were used as culture media to determine the ability of locust gut bacteria to synthesize phenolic components of the locust cohesion pheromone.

Methods and Results: Inoculation of germ-free faecal pellets with Pantoea agglomerans, a species commonly isolated from locusts, resulted in the release of large amounts of guaiacol and small amounts of phenol, both of which are components of the locust cohesion pheromone. Two other locust-derived species, Klebsiella pneumoniae pneumoniae and Enterobacter cloacae, also produced guaiacol from germ-free faecal pellets, but the opportunistic locust pathogen, Serratia marcescens, did not. The most likely precursor for guaiacol is the plant-derived vanillic acid, which is present in large amounts in the faeces of both conventional and germ-free locusts.

Conclusions: These observations are consistent with previous ones, that locust gut bacteria are responsible for the production of components of the locust cohesion pheromone.

Significance and Impact of the Study: These findings illustrate how an insect can adapt to make use of a common bacterial metabolite produced by one or more of its indigenous gut bacterial species. This observation has implications for our appreciation of insect gut microbiota interactions.

INTRODUCTION

The defining feature of the pest status of desert locusts is their ability to change from a grasshopper-like solitary insect to a gregarious form that seeks out the company of other locusts. Swarms of gregarious locusts containing many millions of individuals may travel long distances in search of food (Singh and Singh 1977). Communication between locusts drives the process that causes solitary populations to turn gregarious and is thought to involve an interplay of visual, tactile and chemical stimuli (Byers 1991; Pener and Yerushalmi 1998). Aromatic compounds released from locusts and their faecal pellets maintain aggregation of the insect group, one aspect of the gregarization process (Fuzeau-Braesch et al. 1988; Obeng-Ofori et al. 1994; Schmidt 1997). The phenolic compounds guaiacol and phenol are the predominant electrophysiologically-active components released from juvenile and adult faecal pellets (Obeng-Ofori et al. 1994). Until recently, the origin of the pheromonal compounds guaiacol and phenol was unknown.

An investigation of the origin of the phenolic pheromone compounds revealed that the locust had adapted to use metabolites produced by gut bacteria acquired serendipitously with the food (Dillon et al. 2000). The possible involvement of the locust gut microbiota in the production of the locust aggregation pheromone was investigated by rearing locusts from surface-sterilized eggs in a sterile isolator system, and establishing a breeding colony of axenic locusts using a diet of γ-irradiated freeze-dried grass and bran. Faecal pellet volatiles from open conventional adult and juvenile locusts had similar profiles to those found by Obeng-Ofori et al. (1994), and included the pheromone components guaiacol and phenol. However, guaiacol was absent and phenol present at a reduced level in faecal pellets from germ-free insects.

The introduction and establishment of the bacterium Pantoea agglomerans in the gut of axenic locusts resulted in the re-appearance of the two phenolics in the faeces.

In the present study, faecal pellets from germ-free locusts were used as a culture medium to determine the extent of the guaiacol synthetic capability among bacteria associated with locusts.

MATERIALS AND METHODS

Production of germ-free locusts

A breeding colony of bacteria-free locusts, initiated from surface-sterilized eggs, was maintained in flexible plastic isolators on irradiated freeze-dried grass and bran with vitamin supplement (Charnley et al. 1985). The bacteria-free status of the insects was checked by a combination of aerobic and anaerobic growth assays, direct microscopy and scanning electron microscopy (Charnley et al. 1985).

Bacteria and growth media

Pantoea agglomerans (Sga40) was isolated from an adult Schistocerca gregaria from Addis Ababa, Ethiopia.Klebsiella pneumoniae pneumoniae (Sg16), Enterobacter cloacae (Sg8) and Enterococcus casseliflavus (Sgs32) were isolated from the gut of adult S. gregaria taken from our conventional colony. Bacteria were previously identified using standard bacteriological methods and the API system (bioMerieux UK Ltd, Basingstoke, Hampshire, UK). A strain of Serratia marcescens (ATCC 13880), pathogenic to locusts, was also used in the experiments. Faecal pellet medium was produced using dried pellets collected from cages containing bacteria-free locusts. The media contained 30 g faecal pellets mixed with 50 ml sterile distilled water in a 250 ml conical screw-cap flask. Locust diet medium consisted of 15 g irradiated bran and 15 g irradiated freeze-dried grass in 50 ml sterile distilled water. Vanillic acid broth consisted of glucose (10 g l−1), bacteriological peptone (5 g l−1), yeast extract (5 g l−1), NaCl (3 g l−1) and 1·5 mmol l−1 vanillic acid. Vanillic acid was added after the broth was autoclaved. All media were inoculated with 1 × 108 bacteria from an 18 h culture and incubated for 6 days at 27°C. Bacterial growth in faecal pellet cultures was estimated by serial dilution of a homogenate of the culture on nutrient agar.

Volatile collection and analysis

The collection method was developed from that of Obeng-Ofori et al. (1994). Volatiles released from microbial cultures were collected directly from the headspace above the cultures in flasks via glass tubing passed through a rubber bung. Air (50 ml min−1) was drawn through the tube via a graphitized carbon adsorbent (100 mg Carbotrap ORBO tube, Supelco UK, Poole, Dorset, UK). Incoming air was pre-filtered via an identical carbon filter. For the frass rehydration experiment, 1 g pellets was placed into a 300 × 15 mm glass tube and air (50 ml min−1) was drawn through the tube via a graphitized carbon adsorbent. Volatiles were collected for 24 h at 27°C, eluted with HPLC grade dichloromethane and concentrated under nitrogen gas to 25 μl. An internal standard, 4-methylacetophenone (300 ng), was added prior to the elution and concentration process. Samples (1 μl) were initially analysed on a Pye Unicam gas chromatograph, fitted with a BP20 (25 m × 0·25 mm, SGE Ltd, Milton Keynes, UK) capillary column, programmed from 70 to 220°C at 6°C min−1, held at 220°C for a further 10 min and the peaks integrated with a Trio II (TriVector Systems Int. Ltd, Bedfordshire, UK); the system was calibrated with authentic standards. Confirmation of peak identity was by analysis on a Carlo Erba GC (50 m × 0·32 mm Superox 2 column, Alltech Ltd, Lancashire, UK) linked to a Kratos MS80RFA mass spectrometer (Kratos Analytical Ltd, Manchester, UK). Estimates of the amounts of volatile compound released were adjusted according to the recovery of the internal standard. All experiments were repeated and representative results are provided.

Decarboxylation of vanillate

Pantoea agglomerans was grown overnight in vanillic acid broth containing 1·5 mmol l−1 vanillic acid. Cells were centrifuged at 12 000 g for 10 min, washed in 10 mmol l−1 MES-buffered (pH 6·5) saline, resuspended in 2 mmol l−1 vanillic acid solution in buffered saline and incubated at 27°C. Aliquots of 0·2 ml were removed during the incubation and diluted five times with 0·3% (w/v) SDS to stop the reaction; bacteria were removed by centrifugation and the solution was diluted 10 times with buffer. Absorbance of the solution was measured at 274 and 255 nm. Background absorbance was prepared at each time interval using bacteria incubated in saline without vanillic acid. Concentrations of guaiacol and vanillic acid in the solutions were calculated using the absorption coefficients at 274 and 255 nm obtained for standard solutions of the two compounds.

RESULTS

A previous study detected guaiacol and phenol in the volatiles from faecal pellets of 5th instar, immature and mature adults mono-associated with the bacterium Pantoea (=Enterobacter) agglomerans (commonly isolated from locusts (Dillon and Charnley 1996)). Volatiles from germ-free faecal pellets were lacking in guaiacol and contained reduced amounts of phenol (Dillon et al. 2000). Inoculation of germ-free faecal pellets with P. agglomerans in vitro resulted in the production of large amounts of guaiacol and small amounts of phenol (Table 1). Two other species of the Enterobacteriaceae, Klebsiella pneumoniae pneumoniae and Enterobacter cloacae, which were isolated from the locust gut, also synthesized guaiacol from germ-free faecal pellets (Table 1). A strain of Serratia marcescens pathogenic to locusts produced no detectable guaiacol or phenol (Table 1). There are substantial numbers of enterococci in the locust gut (Hunt and Charnley 1981) but a representative species, Enterococcus casseliflavus, produced only very small amounts of guaiacol and no phenol (Table 1). The lack of volatile production was not due to the poor growth of the bacteria; in all cultures, the estimated number of bacteria was never less than 1 × 106 ml−1.

Table 1.   Volatile phenolic compounds released from bacterial cultures Thumbnail image of

The quantity of guaiacol released was dependent on the moisture content of the faecal pellets. Germ-free faecal pellets inoculated with P. agglomerans produced 359 μg−1 g−1 24 h−1 guaiacol and 8·9 μg−1 g−1 24 h−1 phenol. After drying, the same pellets emitted only 17·6 μg−1 g−1 24 h−1 guaiacol and no detectable phenol. Rehydration to the original moisture content resulted in an increase in guaiacol emission to 223·0 μg−1 g−1 24 h−1 and the reappearance of phenol at 7·6 μg−1 g−1 24 h−1. These results suggest a continual production of guaiacol by gut bacteria in frass post-evacuation, depending on water availability. Substantial release of guaiacol from hydrated faeces is an apt aggregation stimulus, as it signals optimal conditions (presence of water) for plant growth and egg development that are both essential to the survival of the desert locust.

The cohesion pheromone is a complex of chemicals that includes guaiacol and phenol (the dominant and most biological active constituents) (Pener and Yerushalmi 1998). It is possible that some of the other chemicals in the blend are produced by the gut microbiota. Developmental changes in the composition and size of the microbiota could thus account for life stage-dependent changes in blend composition.

Guaiacol production was dependent on the locusts' diet; considerably more guaiacol was present when conventional locusts were fed fresh wheat seedling than the freeze-dried γ-irradiated grass (normally fed to the germ-free locusts) (Dillon et al. 2000). However, incubation of bacteria with either diet resulted in the production of only minor amounts of guaiacol or phenol (Table 1). Thus, digestion/processing of the plant material in the locust gut appears to be a prerequisite for production of guaiacol by bacteria, unless the precursor for guaiacol synthesis is a by-product of locust metabolism. The latter seems unlikely since the most obvious precursor is the plant-derived vanillic acid (4-hydroxy-3-methoxybenzoic acid) that was detected previously in the faeces of both germ-free and conventional locusts (Dillon and Charnley 1988). Evidence of a role for plant-derived vanillic acid in guaiacol production was obtained by feeding conventional locusts with filter paper impregnated with a vanillic acid solution. The resulting faecal pellets yielded large amounts of guaiacol compared with the controls (Dillon et al. 2000).

The ability of a bacterial species to produce guaiacol from faecal pellets correlated with its ability to produce guaiacol from vanillic acid in broth culture. Three species of the Enterobacteriaceae, but not the pathogen S. marcescens (Table 1), produced guaiacol. The enterococcal species did not produce guaiacol under the same conditions. Microbial transformation of vanillic acid to guaiacol requires a simple decarboxylation step (Crawford and Olson 1978; Chow et al. 1999). A decline in vanillic acid concomitant with the production of guaiacol in short-term incubations (5 h) of P. agglomerans is consistent with this synthetic pathway (Fig. 1). Prior growth of the bacterium in vanillic acid was required for subsequent synthesis of guaiacol during the experiment, suggesting induction of the decarboxylase by the substrate.

Figure 1.

 Vanillic acid decarboxylation by Pantoea agglomerans. Vanillic acid (□) decarboxylation to guaiacol (•). Decarboxylation after pre-incubation with vanillic acid (), decarboxylation after pre-incubation without vanillic acid (---)

DISCUSSION

Microbial non-oxidative decarboxylation of aromatic acids has been reported many times and the gene cluster for Streptomyces vanillic acid decarboxylase has been characterized (Chow et al. 1999). Although the enzymes are of interest in the conversion of renewable resources for chemical production, the function of these decarboxylation products is unclear. The work presented here, and in previous studies, suggests adaptive functions for these enzymes in the life of the locust. Gut bacteria, in addition to their involvement in the production of the locust aggregation pheromone, also contribute to locust defence against microbial insect pathogens, in particular by producing antimicrobial phenols (Dillon and Charnley 1988,1995,1996). The most obvious synthetic route for one of these phenols, protocatechuic acid (3,4 dihydroxybenzoic acid), is demethylation of vanillic acid (Sutherland et al. 1981).

Insect gut microbiota often reside in the hindgut of plant-feeding insects. This site is low in nutrients but rich in refractory remnants of insect digestion, such as lignin and related aromatics. Thus, microbial metabolism of secondary plant chemicals in the guts of locusts and other insects could potentially produce many compounds of value to the host. It is suggested that the insect hindgut, where the faecal pellets are produced, is an environment which is selecting for micro-organisms capable of firstly, withstanding and secondly, catabolizing aromatic compounds. This work provides evidence for a moderately mutualistic association between the locust and its microbiota. The bacterial community of the locust gut is adapted to metabolize plant allelochemicals into antimicrobial compounds with increased activity against pathogens or allochthonous microbes and provision of pheromonal compounds. This dual benefit for the insect suggests a closer degree of integration between the locust and its microbial community than was previously suspected. Surprisingly, this has not resulted in the development of an obligately mutualistic association. Instead, the locust has minimized the consequences of mutualist loss by not relying on a single microbial species. Locusts, in common with many other species of insects, have a variable gut biota comprising species commonly encountered in their environment, particularly the phylloplane biota on food plants (Hunt and Charnley 1981). Lignin-derived vanillic acid is widespread in plants and is found in the faeces of both germ-free and conventional locusts (Dillon and Charnley 1988). Guaiacol production by vanillic acid decarboxylation is an attribute of some plant and soil saprophytes (Crawford and Olson 1978; Chow et al. 1999), and it is suggested that locust faecal pellets will always contain guaiacol though the bacteria producing it may differ.

Previous studies which suggested microbial involvement in the production of insect aggregation pheromones (Nolte et al. 1970; Brand et al. 1976) have received serious criticism (Fuzeau-Braesch et al. 1988; Obeng-Ofori et al. 1994Pener and Yerushalmi 1998; Tillman et al. 1999). This paper and a previous study (Dillon et al. 2000) provide the most complete evidence to date for bacterial involvement in the synthesis of an insect aggregation pheromone.

ACKNOWLEDGEMENTS

The Biotechnology and Biological Sciences Research Council funded this work. The authors thank M. Lewis of Long Ashton Research Station, Bristol for the mass spectrometer analysis.

Ancillary