Introduction

Insect parasitoids that attack and develop within non-feeding host stages, such as eggs or pupae, have evolved the ability to exploit chemical cues related to the adult stage. This behavior can help to overcome the so-called reliability-detectability dilemma (infochemical detour) (Vet and Dicke, 1992). In the case of hymenopteran egg parasitoids, an extensive body of literature describes how parasitoids can find patches where eggs have recently been deposited or where this is soon to occur, chiefly by exploring the semiochemicals produced by host adults (reviewed by Fatouros et al., 2008; Colazza et al., 2010). Within the host’s habitat, volatile pheromones may be detected by female egg parasitoids to actively reach the mating sites of hosts (Lewis et al., 1982; Colazza et al., 1997; Boo and Yang, 2000; Conti et al., 2010) or to recognize mated female hosts ready to lay eggs. Such gravid females also may be used for phoretic transportation (Fatouros et al., 2008; Huigens et al., 2010). Contact kairomones left by mated female hosts allow egg parasitoids to locate recently laid eggs more efficiently (Colazza et al., 1999; Chabi-Olaye et al., 2001; Fatouros et al., 2007). Furthermore, plants can modify the synthesis of organic defensive compounds as a consequence of the insect footprints (Bown et al., 2002).

From a tri-trophic perspective, chemical substances deposited by many insects while walking on a plant surface can mediate both intra- and interspecific relationships (Klomp, 1981; Colazza et al., 1999; Witjes and Eltz, 2009). The physicochemical characteristics of plant surfaces should, therefore, strongly influence these interactions, as the chemistry of the epicuticular wax components may affect how exogenous substances are adsorbed and released (Müller and Riederer, 2005). For example, the larval parasitoid Cotesia marginiventris (Cresson) (Hymenoptera: Braconidae) responds differently to the chemical footprints of its host Spodoptera frugiperda (J.E. Smith) (Lepidoptera: Noctuidae), when footprints that mainly consist of linear and monomethyl-branched alkanes are left on the leaf surface of either wild-type or eceriferum mutants of barley. The mutants were characterized by lower amounts of the main wax component hexacosanol and a high aldehyde fraction (Rostás et al., 2008; Rostás and Wölfling, 2009).

In recent years, the role of footprint cues deposited by true bugs (Hemiptera: Heteroptera) as an indirect host-related contact cue for scelionid egg parasitoids (Hymenoptera: Scelionidae) has been investigated (reviewed by Colazza et al., 2010). Extensive studies have been carried out in particular on the host-parasitoid association of Nezara viridula (L.) (Heteroptera: Pentatomidae) and Trissolcus basalis (Wollaston) (Hymenoptera: Scelionidae). In these experiments, the behavioral responses of the parasitoid were observed in arenas made of filter paper, which is an amorphous artificial substrate that does not interfere with parasitoid behaviors and/or the chemical properties of host cues (Colazza et al., 1999; Peri et al., 2006; Salerno et al., 2006). Under such conditions, T. basalis females show an initial motionless period of a few minutes (arrestment response) followed by locomotion and increased antennal movements while searching within the patch (motivated search). During the motivated search, wasp females adjust both turning rate (klinokinesis) and locomotion rate (orthokinesis) depending on the host’s sex. The presence of contact kairomones eventually leads to a longer and more intensive exploration in an area contaminated by the host adult females, as the cues associated with this host sex are spatially and temporally closer to the host eggs (Colazza et al., 1999).

Under natural conditions, the interactions between N. viridula and T. basalis are influenced by the presence of plant surface waxes. Previous experiments demonstrated that the wasps responded to footprints from female N. viridula walking on Vicia faba L. plants only if the leaf’s wax layer was intact, but were unable to detect host kairomones when the bugs have walked over mechanically dewaxed leaves (Colazza et al., 2009). Clearly, the physicochemical characteristics of the leaf surface can hamper the detectability of host footprints by wasp females, as active compounds could be part of the components that make up the plant’s epicuticular wax surface. Recent observations of T. basalis behavior using filter paper arenas demonstrated that the linear alkane n-nonadecane (nC19) is an active component among the cuticular hydrocarbons of N. viridula males that allows wasp females to distinguish between residues left by male or female bugs (Colazza et al., 2007). As linear alkanes are a major class of compounds in plant epicuticular waxes (Eigenbrode and Espelie, 1995; Müller and Riederer, 2005)—although alkanes with less than 21 carbon atoms don’t seem to be widely present—possible masking effects by the alkanes present in the plant waxy background on N. viridula male-specific hydrocarbons need to be assessed.

In this study, we investigated the ability of T. basalis females to distinguish between the footprint cues left by male and female N. viridula on leaves of broccoli plants (Brassica oleracea L., var. Italica, cv. Marathon). Then, we evaluated the responses of female wasps to the male bug-specific hydrocarbon n-nonadecane. Finally, we studied the chemical composition of the broccoli leaf surface to evaluate its role in mediating the interactions between host footprints and female wasps and specifically to answer the question whether n-nonadecane was present or absent.

Methods and Materials

Plants

In all experiments, broccoli plants grown from certified seed material (Esasem Spa, Casaleone -VR- Italy) were used. In experiment 1, plants were cultivated in 14-cm plastic pots with fertilized commercial potting soil (Trflor, HOCHMOOR) in a climate-controlled greenhouse of the Department of DEMETRA, University of Palermo, Italy from 08/2008 to 12/2008. In experiments 2 and 3, plants were grown in 9-cm-pots containing soil mixture ED 73, and kept in a growth chamber with a day/night cycle of 16:8 hr, 20°C temperature, and 60% relative humidity at the Department of Botany II, University of Würzburg, Germany.

Insects

Colonies of N. viridula and T. basalis were established from material collected in the fields around Palermo, Italy, and were maintained in separate climate chambers at 26 ± 1°C, approximately 60% R.H., and a photoperiod of 16:8 hr (L:D). Immatures and adults of N. viridula were reared in wooden cages (50 × 30 × 35 cm with 5-cm circular mesh-covered holes) and fed with seasonal fresh vegetables, and sunflower seeds. Mated adults of known age were available for the bioassays by keeping pairs of N. viridula in different cages, and checking daily until mating was assured. Females and males then were separated and placed into different rearing cages. Egg masses of N. viridula were collected from the rearing cages on a daily basis and used to maintain both N. viridula and T. basalis colonies. Newly laid egg masses of N. viridula were individually exposed to 3–4 female wasps for about 3 d in 16-ml glass tubes. Then, parasitoids were removed and parasitized egg masses were held in the rearing room until adults emerged. Adults of T. basalis were fed with a 10% honey-water solution streaked on the inside surface of the glass tubes.

Experiment 1

The aim of this experiment was to determine whether female parasitoids were able to detect and discriminate the chemical footprints of male and female bugs on the surface of broccoli leaves. Four-wk-old plants with 5–6 fully expanded leaves were exposed to 2–3 N. viridula adults. The stink bugs were allowed to walk over the leaves and produce chemical footprints for about 30 min according to the method of Colazza et al. (2009). Plants not exposed to insects were used as controls. Male and female bugs were 8–10 d-old kept isolated for 3–4 d after mating. Trissolcus basalis females were 4 d-old, mated and naive with respect to both, oviposition experience and contact with host chemical traces. The day before the experiments were carried out, single wasps were isolated in 2-ml glass vials closed with a cotton plug. Parasitoids were fed with a drop of honey provided on the inside surface of the vials. Thirty to sixty min prior to the experiments, female wasps were transferred to the bioassay room to acclimatize. A single wasp was released in the center of an adaxial leaf surface by gently tapping on the vial. Wasps were scored either as “responding” if they immediately displayed the typical arrestment posture, i.e., remaining motionless with the antennae in contact with the leaf surface, or as “non-responding” where they did not show this posture after three trials. For responding females the time spent on the adaxial leaf surface was measured as “retention time” with a stopwatch. Each leaf was used for testing five wasp females. Thirty replicates were carried out for each treatment. All experiments were conducted at room temperature (20–25°C) and under natural daylight conditions between 9:00 a.m. and 2:00 p.m.

Experiment 2

In a series of bioassays we evaluated the effect of n-nonadecane (nC19), the compound specific for male bugs, on the discriminatory behavior of female parasitoids. Females of N. viridula were immobilized by cooling the insects down at −20°C for 5 min. Then, three tarsi on one side of the insect body were treated with hexane (nC6) or with 4 μl of the test alkanes (0.1% w/v in hexane), while the remaining three tarsi were left untreated. Tested alkanes were: n-nonadecane (nC19) (Fluka, 99.8% purity) to mimic footprints of male bugs, and n-eicosane (nC20) (Sigma-S. Louis, MO, USA, 99% purity) a saturated hydrocarbon that is not present in adult N. viridula cuticle but was used as a control for any masking or modifying effects on host footprint detection by T. basalis. The pure solvent was applied as a second control. Female bugs retained normal activity 10–15 min after treatment, and then they were left to walk over the adaxial surface of a leaf disk (5 cm diam) cut out from a fully expanded broccoli leaf for 30 min. Finally, a female parasitoid was released in the center of the leaf disk and its behavior was scored as described in the previous experiment. Each leaf disk was used for testing five wasps. Thirty replicates were carried out for each treatment. All experiments were conducted between 9:00 a.m. and 2:00 p.m. at room temperature (20–25°C) and under natural daylight conditions.

Experiment 3

The compounds that constitute the epicuticular wax layer of broccoli leaves were assessed, in particular, to answer the question whether n-nonadecane was present. Leaf surfaces were prepared by mechanically removing the outermost wax layer, using the cryo-adhesive based method described by Riedel et al. (2003). Disks were cut out from broccoli leaves taken from 4-wk-old plants with 5–6 fully expanded leaves. Leaf disks were placed on an aluminum holder. Then, cylinders (inner diam: 1.4 cm) were centered on the leaf disks and filled with water, completely covering the surface of the disks. The whole apparatus was frozen in liquid nitrogen and the resulting ice blocks were taken off the leaf disks by lifting the cylinders, thereby stripping off the wax. The cylinder walls were rapidly warmed up thus allowing the ice blocks to slip out. The wax-holding ice blocks were immediately collected in a vial containing chloroform (4 ml) and water (4 ml). Preparations from 3 to 5 leaf disks from the same leaf were pooled and n-tetracontane (nC40) was added as an internal standard. The two-phase system was heated to approximately 40°C in a water bath and agitated repeatedly. Then, the organic phase was removed and the aqueous phase was extracted × 3 with chloroform. All three chloroform fractions were combined. The wax extracts in chloroform were dried over Na2SO4 to eliminate all remaining water and were then filtered. Prior to gas chromatographic (GC) analysis, chloroform was evaporated from all samples under a gentle stream of N2 while heating the sample-vials to 50°C. Then, all samples were treated with bis-N,N-(trimethylsilyl)-trifluoroacetamide in pyridine (BSTFA, Supelco, Sigma-Aldrich) (30 min at 70°C) to transform all hydroxyl-containing compounds to their corresponding trimethylsilyl (TMSi) derivatives. The qualitative composition was analyzed using a capillary GC (6890 N, Agilent Technologies, Santa Clara, CA, USA; column: 30 m × 0.32 mm inner diam, Zebron ZB-1HT, df = 0.1 μm, Phenomenex, Torrance, CA, USA) and mass spectrometric detector (MSD 5973 N, Agilent Technologies, Santa Clara, CA, USA; 70 eV, m/z 50–650). Gas chromatography was carried out with cool-on-column injection at 50°C. Oven temperature was held at 50°C for 2 min, raised by 3°C min−1 to 320°C and held for 20 min at 320°C. Helium carrier gas inlet pressure was 50 kPa at injection, held for 5 min, raised by 3 kPa min−1 to 150 kPa and held at 150 kPa for 90 min. Wax components were identified by comparison of their mass spectra with those of authenticated standards and literature data. Quantification of wax compounds was accomplished by capillary GC (6850 N, Agilent Technologies, Santa Clara, CA, USA; same column as above) and flame ionization detection under the same gas chromatographic conditions as described above but with H2 as carrier gas. The wax composition of six individual plants was analyzed.

Statistical Analysis

The total numbers of “responding” and “non-responding” T. basalis females for each treatment were compared with Pearson’s χ 2 test. Goodman’s post hoc procedure was used for internal contrasts among different treatments. The wasp’s resident time on leaves and leaf disks, recorded only from responding individuals, led to unequal numbers of replicates. Therefore, these unbalanced data were analyzed using one-way ANOVA with type III sums of squares for log-transformed data. Comparisons among means were made using Fisher’s LSD post hoc tests. Statistical analyses were performed using XLSTAT add-in software for Excel.

Results

Experiment 1

Wasp responses to broccoli leaves were strongly affected by the presence or absence of host chemical footprints (χ 2(2)  = 32.76; P < 0.001; Pearson’s χ 2) (Fig. 1 a). On leaves contaminated with N. viridula footprints, about 75% of T. basalis females displayed the typical arrestment posture, while on non-contaminated leaves only ca. 13% responded. No statistical differences could be observed when comparing the responses of wasps towards male or female hosts. In contrast, the average leaf residence time of responding wasps was clearly influenced by the host’s sex (F (2,47) = 6.38, P = 0.002; ANOVA Type III Sum of Squares analysis) (Fig. 1 b). On average, responding wasps showed longer retention time on leaves when footprints originated from female rather than male N. viridula. No statistical difference in retention time was found between uncontaminated leaves and leaves contaminated with male footprints.

Fig. 1
figure 1

Effects on the host sex discrimination ability of Trissolcus basalis females exploring the adaxial leaf surface of broccoli. Leaves were contaminated with chemical footprints laid by adult males and females of Nezara viridula. The number of tested wasp females was 30 for each treatment. a Bars indicate the number of the wasps that showed the arrestment response (responding, grey bars) and those that did not show this behavior (non-responding, white bars). Statistically significant differences are indicated with asterisks within treatments, and by carets between treatments. NS = not significant. Karl Pearson χ 2 test and Goodman’s post hoc procedure. b Bars indicate mean values, whiskers are ± SE, of the leaf resident time of T. basalis females that responded to the different treatments. Bars followed by the same letters are not significantly different. One-way analysis of variance with type III sums of squares, followed by Fisher’s LSD post hoc tests

Full size image

Experiment 2

About 80% of all parasitoids displayed the typical arrestment posture on encountering leaf disks treated with chemical footprints from N. viridula females. This response was irrespective of whether bug legs had been treated with n-nonadecane, n-eicosane, or hexane (χ 2(2)  = 0.53, P = 0.76; Pearson’s χ 2) (Fig. 2 a). In contrast, tarsal treatments significantly affected the retention time of the wasps exploring the treated leaf disks. Host footprints from females that had been treated with n-nonadecane induced parasitoids to search for a significantly shorter period than the footprints from bugs treated with n-eicosane or hexane (F (2,68) = 5.62, P < 0.001; ANOVA Type III Sum of Squares analysis) (Fig. 2 b). The retention time of the latter two treatments was similar with no significant difference between them.

Fig. 2
figure 2

Effects on the host sex discrimination ability of Trissolcus basalis females exploring leaf disks (5 cm Ø) of broccoli. Leaves were contaminated with chemical footprints of N. viridula females with legs treated with n-nonadecane, n-eicosane, and hexane, respectively. The number of tested wasp females was 30 for each treatment. a Bars indicate the number of wasps that showed the arrestment response (responding, grey bars) and those that did not show this behavior (non-responding, white bars). Statistically significant differences are indicated by asterisks within treatments, and by carets between treatments. NS = not significant. Pearson’s χ 2 test and Goodman’s post hoc procedure. b Bars indicate mean values, whiskers are ± SE, of the leaf resident time of T. basalis females that responded to the different treatments. Bars followed by the same letters are not significantly different. One-way analysis of variance with type III sums of squares, followed by Fisher’s LSD post hoc tests

Full size image

Experiment 3

The total amount of epicuticular waxes on the adaxial side of Brassica leaves averaged 11.60 μg cm−2 ± 1.50 (mean ± SE). The mixture was composed mainly of linear hydrocarbons (38.70%), ketones (28.11%), and secondary alcohols (12.06%). Small amounts of esters (4.86%), primary alcohols (1.75%), aldehydes (1.56%), and traces of fatty acid esters and saturated and unsaturated fatty acids (less than 1%) were found. Within the class of linear hydrocarbons, n-nonacosane (nC29) was the most abundant component, followed by n-hentriacontane (nC31), while n-nonadecane was not present at the detection level in the extracts (Fig. 3).

Fig. 3
figure 3

Composition (%) of the outer epicuticular wax layer on the adaxial leaf surface of Brassica oleracea var. Italica cv. Marathon. Relative amounts of compounds are given as medians (bars) and 25th and 75th percentile (whiskers) (N = 6). Compounds are listed according to chain lengths in the homologous series of aliphatic compound classes. *Compounds not separable in GC peaks

Full size image

Discussion

Here, we provide further evidence that broccoli epicuticular wax retains chemical footprints of adult N. viridula and show that these cues allow T. basalis females to distinguish between the sexes of their host. On encountering the chemical footprints of adult bugs, motivated egg parasitoids displayed a well-described arrestment response that was followed by extensive host searching behavior. More time was spent investigating leaves contaminated by female than by male hosts or uncontaminated control leaves. By using an approach based on tarsal treatments of adult bugs, our results support previous experiments conducted on filter papers, which indicated that n-nonadecane (nC19), a compound present only in N. viridula males, plays a role as a host sex recognition cue for wasp females. This conclusion can be drawn because treating the legs of female N. viridula with n-nonadecane resulted in footprints that apparently resembled male footprints. Female parasitoids were unabale to differentiate between male footprints and female footprints mimicking those of males. It remains to be investigated for how long T. basalis can detect the chemical footprints of its host. Studies on a larval parasitoid showed that the wasps recognize footprints from caterpillars for a period of at least 2 days (Rostás and Wölfling, 2009). This suggests that these kairomones can be retained for a relevant length of time and should help to improve the efficiency in finding a suitable host.

So far, the linear alkane n-nonadecane has been found in few pentatomid species (Aldrich, 1996; Durak and Kalender, 2007; Durak, 2008). In N. viridula, it is present only in adult males. It occurs in the sex pheromone blend where it represents about 7% of the total volatiles (Aldrich et al., 1987), and it constitutes about 3% of the amount of cuticle linear hydrocarbons (Colazza et al., 2007). Long-chain hydrocarbons are of high volatility, thus n-nonadecane from N. viridula can disperse into the environment as a component of the pheromone blend and adsorb onto plant surfaces at the time the adults touch the plants. Earlier experiments showed that both male volatile pheromone and male contact footprints stimulate wasp females. However, when compared, T. basalis clearly preferred the chemical footprints produced by host females (Colazza et al., 1999). Therefore, the presence of this linear alkane in different sources, i.e., in the pheromone glands and the cuticle of the male bug, and the similar responses of T. basalis females to pheromone and footprints, suggests a function for n-nonadecane as a compound to differentiate between host sexes. Linear hydrocarbons most likely are secreted from cuticle tissue to protect the insect from desiccation (Howard and Blomquist, 2005). However, whether n-nonadecane is passively dispersed as a component of the cuticular hydrocarbons or if it is released through active secretion from the pheromone glands, is as yet unknown.

All compound classes and homologues identified from the wax extracts of broccoli leaves are found commonly in plant waxes (Müller and Riederer, 2005). The wax mixture contained mainly linear hydrocarbons, ketones, and secondary alcohols. Linear hydrocarbons were dominated by n-nonacosane and n-hentriacontane, while n-nonadecane was absent. Based on the literature, n-nonadecane is not present in the epicuticular waxes of several of N. viridula host plants belonging to different families, such as Solanaceae, e.g., Lycopersicum esculentum (Zygadlo et al., 1994), Solanum tuberosum (Szafranek and Synak, 2006); Brassicaceae, e.g., B. oleracea (Baker, 1974; Hunsche et al., 2006; Purdy and Truter, 2010); Fabaceae, e.g., Vicia faba (Powell et al., 1999), Glycine max (Kim et al., 2007), Phaseolus vulgaris (Hunsche et al., 2006), and Pisum sativum (Gniwotta et al., 2005). Therefore, the absence of n-nonadecane among the hydrocarbons of epicuticular plant surfaces of these N. viridula host plants avoids possible masking effects and may enhance T. basalis host location.

Our study represents a starting point in elucidating the mechanisms by which egg parasitoids can distinguish host footprint compounds from the waxy background of the plant cuticle, but further studies are necessary to clarify the role that plant cuticular hydrocarbons play in the host finding process of parasitoids. Recent experiments indicate that footprints of N. viridula females also contain diglycerides and triglycerides of high molecular weight, and long-chain alcohols and fatty acids (Lo Giudice, 2009). On the various host plant species, these compounds can be retained to different degrees as a result of the absorptive properties of epicuticular waxes. As a consequence, the host recognition behavior of T. basalis females may vary among plants. Another interesting question that awaits better clarification is, whether T. basalis is not only able to recognize the adult host’s sex but also the physiological state of the female or distinguish adults from immatures. Such capabilities were demonstrated in the related species Trissolcus brochymenae (Hymenoptera: Scelionidae) and its host Murgantia histrionica (Heteroptera: Pentatomidae). Salerno et al. (2009) found that footprints of mated female bugs elicited longer periods of host searching than kairomones from virgin bugs. However, as in most experiments, the chemical footprints were tested on filter paper substrate, while future studies on parasitoid-host interactions should take into account the great variety of plant-specific epicuticular wax chemistries and how they may potentially influence the perception of kairomones.