Introduction

Mate location is a challenging phase for sexually reproducing insect species and the mechanism can involve visual, auditory and/or olfactory cues. Olfactory cues in the form of pheromones are very common in the order Coleoptera (El-Sayed 2016). The ability of nocturnal insects to find mates at low population densities has puzzled observers, although the first scarab sex pheromone was identified in 1969–1970, from the New Zealand grass grub Costelytra zealandica (Osborne and Hoyt 1969; Henzell and Lowe 1970). Since then, many more scarab pheromones have been elucidated and identified, and more than 150 compounds have been reported with activity against at least 40 species in several subfamilies (El-Sayed 2016). Some of the isolated pheromones resemble phenolic compounds (Ruther et al., 2001), and phenol in combination with p-cresol is reported to be a female-released pheromone for Phyllophaga cuyabana (Zarbin et al. 2007). Phenol is also suggested as a tentative pheromone component for the cockchafer Melolontha melolontha (Ruther et al. 2002).

Hoyt et al. (1971) attributed the source of pheromone to a symbiotic bacterium present in the accessory or colleterial glands located on the sides of the terminal portion of the common oviduct, but the bacterium was never identified and no cultures were obtained. The accumulations of bacteria between the spines in the glands and their ultrastructure were illustrated by Stringer (Stringer 1988). Nevertheless, phenol has been tested for grass grub population monitoring and suppression by means of mass trapping or even mating disruption (Fenemore et al. 1972; Henzell and Kain 1972; East et al. 1982).

Phenol was first reported from marine Pseudomonas bacteria (Updegraff 1949) and later from Citrobacter intermedius by cleavage of tyrosine by a tyrosine lyase (Nagasawa et al. 1981). The bacterial-mediated pheromone biosynthesis in C. zealandica may follow the similar mechanism and pathway, as suggested by Leal (Leal 1998). It has been noted that microbial-induced sex attractants have been generally unexplored, but could prove significant in speciation (Brucker and Bordenstein 2012).

We aimed to (a) confirm that phenol is attractive to male grass grubs, and show that (b) phenol is released by females, (c) phenol is released by bacteria isolated from virgin females, (d) bacterial isolates from females attract males and finally (e) bacteria use tyrosine to make phenol. We also aimed to identify which type of bacteria produce the phenol.

Materials and methods

Harvest of colleterial glands

Adult grass grubs were collected during the flight season from field populations in the vicinity of Lincoln, New Zealand (October–November). Pre-emergent virgin females were separated and maintained in soil in the laboratory at ambient conditions for 7–10 days to allow them to develop to sexual maturity. The colleterial gland was separated and removed from the reproductive tract with fine scissors, following the descriptions of Stringer (Stringer 1988).

Isolation, identification and culture of bacterial isolates

Colleterial glands from 10 females were individually macerated in 100 μl phosphate-buffered saline (PBS) and streaked onto plates of Luria Bertani (LB) agar (Merck, Auckland, New Zealand) for growth overnight at 30 °C. Sub-colonies were then transferred to the Nutrient Broth (NB) (Merck, Auckland, New Zealand) without purification and cultured for 3 days at 30 °C to test for attraction to adult males in field test 1. In addition, purified bacteria were obtained by repeated streaking and selection from overnight cultures on LB agar. The single cell purified and isolated strain was then cultured in NB as described above for use in further field attraction to male grass grubs. Two numbered isolates were from broth grown from colonies isolated on LB agar with macerated insects, with each isolate from a different broth. The process was (1) glands macerated and streaked onto LB, (2) individual colonies inoculated into broth for use as attractants, (3) isolates taken from broth and transferred to the PIM agar to confirm it produced phenol, and (4) identification of isolates named 1187 and 1190 by sequencing.

Chemicals and preparation of lures

Phenol (99.5 % purity) was purchased from BDH Chemicals (Poole, England). 13C15N-labelled tyrosine (>98 % isotopic purity) was purchased from Silantes GmbH (München, Germany). The phenol resin lure was prepared by immersing 50 g of the resin (Amberlite IR-120H ion-exchange resin, Aldrich, Australia) in 500 ml of water and adding 5 g of phenol. After stirring for 1 h, the excess water was filtered off by vacuum suction and the impregnated resin was stored in an airtight container. The lure was prepared by agitating 1 g of the impregnated resin in 25 ml of water. Individual doses of a phenol resin suspension lure were prepared by dissolving 100 mg of phenol in 25 ml of water in a 70-ml plastic jar. Then, 1 g of resin was added and the mixture was thoroughly agitated and stored in airtight containers. Lures were either used fresh or aged under field conditions during late October–early November, and stored until needed in airtight containers.

Field testing method

Field attraction bioassays were conducted by placing samples in a vial at the centre of a 2-l water trap, a 20 × 20 × 8 cm plastic container with water to a depth of 1–3 cm. Traps were placed, before dusk, at 5 m spacing on a short rye grass/clover pasture near Lincoln, Canterbury, New Zealand (43.65 S 174.5 E) that had contained a population of approximately 200 grass grub larvae/m2 in the previous winter following a randomised block design at right angles to the prevailing wind. Adult beetles were collected and counted from each trap 1 h after sunset when the emergence flight had terminated. A representative sample of adults collected was later sexed in the laboratory (Given and Hoy 1952).

Field test 1

Traps were baited with seven different treatments, i.e. virgin females, five unpurified broth cultures (1–5), and a blank (control), and traps were operated nightly for three nights (10–12 November 2006). For the unpurified broth cultures, bacterial suspensions were applied as 1-ml aliquots to tissue paper held in a 75-ml plastic tube in the centre of each trap. The virgin female treatment consisted of five beetles in a 5-cm diameter tube containing soil, and the beetles were confined within the tube by light gauze held in place by a rubber band.

Field test 2

After purification of the crude bacterial cultures found attractive in field test 1, two bacterial isolates (#1187 and #1190, later sequenced) were further tested for attractiveness to male grass grub and compared with the positive control; (100 mg phenol dispensed as crystal in a polyethylene bag) (Unelius et al. 2008) and unbaited control (negative control) traps, using methods as described in field test 1. The mean release rate of phenol from this lure was gravimetrically determined as 2.74 mg/day (±SEM, 0.192) (Harper 2014, unpublished data). The test (n = 3 replicates per treatment) was repeated for two nights (20–21 November 2006), and the results were pooled across nights for analysis of treatment effects.

Bacterial identification by sequencing

Pure cultures of the bacteria strains were grown in nutrient broth (10 ml) at 30 °C for 18 h. Cultures were diluted 1:1 in 30 % glycerol and 1 ml sent to Macrogen Inc., Korea, in sealed 1.7-ml tubes, for DNA extraction, amplification and sequencing. Partial sequencing of the 16SrRNA region was carried out using primers 518F (5′-CCAGCAGCCGTAATACG-3′) and 800R (5′-TACCAGGGTATCTAATCC-3′). Sequences were assembled to form a contig (contiguous, overlapping sequence read) of 1794 bp. Sequences were then compared with a database of sequences to identify the taxa (National Center for Biotechnology Information nucleotide database (http://www.ncbi.nlm.nih.gov).

Headspace sampling of pure bacterial isolates and beetles

Male and female beetles in a group of 9–10 insects were housed separately in cleaned glass chambers (ca. 30 ml). Cleaned and humidified air was pumped into the chambers through an inlet and over the beetles at a rate of 100 ml/min from 17:00 to 22:00 h, which is the active period for mating flight. Volatiles emitted by the insects were adsorbed onto a Tenax-GR cartridge (~100 mg) fitted at the outlet end. All joints were connected by silicone tubing. Upon the completion of the 5-h sampling, the adsorbed volatiles in the Tenax-GR were eluted using 1 ml of absolute ethanol. A total of 11 and 9 replications were conducted for female and male beetles, respectively. Similar procedures were repeated for bacterial isolates 1187 and 1190 (n = 4 each).

During analysis, 100 ng octanol was added as an internal standard for quantification purposes. Prior to injection into gas chromatograph (GC), the samples were reduced to approximately 100 μl under a gentle argon flow. One microlitre aliquot was injected into GCMS for chemical analysis.

Gas chromatography mass spectroscopy (GCMS) was used for quantification using a Saturn 2200 GC/MS (Varian Inc., Walnut Creek, CA, USA). The GC column was a VF5-MS 30 m × 0.25 mm × 0.25 μm. The programme was 40 °C (hold 5 min) and increased to 100 °C at a rate of 5 °C/min and holding time of 5 min. The spectra were recorded at an ionisation voltage of 70 eV over a mass range m/z of 20 to 499. Compounds were identified by comparing their mass spectra with the authentic phenol standard and NIST MS library. Under the operating conditions above, the retention time of phenol was at 13.2 min.

Phenol biosynthesis

Bacterial strains isolated above were tested to determine the biochemical origin of phenol in experiment 1. Bacterial broth suspensions were tested by adding 40 mg tyrosine to 50 mg of meat peptone and 30 mg meat extract (Merck, Auckland, New Zealand) in 10 ml of water. The resulting broth was divided in 0.5-ml aliquots and these samples were sampled using solid phase micro-extraction (SPME) and analysed as above after 3 days of incubation No extra tyrosine was added to the control samples. In experiment 2, the tyrosine was replaced by 40 mg 13C15N-labelled tyrosine (Silantes GmbH, Germany), but otherwise similar sampling procedures were used to those above and the headspace products were analysed in the same manner.

Statistical analyses

Data obtained from field test 1 were characterised by descriptive statistics only, due to a lack of true replication. Data from field test 2 were pooled over two nights for each replicate and log transformed before one-way analysis of variance.

Results

Field test 1: testing of unpurified cultures

Males were attracted to three of the five bacterial cultures and to the virgin females on the three evenings of the test (Fig. 1), in similar numbers. The determination of phenol presence by HPLC aligned with these positive results. Low numbers of adults were caught in the blank traps and in the traps baited with bacterial cultures labelled 1 and 2, in which phenol could not be detected. Laboratory examination revealed that all trapped adults were males. A total of 264 beetles were trapped in this experiment, to bacterial isolates or females.

Fig. 1
figure 1

Attractiveness of five virgin female Costelytra zealandica and unpurified bacterial cultures to male Costelytra zealandica, caught in water traps in Canterbury, New Zealand. Error bars indicate one standard error. Qualitative presence of phenol was conducted by HPLC

Full size image

Field test 2: testing of purified bacterial isolates

After purification of the bacterial cultures found attractive in the field, the resulting bacterial isolate (isolates 1187 and 1190) tested positively for attractiveness to male grass grub beetles (Table 1). There was a significant difference between treatments (F 2,6 = 67.5, P < 0.001). There was higher attraction to the positive control (100 mg phenol) than to the bacterial isolates, which could be a function of cell density or other factors we were not able to control. A total of 1168 male beetles were trapped in this experiment.

Table 1 Effect of virgin females and purified bacterial cultures on attraction of male grass grub Costelytra zealandica to water traps in Canterbury, New Zealand, and phenol quantitation by GCMS from headspace sampling on beetles and purified bacterial isolates
Full size table

Biosynthesis of phenol

The broth containing the bacteria isolated from grass grubs that was found attractive in field test 2 was incubated (a) without additional tyrosine, (b) with extra tyrosine or (c) with 13C15N-labelled tyrosine added. In the broth containing no extra tyrosine, a small amount of phenol was produced, but when extra tyrosine was added considerably more phenol was produced (94 m/z), and in broths containing the labelled tyrosine, 13C-labelled phenol was produced (100 m/z) (Figs. 2 and 3). This provides clear and direct evidence that the phenol is produced by the bacteria and it was biosynthesized from tyrosine.

Fig. 2
figure 2

GC-chromatograms of volatile odours released from bacteria when unlabelled tyrosine (middle) or 13C-labelled tyrosine (bottom) was added. The top trace is the control

Full size image
Fig. 3
figure 3

Mass spectra fragmentation of unlabelled phenol (left) and 13C-labelled phenol (right) derived from volatile emission from bacteria grown in nutrient broths, to which were added labelled or unlabelled tyrosine, respectively

Full size image

Bacteria characteristic and identification

BLAST search analysis of consensus sequences for two bacterial strains in the National Center for Biotechnology Information database found that one strain had 99 % similarity to Morganella morganii (KF611894.1), and another strain had 99 % similarity to M. morganii subsp. Morganii (CP004345.1). The bacteria are gram-negative, non-spore forming, aerobic and facultatively anaerobic, have motile rods and capable of producing phenol from tyrosine.

Discussion

The unusual finding that a symbiotic bacterium of the female grass grub, C. zealandica, is responsible for mate attraction through phenol release was reported earlier (Hoyt et al. 1971), but it is not until this study that the bacteria responsible for the sex attraction in this species has been isolated, purified and identified. M. morganii (Enterobacteriaceae) is a widely distributed gram-negative bacterium commonly found in the environment and in the intestinal tract of humans, mammals and reptiles as normal flora, but can be an opportunistic human pathogen associated with nosocomial infections (McDermott and Mylotte 1984; O’Hara et al. 2000). Other phenol-producing gut bacteria associated with insects have been identified, such as Pantoea agglomerans that is known to produce guaiacol and phenol as the cohesion pheromone that changes the solitary to gregarious status of the desert locusts (Dillon and Charnley 2002; Dillon et al. 2002).

Phenol was isolated from liquid exudates from female grass grubs, and its role as a male attractant sex pheromone is well established through the use of the chemical in grass grub flight monitoring (Henzell and Lowe 1970; Unelius et al. 2008). Phenol is produced in certain bacteria (Updegraff 1949) so the production of phenol per se is not surprising. However, the adoption of phenol-producing bacteria as the source of mate attraction remains unique so far to the New Zealand grass grub. The greatest attraction of beetles occurred when the air temperature was ≥11 °C, evident as bacterial cultures were relatively unattractive at cooler temperatures.

Although there is no general chemical structure of insect pheromones, insects within the same order and family normally use the same substance class (El-Sayed 2016). The structure of the volatile and potentially toxic phenol molecule used for communication by the New Zealand grass grub is unique in its simplicity and, at first glance, is not related to other scarab pheromones. But a review (Leal 1998) makes the point that some scarab beetles may use amino acids as biosynthetic starting materials for their pheromones (Leal 1997; Leal 1998). This speculation was exemplified by the suggested possible production of phenol from tyrosine by bacterial deaminating lyases (Leal 1997). So although indole, phenol and methyl isoleucinate are not very structurally related compounds, an amino acid (i.e. tyrosine, tryptophan and isoleucinate, respectively) is quite a plausible biosynthetic precursor for all of them.

The role of the colleterial gland in scarabs is unclear, although in other insect orders similar glands provide the secretions for the ootheca or mucus for sticking the eggs to surfaces. However, in C. zealandica, this function is unnecessary as eggs are laid in the soil either singly or in loose groups with no need of an adhesive (Stringer 1988). The colleterial gland in female C. zealandica is filled with bacteria (Stringer 1988), and in mature virgin females, our results indicate that the bacteria attractive to males conform to M. morganii. Thus, the role of the accessory gland in the grass grub does not appear similar to that of most other insects. There is a case to suggest that the gland provides a matrix for these symbiotic bacteria. Bacterial mutualisms with insects are in fact very widespread (Brucker and Bordenstein 2012). Thus, colonisation of the colleterial gland by phenol-producing M. morganii appears to provide a mechanism for male attraction by release of phenol as a sex pheromone. That is not an entirely unknown phenomenon. For example, bark beetles process the resin they encounter while attacking trees, and microorganisms in their hindgut produce oxygenated monoterpenes as aggregation pheromone components (Leufven, Bergstrom et al. 1984).

Phenol is an antibacterial agent, formerly used as a disinfectant in carbolic soap. Our finding that the phenol is produced from tyrosine is therefore also interesting from a microbial-ecological perspective. Apart from providing energy to the bacteria, this raises the question of whether the sex pheromone could be a weapon in bacteria versus bacteria warfare? (Ruther, Reinecke et al. 2001). The potential of phenol toxicity to competing bacteria is intriguing, but was not addressed in this study. The presence of the unidentified non-phenol-producing bacteria could indicate some ability to compete with M. morganii and this warrants further investigation.