May, 2006

The epithet "secondary" connotes an ecological function for a metabolite. Given the universality of this connotation, it is surprising that the weight of the evidence for an ecological function of most "secondary" metabolites stems from only inferences about the type of metabolite (if it's an alkaloid, it must be defensive) and its patterns of production: spatial (metabolites from flowers attract pollinator) or temporal (herbivore-elicitation signifies defense). Such inferential evidence of function is weak. To rigorously establish an ecological function, one must manipulate the expression of a metabolite and verify that an ecological interaction is subsequently altered. Such manipulative experiments are possible with conventional chemical techniques when metabolites are externalized, but for plants that store, induce or express metabolites in a highly tissue-specific manner with particular temporal patterns, manipulation is only possible when genes regulating secondary metabolite production can be altered (silenced or over-expressed). Unfortunately, the plant systems that are currently readily transformable are frequently not optimal for ecological research. Crop plants have been selected for yield maximization, a process that has likely altered ecological responses, and Arabidopsis thaliana (L.) lacks important ecological interactions which can be analyzed in near relatives, but transformation systems are not currently available for these species.

The Dept. of Molecular Ecology at the Max Planck Institute for Chemical Ecology in Jena, Germany, has developed Nicotiana attenuata as an ecological expression system for understanding the ecological function of secondary metabolites, with a particular emphasis on those that mediate plant-insect interactions in the 'agricultural niche.' N. attenuata, an annual tobacco plant native to the Great Basin Desert of North America, has the unusual tendency to chase fires. Namely, it colonizes pine-juniper-sage woodlands for 1–3 years after wildfires have consumed the vegetation, and is occasionally found in roadsides or after new construction has disturbed native vegetation. This habitat selection is primarily determined by its germination behavior. Producing a dormant, long-lived seedbank, it synchronizes its germination with the post-fire environment by responding to a combination of stimulants found in wood smoke and inhibitors from the unburned litter of the dominant vegetation. Consequently, seeds germinate synchronously into nitrogen rich soils and therefore, are selected to grow rapidly and compete with each other when water is readily available. Herbivores from more than 20 different taxa attack the plants at a variety of spatial scales, from mammalian browsers that consume entire plants to intracellular-feeding insects. The most damaging herbivores for a given population differ from year to year, as they too must re-colonize these habitats after fires. How successfully a given genotype of N. attenuata alters its defensive phenotype in response to these highly variable biotic selection regimes and translates vegetative growth into seed production, will determine its representation in the seedbank, and, hence, its Darwinian fitness. In short, we selected this species for two reasons 1) it induces the expression of various secondary metabolites in response to different herbivore species and 2) it evolved to grow in the primordial agricultural niche—the immediate post-fire environment. Understanding the genetic basis of N. attenuata's phenotypic plasticity might provide genetic tools to engineer more ecologically sophisticated crop plants.

To illustrate the specificity with which this plant tailors its defense responses to different herbivores, I contrast the responses to mechanical damage (simulates the response to some mammalian browsers) and to jasmonate (JA) elicitation with the response to attack from a specialized lepidopteran herbivore, Manduca sexta. Wounding elicits a massive metabolic commitment to nicotine production, a potent direct defense that is produced and distributed throughout the plant in a manner that optimizes plant fitness. Wounding elicits the JA-cascade and JA-elicitation produces 1) durable resistance to a suite of herbivores, 2) increases in secondary metabolites, which function as direct and indirect defenses, and 3) substantial fitness costs when plants grow among competitors in herbivore-free environments. Plants profoundly alter their wound-responses when M. sexta larvae attack them; the details of this alteration illustrate the importance of intimately understanding a plant's natural history in order to understand its metabolism.

Damage to leaves caused by browsing herbivores, dramatically increases de novo nicotine biosynthesis and whole-plant nicotine accumulation 2–10 fold. Transcripts of the rate-limiting enzyme in its synthesis, putrescine N-methyltransferase (pmt), are only found in the roots, as are protein and activity measures. N. attenuata has two pmt genes, both are tightly co-regulated and transcript levels correlate with rates of de novo biosynthesis. After its synthesis in the roots, nicotine is transported to the shoots in the xylem stream and accumulates in tissues. This pattern is consistent with predictions of optimal defense theory, which argues that defense metabolites are allocated preferentially to tissues with high fitness value and a high probability of being attacked. Young leaves, stems, and reproductive parts tend to have the highest concentrations; roots and old leaves, the lowest. Our current working model for the signal transduction cascade activating this defense is that wounding transiently increases JA pools in shoots, which, in turn, either directly through transport or indirectly through a secondary signal such as systemin, increases JA pools in roots; nicotine synthesis is then stimulated in the roots and nicotine pools increase throughout the plant.

The fitness benefits of JA elicitation for plants under attack are readily seen in field and laboratory studies. N. attenuata plants growing in native populations induced with a root JA treatment early in the growing season had higher nicotine concentrations for the duration of the growing season, lost less leaf area to mammalian browsers, had a lower mortality rate, and produced more viable seed than size-matched controls. Similarly, in laboratory studies, survivorship and growth of M. sexta larvae on JA-treated plants are dramatically lower than on untreated control plants. Given the opportunity, larvae move from induced plants to feed on neighboring controls. In these experiments, JA elicitation clearly increased a plant's direct defenses, possibly accounting for the plants' increased resistance.

JA-elicited nicotine production is likely to account for some of the observed JA-induced resistance. M. sexta larvae, despite their nicotine-resistant physiology, grow faster on leaves with low nicotine compared to leaves cultured in xylem solutions with induced nicotine concentrations, and on plants whose constitutive and induced nicotine production was suppressed by the anti-sense expression of a pmt transcript. However, many other secondary metabolites are induced by JA elicitation of N. attenuata (including phenolics, flavonoids, phenolic putrescine conjugates and diterpene sugar esters), some are known to influence herbivore performance. For example, proteinase inhibitors (PI) are up-regulated by herbivore attack and JA treatment, and are powerful anti-feedants. Moreover, a study that incorporated leaf material from plants flash-frozen at different times after JA elicitation into artificial diets in order to "freeze" the JA-induced chemical dynamics and examine their effects on M. sexta larvae performance, found that rapidly-induced but uncharacterized changes in direct defenses were as important as the induced changes in PIs and nicotine. Hence, while a number of the chemical changes responsible for induced resistance have been identified, many additional ones remain undiscovered. A major challenge will be to understand how induced resistance emerges from all of the chemical changes affected by elicitation.

In contrast to direct defenses, how indirect defenses function chemically is well understood. JA elicitation and herbivore attack from 4 different species of insects (not mechanical wounding) cause plants to systemically release a bouquet of mono-and sesquiterpenes, in addition to the green leaf volatiles that are primarily released from the wounded leaves (Figure 1). The volatile release has been verified in plants grown in native populations, where it functions as an indirect defense in two distinct ways. The volatile release attracts predatory bugs to Manduca eggs and feeding larvae, dramatically increasing predation rates. Second, the volatile release decreases oviposition rates from adult moths. These ovipositing adults are likely relying on the volatile release to identify host plants without competitors (a single Manduca larvae requires many host plants to complete development) and to avoid plants on which predators are probably present. The profound effectiveness of this indirect defense was demonstrated by a field study, in which the volatile release was estimated to decrease herbivore loads by 90%. By synthesizing single components of the herbivore-induced volatile bouquet and applying them to unattacked plants in quantities naturally emitted, it was demonstrated that individual components from all three major biosynthetic pathways contributing to the volatile bouquet, namely a monoterpene (linalool), a sesquiterpene (bergamotene), and a green leaf volatile (cis-3 hexenanol). Each actively attracted predatory bugs. The observation that enhancing the release of single components of the complex blend was sufficient to attract predators in nature makes the engineering of this type of indirect defense in crop plants, perhaps in conjunction with direct defenses, a tractable proposition.

Figure 1   Alteration of the wound response of wild tobacco plants (Nicotiana attenuata) by Manduca caterpillar feeding. (A) Wounding of the leaf tissue results in a jasmonic acid burst, which is amplified by caterpillar feeding and the application of FACs from larval oral secretions to the wound. (B) Caterpillar attack, or the application of FACs to wounds, but not wounding alone induces the production of volatiles that function as predator attractants in the plants’ indirect defense. (C) Caterpillar feeding and the application of their oral secretions to wounds cause an ethylene burst that (D) attenuates the wound- and jasmonic acid-induced accumulation of nicotine by suppressing the accumulation of transcripts for a key regulatory step in nicotine biosynthesis (pmt: putresine N-methyl transferase). The attenuation of the direct defense, nicotine, may be an adaptation to the feeding of a specialized herbivore that is able to tolerate high alkaloid concentrations and can potentially use them for its own defense. (E) Caterpillar attack and the addition of FACs to plant wounds also result in a transcriptional reconfiguration of the plant’s wound-response. This reconfiguration consists of three temporal and spatial alterations. Addition of FACs antagonizes the wound-induced increase (W) of transcripts encoding threonine deaminase (TD), representing a type-I expression pattern that spreads systemically throughout the plant from the wound site. The wound-induced increase in transcripts in a type-IIa expression pattern (exemplified by PIOX: pathogen-induced peroxidase) is further amplified after application of FACs to wounds. In contrast, the genes with a type-IIb expression pattern are suppressed after wounding and further suppressed with the addition of FACs, as exemplified by the gene encoding the light harvesting complex subunit LHB C1. Both type IIa and IIb patterns are found only in the leaves directly suffering the herbivore attack (Redrawn from Kessler, A. and I. T. Baldwin [2002] Plant responses to insect herbivory: The emerging molecular analysis. Annual Review of Plant Biology (in press).

When attacked by the nicotine-tolerant tobacco specialist, M. Sexta (Figure 1), N. attenuata "recognizes" the attack, as evidenced by alterations in a number of its wound- and JA-elicited responses. The induced increase in JA levels that is normally proportional to the amount of mechanical wounding erupts into a JA burst that increases concentrations 2–10 times those of wound-induced levels and is propagated throughout the damaged leaf ahead of the rapidly foraging herbivore. Wounding and JA-elicitation do not result in ethylene emissions, but an M. sexta attack produces a rapid ethylene burst, which is sustained during larval feeding. The ethylene burst suppresses the wound- and JA- induced accumulation of nicotine biosynthetic genes, NaPMT1 and 2, and the associated nicotine accumulations. The ethylene burst does not suppress the volatile release. Manduca attack, therefore, down-regulates a major direct defense, nicotine, while up-regulating an indirect defense, the volatile release (Figure 1).

Applying larval oral secretions and regurgitants to mechanical wounds can mimic all of the Manduca-induced changes in N. attenuata's wound responses. A suite of 8 fatty acid amino acid conjugates (FACs; Figure 1) in the oral secretions is necessary and sufficient for not only the transcriptional changes mentioned below but also the JA burst and the volatile release. If these FACs are removed from the oral secretions by ion-exchange chromatography, eliciting activity is lost and regained when synthetic FACs are added back to the ion-exchanged, inactive, oral secretions. Since the oral secretions of other tobacco-feeding insects (e.g. Heliothis virescens) also elicited similar changes in wound-induced mRNA, maybe all tobacco feeding lepidoptera are "recognized" by a similar suite of FACs.

In addition to the changes in defense phenotype after an M. sexta attack, N. attenuata also undergoes a major transcriptional re-organization. DDRT-PCR was used to gain an unbiased view of the transcriptional changes, and from this study, it was estimated that more than 500 genes respond to an herbivore attack. The herbivore-regulated genes could be crudely classified as being related to photosynthesis, electron transport, cytoskeleton, carbon and nitrogen metabolism, signaling, and a group responding to stress, wounding, or invasion of pathogens. Overall, transcripts involved in photosynthesis were strongly down-regulated, while those responding to stress, wounding, and pathogens, and involved in shifting carbon and nitrogen to defense were strongly up-regulated. To separate the wound-induced changes from the changes elicited by the M. sexta oral secretion and regurgitant, a subset of the differentially expressed transcripts was analyzed, and three discrete patterns of expression were identified. Regurgitant modified the wound-induced responses by suppressing wound-induced transcripts systemically in the plant (type I) or amplifying the wound response in the attacked leaves (type II). This amplification was either a down-regulation of wound-suppressed transcripts (type IIb) or an up-regulation of wound-increased transcripts (type IIa) (Figure 1). Interestingly, all three patterns of Manduca-induced transcriptional changes of N. attenuata's wound response could be fully mimicked by adding minute amounts of FACs to wounds. The amounts of FACs required are so small that they may be transferred to the plant during normal feeding. These coordinated changes point to the existence of central herbivore-activated regulators of metabolism, which in turn are activated by minute amounts of FACs in Manduca's oral secretions. Identification of these putative trans-activating factors would provide important insights into how plants recognize and coordinate their metabolic responses to an herbivore attack.

The significance of this research is that a plant's ecological interactions are played out in an arena that is larger than the plant itself and as indirect defenses so clearly illustrate, incorporates many community-level components. These higher-order interactions can reverse the fitness outcome of a trait, as occurs when plant chemical defenses are sequestered by herbivores and used against their own predators, or establish linkages between responses which are incomprehensible unless a plant's natural history is taken into account. Only by manipulating the production of a secondary metabolite and exposing plants to the selection pressures of their native habitats can one determine the raison d'etre of these intriguing metabolites.

References

Baldwin, I. T. (2001) An ecological-motivated analysis of plant-herbivore interactions in native tobacco Plant Physiology 127:1449–1458.

Baldwin, I. T., Halitschke, R., Kessler, A., and Schittko, U. (2001) Merging molecular and ecological approaches in plant-insect interactions. Current Opinions in Plant Biology 4(4): 351–358.

Heil, M., and Baldwin, I. T. (2002) Fitness Costs of Induced Resistance: the emerging experimental support for a slippery concept. Trends in Plant Science [in press].