Forest tent caterpillars (Malacosoma disstria) induce local and systemic diurnal emissions of terpenoid volatiles in hybrid poplar (Populus trichocarpa × deltoides): cDNA cloning, functional characterization, and patterns of gene expression of (−)-germacrene D synthase, PtdTPS1

Authors

  • Gen-ichiro Arimura,

    1. Biotechnology Laboratory, University of British Columbia, 6174 University Boulevard, Vancouver, BC, Canada V6T 1Z3,
    Search for more papers by this author
  • Dezene P. W. Huber,

    1. Biotechnology Laboratory, University of British Columbia, 6174 University Boulevard, Vancouver, BC, Canada V6T 1Z3,
    Search for more papers by this author
  • Jörg Bohlmann

    Corresponding author
    1. Biotechnology Laboratory, University of British Columbia, 6174 University Boulevard, Vancouver, BC, Canada V6T 1Z3,
    2. Department of Botany, University of British Columbia, 6174 University Boulevard, Vancouver, BC, Canada V6T 1Z3, and
    3. Department of Forest Sciences, University of British Columbia, 6174 University Boulevard, Vancouver, BC, Canada V6T 1Z3
    Search for more papers by this author

For correspondence (fax +1 604 822 2114; e-mail bohlmann@interchange.ubc.ca).
These authors contributed equally to this work.

Summary

Feeding forest tent caterpillars (FTCs) induced local and systemic diurnal emissions of (−)-germacrene D, along with (E)-β-ocimene, linalool, (E)-4,8-dimethyl-1,3,7-nonatriene (DMNT), benzene cyanide, and (E,E)-α-farnesene, from leaves of hybrid poplar. FTC feeding induced substantially higher levels of volatiles in local and systemic leaves than did mechanical wounding. A full-length poplar sesquiterpene synthase cDNA (PtdTPS1) was isolated and functionally identified as (−)-germacrene D synthase. Expression of PtdTPS1, expression of genes of early, intermediate and late steps in terpenoid biosynthesis, and expression of a lipoxygenase gene (PtdLOX1) were analyzed in local FTC-infested and systemic leaves. Transcript levels of PtdTPS1 and PtdLOX1 were strongly increased in response to herbivory. PtdTPS1 was also induced by mechanical wounding or by methyl jasmonate (MeJA) treatment. FTC feeding did not affect transcript levels of 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR), 1-deoxy-d-xylulose 5-phosphate reductoisomerase (DXR), and isoprene synthase (IPS). Two other TPS genes, PtdTPS2 and PtTPS3, and farnesyl diphosphate synthase were only very transiently induced. These results illustrate differential expression of terpenoid pathway genes in response to insect feeding and a key function of (−)-germacrene D synthase PtdTPS1 for herbivore-induced local and systemic volatile emissions in hybrid poplar. FTC-induced transcripts of PtdTPS1 followed diurnal rhythm. Spatial patterns of FTC-induced PtdTPS1 transcript accumulation revealed acropetal but not basipetal direction of the systemic response. Implications for tritrophic poplar-FTC-predator/parasitoid interactions are discussed.

Introduction

Forest tent caterpillars (FTCs) are distributed throughout North America and Eurasia. FTC hatch in early spring and immediately begin to feed on the leaves of their hosts, which include species of poplar and aspen. By their final instar, larvae grow to over 1000 times their mass at hatching and consume more than 15 000 times their initial body weight in leaf tissue (Fitzgerald, 1995). Even small infestations result in noticeable defoliation on single trees, and outbreaks of FTC can cause large-scale defoliation of natural or plantation forests. Repeated and prolonged attack by FTC makes individual trees more susceptible to other deleterious factors (Churchill et al., 1964; Hogg et al., 2002) and results in reduced tree growth (Hindahl and Reeks, 1960). FTC defoliation, combined with climatic factors, can result in the large-scale dieback of poplar forests (Hogg et al., 2002).

In recent years, much attention has focused on indirect defenses of plants against herbivores (Dicke and Vet, 1999; Kessler and Baldwin, 2002; Paré and Tumlinson, 1999; Takabayashi and Dicke, 1996). Indirect defense often involves the induced release of volatile organic compounds in response to feeding by arthropods. These volatiles can serve in tritrophic systems to attract enemies of herbivores, such as parasitic wasps and flies or predatory mites, which can protect the signaling plant from further damage (Kessler and Baldwin, 2001). Among the most common volatile signals in indirect defense are metabolites of the lipoxygenase (LOX) pathway, metabolites of the shikimic acid pathway, and products of the terpenoid pathway (monoterpenes, sesquiterpenes, homoterpenes) (Pichersky and Gershenzon, 2002). Various aspects of formation and release of herbivore-induced terpenoid volatiles have been elucidated in studies with herbaceous species. Although some of the early, as well as recent, studies of induced plant volatile signaling have utilized trees (e.g. Baldwin and Schultz, 1983; Martin et al., 2003; Tscharntke et al., 2001; Wegener et al., 2000), trees have not been well studied with regard to molecular or biochemical mechanisms of insect-induced volatile emissions.

As there is evidence that poplars do respond to herbivory by attracting predators and parasitoids (Havill and Raffa, 2000; Mondor and Roland, 1997, 1998), we investigated the emission of volatiles and their molecular regulation in rooted ramets of hybrid poplar interacting with FTC. In this paper, we report: (i) release of a suite of FTC-induced volatiles from local, attacked leaves and from systemic, undamaged leaves; (ii) rhythmic release of (−)-germacrene D as a major, systemically FTC-induced sesquiterpene volatile; (iii) cDNA cloning and functional identification of an insect-, wound-, and methyl jasmonate (MeJA)-inducible (−)-germacrene D synthase gene PtdTPS1; (iv) analysis of isoprenoid pathway and LOX transcript levels in local and systemic leaves; and (v) analysis of the spatial distribution of transcript accumulation of PtdTPS1 following FTC feeding. We identified a key role of the PtdTPS1 gene for regulation of insect-induced diurnal and systemic sesquiterpene emisson. This paper describes a foundation for future molecular ecological studies of volatile signaling and systemic defense in poplar, the emerging genomic model system for tree biology.

Results

Qualitative and quantitative analysis of local and systemic volatiles released from FTC-infested poplars

Local, FTC-infested poplar leaves and systemic non-infested leaves released very similar blends of volatiles consisting of (−)-germacrene D, (E)-β-ocimene, linalool, (E)-4,8-dimethyl-1,3,7-nonatriene (DMNT), benzene cyanide, and (E,E)-α-farnesene, and sometimes phenethyl acetate, during the light period of the first day after initiation of FTC feeding (Figure 1; Supplementary Material, Figure S1). FTC-induced volatile release from local or systemic leaves was almost nil during dark hours (data not shown). Local FTC-infested leaves released statistically significantly higher (P = 0.01) amounts of DMNT than mechanically wounded local leaves and substantially higher (P = 0.07 in both cases) levels of (E)-β-ocimene and (−)-germacrene D, over five replicates. Sampling of systemic emissions of mechanically wounded trees showed very little of these volatiles (Figure 2 for (−)-germacrene D). Unwounded trees did not produce detectable amounts of any of these compounds.

Figure 1.

Volatile emissions (mean µg h−1 g−1 of dry leaf tissue ± 1 SE for five replicates) from local, FTC-infested portions of hybrid poplar after 24 h of insect feeding compared to mechanical wounding of leaves.

Volatiles were collected on superq between 12.00 and 18.00 hours during second half of the light period using an automated headspace volatile collection system (Martin et al., 2003; Röse et al., 1996). Volatiles were identified by GC–MS analysis and by comparison with authentic standards and library matches (chemstation software, Hewlett Packard, USA; Wiley library; and Adams, 2001). Details of the volatile collection system and GC–MS analysis are provided as Supplementary Material. The number above each pair of bars shows the P-value of a t-test comparing the emissions of that particular compound between the two treatment groups.

Figure S1.

GC–MS analysis of volatiles collected on superq from systemic portions of FTC, M. disstria, induced P. trichocarpa × deltoides H11-11 between 12.00 and 15.00 hours (third quarter of light period).

Numbers beside each peak refer to the associated mass spectra. In each case the top spectrum is the peak of interest and the bottom spectrum corresponds to the best library match (chemstation software, Hewlett Packard, USA; Wiley library; and Adams, 2001). (E)-β-Ocimene, linalool, DMNT, and (−)-germacrene D were also identified by retention time matching with authentic standards. The possibility that the peak identified as benzene cyanide was actually indole ruled out when an authentic indole standard did not match retention time. Phenethyl acetate was sometimes present in diurnal samples of systemic volatiles as well, and might be present in this GC trace as the small peak at approximately 32 min. Concurrent volatile collections of systemic volatiles of mechanically wounded trees showed virtually none of any of these volatiles.

Figure 2.

Insect-induced systemic emission of (−)-germacrene D from poplar during and after feeding by the FTC and systemic emission upon mechanical wounding of the leaves with scissors.

(a) Volatile emissions were collected from systemic upper sections of the plants. Lower parts of the plants where the treatments were applied were not included in the volatile collection.

(b) Black bars represent emissions from FTC-infested trees while white bars correspond to scissor wounding. Four-hour collection periods are shown across the bottom x-axis, while total elapsed time is shown on the top x-axis. Alternating light and shaded sections along the graph correspond to precise 12-h light and dark periods. The double-headed arrow covers the time in which FTCs were in contact with the trees. Time periods in which (−)-germacrene D emissions were not detected in GC–MS analyses are denoted by very small bars. A detailed view of each of the three replicates is available as Supplementary Material.

(−)-Germacrene D release (mean µg h−1 g−1 of dry leaf tissue ± 1 SE) from three independent replicates is shown.

Systemic diurnal emission of (−)-germacrene D induced by FTC

In order to study molecular and biochemical mechanisms of FTC-induced volatile emissions, we focused on one major FTC-induced sesquiterpene (−)-germacrene D. We analyzed temporal patterns of systemic emissions of this compound in response to FTC over time courses of up to 7 days after initiation of insect infestation. Release of (−)-germacrene D followed a diurnal cycle with maximum emissions during the light period of each day (06.00–18.00 hours) and strongly reduced emissions (often at undetectable levels) during darkness (Figure 2). (−)-Germacrene D was released from uninfested portions of the trees within 24 h after first exposure to FTC. Diurnal, systemic release remained elevated over a period of 3 days while insects were in contact with the trees. After removal of FTC, some systemic release of (−)-germacrene D was sustained during the next 24 h. Volatile collections at time points more than 24 h after removal of FTCs did not reveal any (−)-germacrene D. The poplar/FTC system exhibited some variability in the amount of (−)-germacrene D released from trees in three independent experiments (see Supplementary Material, Figure S2) that might be because of some combination of tree-to-tree variation and variation in the FTC feeding behavior. For comparison with continuous feeding by FTC, a single mechanical wounding event resulted in weak and brief systemic (−)-germacrene D release (Figure 2). The emission patterns of the other FTC-induced systemic volatiles were similar to that of (−)-germacrene D (results not shown).

Figure S2.

Insect-induced systemic emission of (−)-germacrene D from poplar during and after feeding by the FTC, M. disstria, and systemic emission upon mechanical wounding of the leaves with scissors.

(a) Volatile emissions were collected from systemic upper sections of the plants. Lower parts of the plants where the treatments were applied were not included in the volatile collection.

(b) Black bars represent emissions from FTC-infested trees while white bars correspond to scissor wounding. Four-hour collection periods are shown across the bottom x-axis, while total elapsed time is shown on the top x-axis. Alternating light and shaded sections along the graph correspond to precise 12-h light and dark periods. The double-headed arrow covers the time in which FTCs were in contact with the trees. Time periods in which (−)-germacrene D emissions were not detected in GC–MS analyses are denoted by very small bars.

Three independent replicates are shown.

Isolation of sesquiterpene cDNA PtdTPS1

A full-length cDNA of 1877 bp for poplar (−)-germacrene D synthase (PtdTPS1) was isolated by reverse transcriptase-polymerase chain reaction (RT-PCR) using degenerate oligonucleotide primers followed by rapid amplification of cDNA ends (RACE) with RNA from FTC-induced leaves. The largest open-reading frame of PtdTPS1 encodes a predicted protein of 561 amino acids, a molecular mass of approximately 65 kDa, and a calculated isoelectric point of 5.36 (Figure 3a). The deduced protein sequence of PtdTPS1 showed highest similarity with known angiosperm sesquiterpene synthases of the TPS-a group (Bohlmann et al., 1998a; Figure 3b). PtdTPS1 shares 64% identity (ID) and 78% similarity (SI) with rose (Rosa hybrida) germacrene D synthase, 48% ID and 69% SI with tomato (Lycopersicon hirsutum) germacrene D synthase, 47% ID and 70% SI with tomato (L. esculentum) germacrene C synthase (Figure 3a). Sequence similarity was lower with members of the TPS-b group (Figure 3b) that contains angiosperm monoterpene synthases and poplar (Populus alba × P. tremula) isoprene synthase (IPS), a hemiterpene synthase (Miller et al., 2001), or with members of other TPS-subfamilies.

Figure 3.

Sequence relatedness of poplar PtdTPS1 and other plant TPS.

(a) Alignment of deduced amino acid sequences of poplar PtdTPS1 and selected angiosperm sesquiterpene synthases. Sequences are PtdTPS1 (this study), R. hybrida germacrene D synthase (RhGDS; Guterman et al., 2002), L. hirsutum germacrene D synthase (LhGDS; AF279456), and L. esculentum germacrene C synthase (LeGCS; T06265). Amino acid residues that are identical among at least two sequences are indicated by white letters on black background. Other conserved residues are shaded in gray. Additional residues are pair-wise identical among PtdTPS1 and RhGDS and among LhGDS and LeGCS. Dashes indicate sequence gaps introduced to optimize the alignment. Conserved motifs R(P/R)x8W and DDxxD are indicated. Regions used for degenerate primer design are shown with asterisks.

(b) Dendrogam of deduced amino acids of PtdTPS1 and other angiosperm TPS of known function. Accessions for TPS represented in this tree are: A. thaliana myrcene/(E)-β-ocimene synthase (AF17835) and (E)-β-ocimene synthase (AY151086), C. intybus germacrene A synthase (AF497999), G. arboreum (+)-δ-cadinene synthase (CAA77191), L. esculentum germacrene C synthase (T06265), L. hirsutum germacrene B synthase (AF279455) and germacrene D synthase (AF279456), L. sativa germacrene A synthase (AAM11627), M. spicata 4S-limonene synthase (L13459), P. alba × tremula IPS (CAC35696), and R. hybrida germacrene D synthase (Guterman et al., 2002).

PtdTPS1 contains the arginine–arginine/tryptophan conserved motif (RRx8W) at 20–30 amino acids from the N-terminus and the DDxxD active-site motif located at 314–318 amino acids from the N-terminus (Figure 3a). These motifs, and variations thereof, are commonly found in plant TPS (Aubourg et al., 2002) and are important for enzyme function (Davis and Croteau, 2000). The short N-terminal region upstream of the RRx8W motif is indicative of the absence of a plastid transit peptide (Bohlmann et al., 1998a).

Functional expression and characterization of PtdTPS1

For functional identification, PtdTPS1 was expressed in Escherichia coli transformed with pET100/D-PtdTPS1. Isopropyl-1-thio-β-d-galactopyranoside (IPTG)-induction resulted in increased expression of a recombinant protein of molecular mass of approximately 65 kDa (Figure 4a), consistent with the predicted molecular mass of PtdTPS1. Protein extracts of induced E. coli strain BL21-CodonPlus(DE3)/pET100/D-PtdTPS1 assayed with farnesyl diphosphate (FPP) yielded germacrene D as the major product (79.4% of total olefins) as identified by gas chromatography–mass spectrometry (GC–MS) and co-elution with an authentic standard (Figure 4b,c). Stereochemistry of the (−)-germacrene D product was identified by GC–MS and by gas chromatography–flame ionization detection (GC–FID) and by comparison of retention time with that of an authentic standard (Supplementary Material, Figure S3). Minor products of PtdTPS1 were β-caryophyllene (peak 2, 2.2% of total product), α-humulene (peak 3, 4.5% of total product), and alloaromadendrene (peak 4, 3.4% of total product; Figure 4). Four other, minor sesquiterpene products (peak 1 (1.9%), peak 6 (6.0%), peak 7 (0.8%), and peak 8 (1.8%)) were not identified. A control extract prepared from E. coli BL21(DE3) transformed with pET100/D-TOPO without the PtdTPS1 insert did not produce any detectable sesquiterpenoid products. Recombinant PtdTPS1 enzyme was not active with geranyl diphosphate or geranylgeranyl diphosphate as substrate.

Figure 4.

Expression in E. coli of recombinant PtdTPS1 protein and GC–MS analysis of sesquiterpene products formed by recombinant PtdTPS1 enzyme activity in vitro.

(a) SDS–PAGE analysis of total protein extracts from bacteria transformed with pET100/D vector or with plasmid pET100/D-PtdTPS1 with (+) or without (−) IPTG induction. Coomassie Blue staining reveals a polypeptide of approximately 65 kDa in IPTG-induced bacteria transformed with pET100/D-PtdTPS1 (arrow).

(b, c) Total ion chromatogram (b) and mass spectra (c) of products of PtdTPS1 in vitro enzyme activity with FPP as substrate. The mass spectrum of the most abundant product (peak 5), identified as germacrene D, is shown together with that of an authentic germacrene D standard.

Figure S3.

Identification of stereochemistry of (−)-germacrene D produced by PtdTPS1 and emitted from poplar leaves by GC–MS and GC–FID.

(a) Mass spectrum of the FTC-induced single major volatile, germacrene D, released by poplar clone H11-11 shown (top) together with that of a library standard (bottom). The spectrum of an authentic germacrene D standard is shown in Figure 4.

(b) Stereochemistry of FTC-induced (−)-germacrene D was identified by comparison of retention time with that of authentic standards of (−)-germacrene D and (+)-germacrene D using GC–FID. Also shown is the sesquiterpene product formed by recombinant PtdTPS1 enzyme in vitro.

Local expression of PtdTPS1 in response to FTC, MeJA, and wounding

In order to measure transcripts of PtdTPS1, first in local leaves, we probed Northern blots of total RNA isolated from leaves exposed to FTC feeding, from MeJA-treated leaves, or from mechanically wounded leaves. Irrespective of the developmental stage of these leaves, constitutive expression of PtdTPS1 was below detection limits (Figure 5). Similarly, PtdTPS1 transcripts were not detected in green stem tissues. PtdTPS1 transcript levels were greatly increased at 8- and 24-h after onset of FTC feeding, after mechanical wounding, or after treatment with MeJA (Figure 5a). Because MeJA and FTC induced similar PtdTPS1 transcript accumulation in local leaves, we tested the possible involvement of the oxylipin pathway in FTC-induced expression of PtdTPS1 with a pharmacological inhibitor. Treatment of leaves with salicylhydroxamic acid (SHAM), an inhibitor of LOX (Macríet al., 1994; Peña-Cortes et al., 1993), reduced the level of PtdTPS1 transcripts induced by FTC feeding (Figure 5b).

Figure 5.

Local effect of FTC feeding, wounding or treatment with MeJA on transcript levels of (−)-germacrene D synthase (PtdTPS1) in poplar leaves.

(a) Northern blot analysis of PtdTPS1 transcripts in untreated green bark and leaves of young poplars, and in leaves 8 and 24 h after treatment with: 0.1% Tween 20 solution (control), after treatment with MeJA (0.5 mm) in 0.1% Tween 20, after mechanical wounding, or after feeding by FTCs. Results are shown for two independent sets of experiments.

(b) Northern blot analysis of PtdTPS1 in detached leaves with (+) and without (−) exposure to FTC for 24 h after prior treatment of leaves with (+) or without (−) SHAM.

Systemic and diurnal expression of PtdTPS1 in response to FTC

In Northern blot analysis of RNA from systemic and local leaves, we followed the time course of PtdTPS1 transcripts upon FTC feeding on lower parts of trees. In local and systemic leaves, PtdTPS1 was induced by FTC within 3–6 h and reached a maximum at 24 h (Figure 6), coinciding with the time of the first diurnal maximum of (−)-germacrene D emission (Figure 2). Interestingly, in systemic leaves, PtdTPS1 transcripts declined at 32 h (at the end of the second light period) were reduced to very low levels at 40 h (second dark period) and were present again at 48 h (third light period) and 56 h (end of light period; Figure 6). Curiously, at 18 h (first dark period) PtdTPS1 transcripts remained elevated in systemic and local leaves.

Figure 6.

Northern analysis of transcripts in local, FTC-infested and in systemic undamaged leaves.

(a) Transcript levels were analyzed for PtdTPS1, PtdTPS2, PtdTPS3, PtdFPPS1, PtdHMGR1, PtdDXR1, and PtdLOX1.

(b) Hybridization signals for PtdTPS1, PtdTPS2, PtdTPS3, PtdFPPS1, PtdHMGR1, PtdDXR1, and PtdLOX1 in local leaves at 0, 1, 3, 6, 12, 18, 24, and 48 h after first contact with FTCs and in systemic leaves at 0, 1, 3, 6, 12, 18, and 24 h after first contact with FTCs.

(c) Extended time course of transcript analysis in systemic leaves at 24, 32, 40, 48, and 56 h after first contact with FTCs. For each time point and sample, two leaves were used for RNA isolation.

Leaves on lower parts of poplar trees were exposed to FTCs with 20 larvae per tree caged in mesh bags. Local and systemic leaves were collected over a time course with regular light and dark intervals.

Expression of other poplar TPS genes in response to FTC

To test if FTC feeding affects other TPS genes in a similar fashion as PtdTPS1, we investigated transcript levels of two additional PtdTPS genes of as yet unknown biochemical functions and of IPS (PtdIPS), a hemiterpene synthase (Miller et al., 2001). Probes for PtdTPS2 and PtdTPS3 were generated by RT-PCR based on expressed sequence tags (ESTs; http://poppel.fysbot.umu.se/: PtdTPS2 (Q033A03), PtdTPS3 (P003C04)). ESTs PtdTPS2 and PtdTPS3 are 61.4 and 58.0%, respectively, identical with PtdTPS1. Unlike PtdTPS1, transcripts of PtdTPS2 and PtdTPS3 were only very transiently induced in local leaves at 1 h after first contact with FTC and rapidly declined despite continuous feeding by FTC (Figure 6). In systemic leaves only very low, if any, expression of PtdTPS2 and PtdTPS3 was detected over the entire time course of FTC exposure. PtdIPS was also not affected by FTC in local or systemic leaves, but transcripts showed the same diurnal profiles in control leaves and in FTC-induced local and systemic leaves (Figure 7). These results demonstrate differences in the response of four TPS genes in poplar. Induced transcript levels of PtdTPS1 were reflected in a corresponding FTC-induced volatile emission of (−)-germacrene D.

Figure 7.

Northern analysis of poplar IPS (PtdIPS) transcript accumulation in local, FTC-infested leaves compared to systemic, undamaged and control leaves.

Times of harvest are in hours (h) after first contact with FTC. The times of harvest are on the same time of the day or night as those shown in Figure 6. PtdIPS transcript accumulation follows very similar diurnal cycles in the control and in both treatment groups.

Expression of early and intermediate steps of terpenoid biosynthesis in response to FTC

Early and intermediate steps in terpenoid biosynthesis can be involved in the regulation of induced terpenoids (Korth and Dixon, 1997; Martin et al., 2002). Poplar cDNA probes for three additional genes of terpenoid biosynthesis, 1-deoxy-d-xylulose 5-phosphate reductoisomerase PtdDXR1, 3-hydroxy-3-methylglutaryl-CoA reductase PtdHMGR1, and farnesyl diphosphate synthase PtdFPPS1, were generated by RT-PCR based on ESTs (PtdHMGR1 (025F04), PtdDXR1 (V052H05), PtdFPPS1 (G004P36Y and A032P02U)). Similarly, a probe was obtained for a poplar candidate lipoxygenase gene (PtdLOX1, C017P54U) of oxylipin formation, which could be involved in the regulation of herbivore-induced responses (Feussner and Wasternack, 2002).

In Northern analysis, probes for PtdHMGR1, PtdDXR1, PtdFPPS1, and PtdLOX1 revealed expression profiles in local and systemic poplar leaves upon feeding by FTC that were different from that of (−)-germacrene D synthase (PtdTPS1; Figure 6). In local leaves, expression of PtdFPPS1 was rapidly induced at 1 h, but was already reduced after 3 h. In systemic leaves, PtdFPPS1 transcript levels followed a similar time course, but expression was weaker. Transcripts of PtdFPPS1 were also increased at 24 h, although to a lesser extent than during the initial response at 1 h in local and systemic leaves. FTC did not affect expression of PtdHMGR1 and PtdDXR1 in local or systemic leaves in any samples analyzed. These data reveal differential gene expression of early (PtdHMGR1, PtdDXR1), intermediate (PtdFPPS1), and late steps (PtdTPS1) in FTC-infested poplar (Figure 6).

Expression of PtdLOX1 was induced after 24 and 48 h in local leaves, and increased gradually, but strongly over a 24-h time course in systemic leaves (Figure 6). In systemic leaves, PtdLOX1 levels were highest at 24 h and were at similar levels at 32 h (during the third light period; Figure 6). PtdLOX1 transcripts declined to very low levels at 40 h (during the second dark period) but were detectable again at 48 and 56 h (during the third light period) similar to (−)-germacrene D synthase PtdTPS1 (Figure 6). Also similar to PtdTPS1, PtdLOX1 transcripts remained elevated in systemic leaves at 18 h (towards the end of the first dark period).

Spatial patterns and acropetal orientation of systemic induction of PtdTPS1 transcripts

We investigated spatial patterns of local and systemic induction of PtdTPS1 in two sets of experiments with FTC feeding either on the lower parts of trees or on the upper part of trees (Figure 8). At 24 h after initiation of FTC feeding, when maximum local and systemic PtdTPS1 transcript levels were found (Figure 6), leaves were harvested at defined positions along the length of trees and PtdTPS1 transcripts were measured. These experiments were repeated with two groups of trees that differed in branching patterns. First, we used 110–120-cm tall trees that were regularly pruned up to 2 weeks prior to experiments to remove developing lateral branches (Figure 8a). A second group of trees (155–185 cm tall) had lateral branches of 10–50 cm length along the main stem (Figure 8b). In both types of trees, transcript accumulation of PtdTPS1 was increased only in the local infested leaves and only very weakly in nearest lower systemic leaves after infestation of the top of the tree with FTC for 24 h (Figure 8c,d). No long-range, basipetal systemic response was observed using PtdTPS1 as a probe, when FTC were placed on the top of trees.

Figure 8.

Spatial patterns of PtdTPS1 transcript accumulation in two sets of poplar trees after FTC feeding.

(a) FTC feeding locations and locations of leaves collected for Northern analyses in the experiment using trees without branches.

(b) FTC feeding locations and locations of leaves collected for Northern analyses in the experiment using trees with branches.

(c) PtdTPS1 transcript accumulation in leaf positions shown in (a) for top- and bottom-infested trees.

(d) PtdTPS1 transcript accumulation in leaf positions shown in (b) for top- and bottom-infested trees.

IM refers to immature leaves, 5–10 cm across at the widest point. FTCs were on either the top or the bottom portion of trees.

In contrast, when bottom leaves were exposed to FTC for 24 h, PtdTPS1 transcript accumulation occurred not only in local, infested leaves (leaf positions 9 and 10 in pruned trees, Figure 8a,c; leaf position 5 on a lower branched of branched trees, Figure 8b,d), but transcripts of PtdTPS1 were also increased in systemic undamaged leaves up to the top of the trees. In trees without lateral branches, all undamaged leaves along a distance of 50 cm (leaf positions 1–8 and immature leaves, Figure 8a,c) showed strong PtdTPS1 transcript accumulation. In the larger branched trees (155–185 cm tall) induced PtdTPS1 transcript levels appeared lower in systemic leaves on branches proximal to and approximately 50 cm above the FTC-infested region (leaf positions 3 and 4, Figure 8b,d), but the most distal mature leaves on the main stem (leaf position 1, Figure 8b) showed strong systemic FTC-induced PtdTPS1 transcript accumulation (Figure 8d). In summary, independent of branching pattern of trees, we did not detect significant basipetal systemic induction of PtdTPS1 transcripts over distances of 50–180 cm. However, systemic, acropetal PtdTPS1 transcript accumulation was found over distances of the entire length of the trees used in these experiments.

Discussion

Local and systemic FTC-induced emissions of volatiles from hybrid poplar

Indirect plant defense against arthropod herbivores often involves the induced release of volatile signals. Induced volatile emission can be the result of transcriptional upregulation and increased pathway activities in local or systemic parts of infested plants. Previously accumulated compounds may also be released following local damage. Using FTC-infested poplar, we analyzed molecular and biochemical events of induction of local and systemic leaf emissions. A blend of volatiles induced in response to FTC infestation contained the sesquiterpene (−)-germacrene D along with five or six other compounds, mainly mono-, sesqui-, and homoterpenes. Our methods did not allow the detection of isoprene, a common hemiterpene emitted from poplars. Northern analysis showed that IPS is expressed on a diurnal cycle and is not affected by FTC feeding in either local or systemic leaves. Analyses of volatiles from systemic portions of the plant excluded insects as the source for the volatiles. FTC-induced volatile emissions from local and systemic leaves were very similar. This could suggest that the same or similar signaling events lead to volatile emission in the local herbivore-attacked leaf and to the activation of a systemic response in distant leaves. While spatial patterns of PtdTPS1 transcript accumulation (Figures 6 and 8) provide information pertaining to local and systemic regulation of (−)-germacrene D emissions across a tree, future work will determine if leaves close to an infestation are stronger emitters of induced volatiles than those some distance away, or if the upper most mature systemic leaves are stronger emitters than lower systemic leaves.

Our findings of FTC-induced emission of (−)-germacrene D and corresponding gene expression provide molecular and chemical evidence to support the observation that FTC-infested poplar leaves are able to provide honest signals to potential antagonists of the herbivore (Havill and Raffa, 2000). Responses of parasitoids of FTC to (−)-germacrene D and to other induced volatiles remain to be tested, which could elucidate preference of certain parasitoids for some species of FTC-infested poplars compared to others (Mondor and Roland, 1997, 1998). It seems likely that diurnal volatiles could be behaviorally active for FTC antagonists that forage during daytime, as the trees emit little to none of these compounds at night. Future work will address to what extent diurnal patterns of emissions and transcript levels are regulated by possible, diurnal FTC-activity, circadian clock, or other factors.

Temporal and spatial patterns of FTC-induced systemic (−)-germacrene D emissions and PtdTPS1 transcript accumulation

Local and systemic diurnal emissions of (−)-germacrene D in poplar involves changes in transcripts of the PtdTPS1 gene (Figures 6 and 8). Volatile release and PtdTPS1 transcripts are able to rapidly increase upon onset of FTC feeding but also rapidly shut down after removal of herbivores. In a natural setting, this may reflect selection for trees that tightly link the beginning and end of an infestation with, respectively, initiation and termination of volatile signaling by regulation of TPS transcripts. Volatile emission is also reduced during the dark phase, while FTC remained on the leaves (Figure 2). The disappearance of systemically induced PtdTPS1 transcripts in the second dark phase and the corresponding reduction of (−)-germacrene D emission support the concept of a tight control of transient and diurnal volatile emissions that are regulated, at least in part, at the level of transcription of PtdTPS1.

The spatial patterns of PtdTPS1 transcript accumulation (Figures 6 and 8) provide information about systemic herbivore-induced signaling in poplars. When FTC feed on lower portions of a small tree, increase of PtdTPS1 transcripts spread out toward the upper parts of the tree. Systemic emissions of FTC-induced volatiles (Figure 2; Supplementary Material, Figure S1) support the acropetal movement and effect of herbivore-induced signals. However, no similar basipetal induction of PtdTPS1 transcripts was observed when FTC feed on the upper portions of a tree. Acropetal orientation of transcript accumulation also occurs in the case of systemic inductions of poplar proteinase inhibitor promoter in transgenic tobacco plants (Hollick and Gordon, 1995). It is conceivable that a systemic, insect-induced defense signal of as yet unknown nature in poplar moves predominantly acropetally, thus activating protection to the apical leader tissues. While work described here was on relatively small trees (less than two meters tall), poplars, because of their large physical dimensions and amenability to biochemical, molecular, and genomic approaches, provide a highly suitable system for further analysis of long-distance signaling in herbivore-induced systemic defense.

Differential responses of genes of terpenoid biosynthesis in response to FTC feeding

In plants, two pathways produce the 5-carbon precursors for terpenoids (Croteau et al., 2000). HMGR is a key step in the mevalonate pathway, while DXR is a key enzyme in the methylerythritol 4-phosphate pathway. Expression of at least one HMGR gene and one DXR gene are constitutive in poplar, regardless of exposure to FTC feeding (Figure 6). This differs from findings with potato (Solanum tuberosum) where HMGR transcription was induced in response to mechanical wounding, insect feeding, or insect oral secretions (Korth and Dixon, 1997). FPP synthase forms the prenyl diphosphate substrate for (−)-germacrene D synthase. PtdFPPS1 transcripts increased rapidly but more transiently than PtdTPS1 in local and systemic leaves after onset of FTC feeding. Systemic FTC-induced transcript accumulation of (−)-germacrene D synthase (PtdTPS1) and LOX (PtdLOX1) in poplar leaves is similar to profiles of transcript accumulation of genes of direct defense in poplars and aspens, namely genes of phenylpropanoid formation and condensed tannin biosynthesis, and polyphenol oxidase (Constabel et al., 2000; Peters and Constabel, 2002). Taken together, our data reveal differential gene expression of early (DXR and HMGR), intermediate (FPP synthase), and late (germacrene D synthase) steps in the biosynthesis of insect-induced terpenoid emission in poplar, with the expression of the (−)-germacrene D synthase gene, most closely matching the actual pattern of FTC-induced volatile release and, hence, supporting the importance of this TPS as a control point for regulation of herbivore-induced sesquiterpenoid volatile emission and a possible important role in indirect defense.

Additional control of volatile emission

In comparison with other systems of terpenoid volatile emission in plant biology, the diurnal patterns of emission and control at the level of TPS genes are intriguingly similar in poplar leaves and in snapdragon flowers (Antirrhinum majus, Dudareva et al., 2003), suggesting some similar mechanisms of regulation of herbivore-induced volatile release from green leaves and developmentally controlled scent emission from flowers. As with floral scent emission (Dudareva et al., 2003), other mechanisms in addition to TPS transcript accumulation undoubtedly influence release of terpenoids such as (−)-germacrene D from poplar leaves. For instance, despite similarly strong induction of PtdTPS1 transcripts by mechanical wounding and FTC in local leaves at 8 and 24 h post-treatment (Figure 6), herbivores induce substantially higher emission than that induced by mechanical wounding (Figure 1). Translation of TPS or substrate availability could be important in the control of FTC-induced volatile release. Furthermore, the concerted upregulation of transcripts of (−)-germacrene D synthase (PtdTPS1) and LOX, together with the induction of PtdTPS1 by MeJA and inhibition of FTC-induced PtdTPS1 transcript accumulation by SHAM, suggests a role of the oxylipin pathway in FTC-induced volatile emission in poplar.

Experimental procedures

Trees

Populus trichocarpa Torr. & Gray ×P. deltoides Bartr. (Salicaceae), clone H11-11, was grown on University of British Columbia South Campus farm. Cuttings of 30–200 cm were taken in February of 2002 and 2003 from previous year shoots, placed in soil (35% peat, 15% perlite, 50% pasteurized mineral soil, 250 g m−3 Osmocote™ 13-13-13 plus micronutrients) in 656 ml or 3-gal. pots, and regenerated into rooted trees. Trees were maintained in a greenhouse under constant summer conditions (daytime temperature: 18.3–21.8°C; relative humidity: 24.5–47.2%). Trees of 1–2 m height were used in experiments in June–October 2002 and in June–July 2003.

Insects

Forest tent caterpillars, Malacosoma disstria Hübner (Lepidoptera: Lasiocampidae), were from the Great Lakes Forestry Centre (NRCan, Sault Ste. Marie, Ont., Canada). FTC were reared and maintained on artificial diet (Addy, 1969) at 27°C, 50–60% RH, 16 : 8 light:dark. Groups of third to fifth instar larvae were used in experiments. Prior to being placed on plants, insects were starved for 24 h on moist filter paper.

Volatile collection and volatile analysis

Volatiles were collected with an automated system (Analytical Research Systems Inc., Gainesville, FL, USA; Röse et al., 1996) using procedures for sampling and volatile analysis by GC–MS modified after Martin et al. (2003). Description of the volatile collection system, full details of collection procedures from FTC-infested and wounded trees, and detailed description of GC–MS analysis of volatiles are provided as Supplementary Material.

cDNA cloning of partial PtdTPS1

First-strand cDNA was synthesized using SuperScript II Reverse Transcriptase (Invitrogen, Burlington, Ont., Canada), oligo-dT primer (Takara, Otsu, Japan), and 5 µg of total RNA from FTC-infested leaves at 42°C for 50 min. Reaction was stopped by incubating at 70°C for 15 min. Degenerate primers were designed from conserved regions of known sesquiterpene synthases (Bouwmeester et al., 2002): forward primer (5′-GAYGARAAYGGNAARTTYAARGA-3′) and reverse primer (5′-TYRTTDATRTCYTTCCANGC-3′). PCR was performed in a total volume of 50 µl containing 0.25 µm primers, 0.2 mm 2'-deoxynucleoside S'-triphosphates (dNTPs), 1.5 mm MgCl2, 2.5 units Platinum Taq DNA polymerase (Invitrogen), 1× PCR buffer (Invitrogen), and 2 µl of first-strand cDNA. Temperature programs were: 2 min at 94°C; 38 cycles of 30 sec at 94°C, 90 sec at 42°C, 60 sec at 72°C; and 5 min at 72°C. PCR products were ligated into pCR2.1-TOPO (Invitrogen), transformed into E. coli TOP10 cells (Invitrogen), and inserts sequenced using DyeDeoxy Terminator Cycle Sequencing method (Applied Biosystems, Streetsville, Ont., Canada).

5′- and 3′-RACE of PtdTPS1

Cloning of 5′- and 3′-ends of PtdTPS1 cDNA was accomplished by RACE using the First Choice™ RLM-RACE Kit (Ambion, Austin, TX, USA) following the manufacturer's protocol. Poly(A) RNA was isolated with a MicroPoly(A) Pure™ Small Scale mRNA Purification Kit (Ambion) from total RNA (Wang et al., 2000) from poplar leaves infested by FTCs. For 5′-RACE, cDNA was amplified using Turbo Pfu polymerase (Stratagene, La Jolla, CA, USA), first with PtdTPS1-reverse outer primer (5′-TGGTATGTTCTTGCGAACAGGC-3′) and 5′-RACE outer primer (included with the kit). The second PCR was with PtdTPS1-reverse inner primer (5′-TGGGAAGTAGTGAACACAAGTG-3′), 5′-RACE inner primer (kit), and PCR products from the first PCR. PCRs were 2 min at 95°C; 35 cycles of 30 sec at 95°C, 30 sec at 55°C, and 60 sec at 72°C; and 5 min at 72°C. First and second 3′-RACE PCRs were accomplished using PtdTPS1-outer forward primer (5′-GCTGTTGCATTGGTCACTTCGG-3′) with 3′-RACE outer primer (kit), and PtdTPS1-inner forward primer (5′-CAGGCTCATGGATGACATCGTG-3′) with 3′-RACE inner primer (kit), respectively. 5′- and 3′-RACE products were ligated into pCR-Blunt TOPO (Invitrogen) and sequenced.

Cloning of full-length PtdTPS1 cDNA, functional expression and enzyme assays

Full-length cDNA PtdTPS1 was cloned by PCR using Turbo Pfu polymerase with forward primer (PtdTPS1-ORFS + CACC: 5′-CACCATGTCTGTTGAAGGTTCTGC-3′) and reverse primer (PtdTPS1-ORFA: 5′-TCACATTGGCACAGGATCAATG-3′) from the cDNA pool used for 3′-RACE. PCR was performed for 2 min at 95°C; 35 cycles of 30 sec at 95°C, 30 sec at 60°C, 90 sec at 72°C; and 5 min at 72°C. PtdTPS1 cDNA was cloned into pET100/D-TOPO vector (Invitrogen) and transformed into E. coli TOP10 cells. Plasmid pET100/D-PtdTPS1 was purified, its insert sequenced, and transformed into E. coli BL21-CodonPlus(DE3) (Stratagene). Expression of PtdTPS1 in E. coli BL21-CodonPlus(DE3)/pET100/D-PtdTPS1, extraction of protein and enzyme assays followed published procedures (Bohlmann et al., 1998b; Fäldt et al., 2003) detailed in Supplementary Material. SDS–polyacrylamide gel electrophoresis (PAGE) analysis was performed on 10% (w/v) polyacrylamide gel. Proteins were detected with Coomassie Blue.

Sequence analysis

Sequence analysis used programs from the lasergene DNA-Star Package version 5, clustalx (http://www.igbmc.u-strasbg.fr/BioInfo/ClustalX/Top.htm), and genedoc (http://www.psc.edu/biomed/genedoc). Sequence relatedness was analyzed using clustalx using the Neighbor Joining method, and visualized using treedraw (http://taxonomy.zoology.gla.ac.uk/rod/treeview.html).

Analyses of temporal patterns of transcript accumulation

Insect and mechanical wounding treatments were as described for volatile analyses (see Supplementary Material). For chemical treatments, 150 ml of 0.5 mm MeJA (Aldrich, Oakville, Ont., Canada) in 0.1% Tween 20/water was sprayed onto six potted trees. Trees were covered with plastic bags for 1 h after treatment. RNA from leaves was isolated (Wang et al., 2000), and samples of 20 µg of total RNA were separated under denaturing conditions using formaldehyde in 1.2% agarose gels, and transferred overnight by capillary action to nitrocellulose membrane (Hybond-N+, Amersham Biosciences, Piscataway, NJ, USA). Probes for PtdTPS2, PtdTPS3, PtdFPPS1, PtdHMGR1, PtdDXR1, and PtdLOX1 were isolated by PCR using Taq DNA polymerase (New England Biolabs, Mississauga, Ont., Canada) using first-strand cDNA from 3′-RACE of PtdTPS1 as the template. Primers were designed using partial DNA sequences obtained from the PopulusDB sequence database (http://poppel.fysbot.umu.se/). Sequences used for blast searches and primers designed for amplification of poplar cDNAs are listed in Supplementary Material. A probe for PtdTPS1 was obtained by PCR from pET100/D-PtdTPS1, using primers, PtdTPS1-ORFS and PtdTPS1-ORFA. A cDNA probe for IPS was obtained by PCR from pBSiso (Miller et al., 2001), using primers 5′-TATTGTTCTAGAACCGTCCAATC-3′ and 5′-GAGCAGTAGCATGAAGCTTAG-3′. Probes were labeled using Rediprime™ II Random Prime Labeling (Amersham Biosciences) with (α-32P)dCTP (3000 Ci mmol−1, 10 mCi ml−1, Perkin Elmer Applied Biosystems, Streetsville, Ont., Canada). Pre-hybridization of membranes was for 1 h at 65°C in hybridization buffer (0.05 m Na4P2O7 × 10 H2O, 0.115 m NaH2PO4, 7% SDS, 1 mm EDTA, 100 µg ml−1 denatured salmon sperm DNA), followed by addition of heat-denatured cDNA probes and hybridization for 16 h at 65°C. Membranes were rinsed once, washed two times at 65°C for 30 min, and once at 68°C for 30 min in wash buffer (0.05 m Na4P2O7 × 10 H2O, 0.115 m NaH2PO4, 1% SDS, 1 mm EDTA). Signals were detected using a Storm 860 phosphorimager (Amersham Biosciences).

Analyses of spatial pattern of PtdTPS1 accumulation

Spatial patterns of PtdTPS1 transcript accumulation in FTC-induced local and systemic leaves were investigated in two sets of experiments. First, trees without branches (110–120 cm tall) were infested with 10 larvae placed either on the top or at the bottom of the tree (Figure 8a). The FTCs were enclosed in a mesh bag on each infested tree for 24 h prior to collection of leaf material. On top-infested trees, caterpillars were enclosed with immature leaves (5–10 cm across at widest point) and two mature leaves (leaf positions 1 and 2 in Figure 8a). On bottom-infested trees, caterpillars were enclosed with three mature leaves (leaf positions 9, 10, and 11 in Figure 8a). An unexposed tree served as a control. Leaves were collected at positions shown in Figure 8(a). In a second set of experiments trees (155–185 cm) with lateral branches (10–50 cm) were exposed to 20 FTCs. Trees were infested on the top or the bottom approximately 40 cm for 24 h. Two leaves were collected at each of the following positions (shown in Figure 8b): four positions were collected from the top-infested tree (two mature leaves from FTC-infested area; two leaves on branches directly adjacent to, but excluded from, the FTC-infested area; two leaves on branches 50 cm below the bottom of FTC-infested area; and the bottom two leaves of tree), four positions were collected from the bottom-infested tree (two mature leaves on branches from FTC-infested area; two leaves on branches directly adjacent to, but excluded from, the FTC-infested area; two leaves on branches 50 cm above the top of FTC infested area; and the top two mature leaves of tree). Unexposed trees served as the control with two leaves collected at each of the three positions: lowest leaves, leaves from the middle, and the top mature leaves. In all spatial analysis experiments, RNA extraction with pooled leaves from each leaf position and treatment and Northern hybridizations with a PtdTPS1 probe were conducted as described above for analyses of temporal patterns of expression.

Inhibitor treatment

Salicylhydroxamic acid (Sigma, St Louis, MO, USA) was applied by placing the petioles of detached poplar leaves in 2 ml of 1 mm SHAM solution in 0.1% (v/v) methanol for 2–3 h. Controls were placed in 0.1% (v/v) methanol. After completely absorbing SHAM or control solutions, petioles were immediately placed into 50 ml of distilled water and each leaf was exposed to three FTC larvae for 24 h. Inhibitor treatment did not cause any obvious damage to leaves for up to 5 days after treatment.

Acknowledgements

We thank Drs Wilfried König, Stefan Bartram, and Wilhelm Boland for standards; Dr John Crock, Dr Anna-Karin Borg-Karlson, and Diane Martin for technical advice; Ryan Philippe and Sarah Martz for excellent technical assistance; David Kaplan for plant maintenance; Dr Ian T. Baldwin and two reviewers for helpful comments. D.P.W.H. is a recipient of the Natural Sciences and Engineering Research Council of Canada (NSERC) postdoctoral fellowship. G.A. is a recipient of a fellowship of the Japan Society for the Promotion of Science. This work was supported by grants to J.B. from NSERC, and the Canada Foundation for Innovation (CFI).

Supplementary Material

The following material is available from http://www.blackwellpublishing.com/products/journals/suppmat/TPJ/TPJ1987/TPJ1987sm.htm

Figure S1.  GC–MS analysis of volatiles collected on superq from systemic portions of FTC, M. disstria, induced P. trichocarpa × deltoides H11-11 between 12.00 and 15.00 hours (third quarter of light period).

Numbers beside each peak refer to the associated mass spectra. In each case the top spectrum is the peak of interest and the bottom spectrum corresponds to the best library match (chemstation software, Hewlett Packard, USA; Wiley library; and Adams, 2001). (E)-β-Ocimene, linalool, DMNT, and (−)-germacrene D were also identified by retention time matching with authentic standards. The possibility that the peak identified as benzene cyanide was actually indole ruled out when an authentic indole standard did not match retention time. Phenethyl acetate was sometimes present in diurnal samples of systemic volatiles as well, and might be present in this GC trace as the small peak at approximately 32 min. Concurrent volatile collections of systemic volatiles of mechanically wounded trees showed virtually none of any of these volatiles.

Figure S2.  Insect-induced systemic emission of (−)-germacrene D from poplar during and after feeding by the FTC, M. disstria, and systemic emission upon mechanical wounding of the leaves with scissors.

(a) Volatile emissions were collected from systemic upper sections of the plants. Lower parts of the plants where the treatments were applied were not included in the volatile collection.

(b) Black bars represent emissions from FTC-infested trees while white bars correspond to scissor wounding. Four-hour collection periods are shown across the bottom x-axis, while total elapsed time is shown on the top x-axis. Alternating light and shaded sections along the graph correspond to precise 12-h light and dark periods. The double-headed arrow covers the time in which FTCs were in contact with the trees. Time periods in which (−)-germacrene D emissions were not detected in GC–MS analyses are denoted by very small bars.

Three independent replicates are shown.

Figure S3.  Identification of stereochemistry of (−)-germacrene D produced by PtdTPS1 and emitted from poplar leaves by GC–MS and GC–FID.

(a) Mass spectrum of the FTC-induced single major volatile, germacrene D, released by poplar clone H11-11 shown (top) together with that of a library standard (bottom). The spectrum of an authentic germacrene D standard is shown in Figure 4.

(b) Stereochemistry of FTC-induced (−)-germacrene D was identified by comparison of retention time with that of authentic standards of (−)-germacrene D and (+)-germacrene D using GC–FID. Also shown is the sesquiterpene product formed by recombinant PtdTPS1 enzyme in vitro.

The complete full-length cDNA sequence for PtdTPS1 (−)-germacrene D synthase is available in GenBank under the Accession number AY438099.

Ancillary