J Chem Ecol (2010) 36:620–628 DOI 10.1007/s10886-010-9802-6

Present or Past Herbivory: A Screening of Volatiles Released from Brassica rapa Under Caterpillar Attacks as Attractants for the Solitary , vestalis

Soichi Kugimiya & Takeshi Shimoda & Jun Tabata & Junji Takabayashi

Received: 20 February 2010 /Revised: 7 April 2010 /Accepted: 11 May 2010 /Published online: 20 May 2010 # Springer Science+Business Media, LLC 2010

Abstract Females of the solitary endoparasitoid Cotesia decreased after removal of the host larvae, whereas vestalis respond to a blend of volatile organic compounds terpenoids and their related compounds continued to be (VOCs) released from plants infested with larvae of their released at high levels. Benzyl cyanide and dimethyl host, the (Plutella xylostella), which is trisulfide attracted in a dose-dependent manner, an important pest of cruciferous plants. We investi- whereas the other compounds were not attractive. These gated the flight response of female parasitoids to the results suggest that nitrile and sulfide compounds tempo- cruciferous plant Brassica rapa, using two-choice tests rarily released from plants under attack by host larvae are under laboratory conditions. The parasitoids were more potentially more effective attractants for this parasitoid than attracted to plants that had been infested for at least 6 hr by other VOCs that are continuously released by host- the host larvae compared to intact plants, but they did not damaged plants. distinguish between plants infested for only 3 hr and intact plants. Although parasitoids preferred plants 1 and 2 days Key Words -induced plant volatiles . after herbivory (formerly infested plants) over intact plants Indirect defense . Tritrophic interaction . Brassica rapa . they also preferred plants that had been infested for 24 hr Plutella xylostella . Cotesia vestalis over formerly infested plants. This suggests that parasitoids can distinguish between the VOC profiles of currently and formerly infested plants. We screened for differences in Introduction VOC emissions among the treatments and found that levels of benzyl cyanide and dimethyl trisulfide significantly Many plant species release specific blends of volatile organic compounds (VOCs) in response to attack by . The release of VOCs induced by herbivory can attract natural enemies of herbivores and may guide : parasitoids or predators to their hosts or prey (Turlings et al. S. Kugimiya (*) J. Tabata 1990; Takabayashi and Dicke 1996). Chemical information National Institute for Agro-Environmental Sciences, of this nature has attracted considerable attention as an Kannondai 3-1-3, Tsukuba, Ibaraki 305-8604, Japan indirect defense of plants against herbivorous e-mail: [email protected] mediated by their natural enemies at a higher level in a tritrophic interaction system (Karban and Baldwin 1997; T. Shimoda Sabelis et al. 2007). However, the plant defense may not be National Agricultural Research Center, Kannondai 3-1-1, constant, since the composition of plant VOCs can vary Tsukuba, Ibaraki 305-8666, Japan both in quality and in quantity, depending on various biotic and abiotic factors, and these changes can affect the J. Takabayashi attractiveness of the plants to natural enemies of the Center for Ecological Research, Kyoto University, Hirano 2-509-3, herbivores (Takabayashi et al. 1994; Maeda et al. 2000; Otsu, Shiga 520-2113, Japan Gouinguené and Turlings 2002). J Chem Ecol (2010) 36:620–628 621

The profile of VOCs also can change during the course after the start of herbivory and after removal of the larvae, of herbivory (Loughrin et al. 1994; Turlings et al. 1998; and tested synthetic versions of VOCs to evaluate their Scascighini et al. 2005). Diurnal cycles in the release of effectiveness as attractants for parasitoids. VOCs are related to the attraction of natural enemies of herbivores or to deterrence of herbivores (Loughrin et al. 1994; De Moraes et al. 2001; Shiojiri et al. 2006a), and the Methods and Materials importance of the effects of light in a diurnal cycle have been reported (Gouinguené and Turlings 2002). However, Plants and Insects Japanese mustard spinach, Brassica there is a little information on changes in VOCs of infested rapa L. var. perviridis (Capparales: Brassicaceae), was plants during and after herbivory, particularly from the cultivated in a greenhouse (25±3°C, 60±10% relative perspective of potential differences in plant-insect inter- humidity (RH), 16L:8D photoperiod). Five plants were actions between the two phases (Mattiacci et al. 2001; reared in a plastic pot (90 mm diam, 70 mm depth) for 4– Hoballah and Turlings 2005). In natural ecosystems, 5 wk, and were used for both insect rearing and flight herbivores may not stay on a single plant, but may move preference tests. Diamondback moths, P. xylostella, origi- among plants, be eliminated by other predators (Shiojiri and nally collected from fields in Ayabe, Kyoto Prefecture, Takabayashi 2005), or pupate and become unsuitable Japan, in 2001, were mass-reared on potted plants in a targets for parasitoids to oviposit. Thus, parasitoids may climate-controlled room (25±3°C, 60±10% RH, 16L:8D be able to distinguish between plants currently being photoperiod). Eggs were collected every day, and hatched attacked (hereafter, “infested plants”) and plants that were larvae were reared on cut plants in small cages (width formerly attacked by herbivores (hereafter, “formerly 25 cm, depth 15 cm, height 10 cm). The solitary parasitoids infested plants”). By focusing on the difference in compo- C. vestalis, which parasitize mainly P. xylostella larvae, sitions of VOCs released from infested and formerly were obtained from their hosts collected in the same fields infested plants, we may be able to identify effective and were reared on P. xylostella-infested plants under the attractants for parasitoids. same conditions as their hosts. For use in the two-choice In Japan, Brassica plants (Capparales: Brassicaceae) are test, cocoons were collected from the stock culture, grown during the spring, close enough together for adjacent and newly emerged adults were maintained with 50% plants to touch. A tritrophic system forms among the plant aqueous honey in acryl cages (width 35 cm, depth 25 cm, species, herbivorous insects, and their parasitoids. Cater- height 30 cm) separately from the host-infested plants until pillars of the diamondback moth, Plutella xylostella L. the experiments. (Lepidoptera: Yponomeutidae), are oligophagous on crucif- erous plants. The specialist parasitoid wasp Cotesia vestalis Flight Response of Parasitoids to Host-infested Plants and (Haliday) [= C. plutellae (Kurdjumov)] (: to Synthetic Compounds The flight response of C. vestalis ) oviposits on P. xylostella larvae. The female females was assessed by using a two-choice test in an acrylic parasitoids are attracted to crucifers infested by P. xylostella chamber (width 35 cm, depth 25 cm, height 30 cm, 3,000 larvae through a specific blend of VOCs (Shiojiri et al. lux) in a climate-controlled room (25±3°C, 60±10% RH) 2000) and lay a single egg per host. (Shiojiri et al. 2000). Infested plants were prepared by Under laboratory conditions, infested cruciferous plants allowing 15 third-instar P. xylostella to feed on potted plants attract various parasitoids by emitting specific blends of of B. rapa for 3, 6, and 24 hr (hereafter, 3, 6, and 24 hr- VOCs (Mattiacci et al. 1994; Shiojiri et al. 2000;Van infested plants, respectively; Fig. 1). The infestation began at Poecke et al. 2001), and the mechanisms for the induction, 12:00 am, during the middle of the light phase of the regulation, and biosynthesis of those VOCs have been photoperiod. Formerly infested plants were prepared by characterized (Mattiacci et al. 1995; Van Poecke and Dicke removing the host larvae from 24 hr-infested plants and then 2002; Shiojiri et al. 2006b;D’Auria et al. 2007; Herde et al. by using the plants for experiments 48 and 72 hr after the 2008). However, it is unclear which VOC components start of the infestation (1 and 2d-after plants, respectively; attract C. vestalis. In the present study, we screened for Fig. 1). In every treatment, we used almost the same size potential parasitoid attractants, by investigating the flight larvae, showing similar feeding activity, and the areas of response of female parasitoids to plants that had been damage caused by the larvae were identical for a given infested by their host larvae, P. xylostella, for different 24 hr-infested plant and the 1d-after and 2d-after versions of durations (infested plants), and their response to plants at that plant. Intact plants were prepared as controls, without various times after herbivory (formerly infested plants). We any treatment. To produce a two-choice test design, one then analyzed the headspace volatiles released from potted plant from a given treatment was placed ca. 10 cm infested and formerly infested plants to establish the time from a plant that had been subjected to a different treatment course of changes in the relative amounts of plant VOCs inside the acrylic chamber, with the leaves not overlapping. 622 J Chem Ecol (2010) 36:620–628

Hosts attacking Collection of Volatiles Headspace volatiles were sampled Hosts removed from whole potted plants subjected to different treatments 2d-after plants (infested plants, formerly infested plants, and intact plants) 1d-after plants using a dynamic headspace collection system in a climate- 24h-infested plants controlled room (25±3°C, 60±10% RH, 16L:8D photope- 6h-infested plants riod). Each pot of plants was placed in a 2-l glass container, and the container was sealed with a glass lid that contained 3h-infested plants an air inlet and an air outlet. The container then was tightly 0h-infested plants (Intact plants) sealed with metal clamps on the lid. Incoming air was 7248 246 3 0 purified by filtration through silica gel and activated charcoal, and was actively pumped through the container Time (h) at a flow rate of 300 ml min−1, constantly monitored and Fig. 1 Experimental time line for infestation of Plutella xylostella. controlled by a flowmeter. Volatiles were collected for 3 hr Data were collected at various times following infestation (0, 3, 6, and on 180 mg of Tenax TA (60/80 mesh; Gerstel GmbH & Co. 24 hr-infested plants) and following removal of larvae (1 and 2d-after plants) KG, Mülheim an der Ruhr, Germany) packed in a glass tube, which was directly connected to the outlet. Volatiles At the start of each trial, a group of 10 female parasitoids were collected from 1.5–4.5 hr after the start of the was released from glass tubes (25 mm diam, 120 mm infestation (3 hr-infested), from 4.5–7.5 hr after the start of height) at a position centered between the two pots. After the infestation (6 hr-infested), or for 3 hr beginning 24, 48, hovering between the two potted plants inside the chamber, and 72 hr after the start of the infestation (24 hr-infested, 1 the females landed on one of the plants. The total numbers and 2d-after plants, respectively). For every treatment, of first landings on each potted plant by the parasitoids collection of volatiles was replicated using 5 plants. were counted for 30 min. After landing, parasitoids were carefully removed from the chamber with an insect Chemical Analysis of Headspace Volatiles The headspace aspirator. A few females that did not land on any potted volatiles collected in the Tenax tubes were analyzed using a plants within 30 min were recorded as no-choice subjects. gas chromatography-mass spectrometry (GC-MS) system At the end of each test, the test chamber was wiped with consisting of an Agilent 6890N gas chromatograph (Agilent 70% ethanol aq., and the inside air was flushed out with a Technologies, Inc., Palo Alto, CA, USA) coupled with an fan to avoid potential contamination by volatiles. Another Agilent 5973N quadrupole mass selective detector. The group of 10 parasitoids was tested on the same day, with the collected volatiles were desorbed from the Tenax in a positions of the plants subjected to different treatments Thermodesorption System (Gerster, Inc.) by heating the switched to counteract any potential positional effects. tube from 20°C (1 min hold) to 200°C (4 min hold) at a rate These data were pooled as a single replicate. Three of 60°C min−1. The released volatile compounds were replications were carried out on different days. Individual carried through a transfer line (250°C) to a Cold Injection insects (60 parasitoids in all) were used only once. System (Gerster, Inc.) and cryofocused at −100°C (30 sec To test the flight response of parasitoids to compounds hold). They then were heated at a rate of 12°C min−1 to that were identified from infested plants and formerly 260°C (5 min hold) and injected by a splitless mode into an infested plants (as described in the next section), we used analytical capillary column (HP-5MS, 30 m×0.25 mm i.d., triethyl citrate (Wako Pure Chemical industries, Ltd., 0.25 μm film thickness; Agilent Technologies Inc.). Helium Osaka, Japan) as a solvent and prepared 1–100 mg/l was used as carrier gas at a flow rate of 1.2 ml min−1. The solutions of the test compounds. Each solution (200 μl) GC oven was programmed to start at 40°C (9 min hold) and was applied to two square pieces of filter paper (20× then to rise at a rate of 10°C min−1 to 280°C (5 min hold). 20 mm, Advantec No. 1, Japan) that were placed together The column effluents were ionized by electron impact on a cover glass (24×24 mm, Matsunami No. 1, Japan), ionization (70 eV) with an ion source temperature of 250°C which was then placed in the pot beside the intact plants. and monitored in the mass selective mode with a scan range As a control, solvent only (200 μl) was applied to filter from 35 to 350 m/z. Volatile compounds were identified papers on the cover glass and placed beside intact plants. tentatively by comparing their mass spectra with those in The preference of the parasitoids for intact plants presented the Wiley Library database and further confirmed by together with the synthetic compounds vs. control plants comparing retention times and mass spectra to commer- was tested in the acrylic chamber. The amount of each cially available authentic standards (Wako Pure Chemical compound in the headspace of the plants was checked and Industries, Ltd., Osaka, Japan and Tokyo Chemical Industry compared with the amount in the headspace of infested Co., Ltd., Tokyo, Japan). (E)-β-Ocimene was synthesized plants. from (Z)-β-ocimene (available as mixture of ocimene J Chem Ecol (2010) 36:620–628 623

(a) isomers; SAFC Supply Solutions, St. Louis, MO, USA) by Infested Intact isomerizing the cis to the trans configuration (Sgoutas and 3h vs. Intact NS 35 23 (2) Kummerow 1967). 6h vs. Intact ** 40 18 (2) Statistics In the flight response test, significant differences 24h vs. Intact 46 14 (0) between the numbers of that landed on either of the *** two potted plants and the null hypothesis of an expected (b) Formerly-infested Intact ratio of 0.5:0.5 were analyzed by using a replicated G-test to account for suspected heterogeneities among the repli- 1d-after vs. Intact * 38 22 (0) cations (Sokal and Rohlf 1995). To compare the relative 2d-after vs. Intact * 37 21 (2) amounts of each volatile compound released from the plants after different treatments, the total ion peak area of (c) Infested Formerly-infested each compound detected by GC-MS analysis was calculat- ed by integration. We applied one-way analysis of variance 24h vs. 1d-after *** 43 17 (0) ’ (ANOVA), followed by Tukey s HSD test, to the results of 24h vs. 2d-after ** 42 18 (0) the chemical analysis. When necessary, data were normal- ized using logarithmic transformation [ln (10x + 1)] to meet 3h vs. 2d-after NS 35 25 (0) the assumptions for ANOVA. These analyses were per- (d) formed using JMP software (version 7.0.1, SAS Institute 3h-Infested 6h-infested Inc., Cary, NC, USA). 3h vs. 6h NS 32 28 (0)

50 005 Flight preference (%)

Results Fig. 2 Flight preferences (%) of female parasitoids Cotesia vestalis (N=60) among intact plants (open bars), Plutella xylostella larvae- Flight Response of Parasitoids to Infested and Formerly infested plants (3, 6, and 24 hr: filled bars) and formerly infested Infested Plants Female parasitoids preferred 6 and 24 hr- plants (1 and 2d-after: dotted bars). Numbers in bars indicate the number of parasitoids that landed on each plant. Numbers in infested plants over intact plants in the two-choice test parentheses indicate parasitoids that did not choose any plants. (Fig. 2a) but 3 hr -infested plants were not distinguished Asterisks mean significant differences within each preference test set from intact plants (Fig. 2a) or from 6 hr-infested plants (replicated G-test; *: P<0.05, **: P<0.01, ***: P<0.001, NS = no (Fig. 2d). Female parasitoids also preferred 1 and 2d-after significance) plants over intact plants, respectively (Fig. 2b). When offered both 24 hr-infested plants and formerly infested cyanide (Fig. 3); the sulfur-containing compounds, aryl plants (1 or 2d-after) at the same time, parasitoids preferred isothiocyanate and dimethyl trisulfide (Fig. 3); the mono- the 24 hr-infested plants (Fig. 2c). When offered 3 hr- terpenes, myrcene, limonene, (E)-β-ocimene, and α-pinene infested plants and 2d-after plants in the two-choice test, (Fig. 4); the sesquiterpenes, (E,E)-α-farnesene and junipene parasitoids showed no significant preference (Fig. 2c). (Fig. 4); and the terpenoid related ketones, 6-methyl-5- There was no significant heterogeneity among replicates hepten-2-one and 6,10-dimethyl-5,9-undecadien-2-one in any tested sets (replicated G-test, df=2, P>0.05 for each (Fig. 4).

Gh), suggesting good reproducibility of the two-choice test. The relative amounts of benzyl cyanide (ANOVA, MS= 1.885, df=5, F=27.415, P<0.001) and dimethyl trisulfide Volatiles Released from Infested and Formerly Infested (MS=0.923, df=5, F=22.998, P<0.001) in the headspace Plants Profiles of VOCs released by infested plants, increased after the start of the infestation (Fig. 3). The formerly infested plants, and intact plants were quantita- relative amounts of the four monoterpenes (myrcene, MS= tively but not qualitatively different. In the headspace of 0.148, df=5, F=4.662, P=0.004; limonene, MS=0.314, these plants, 11 major compounds were identified by GC- df=5, F=8.494, P<0.001; (E)-β-ocimene, MS=0.369, df= MS analysis, and other unidentified minor compounds were 5, F=9.901, P<0.001; α-pinene, MS=1.144 df=5, F= ignored. Various alkanes and aldehydes with lengths of C9 9.385, P<0.001; Fig. 4), and one sesquiterpene [(E,E)-α- to C24 were ignored as impurities, since they were also farnesene, MS=0.372, df=5, F=5.134, P=0.0024; Fig. 4] detected from the blank headspace. Ten compounds were increased following infestation. However, no significant identified by using authentic standards, and junipene was differences were observed in the relative amounts of the tentatively identified although an authentic standard was remaining compounds, aryl isothiocyanate (MS=0.401, df= not available. The compounds included the nitrile, benzyl 5, F=2.189, P=0.089; Fig. 3), junipene (MS=0.069, df=5, 624 J Chem Ecol (2010) 36:620–628

Hosts attacking Hosts removed x 105 Dose-response Relationships in Flight Preferences of the 100 Benzyl cyanide a Parasitoids for Synthetic Compounds Female parasitoids 80 preferred intact plants presented with 10 or 100 mg/l benzyl 60 cyanide solution over control plants with solvent alone 40 ab (Fig. 5a). However, no significant preference was observed bc between plants with 1 mg/l benzyl cyanide solution and 20 cd e de control plants (Fig. 5a). Similarly, female parasitoids 0 5 preferred intact plants with 10 or 100 mg/l dimethyl 80 x 10 Aryl isothiocyanate trisulfide solutions over control plants, respectively NS 60 (Fig. 5b). However, no significant preference was observed 40 between plants with 1 mg/l dimethyl trisulfide solution and 20 control plants (Fig. 5b). On the other hand, at the same

Relative intensity 0 concentrations, parasitoids showed no significant prefer- x 105 β 40 ences for (R)-limonene, (S)-limonene, (E)- -ocimene, (R)- Dimethyl trisulfide α-pinene, (S)-α-pinene, or (E,E)-α-farnesene (Fig. 5c–h). a ab The absence of significant heterogeneity detected among 20 replicates in every tested set (replicated G-test, df=2, P> cd bc e de 0.05 for each Gh) suggested good reproducibility of the 0 two-choice test. 0362448 72 (1 d) (2 d) The highest levels of benzyl cyanide (3.68±0.59 ng) and Time (h) dimethyl trisulfide (0.50±0.16 ng) were released by 24 hr- Fig. 3 Time-course indicating relative amounts of a nitrile and sulfur- infested plants. When intact plants were treated with containing compounds detected by GC-MS analysis of the headspace 10 mg/l benzyl cyanide, levels in the headspace were volatiles released from intact plants (0 hr), Plutella xylostella larvae- approximately three-fold higher than the levels from 24 hr- infested plants (3, 6, and 24 hr) and formerly infested plants (1 and 2d- infested plants. Even higher levels were detected with benzyl after) (N=5). Different small letters mean significant differences in relative amounts of each compound (ANOVA followed by Tukey’s cyanide applied at 100 mg/l. When dimethyl trisulfide was HSD test; P<0.05, NS = no significance) applied at concentrations of 10 and 100 mg/l, 200–500 times the amount released from 24 hr-infested plants were detected in the headspace. The tested concentrations of other synthetic F=1.968, P=0.120; Fig. 4), 6-methyl-5-hepten-2-one VOCs were similar to the amounts released by the infested (MS=0.281, df=5, F=1.354, P=0.276; Fig. 4), and 6,10- plants detected in the headspace. dimethyl-5,9-undecadien-2-one (MS=0.133, df=5, F= 1.118, P=0.377; Fig. 4). The compounds with significant differences during the time-course measurements were analyzed further by using a Discussion pairwise test. In particular, benzyl cyanide, dimethyl trisulfide, (E)-β-ocimene, and α-pinene were induced as Female C. vestalis parasitoids showed a significant prefer- soon as 3 hr after the start of the infestation (Figs. 3 and 4). ence for B. rapa plants that had been infested for 24 hr with Limonene and (E,E)-α-farnesene were not induced until larvae of their host herbivores, P. xylostella, over formerly 24 hr after infestation (Fig. 4). After removal of the larvae, infested plants from which the larvae had been removed benzyl cyanide and dimethyl trisulfide decreased rapidly, (Fig. 2c). GC-MS analysis revealed that benzyl cyanide and and the amounts of both compounds released from 1 and dimethyl trisulfide were released at significantly higher 2d-after plants were smaller than those released from 24 hr- levels from 24 hr-infested plants than from formerly infested plants (Fig. 3). The relative amounts of the two infested plants (Fig. 3). These results suggest that the two compounds from 1d-after plants were larger than those compounds released from 24 hr-infested plants are poten- from intact plants (Fig. 3), but the amounts in 2d-after tially better attractants for parasitoids than the other VOCs plants did not differ significantly from intact plants (Fig. 3). released from formerly infested plants. In contrast, most terpenoids were released from the plants at We further tested synthetic versions of benzyl cyanide significantly higher levels than from intact plants after the and dimethyl trisulfide as strong candidates for parasitoid larvae had been removed (Fig. 4), and the amounts of attraction, and found that intact plants presented together limonene, (E)-β-ocimene, α-pinene, and (E,E)-α-farnesene with the single compounds attracted parasitoids in a dose- released from 1 and 2d-after plants were not significantly dependent manner (Fig. 5a,b), in contrast to the other different from those of 24 hr-infested plants (Fig. 4). compounds found in the headspace (Fig. 5c–h). When J Chem Ecol (2010) 36:620–628 625

Fig. 4 Time-course indicating Hosts attacking Hosts removed Hosts attacking Hosts removed relative amounts of terpenoids x 105 detected by GC-MS analysis of Myrcene ab Limonene the headspace volatiles released 40 abc a from intact plants (0 hr), Plu- abc a tella. xylostella larvae-infested bc 20 c ab ab plants (3, 6, and 24 hr) and bc formerly infested plants (1 and c bc 2d-after) (N=5). Different small 0 letters mean significant differ- x 105 10 ences in relative amounts of (E)-β-Ocimene a α-Pinene a each compound (ANOVA fol- a a lowed by Tukey’s HSD test; P< 5 a a a a 0.05, NS = no significance) a a b b 0 x 105 20 (E,E)-α-Farnesene Junipene a a a Relative intensity 10 NS b ab ab 0 x 105 6-methyl-5-hepten-2-one 6,10-dimethyl-5,9-undecadien-2-one 20

NS NS 10

0 0362448 72 0362448 72 (1 d) (2 d) (1 d) (2 d) Time (h) Time (h) benzyl cyanide was applied to intact plants at the most Female parasitoids preferred 6 or 24 hr-infested plants attractive concentration (10 mg/l; Fig. 5a), its concentration over intact plants (Fig. 2a). There was also a slight, but not in the headspace was only a few-fold higher than the significant, trend in the preference for 3 hr-infested plants concentration detected in the headspace of 24 hr-infested over intact plants. This trend is supported by the observa- plants. On the other hand, when dimethyl trisulfide was tion that parasitoids did not discriminate between 3 and applied at the attractive concentrations (10 and 100 mg/l; 6 hr-infested plants (Fig. 2d). GC-MS analysis revealed that Fig. 5b), its concentration in the headspace was far larger the emission of some VOCs was induced within a few than the concentration released from 24 hr-infested plants. hours of the initial damage by larvae. Together with benzyl These results suggest that although either compound can act cyanide and dimethyl trisulfide, (E)-β-ocimene and α- as a parasitoid attractant, benzyl cyanide is most likely to be pinene were significantly increased in 3, 6, and 24 hr- responsible for parasitoid discrimination between 24 hr- infested plants (Fig. 4). (E)-β-Ocimene and α-pinene were infested plants and formerly infested plants. Dimethyl not attractive to parasitoids at the tested concentrations trisulfide may have little effect at the levels normally (Fig. 5e–g), which are similar to the levels these com- released by these plants. pounds released by the infested plants. Benzyl cyanide may Significant differences in the amounts of VOCs detected be the main compound that attracts parasitoids to infested by chemical analysis may not always result in significant plants, and dimethyl trisulfide may be of less importance. preferences by parasitoids. For example, significantly larger However, the relative attractiveness for parasitoids of amounts of benzyl cyanide were analytically detected from mixtures of terpenoids with benzyl cyanide and dimethyl 3 hr-infested plants than from 2d-after plants (Fig. 3), but trisulfide has not been tested, but may be important in light no significant preference was shown by parasitoids between of the potential ability of the parasitoids to use associative these plants (Fig. 2c). Differences in the amounts of VOCs learning about VOCs (Turlings et al. 1993; Vet et al. 1995). are necessary for parasitoids to discriminate between plants, Formerly infested plants, which no longer had larvae, but small differences sometimes may not be enough to be were still preferred by the parasitoids over intact plants perceived by parasitoids with their limited sensitivity to (Fig. 2b). It appears that the plants kept releasing VOCs VOCs. that attracted the parasitoids for 2 d after herbivory had 626 J Chem Ecol (2010) 36:620–628

(a) Benzyl cyanide Treatment Control Benzyl cyanide was emitted by cabbage (Brassica 1 mgl-1 NS 29 30 (1) oleracea var. capitata) infested by Pieris brassicae or P. 10 mgl-1 *** 44 15 (1) 100 mgl-1 * 39 20 (1) rapae caterpillars (Geervliet et al. 1997), and was induced (b) Dimethyl trisulfide within a few hours in Brussels sprouts (B. oleracea var. 1 mgl-1 NS 34 26 (0) gemmifera) (Scascighini et al. 2005). In the parasitoids 10 mgl-1 * 39 21 (0) Cotesia glomerata and C. rubecula, benzyl cyanide evokes 100 mgl-1 * 39 20 (1) (c) (R)-Limonene an electro-antennogram response (Smid et al. 2002). Benzyl 1 mgl-1 NS 26 30 (4) cyanide emitted from mated females of P. brassicae attracts 10 mgl-1 NS 30 29 (1) their egg parasitoid, Trichogramma brassicae,which -1 100 mgl NS 29 29 (1) subsequently reaches the oviposition site of the butterfly (d) (S)-Limonene 1 mgl-1 NS 30 29 (1) by means of phoresy (Fatouros et al. 2005). Dimethyl 10 mgl-1 NS 30 26 (4) trisulfide is one of many sulfide compounds that are 100 mgl-1 NS 28 29 (3) characteristic of Allium plants (Dugravot et al. 2004; E β (e) ( )- -Ocimene Tatemoto and Shimoda 2008), but is also found in infested 1 mgl-1 NS 31 27 (2) 10 mgl-1 NS 29 30 (1) cruciferous plants such as B. oleracea and B. napus 100 mgl-1 NS 31 28 (1) (Geervliet et al. 1997; Ferry et al. 2007). However, to our (f) (R)-α-Pinene knowledge, there have been no reports that parasitoids are -1 1 mgl NS 31 28 (1) attracted to sulfide compounds. Green leaf volatiles (GLVs) 10 mgl-1 NS 32 27 (1) 100 mgl-1 NS 31 26 (3) generally are released soon after herbivory, and various (g) (S)-α-Pinene parasitoids respond to them (Whitman and Eller 1992; 1 mgl-1 NS 29 27 (4) Birkett et al. 2003; Gouinguené et al. 2005). However, they -1 10 mgl NS 29 28 (3) were not analyzed in the present study, because single GLV 100 mgl-1 NS 28 29 (3) (h) (E,E)-α-Farnesene compounds do not attract C. vestalis females (Shiojiri et al. 1 mgl-1 NS 33 26 (1) 2006b). 10 mgl-1 NS 28 30 (2) It is reasonable for parasitoids to use benzyl cyanide as a -1 NS 31 28 100 mgl (1) host-searching cue, since it is common in cruciferous plants 500 50 as a breakdown product of glucosinolates through a rapid Flight preference (%) reaction catalyzed by myrosinase at damaged sites (Wittstock and Halkier 2002). Benzyl cyanide was not detected in feces Fig. 5 Flight preferences (%) of female parasitoids Cotesia vestalis (N=60) between intact plants presented together with different of P. xylostella larvae, which can disarm consumed concentrations of synthetic compounds (filled bars) and control plants glucosinolates by means of glucosinolate sulfatase (Ratzka (open bars). Numbers in bars indicate the number of parasitoids that et al. 2002). Nonetheless, female parasitoids could detect landed on each plant. Numbers in parentheses indicate parasitoids that benzyl cyanide released from damaged plant tissue. In did not choose any plants. Asterisks mean significant differences within each preference test set (replicated G-test; *: P<0.05, ***: P< addition, benzyl cyanide could be systemically inducible, 0.001, NS = no significance) like terpenoids and GLVs, through the jasmonic acid (JA) pathway as suggested for other Brassica species (Bruinsma stopped. Benzyl cyanide and dimethyl trisulfide cannot et al. 2009). Further physiological studies are needed to explain the observed preferences, since the two compounds investigate JA-dependent induction of benzyl cyanide and decreased rapidly after removal of the larvae (Fig. 3). On dimethyl trisulfide in B. rapa. the other hand, terpenoids were emitted from formerly In the present study, we clarified the attractants for infested plants at significantly higher levels than from intact parasitoids by screening for candidates based on the relative plants (Fig. 4). Ibrahim et al. (2005) reported that C. amounts of VOCs released from plants currently under vestalis preferred intact cabbages plus synthetic (R)-(+)- attack and from plants that were attacked in the past. limonene to control cabbages, depending on the cabbage Further comparative studies to identify the VOCs that subspecies. In this study, intact plants plus individual attract foraging parasitoids to various plant species are synthetic terpenoids, including (R)-(+)-limonene, did not necessary, to effectively use VOCs for biological control. attract parasitoids significantly (Fig. 5c–h). These results Further studies could also reveal details of the chemical imply that a certain blend of VOCs, especially of terpenoid basis for the tritrophic systems that are formed in ecological compounds, was responsible for the preference for formerly processes. infested plants over intact ones. There was also another possibility that the hosts themselves and visual cues, such Acknowledgements We are grateful to Kimiko Kanbe and Yumiko as damage holes, might have affected the preference for Togashi (NARC) for their help in rearing insects and cultivating plants infested and formerly infested plants over intact plants. used in the experiments. This study was partly supported by a Grant- J Chem Ecol (2010) 36:620–628 627 in-Aid for Young Scientists (B) [No. 21710241 for SK] and by a MAEDA, T., TAKABAYASHI, J., YANO, S., and TAKAFUJI, A. 2000. Grant-in-Aid for Scientific Research (B) [No. 19380188 for TS] from Effects of light on the tritrophic interaction between kidney bean the Japan Society for the Promotion of Science (JSPS). plants, two-spotted spider mites and predatory mites, Amblyseius womersleyi (Acari: Phytoseiidae). Exp. Appl. Acarol. 24:415– 425. MATTIACCI, L., DICKE, M., and POSTHUMUS, M. A. 1994. Induction References of parasitoid attracting synomone in Brussels sprouts plants by feeding of Pieris brassicae larvae: Role of mechanical damage BIRKETT, M. A., CHAMBERLAIN, K., GUERRIERI, E., PICKETT, J. A., and herbivore damage. J. Chem. Ecol. 20:2229–2247. WADHAMS, L. J., and YASUDA, T. 2003. Volatiles from whitefly- MATTIACCI, L., DICKE,M.,andPOSTHUMUS, M. A. 1995. β- infested plants elicit a host-locating response in the parasitoid, Glucosidase: an elicitor of herbivore-induced plant odor that Encarsia formosa. J. Chem. Ecol. 29:1589–1600. attracts hosts-searching parasitic wasps. Proc. Natl. Acad. Sci. U. BRUINSMA, M., POSTHUMUS, M. A., MUMM, R., MUELLER, M. J., S. A. 92:2036–2040. VAN LOON, J. J. A., and DICKE, M. 2009. Jasmonic acid-induced MATTIACCI,L.,AMBÜHL-ROCCA, B., SCASCIGHINI,N.,D’ALESSANDRO, volatiles of attract parasitoids: effects of time M., HERN,A.,andDORN, S. 2001. Systemically induced plant and dose, and comparison with induction by herbivores. J. Exp. volatiles emitted at the time of “danger”. J. Chem. Ecol. 27:2233– Botany 60:2575–2587. 2252. D’AURIA, J. C., PICHERSKY, E., SCHAUB, A., HANSEL, A., and RATZKA, A., VOGEL, H., KLIEBENSTEIN, D. J., MITCHELL-OLDS,T., GERSHENZON, J. 2007. Characterization of a BAHD acyltransfer- and KROYMANN, J. 2002. Disarming the mustard oil bomb. Proc. ase responsible for producing the green leaf volatile (Z)-3-hexen- Natl. Acad. Sci. U. S. A. 99:11223–11228. 1-yl acetate in Arabidopsis thaliana. Plant J. 49:194–207. SABELIS, M. W., TAKABAYASHI, J., JANSSEN, A., KANT, M. R., VAN DE MORAES, C. M., MESCHER, M. C., and TUMLINSON, J. H. 2001. WIJK, M., SZNAJDER, B., et al. 2007. Ecology meets plant Caterpillar-induced nocturnal plant volatiles repel conspecific physiology: herbivore-induced plant responses and their indirect females. Nature 410:577–580. effects on communities, pp. 188–217, in T. Ohgushi, T. DUGRAVOT, S., THIBOUT, E., ABO-GHALIA, A., and HUIGNARD,J. P. Craig and P. W. Price (eds.). Ecological Communities: Plant 2004. How a specialist and a nonspecialist insect cope with the Mediation in Indirect Interaction Webs. Cambridge University dimethyl disulfide produced by Allium porrum. Entomol. Exp. Press, Cambridge. Appl. 113:173–179. SCASCIGHINI, N., MATTIACCI, L., D’ALESSANDRO, M., HERN, A., FATOUROS, N. E., HUIGENS, M. E., VAN LOON, J. J. A., DICKE, M., ROTT A. S., and DORN, S. 2005. New insights in analysing and HILKER, M. 2005. Chemical communication—Butterfly anti- parasitoid attracting synomones: early volatile emission and use aphrodisiac lures parasitic wasps. Nature 433:704. of stir bar sorptive extraction. Chemoecology 15:97–104. FERRY, A., DUGRAVOT, S., DELATTRE, T., CHRISTIDES,J.-P.,AUGER, SGOUTAS, D. S., and KUMMEROW, F. A. 1967. Cis-trans isomerization J., BAGNÈRES, A. -G., POINSOT, D., and CORTESERO, A. -M. of unsaturated fatty acid methyl esters without double bond 2007. Identification of a widespread monomolecular odor migration. Lipids 4:283–287. differentially attractive to several Delia radicum ground- SHIOJIRI, K., and TAKABAYASHI, J. 2005. Parasitoid preference for dwelling predators in the field. J. Chem. Ecol. 33:2064–2077. host-infested plants is affected by the risk of intraguild predation. GEERVLIET, J. B. F., POSTHUMUS, M. A., VET, L. E. M., and DICKE, J. Insect Behav. 18:567–576. M. 1997. Comparative analysis of headspace volatiles from SHIOJIRI, K., TAKABAYASHI, J., YANO, S., and TAKAFUJI, A. 2000. different caterpillar-infested or uninfested food plants of Pieris Flight response of parasitoid toward plant-herbivore complexes: species. J. Chem. Ecol. 23:2935–2954. A comparative study of two parasitoid-herbivore systems on GOUINGUENÉ, S., and TURLINGS, T. C. J. 2002. The effects of abiotic cabbage plants. Appl. Entomol. Zool. 35:87–92. factors on induced volatile emissions in corn plants. Plant SHIOJIRI, K., OZAWA, R., and TAKABAYASHI, J. 2006a. Plant volatiles, Physiol. 129:1296–1307. rather than light, determine the nocturnal behavior of a GOUINGUENÉ, S., PICKETT, J. A., WADHAMS, L. J., BIRKETT, M. A., caterpillar. PloS Biol. 4:e164. doi:10.1371/journal.pbio.0040164. and TURLINGS, T. C. J. 2005. Antennal electrophysiological SHIOJIRI, K., OZAWA, R., MATSUI, K., KISHIMOTO,K.KUGIMIYA, S., responses of three parasitic wasps to caterpillar-induced volatiles and TAKABAYASHI, J. 2006b. Role of the lipoxygenase/lyase from maize (Zea mays), cotton (Gossypium herbaceum), and pathway of host-food plants in the host searching behavior of two cowpea (Vigna unguiculata). J. Chem. Ecol. 31:1023–1038. parasitoid species, Cotesia glomerata and Cotesia plutellae. J. HERDE, M., GARTNER,K.,KÖLLNER,T.G.,FODE,B.,BOLAND,W., Chem. Ecol. 32:969–979. GERSHENZON,J.,GATZ,C.,andTHOLL, D. 2008. Identification SMID,H.M.,VAN LOON,J.J.A.,POSTHUMUS,M.A.,andVET,L.E.M. and regulation of TPS04/GES, an Arabidopsis geranyllinalool 2002. GC-EAG-analysis of volatiles from Brussels sprouts plants synthase catalyzing the first step in the formation of the insect- damaged by two species of Pieris caterpillars: olfactory receptive induced volatile C16-homoterpene TMTT. Plant Cell 20:1152– range of a specialist and a generalist parasitoid wasp species. 1168. Chemoecology 12:169–176. HOBALLAH, M. E., and TURLINGS, T. C. J. 2005. The role of fresh SOKAL, R. R., and ROHLF, F. J. 1995. Biometry: The Principles and versus old leaf damage in the attraction of parasitic wasps to Practice of Statistics in Biological Research (3rd ed.). W. H. herbivore-induced maize volatiles. J. Chem. Ecol. 31:2003–2018. Freeman and Company, New York. IBRAHIM,M.A.,NISSINEN,A.,andHOLOPAINEN, J. K. 2005. TAKABAYASHI, J., and DICKE, M. 1996. Plant-carnivore mutualism Response of Plutella xylostella and its parasitoid Cotesia through herbivore-induced carnivore attractants. Trends Plant plutellae to volatile compounds. J. Chem. Ecol. 31:1969–1984. Sci. 1:109–113. KARBAN,R.,andBALDWIN, I. T. 1997. Induced Responses to TAKABAYASHI, J., DICKE, M., and POSTHUMUS, M. A. 1994. Volatile Herbivory. University of Chicago Press, Chicago. herbivore-induced terpenoids in plant-mite interactions: Variation LOUGHRIN, J. H., MANUKIAN, A., HEATH, R. R., TURLINGS,T.C.J., caused by biotic and abiotic factors. J. Chem. Ecol. 20:1329– and TUMLINSON, J. H. 1994. Diurnal cycle of emission of 1354. induced volatile terpenoids by herbivore-injured cotton plants. TATEMOTO, S., and SHIMODA, T. 2008. Olfactory responses of the Proc. Natl. Acad. Sci. U. S. A. 91:11836–11840. predatory mites (Neoseiulus cucumeris)andinsects(Orius 628 J Chem Ecol (2010) 36:620–628

strigicollis) to two different plant species infested with onion noid and the salicylic acid pathway. J. Exp. Botany 53:1793– thrips (Thrips tabaci). J. Chem. Ecol. 34:605–613. 1799. TURLINGS,T.C.J.,TUMLINSON, J. H., and LEWIS, W. J. 1990. VAN POECKE, R. M. P., POSTHUMUS, M. A., and DICKE, M. 2001. Exploitation of herbivore-induced plant odors by host seeking Herbivore-induced volatile production by Arabidopsis thaliana parasitic wasps. Science 250:1251–1253. leads to attraction of the parasitoid Cotesia rubecula: chemical, TURLINGS,T.C.J.,WÄCKERS, F. L., VET, L. E. M., LEWIS, W. J., and behavioral and gene-expression analysis. J. Chem. Ecol. TUMLINSON, J. H. 1993. Learning of host-finding cues by 27:1911–1928. hymenopterous parasitoids, pp. 51–78, in D. R. Papaj and VET,L.E.M.,LEWIS,W.J.,andCARDÉ, R. T. 1995. Parasitoid foraging A. C. Lewis (eds.). Insect Learning. Ecological and Evolutionary and learning, pp. 65–101, in R. T. Cardé and W. J. Bell (eds.). Perspectives. Chapman & Hall, New York. Chemical Ecology of Insects II. Chapman & Hall, New York. TURLINGS, T. C. J., LENGWILER, U. B., BERNASCONI M. L., and WHITMAN, D. W., and ELLER, F. J. 1992. Orientation of Microplitis WECHSLER, D. 1998. Timing of induced volatile emissions in croceipes (Hymenoptera, Braconidae) to green leaf volatiles: maize seedlings. Planta 207:146–152. Dose–response curves. J. Chem. Ecol. 18:1743–1753. VAN POECKE, R. M. P., and DICKE, M. 2002. Induced parasitoid WITTSTOCK, U., and HALKIER, B. A. 2002. Glucosinolate research in attraction by Arabidopsis thaliana: involvement of the octadeca- the Arabidopsis era. Trends Plant Sci. 7: 263–270.