J Chem Ecol (2010) 36:905–913 DOI 10.1007/s10886-010-9825-z

Differential Effects of Indole and Aliphatic Glucosinolates on Lepidopteran Herbivores René Müller & Martin de Vos & Joel Y. Sun & Ida E. Sønderby & Barbara A. Halkier & Ute Wittstock & Georg Jander

Received: 17 March 2010 / Revised: 15 June 2010 / Accepted: 18 June 2010 / Published online: 9 July 2010 # Springer Science+Business Media, LLC 2010

Abstract Glucosinolates are a diverse group of defensive secondary metabolites that is characteristic of the Brassicales. Arabidopsis thaliana (L.) Heynh. (Brassicaceae) lines with mutations that greatly reduce abundance of indole glucosinolates (cyp79B2 cyp79B3), aliphatic glucosinolates (myb28 myb29), or both (cyp79B2 cyp79B3 myb28 myb29) make it possible to test the in vivo defensive function of these two major glucosinolate classes. In experiments with Lepidoptera that are not crucifer-feeding specialists, aliphatic and indole glucosinolates had an additive effect on Spodoptera exigua R. Müller : U. Wittstock Institut für Pharmazeutische Biologie, Technische Universität Braunschweig, Braunschweig, Germany M. de Vos : J. Y. Sun : G. Jander (*) Boyce Thompson Institute for Plant Research, 1 Tower Road, Ithaca, NY 14853, USA e-mail: [email protected] I. E. Sønderby : B. A. Halkier VKR-Research Centre for Pro-Active Plants, Department of Plant Biology and Biotechnology, Faculty of Life Sciences, University of Copenhagen, Copenhagen, Denmark Present Address: M. de Vos Keygene N.V., Agro Business Park 90, 6708 PW Wageningen, The Netherlands Present Address: I. E. Sønderby Department of Medical Genetics, Oslo University Hospital Ullevaal, Postboks 4956 Nydalen, 0424 Oslo, Norway

(Hübner) (Lepidoptera: Noctuidae) larval growth, whereas Trichoplusia ni (Hübner) (Lepidoptera: Noctuidae) and Manduca sexta (L.) (Lepidoptera: Sphingidae) were affected only by the absence of aliphatic glucosinolates. In the case of two crucifer-feeding specialists, Pieris rapae (L.) (Lepidoptera: Pieridae) and Plutella xylostella (L.) (Lepidoptera: Plutellidae), there were no major changes in larval performance due to decreased aliphatic and/or indole glucosinolate content. Nevertheless, choice tests show that aliphatic and indole glucosinolates act in an additive manner to promote larval feeding of both species and P. rapae oviposition. Together, these results support the hypothesis that a diversity of glucosinolates is required to limit the growth of multiple insect herbivores. Key Words Glucosinolates . Lepidoptera . Oviposition . Feeding . Arabidopsis thaliana . Manduca sexta . Trichoplusia ni . Pieris rapae . Plutella xylostella . Spodoptera exigua

Introduction The glucosinolate-myrosinase system is a characteristic defense of the Brassicaceae and related plant families (Halkier and Gershenzon, 2006). In the absence of herbivore attack, glucosinolates, β-thioglucoside-N-hydroxyiminosulfates with diverse, amino acid-derived side chains, are stored separately from the activating enzyme, myrosinase (βthioglucoside glucohydrolase, EC 3.2.1.147). Tissue disruption during herbivory brings glucosinolates into contact with myrosinase, resulting in release of the glucose moiety and formation of biologically active breakdown products, including nitriles, isothiocyanates, thiocyanates, oxazolidine-2thiones, and epithionitriles (Wittstock and Burow, 2010).

906

More than 120 different glucosinolate side chains have been identified (Fahey et al., 2001), and about forty of these are found in natural isolates of the model plant Arabidopsis thaliana (L.) Heynh. (Brassicaceae) (Kliebenstein et al., 2001; Reichelt et al., 2002). The metabolic diversity of glucosinolates, combined with variation in the breakdown pathways (Wittstock and Burow, 2010), likely results in several hundred defensive metabolites that can be formed during herbivory. Genetic modification of the glucosinolate breakdown pathway shows that qualitative changes affect herbivore resistance (Jander et al., 2001; Lambrix et al., 2001; Burow et al., 2006). Less is known about the specific defensive function of different glucosinolate classes, e.g., those derived from tryptophan, methionine, or phenylalanine. Increased production of both tryptophan-derived (indole) and methionine-derived (aliphatic) glucosinolates through overexpression of the MYB51 and MYB28 transcription factors, respectively, reduced herbivory by Spodoptera exigua (Hübner) (Lepidoptera: Noctuidae) on A. thaliana (Gigolashvili et al., 2007a, b). Both total glucosinolate content and, more specifically, methylsulfinylalkylglucosinolate levels were negatively correlated with S. exigua weight gain (Mewis et al., 2005; Rohr et al., 2006). The related Spodoptera littoralis (Boisduval) (Lepidoptera: Noctuidae) showed improved growth in the absence of indole glucosinolates (Schlaeppi et al., 2008). Similarly, reduced aliphatic glucosinolate content increased the growth of the generalist lepidopteran herbivore Mamestra brassicae (L) (Lepidoptera: Noctuidae) (Beekwilder et al., 2008). Some crucifer-specialist insects, including Pieris rapae (L.) (Lepidoptera: Pieridae) and Plutella xylostella (L.) (Lepidoptera: Plutellidae), not only circumvent the release of toxic breakdown products from glucosinolates (Ratzka et al., 2002; Wittstock et al., 2004) but also co-opt these defensive metabolites as attractive signals. Glucosinolates elicit specific responses in maxillary chemoreceptors and stimulate feeding in both P. rapae and P. xylostella larvae (Verschaffelt, 1910; Thorsteinson, 1953; Van Loon et al., 2002; Miles et al., 2005). Among ten glucosinolates with side chains derived from tryptophan, methionine, or phenylalanine, indol-3-ylmethylglucosinolate was identified as the strongest oviposition stimulant for P. rapae oviposition in vitro (Renwick et al., 1992; Städler et al., 1995). Although most experiments show that glucosinolates have positive effects on P. xylostella feeding and oviposition (Sarfraz et al., 2006), individual glucosinolates that were increased in transgenic lines did not have unequivocal effects (Sarosh et al., 2010), and 3-butenylglucosinolate levels in A. thaliana field studies were negatively correlated with female adult weight (Bidart-Bouzat et al., 2005). In many A. thaliana accessions, the side chains of the most abundant foliar glucosinolates are derived from methionine or tryptophan (Wittstock and Halkier, 2002).

J Chem Ecol (2010) 36:905–913

Whereas plants with knockout mutations of both CYP79B2 and CYP79B3, two cytochrome P450s that lead to the production of indole-3-acetaldoxime, are blocked in the production of indole glucosinolates (Zhao et al., 2002), double mutants of the MYB28 and MYB29 transcription factors have very low levels of aliphatic glucosinolates (Sønderby et al., 2007; Beekwilder et al., 2008). Quadruple mutants (cyp79B2 cyp79B3 myb28 myb29) are nearly devoid of both major classes of foliar glucosinolates found in A. thaliana Columbia-0 (Col-0; Sun et al., 2009). Together, these mutant lines provide a unique opportunity to investigate the relative in vivo function of aliphatic and indole glucosinolates in plant defense against different herbivores.

Methods and Materials Plants and Growth Conditions A. thaliana Col-0 wildtype seeds were obtained from the Arabidopsis Biological Resource Center (ABRC, www.arabidopsis.org). The cyp79B2 cyp79B3 mutant (Zhao et al., 2002) was kindly supplied by J. Celenza (Boston University, Boston, MA, USA). Creation of myb28 myb29 and cyp79B2 cyp79B3 myb28 myb29 mutants has been described previously (Sønderby et al., 2007; Sun et al., 2009). All mutations are in the Col-0 genetic background. A. thaliana plants for S. exigua, T. ni, P. xylostella, and Manduca sexta (L.) (Lepidoptera: Sphingidae) experiments were grown in a Conviron (Winnipeg, Canada) chamber in 12× 12 cm pots using Cornell Mix [by weight 56% Peatmoss, 35% Vermiculite, 4% Lime, 4% Osmocoat slow-release fertilizer (Scotts, Marysville, OH, USA), and 1% Unimix (Peters, Allentown, PA, USA)] at 23°C, 60% relative humidity, with a light intensity of 180 μmolm−2 s−1 photosynthetic photon flux density and a 16:8 h light:dark photocycle. Arabidopsis thaliana for P. rapae experiments were grown on soil composed of 80% potting soil (Compo, Münster, Germany), 10% sand, 10% Perligran (Knauf Perlite, Dortmund, Germany), and Triabon (Compo, Münster, Germany) and Sierrablen (Scotts, Heerlen, The Netherlands) as fertilizer at 22°C, 60–70% relative humidity, 300 μmolm−2 s−1 photosynthetic photon flux density, and a 10:14 h light:dark photocycle. Insects and Growth Conditions P. xylostella, T. ni, and S. exigua eggs were purchased from Benzon Research (Carlisle, PA, USA). M. sexta eggs were kindly supplied by M. del Campo (Cornell University) and J. Beal (Boyce Thompson Institute). P. xylostella for feeding choice experiments were reared in paper cups (0.5 L, International Paper, Memphis, TN, USA) on artificial diet (Southland Products, Lake Village, AK, USA) at 23°C. P. rapae were

J Chem Ecol (2010) 36:905–913

maintained on Brussels sprouts (Brassica oleracea var. gemmifera) in a chamber at 24°C, 65% relative humidity and a 16:8 h light:dark photocycle. Larval Feeding no Choice Tests Lepidopteran eggs (T. ni, M. sexta, and S. exigua, and P. xylostella) were hatched at 23°C in paper cups (0.5 l, International Paper, Memphis, TN, USA) with a moist paper towel (C-fold white, Kimberly-Clark, Dallas, TX, USA). Individual neonate larvae were placed onto paired 18-d-old mutant and wildtype plants growing together in the same pot, and each individual plant was covered with a mesh cup. Larvae were harvested at 9 d (T. ni), 8 d (M. sexta), or 10 d (S. exigua). The length of the experiment for each species was chosen based on the differing intrinsic growth and feeding rates of the larvae. In each case, even the largest larvae had not consumed the entire A. thaliana plant at the end of the experiment. Plutella xylostella were harvested at pupation (8–10 d). Larvae and pupae were frozen on dry ice and were lyophilized overnight (∼12 h), and weighed. Each experiment was conducted two to four times, giving similar results. For P. rapae larval growth experiments, one neonate larva was placed in a pot containing one 6-wk-old A. thaliana Col-0 wildtype or mutant plant, respectively, covered with a perforated plastic bag and placed at 24°C and 65% relative humidity. The weight of the larvae was recorded after 10 d of feeding. For P. rapae pupation experiments, groups of ten neonate larvae were placed in a pot containing four 6-wk-old A. thaliana Col-0 wildtype or mutant plants, respectively, covered with a perforated plastic bag and placed at 24°C and 65% relative humidity. Additional food plants were provided on demand. Each experiment was run twice independently. Larval Feeding Choice Tests For P. xylostella choice tests, leaves of similar age and size were harvested from paired 20-d-old mutant and wildtype A. thaliana plants growing together in the same pot. Leaf areas were scanned, and paired leaves were placed in Petri dishes (9 cm diam, ThermoFisher, Waltham, MA, USA), with their petioles embedded in 0.8% Phytagar (Invitrogen, Carlsbad, CA, USA). Larvae for experiments were raised for 3 d on artificial diet. A single larva was placed with each leaf pair and, after 3 d, leaves were collected and scanned again. Leaf area consumed was calculated using ImageJ (Rasband, 1997–2007). Leaf plug choice experiments were conducted in a similar manner. A single 3-d-old larva was placed in a Petri dish (4.5 cm diam, ThermoFisher, Waltham, MA, USA) with paired equal-sized leaf disks of each genotype (0.6 cm2) on moist filter paper. After 24 h, leaves were scanned and leaf areas were calculated with ImageJ (Rasband, 1997–2007). Intact leaf

907

and leaf plug experiments were conducted twice independently. For P. rapae feeding choice tests, one third-instar P. rapae larva was allowed to feed on a pot containing two 6-wk-old plants of two A. thaliana lines that were being compared. After 24 h, the plant rosettes were harvested and scanned using a desktop scanner. The removed leaf area was completed and calculated with ImageJ (Rasband, 1997–2007). These experiments were conducted twice independently. Oviposition Choice Tests For P. rapae oviposition experiments, adults were kept without plants for 6 d after eclosion to ensure mating and maximize egg production. A single P. rapae female then was introduced to a 39×28×28 cm plastic box containing two 7-wk-old plants, each from a different plant genotype. After 24 h, the eggs laid per plant were counted. These experiments were conducted twice independently. Glucosinolate Analysis Glucosinolate extraction and HPLC analysis of desulphoglucosinolates was performed as previously described (Kliebenstein et al., 2001; Burow et al., 2006; Hansen et al., 2007), using half rosettes of 6-wk-old plants. Data Analysis Statistical analyses were conducted with JMP (SAS Institute, Cary, NC, USA) and SigmaStat 3.1 (Systat Software, San José, CA, USA).

Results Under the growth conditions used for insect experiments, cyp79B2 cyp79B3 mutants contain little or no indole glucosinolates, myb28 myb29 mutants have greatly decreased aliphatic glucosinolates, and cyp79B2 cyp79B3 myb28 myb29 quadruple mutants are deficient in both indole and aliphatic glucosinolates (Table 1; Sun et al., 2009). To determine the relative effects of indole and aliphatic glucosinolate decreases on non-crucifer-specialist Lepidoptera, neonate S. exigua, M. sexta, and T. ni were caged on individual mutant and wildtype Col-0 A. thaliana plants. Spodoptera exigua larval weight was increased (Fig. 1a) on both cyp79B2 cyp79B3 and myb28 myb29 mutants compared to wildtype Col-0. An additive effect was observed in that the larvae grew better on the cyp79B2 cyp79B3 myb28 myb29 quadruple mutant than on either of the double mutants (Fig. 1a). In contrast, M. sexta and T. ni weight gain was increased relative to controls only on lines deficient in aliphatic glucosinolates (myb28 myb29 and the quadruple mutant; Fig. 1b, c). Although the effect was not

908

J Chem Ecol (2010) 36:905–913

Table 1 Foliar glucosinolates of Col-0 wildtype and mutants Col-0a

glucosinolate

3MSP 4MTB 4MSB 5MSP 7MTH 7MSH 8MTO 8MSO Total aliphatic I3M 4MOI3M NMOI3M 4OHI3M Total indole

cyp79B2 cyp79B3a

Meanb

SE

Meanb

1.41 3.52 8.52 0.30 0.27

0.23 0.55 1.34 0.05 0.10

1,44 2.36 9.55 0.35 0.13

0.18 0.36 0.89 15.45 1.30 0.02 0.41 0.57 2.29

0.04 0.09 0.16 2.56 0.14 0.01 0.11 0.12 0.37

0.20 0.20 0.80 15.04 0.02 0.01 0.01 0.03 0.06

myb28 myb29a

QKOa

Meanb

SE

Meanb

SE

0.09 0.24 0.71 0.02 0.01

0.07 0.29 0.48 ND 0.02

0.06 0.29 0.48

0.00 0.00 0.00

0.02

0.01 0.00 0.01 ND 0.00

0.01 0.02 0.04 1.15 0.00 0.00 0.00 0.00 0.01

ND 0.03 0.10 0.98 1.60 0.02 0.62 0.58 2.82

0.03 0.07 0.95 0.28 0.01 0.16 0.06 0.50

ND 0.00 ND 0.02 0.00 ND ND 0.00 0.00

SE

0.00 0.00 0.01 0.00

0.00 0.00

a

nmol/mg fresh weight, b significant differences relative to wildtype Col-0 are marked in bold (P