INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS AND DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

Feeding by Manduca sexta (Lepidoptera, Sphingidae) larvae on Nicotiana attenuata Torr. ex Wats. elicits responses clearly different from those induced by careful mechanical simulation of larval feeding. The wound-induced increase in jasmonic acid (JA) levels is amplified by herbivore feeding and by application of larval oral secretions and regurgitant (R) to mechanical wounds (McCloud and Baldwin, 1997; Schittko et al., 2000), whereas the wound-induced increase in nicotine-accumulation, which strongly correlates with wound-induced JA-levels (Baldwin et al., 1994a, 1997) is suppressed (Baldwin, 1988; McCloud and Baldwin, 1997) by an ethylene-burst released by the plant after herbivore attack (Kahl et al., 2000). Furthermore, herbivore feeding and R application to plant wounds are also known to induce the release of several mono- and sesquiterpenes in N. attenuata (Halitschke et al., 2000; Kahl et al., 2000), which, in turn, are thought to function as an indirect defense, guiding parasitoids to feeding larvae. Moreover, whereas both the volatile release and nicotine-accumulation can be elicited by the application of jasmonates to plants (Baldwin, 1999; Halitschke et al., 2000; Kahl et al., 2000), only the wound-induced nicotine-response is suppressed by application of inhibitors of endogenous JA-biosynthesis (Baldwin et al., 1997; Halitschke et al., 2000). In short, R results in direct and indirect defense responses in this specialist herbivore-plant system, and these responses appear to involve the octadecanoid cascade.

In addition to these well-described phenotypic responses to herbivory, extensive transcriptional re-organization was recently revealed by mRNA differential display of N. attenuata in response to M. sexta feeding. In 1/20th of the insect-responsive transcriptome, 27 genes displayed altered expression patterns (Hermsmeier et al., 2001). A subset of seven genes was found to differentially respond to R as compared with mechanical damage. Larval R of M. sexta and Manduca quinquemaculata antagonistically (type I genes) or synergistically (type II genes) modified wound-induced transcriptional responses of these seven genes (Schittko et al., 2001). Given that chemical attributes of larval feeding mediate extensive changes in transcript accumulation and phenotypic responses, characterization of active components of R is of great interest.

Two types of elicitors have been identified in lepidopteran R that result in the release of plant volatiles responsible for attracting parasitic wasps. First, an enzymatic elicitor, -glucosidase, isolated from Pieris brassicae R, was shown to elicit the release of parasitoid-attracting volatile emissions from cabbage leaves. This elicitor is thought to release signal compounds by cleaving stored glycosidic precursors (Hopke et al., 1994; Mattiacci et al., 1995). Second, volicitin, N-(17-hydroxylinolenoyl)-L-Gln, a fatty acid-amino acid conjugate (FAC) identified in the R of Spodoptera exigua induces the release of volatiles in corn plants (Zea mays) comparable with that induced by larval feeding (Alborn et al., 1997; Turlings et al., 2000). Volicitin and several structurally related FACs have been identified in R of different lepidopteran species (Paré et al., 1998; Pohnert et al., 1999a; Alborn et al., 2000).

Because linolenic acid is a precursor of JA in the octadecanoid cascade, the inducing activity of the FACs may be due to the supply of fatty acid substrates introduced to the plant after hydrolytic cleavage of the FAC amide-bond (Koch et al., 1999). This mechanism is supported by investigations with the lima bean (Phaseolus lunatus) in which (a) Treatment of leaves with free linolenic acid results in the release of volatiles comparable with that elicited by treatments with N-linolenoyl-L-Gln (18:3-Gln), and (b) treatment of the leaves with inhibitors of the octadecanoid pathway suppresses the volatile response elicited by the application of free fatty acids (FAs; Koch et al., 1999). Unfortunately, in these studies the FAs and FACs were supplied in concentrations far exceeding those found in larval R. Moreover, other mechanisms that do not invoke substrate supply for the octadecanoid pathway can account for the activity of the conjugates. For example, the conjugates may be recognized by specific receptors that subsequently trigger the octadecanoid pathway.

Here we identify and quantify the FAs and FACs in the R of M. sexta and M. quinquemaculata larvae, synthesize these FACs, and investigate their role in eliciting the volatile release, endogenous JA-accumulation and changes in transcript accumulation of six mRNAs of N. attenuata that are known to be specifically altered by R from M. sexta and M. quinquemaculata. We critically evaluate the biological roles of the identified compounds by removing all FAs and FACs in R by anion-exchange chromatography and add back synthetic FAs and FACs to the ion-exchanged R (exR) at naturally occurring concentrations.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS AND DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

Chemical Analysis of R

Oral secretions and regurgitant of M. sexta larvae fed on N. attenuata foliage were separated by HPLC (gradient C-18: CH₃CN/H₂O; 0.5% [v/v] HAc; 0.7 mL/min: 0% [v/v] CH₃CN, 20-25 min 100% [v/v] CH₃CN) and analyzed by atmospheric pressure chemical ionization-mass spectrometry. Analysis of fragmentation patterns (Pohnert et al., 1999a) revealed the presence of a series of FACs as minor constituents in the medium polar region of the chromatogram (Fig. 1A). Adjustment of the separation conditions (HPLC gradient C18: CH₃CN/H₂O; 0.5% [v/v] HAc; 0.7 mL/min: 40% [v/v] CH₃CN, 7 min 68% [v/v] CH₃CN, 18 min 80% [v/v] CH₃CN, 28 min 100% [v/v] CH₃CN) allowed us to separate the FACs (Fig. 1, B and C) and identify seven structurally related FACs by comparison of their retention times and APCI mass spectra with synthetic references (see "Materials and Methods"). The FACs in the R of M. sexta are dominated by Glu conjugates of C16- and C18-fatty acids (Fig. 1B, compounds 3, 5, and 8), which contrasts with the composition of seven other lepidopteran larvae R, which in turn, are dominated by the Gln conjugates of these fatty acids (Paré et al., 1998; Pohnert et al., 1999a; Alborn et al., 2000). The Gln conjugates (Fig. 1B, compounds 1, 4, and 6) are relatively minor components of M. sexta R compared with the corresponding Glu conjugates. This unusual composition was also found in nearly identical relative ratios in the R of another specialist herbivore of N. attenuata, M. quinquemaculata (Fig. 1C). Remarkably, no functionalized FACs (e.g. volicitin), often present in the R of lepidopteran larvae (Alborn et al., 1997, 2000; Pohnert et al., 1999a; Turlings et al., 2000), could be detected in the R of M. sexta or M. quinquemaculata.

View larger version (32K): [in this window] [in a new window]   Figure 1.   Scheme of experimental setup of the ion-exchange approach and structures of identified FACs (left) and HPLC-MS-Base peak profiles of 10-µL injections of test solutions (right): A, oral secretions and regurgitant (R) from M. sexta larvae. HPLC gradient (C18): CH3CN/H2O; 0.5% (v/v) HAc; 0.7 mL/min: 0% (v/v) CH3CN, 20 to 25 min 100% (v/v) CH3CN. Separation of the FACs in M. sexta (B) and M. quinquemaculata (C) R: HPLC gradient (C18): CH3CN/H2O; 0.5% (v/v) HAc; 0.7 mL/min: 40% (v/v) CH3CN, 7 min 68% (v/v) CH3CN, 18 min 80% (v/v) CH3CN, 28 min 100% (v/v) CH3CN. 1, N-linolenoyl-L-Gln; 2, unidentified; 3, N-linolenoyl-L-Glu; 4, N-linoleoyl-L-Gln; 5, N-linoleoyl-L-Glu; 6, N-palmitoyl-L-Gln; 7, N-oleoyl-L-Gln; and 8, N-palmitoyl-L-Glu. Base peak profiles of ion-exchanged M. sexta R (D) and mixture of synthetic FACs at concentrations found in M. sexta R (E) analyzed with the HPLC gradient as in B and C.

The total concentration of FACs in the R of M. sexta reared on fresh N. attenuata foliage in the laboratory varied from 0.6 to 1.2 mM. Analysis of free fatty acids showed two major FAs, linolenic acid (18:3) at a concentration of 1.3 mM, and linoleic acid (18:2) at a concentration of 0.4 mM. Ion-exchange chromatography on R and the synthetic mixture of FACs removed all detectable amounts of FACs (Fig. 1D, detection limit = 30 nM) and FAs (detection limit = 200 nM).

Induction of cis--Bergamotene Release

Treatment of standard puncture wounds on a single leaf with M. sexta R elicited significant increases in whole-plant (WP) emissions of cis--bergamotene compared with untreated plants or plants that were comparably damaged but had water applied to their puncture wounds (Fig. 2A, ANOVA F_{9, 70} = 14.565, P < 0.0001). Application of M. quinquemaculata R induced elevated WP cis--bergamotene emissions comparable with those elicited by R of the closely related species M. sexta. The volatile-inducing activity of R was completely eliminated by ion-exchange chromatography. Application of the exR, which did not contain any detectable amounts of the analyzed FAs or FACs (Fig. 1D), did not induce cis--bergamotene emissions. However, the addition of synthetic FACs (Fig. 1E) to the ion-exchanged R at their original concentrations completely restored the volatile-inducing activity. Moreover, an aqueous solution of synthetic FACs at concentrations found in R was as active as the larval R. The free fatty acids, when applied as aqueous solution or dissolved in the exR at concentrations found in M. sexta R, did not induce volatile emissions, even though the molar concentrations of the applied test solutions were higher than those of the FAC treatment. From these results, we conclude that the FACs found in M. sexta R, but not the FAs, are necessary and sufficient for the elicitation of the volatile release in N. attenuata plants. However, in excised leaves of lima beans, linolenic acid and its amino acid conjugate 18:3-Gln both induce homoterpene emissions (Koch et al., 1999), which were not detected among the volatiles released by N. attenuata (Halitschke et al., 2000). Even though the compounds were applied at higher concentrations than in this study, different mechanisms of volatile-induction may exist in different plant-herbivore systems.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Mean (±SE) WP cis--bergamotene trapped per hour, per liter air sampled from individual (eight per treatment) N. attenuata (A) plants and mean (±SE) JA concentrations (n.d., not determined) of node two leaves of four replicate plants per treatment 35 min (time of maximum JA induction; B) after the node two leaf was wounded and treated with 20 µL of the following test solutions: water (W), oral secretions and regurgitant from M. sexta larvae (M.s.) or M. quinquemaculata (M.q.), exR, Triton X-100 in water (T), FAC in concentrations found in R and FA in concentrations found in R in the triton solution or in exR. Control plants (C) remained undamaged. Stars represent significantly (P < 0.05) increased emissions as compared with wounded plants treated with water (W) as determined by Fisher's protected least significant difference from ANOVAs.

Induction of Endogenous JA Burst

As previously described (Kahl et al., 2000; Schittko et al., 2000), application of M. sexta R to puncture wounds on N. attenuata leaves transiently elicits higher JA concentrations than does the addition of water to identical puncture wounds (Fig. 2B, ANOVA F_{8, 26} = 14.551, P < 0.0001). The ion-exchange treatment removed the JA inducing activity of R so that application of exR did not amplify the wound-induced JA accumulation. We tested the mixture of synthetic FACs (Fig. 1E) in a triton-containing aqueous solution and in exR at concentrations comparable to those found in R. Both solutions elicited a dramatic amplification of the wound-induced increase in JA concentrations as observed after application of larval R. No amplification of JA induction was observed after treatments with triton control solution and FA mixtures in exR or triton-containing aqueous solution compared with the wound treatment. These results demonstrate that other induction-mechanisms than a simple supply of FA as substrate for endogenous JA biosynthesis must account for the response activation processes in the N. attenuata-M. sexta system. The JA response is known to be very sensitive to M. sexta R, which retain their activity even when diluted to 1/1,000 with water (Schittko et al., 2000). This sensitivity also argues against a substrate supply mechanism, because the quantity of FAs delivered to a leaf as FACs in this highly diluted, but still active R is not sufficient to supply the quantity of fatty acid substrate required for the observed endogenous JA burst.

Changes in Transcript Accumulation

We also investigated the effect of the identified R-components on transcript accumulation. We applied the same test solutions as described for volatile and JA analysis and examined changes in transcript accumulation of a set of genes that specifically respond to R (Schittko et al., 2001). We found that FACs, supplied either in triton-containing aqueous solution or exR, caused specific changes in transcript accumulation exactly as larval R did, whereas transcript accumulation in response to FA solutions (in triton-containing aqueous solution or exR) did not differ from wound-induced transcript levels (Fig. 3). As described by Schittko et al. (2001), two types of expression patterns were distinguished. Wound-induced transcript accumulation was specifically repressed (type I) by R or FAC treatments for Thr deaminase (pDH14.2; Hermsmeier et al., 2001) and an unknown gene encoded by pDH23.5, whereas the wound response of the other four investigated genes was amplified (type II; Fig. 3). Wound-induced transcript accumulation of pathogen-induced oxygenase (pDH41.6; Hermsmeier et al., 2001) was further up-regulated, and wound-suppressed transcript accumulation of genes encoded by pDH61.1 (similar to the tomato gene for a light harvesting complex II subunit, lhb C1; Schwartz et al., 1991), pDH39.1, and pDH68.1 was further down-regulated (Fig. 3). No changes in transcript accumulation compared with the wound-treatment were observed after application of exR or triton control solution (Fig. 3).

View larger version (114K): [in this window] [in a new window]   Figure 3.   Northern analysis of transcript accumulation in response to different test solutions. The node two leaf of five replicate rosette-stage plants was continuously wounded and supplied with water (W), M. sexta larval oral secretions and regurgitant (R), exR, Triton X-100 in water (T), FACs or FAs in concentrations found in R, dissolved in either T or exR for 80 min, creating one row of puncture wounds every 20 min, and harvesting 20 min after the final treatment. Untreated node two leaves were harvested as controls (C). Hybridization with an 18S rRNA probe indicates equal loading.

	CONCLUSION

TOP ABSTRACT INTRODUCTION RESULTS AND DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

In this study we identified FACs in R of the closely related herbivores M. sexta and M. quinquemaculata that, when applied to leaves of N. attenuata, are sufficient to activate the three investigated herbivore-specific plant responses in the signal transduction hierarchy (JA accumulation, changes in transcript accumulation, and volatile release). The chromatographic inactivation of a complex mixture of elicitors by ion exchange, and the restoration of activity by re-addition of synthetic FACs, powerfully demonstrates the biological activity of FACs. Future investigations are necessary to examine the contribution of each individual compound to the activity of R and to elucidate the structure-function relationship of the identified FACs.

Because FAs in Manduca R were not active elicitors of herbivore-induced responses in N. attenuata, the process of their conjugation with amino acids to form FACs in the insect (Paré et al., 1998) suggests that the insect controls the production of its own elicitors. Although the function of FACs in insects is not absolutely clarified, FACs are likely to function as emulsifiers and detergents (Collatz and Mommsen, 1974). Hence, the plant distinguishes the feeding activity of this herbivore from other agents that cause leaf damage by recognizing compounds essential for the insect's digestive processes.

The FACs could serve as useful tools for the study of plant-herbivore interactions because they allow researchers to uncouple herbivore-specific plant responses from herbivory and the damage it causes.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS AND DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

Plant Growth and Insect Rearing

Nicotiana attenuata Torr. ex Wats. seeds (collected at the DI ranch, UT, T40S R19W, section 10, 1988) were germinated in smoke-treated soil (Baldwin et al., 1994b). For JA experiments, seedlings were transferred to soil and grown for 3 to 4 weeks. Plants for volatile experiments and northern analysis were grown as described in Hermsmeier et al. (2001) in no-nitrogen hydroponic solution (Baldwin and Schmelz, 1994). To provide nitrogen, 2 mL of 1 mM KNO₃ solution were added to each 1-L chamber, followed by another 1 mL, 10 to 12 d later (the day before the experiment started). All plants were grown under a 32°C, 16-h/27°C, 8-h day/night regime and were in the rosette-stage of growth at the time of the experiment.

Manduca sexta (Lepidoptera, Sphingidae) larvae were hatched from eggs (Carolina Biological Supply, Burlington, NC) and reared on fresh N. attenuata foliage under a 28°C, 16-h/8-h day/night regime. Eggs of Manduca quinquemaculata Haworth were collected at Pahcoon Springs Burn in Utah in 1999, and the larvae were reared in the laboratory as described for M. sexta.

Analysis of Oral Secretions and Regurgitant

Atmospheric pressure chemical ionization-HPLC-MS analyses of M. sexta and M. quinquemaculata R and synthetic FAC mixtures were performed as previously described (Pohnert et al., 1999a) using a reversed-phase HPLC separation (LiChrospher 100 RP-18, 5 µm, 250 × 4 mm, Merck, Darmstadt, Germany) with acetonitrile, water, and acetic acid as eluent. Detection and identification of the FACs was performed with a Finnigan (San Jose, CA) LCQ ion trap MS (Atmospheric pressure chemical ionization, vaporizer 560°C) by comparison with synthetic standards. Details on the LC-MS procedure, the synthesis and spectroscopic data of FACs 1 and 3 to 7 are published elsewhere (Pohnert et al., 1999a). Synthesis of the newly identified FAC 8, found in Manduca R, proceeded from free palmitic acid and unprotected Glu following a published protocol (Pohnert et al., 1999b).

The following selected spectroscopic data were obtained by MS and NMR analyses of N-palmitoyl-L-glutamate (16:0-Glu, 8): [¹H]NMR (CD₃OD, 500 MHz) : 0.9 (t, J = 7.1, 3H); 1.25 to 1.35 (m, 26H); 1.62 (t, J = 7.2, 2H); 1.89 to 1.97 (m, 1H); 2.14 to 2.22 (m, 1H); 2.25 (t, J = 7.5, 1H; 2.4 (t, J = 7.8, 1H); 4.43 (dd, J = 5, 9.17, 1H); [¹³C]NMR (CD₃OD, 125 MHz) : 14.54; 23.83; 27.01; 27.94; 30.36; 30.56; 30.57; 30.74; 30.83; 30.83; 30.85; 30.86; 30.87; 30.88; 30.89; 30.9; 31.35; 53.01; 175.08; 176.39; 176.58, MS (70 eV); 385(M⁺, 8); 3.67(10); 341(8); 256(7); 239(10); 189(89); 171(30); 130(47); 102(100); 84(59); 57(65); and HR-MS: m/z calculated for C₂₁H₃₉NO₅: 385.2828, observed: 385.2828.

Free fatty acids were extracted from 80 µL of R after addition of 4 µg cis-10-nonadecenoic acid as an internal standard with a ternary solvent composition (water-methanol-chloroform), as described by Bligh and Dyer (1959). The extract was derivatized with 600 µL of freshly prepared solution of diazomethane in ether. The solvent was evaporated and the residue dissolved in 20 µL of N-methyl-N-trimethylsilyltrifluoroacetamide (Macherey-Nagel, Düren, Germany). One-microliter aliquots were injected and analyzed on a Saturn 2000 GC-MS (Varian, Walnut Creek, CA). Methylated FAs were separated on a 30-m × 0.25-mm DB-Wax column (0.25-µm film thickness; J&W Scientific, Folsom, CA). The injector temperature was held at 225°C, and the column oven temperature was programmed as follows: initial column temperature 120°C held for 3 min, ramped from 120°C to 170°C at 10°C/min, held at 170°C for 6 min, ramped from 170°C to 230°C at 3°C/min, ramped from 230°C to 240°C at 20°C/min, and finally held at 240°C for 10 min. The carrier gas flow throughout the program was maintained at 1 mL/min.

Test Solutions

Oral secretions and regurgitant were collected with Teflon tubing connected to a vacuum from 4th to 5th instar M. sexta and M. quinquemaculata larvae reared on N. attenuata leaves and stored under argon at 80°C. They were diluted 1:1 (v:v) with water prior to the treatment. To remove FAs and FACs, 400 µL of R were eluted consecutively through four ion-exchange columns containing 400 mg of the basic anion-exchange resin Amberlite IRA-400 (Sigma, Steinheim, Germany). The final eluate was called "ion-exchanged oral secretions and regurgitant" (exR). For application of FAs and FACs at concentrations similar to those found in R, aqueous solutions containing 0.005% (w/w) Triton X-100 (Fluka, Buchs, Switzerland) were prepared and diluted 1:1 (v:v) with water or exR prior to the treatment. The FA solution contained 120 ng µL¹ (0.4 mM) of linoleic acid and 350 ng µL¹ (1.3 mM) of linolenic acid. A mixture of the four main FACs was prepared at concentrations of 50 ng µL¹ (0.12 mM) N-linolenoyl-L-Gln (18:3-Gln, 1), 138 ng µL¹ (0.34 mM) N-linolenoyl-L-Glu (18:3-Glu, 3), 41 ng µL¹ (0.10 mM) N-linoleoyl-L-Gln (18:2-Gln, 4), and 26 ng µL¹ (0.06 mM) N-linoleoyl-L-Glu (18:2-Glu, 5). The FAC mixture differed from the FAC composition of R in that it contained approximately 4 times the amount of FAC 4 and did not contain FAC 8. An aqueous solution containing 0.0025% (w/w) Triton X-100 was used to control for the potential inducing activity of this detergent.

Volatile and JA Analysis

To determine the JA- and volatile-inducing activity of different test solutions, 20-µL samples were added to the leaf lamina immediately after three rows of puncture wounds were created on each leaf half with a fabric pattern wheel (Dritz, Spartanburg, SC). All treatments were applied to a single leaf node (node two) of each plant with the youngest fully expanded leaf, the leaf that had just completed the source-sink transition (as defined in Wait et al., 1998) defining node one.

Leaves scheduled for JA analysis were harvested 35 min after the induction of four replicate plants per treatment. Jasmonate concentrations were determined with ¹³C_1,2-JA as an internal standard and analyzed by GC-MS as described by Schittko et al. (2000).

Volatile collection commenced 24 h after the treatment and lasted for 8 h. Eight replicate plants per treatment were covered with 1-L open-top WP volatile collection chambers, and volatiles were collected by adsorption on 30 mg of SuperQ at a mean flow rate of 300 mL min¹ through the WP-chamber and analyzed by GC-MS as previously described (Halitschke et al., 2000). Because cis--bergamotene is the most consistently systemically released volatile from different genotypes of N. attenuata (Halitschke et al., 2000), we used the WP emission of this sesquiterpene to quantify the induced volatile response. The released amounts were calculated from peak areas using calibration curves with tetraline as an internal standard and normalized to trapping efficiencies by peak areas of a trapped sesquiterpene that was abundant in the surrounding growth room air (Halitschke et al., 2000). Given that induced volatile emissions are known to be influenced by many environmental factors (Loughrin et al., 1994; Takabayashi and Dicke, 1996; Paré and Tumlinson, 1998; Halitschke et al., 2000), our open-flow trapping system has the distinct experimental advantage of allowing the simultaneous analysis of 80 plants. Statistical comparisons of volatile and JA data were performed with protected contrasts (Fisher's protected least significant difference) from ANOVAs.

Northern Analysis

A fabric pattern wheel (Dritz, Spartanburg, SC) was used to create one row of puncture wounds in parallel to the leaf midrib every 20 min, and 5-µL aliquotes of the respective test solution were applied to the fresh wounds. A total number of five rows were applied to a leaf at node two and the treated leaf of five replicate plants per treatment was harvested 20 min after the last wounding. Total cellular RNA was isolated according to Pawlowski et al. (1994). Agarose gel electrophoresis, northern blotting, probe labeling, and hybridizations were performed as described in Hermsmeier et al. (2001). GenBank accession numbers of the template sequences are AW191811 (pDH14.2), AW191815 (pDH23.5), AW191819 (pDH39.1), AW191821 (pDH41.6), AW191826 (pDH61.1), AW191828 (pDH64.4), and AW191830 (pDH68.1). Hybridization with an 18S rRNA probe (pDH64.4) was used to monitor loading. The wound-induced response of the other six mRNAs of N. attenuata is known to be specifically altered by M. sexta and M. quinquemaculata R (Schittko et al., 2001).