Synthesizing specificity: multiple approaches to ... - Semantic Scholar

TRPLSC-963; No. of Pages 4

Editorial

Special Issue: Specificity of plant–enemy interactions

Synthesizing specificity: multiple approaches to understanding the attack and defense of plants Anurag A. Agrawal1 and Martin Heil2 1 2

Department of Ecology and Evolutionary Biology, Cornell University, E425 Corson Hall, Ithaca, NY 14853-2701, USA Departamento de Ingenierıá Gene´tica, CINVESTAV - Irapuato, Me´xico

The concept of specificity in plant–herbivore and plant– pathogen interactions excites plant pathologists, molecular biologists and animal ecologists alike. This excitement grows out of the notion that individual plant and enemy species (or populations) are reciprocally interacting in a way that shapes their traits and evolution [1,2]. Why is it that most herbivores and pathogens attack a minute fraction of the plants or even plant organs available to them? How do plants manage to defend against diverse enemies? Why are plant enemies specialized at all, given that specialization seems to simply limit the number of available hosts? Are most current plant–enemy interactions the result of a coevolutionary history, and can these be manipulated to protect agricultural crops from pest insects and disease, and ecosystems from invasive species? These are the questions central to this Special Issue of Trends in Plant Science. Here, we combine perspectives of the plant with that of its enemies in order to focus on the traits that allow for successful plant defense versus successful exploitation of plant tissues. Although the topic is often approached from different research traditions (ecology, genetics, physiology, etc.), scientists studying herbivores and plant pathogens have occasionally joined forces and should continue to do so, because there is much to be learned by crossing traditional academic boundaries [3–5]. In addition, we now realize that co-infection, multiple attack, and interactions between herbivores and pathogens are themselves commonplace [6,7]. At the core of issues relating to specificity are two contrasting views, one from the perspective of the plant and one from the perspective of the enemy. First, from a plant’s perspective, there are myriad primary protective barriers, and some of these will be effective against many, if not most attackers [8]. For example, the plant cuticle represents the first barrier encountered by most herbivores and pathogens. Even once this is breached, general strategies, such as the production of hydrogen peroxide, are used to strengthen cell walls. Phenolics occur in most plants, and many of these compounds likely serve some protective role. Even more common, perhaps ubiquitous, are defensive proteins, which are frequently induced upon attack and, depending on their specific structure, can protect plants from pathogens and/or herbivores. Other fairly general forms of defense include the production of hydrogen cyanide and latex, each found in nearly 10% Corresponding author: Agrawal, A.A. ([email protected]).

of all flowering plants, and both activated upon tissue damage [9]. Despite these general barriers against enemies, the diversity of enemies that are likely to attack any given plant begs two important questions: (i) Do some traits that protect the plant against one attacker make the plant more susceptible to attack from others?; and (ii) Given the limited resources available, even if the first question does not apply, could a plant simultaneously defend against all attackers? This leads to further questions on the strategies of the attackers. What are the central traits of enemies enabling them to overcome these multiple defensive strategies? Are enemies that are particularly efficient at exploiting one specific host necessarily suffering from a lower performance on others, and what are the causal mechanisms that underlie this ‘jack of all traits – master of none’ principle? The good news is that we know some of the answers. It is reasonably certain that some plant traits that are expressed in a defensive context are a double-edged sword from the plant’s perspective [10,11], and there is no way to simultaneously deploy even a fraction of the total plant defensive traits in a plant genome. As a result, phenotypically plastic means of defense are the norm in plants. Two articles in this Special Issue by Matthias Erb and colleagues, and Noah Whiteman and colleagues are devoted to understanding the mechanisms and evolution of attacker specific responses and how they may be coordinated. From a more ecological perspective, the extent to which plant traits (driven by genotypic variation) impact single enemies, guilds of attackers and entire communities, is addressed by Thomas G. Whitham and colleagues. The extent to which we expect specific plant responses to specific enemies will depend on the extent of natural selection imposed by each enemy and this, in turn, may be driven by the extent to which enemies are specialized on particular plants (Figure 1). The issue of specificity, from the perspective of the enemy, is tied up in the potential benefits of being able to make a living on a plant resource that is perhaps less available to more generalized enemies or more predictable in its chemical composition [12]. For generalists, in the extreme form, some microbes can infect both plant and animal hosts. In a provocative opinion piece in this issue, Adam Schikora and Heribert Hirt explores the specifics of growth and propagation by Salmonella in plant and animal hosts. Although there are a few super-generalist insect

1360-1385/$ – see front matter ß 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tplants.2012.03.011 Trends in Plant Science xx (2012) 1–4

1


Editorial

Trends in Plant Science xxx xxxx, Vol. xxx, No. x

Effective plant defense? Yes

No

Herbivore diet breadth

Specialist

Cycling coevolution (gene-for-gene) Some pathogens, gallers and plants with hypersensitive responses Plant hormonal signaling used in concert with recognition, indirect defense could be effective, but not well-studied

Arms-race coevolution

Gallers, sequesterers

Many classic cases Specialists are tolerant of some defense, but the plant has effective traits. Hormonal signaling can be effective for direct and indirect defense

No coevolution; plant is well defended (via effective general barriers) and there are minimal reciprocal impacts because the generalist has other options

Generalist

Evolution of reduced herbivore virulence and plant tolerance

A very common interaction, many mobile herbivores, each of which has relatively little impact (e.g., grasshoppers)

Plant hormonal defenses are inactive, ineffective, or manipulated

Unspecialized arms-race coevolution Relatively less mobile generalists (larvae) that suppress or deactivate plant defenses using highly conserved plant traits and pathways Plant hormonal signaling coopted by herbivores

Plant hormonal signaling is effective (indirect defense may be ineffective for mobile herbivores, but generalist natural enemies may work)

TRENDS in Plant Science

Figure 1. A preliminary scheme for conceptualizing specificity in multitrophic interactions. For specialist and generalist herbivores, we outline the predicted evolutionary interaction (purple), types of species involved (blue), and mechanisms and roles of plant defense signaling in mediating such interactions (green). Arms-race coevolution sits in the middle because herbivores are adapted to plant defense, but selection continually favors more effective defenses (including indirect defense).

herbivores (typically some of the worst agricultural pests), most herbivores specialize to species in one plant family, and often one or a few genera, and generalists are typically limited to utilizing defined plant organs [13]. Some of the classic predictions in plant–enemy interactions revolve around not only why organisms specialize, but also how specialists differentially interact with their host plants compared with generalists. This is the subject of the papers by Luke G. Barrett and Martin Heil, and Jared G. Ali and Anurag A. Agrawal. At the extreme, some specialist herbivores use plant defenses for host finding or even to sequester secondary compounds that are then used as defenses against their own enemies [10,11]. Although it has been rarely considered in the past, a sequestering versus non-sequestering specialist herbivore is predicted to have different relations with plants, both in terms of the attacker’s tolerance of plant defenses and the most adaptive defense strategy by plants. Indeed, the paradigms about specialists are changing, and it is now well documented that at least some generalist herbivores and pathogens are highly manipulative of their host plants [14]. The papers in this Special Issue elaborate on these issues, and demonstrate why such interactions may make sense in the evolutionary context, despite a muddled interpretation in past literature. 2

Novel approaches that span phylogenetic to transgenic methods will greatly aid in our progress of understanding specificity in the interactions of different species. Many modern theories of plant–enemy interactions typically invoke three trophic levels, with the potential for several specific linkages between any of the potential pairings (plant–enemy, enemy–predator and plant–predator) [15]. Jonathan Gershenzon and colleagues focus on the issue of indirect defense (i.e. plant–predator signaling), and whether natural selection could well have resulted in specific adaptations across trophic levels. Demonstration of such specificity has long been a holy grail and we still have remarkably few examples of adaptive specificity in plant defense against herbivores. Nonetheless, most researchers believe that there is the potential for such adaptive defense tailoring by plants, as it is commonly seen in effectormediated and highly specialized interactions between plants and pathogens. Similarly, we don’t yet understand why specialist herbivores specialize, and what the consequences may be for interactions with plants or predators. However, all is not lost. Technological advances, such as those outlined by Marcel Dicke and colleagues and, perhaps more importantly, conceptual advances, some of which are outlined in this Special Issue, provide a roadmap for how to proceed.


Editorial An evolutionary framework Among all the papers in this Special Issue, there is an underlying evolutionary framework. To solidify this here, we offer a novel conceptualization for plant–enemy interactions (Figure 1). Although ‘arms race coevolution’ has been the dominant paradigm in plant–enemy interactions for decades, we posit that such interactions only occupy a small space of the conceptual landscape (Figure 1). In particular, arms race coevolution is only expected when plants interact with specialist parasites, where plant traits can directly impact the fitness of the parasite, and when the parasite is virulent and abundant enough to impose fitness losses to the plant. As suggested by several papers in this Special Issue, specialist enemies may be somewhat tolerant of certain plant defenses, but plants nonetheless can defend against them, often using hormonal signaling to upregulate direct and indirect defenses. By contrast, when specialists have truly broken the code of plant defense, most of the secondary compounds that have evolved as induced defenses may be ineffective, and plant parasites typically evolve low virulence (and plants evolve tolerance). The view of coevolution in plant–pathogen interactions is dominated by the gene-for-gene concept [4]; in this alternative to arms race coevolution, virulence and avirulence genes in a population may show stable cycles due to frequency-dependent selection (Figure 1). In particular, plant and pathogen phenotypes do not ‘escalate’ here, but instead persistent attack results in the evolution of a novel plant resistance (typically alleles that are unrecognized or impervious to particular alleles in pathogens). Ultimately, a matching allele evolves in the pathogen and sweeps through the population, conferring virulence against the plant. This interaction cycles via frequency dependence, because the critical phenotypes have to do with allele matching, rather than increasingly virulent parasites and increasingly defended host plants. Although such cycling is a hallmark of coevolutionary interactions between plants and pathogens, relatively little is known about such interactions of plants with herbivores. The evolutionary landscape for generalist plant enemies is often different to that of specialists [11]. Yes, occasionally generalists attack large parts of local host populations, leaving them with reduced fitness to the point where selection favors defense. However, such defenses are likely only maintained by natural selection in the cases where generalized defenses will be effective against several, taxonomically unrelated enemies. Thus, the reciprocal impact of plant traits on generalists is often composed of general barriers (Figure 1). In other words, we argue that the main lines of general defense of plants may well be effective against generalist herbivores, but these are likely to be well conserved among plant species. Hormonally regulated defense expression may indeed be effective against generalists (in an attempt to send these herbivores away), but these same means, because they are highly conserved, may be subject to manipulation and suppression by some generalist enemies [16]. As some of the articles in this Special Issue discuss, it is still too early to know whether such suppression is most common among generalist enemies. However, perhaps the key point to emphasize is that the outcome of plant–enemy interactions depends not


only on the strategy of the enemy, but also of the plant’s ability to recognize that enemy and defend appropriately. Too often we take a single perspective, that of the enemy or the plant, and assume the other party is static. The way it appears however, and this should not surprise anyone, is that coevolution proceeds as a reciprocal evolutionary interaction. Such coevolutionary interactions play out in ecological time as a back-and-forth, as envisioned in Dangl and Jones’ zig-zag model [17]. Plants and their enemies each produce substances involves in recognition and signaling of plant defense [18,19]. Similarly, both plants and their enemies can respond in a highly phenotypically plastic manner to interactions with specific partners, adding a further level of complexity to studies that aim at understanding the reasons of specificity in plant–enemy interactions or its consequences for future evolution [20]. Thus, there will be surprises! In addition, there is a strong need for solid predictions and rigorous analyses that integrate research at the molecular, physiological and ecological level to span the measure of plant traits, resistance to enemies and fitness impacts. We hope that this Special Issue contributes to this needed new synthesis. The Guest Editors References 1 Janzen, D.H. (1980) When is it coevolution? Evolution 34, 611–612 2 Dodds, P.N. and Rathjen, J.P. (2010) Plant immunity: towards an integrated view of plant–pathogen interactions. Nat. Rev. Genet. 11, 539–548 3 Rausher, M.D. (2001) Co-evolution and plant resistance to natural enemies. Nature 411, 857–864 4 Brown, J.K.M. and Tellier, A. (2011) Plant–parasite coevolution: bridging the gap between genetics and ecology. Ann. Rev. Phytopathol. 49, 345–367 5 Erb, M. et al. (2011) Synergies and trade-offs between insect and pathogen resistance in maize leaves and roots. Plant Cell Environ. 34, 1088–1103 6 Thaler, J.S. et al. (2010) Salicylate-mediated interactions between pathogens and herbivores. Ecology 91, 1075–1082 7 Garcia-Guzman, G. and Dirzo, R. (2001) Patterns of leaf-pathogen infection in the understory of a Mexican rain forest: incidence, spatiotemporal variation, and mechanisms of infection. Am. J. Botany 88, 634–645 8 Walters, D. (2011) Plant Defense: Warding Off Attack by Pathogens, Pests and Vertebrate Herbivores, Wiley-Blackwell 9 Agrawal, A.A. (2011) Current trends in the evolutionary ecology of plant defence. Funct. Ecol. 25, 420–432 10 Dobler, S. et al. (2011) Coping with toxic plant compounds – the insect’s perspective on iridoid glycosides and cardenolides. Phytochemistry 72, 1593–1604 11 Lankau, R.A. (2007) Specialist and generalist herbivores exert opposing selection on a chemical defense. New Phytol. 175, 176–184 12 Singer, M.S. (2008) Evolutionary ecology of polyphagy. In The Evolutionary Biology of Herbivorous Insects. Specialization, Speciation, and Radiation (Tilmon, K., ed.), pp. 29–42, University of California Press 13 Schoonhoven, L. et al. (2005) Insect–Plant Biology, (2nd edn), Oxford University Press 14 Musser, R.O. et al. (2005) Evidence that the caterpillar salivary enzyme glucose oxidase provides herbivore offense in Solanaceous plants. Arch. Insect Biochem. Physiol. 58, 128–137 15 Heil, M. (2008) Indirect defence via tritrophic interactions. New Phytol. 178, 41–61 16 Zarate, S.I. et al. (2007) Silverleaf whitefly induces salicylic acid defenses and suppresses effectual jasmonic acid defenses. Plant Physiol. 143, 866–875 3


Editorial 17 Jones, J.D.G. and Dangl, J.L. (2006) The plant immune system. Nature 444, 323–329 18 Heil, M. et al. (2012) How plants sense wounds: damaged-self recognition is based on plant-derived elicitors and induces octadecanoid signaling. PLoS ONE 7, e30537

4


19 Alfano, J.R. and Collmer, A. (2004) Type III secretion system effector proteins: double agents in bacterial disease and plant defense. Ann. Rev. Phytopathol. 42, 385–414 20 Agrawal, A.A. (2001) Phenotypic plasticity in the interactions and evolution of species. Science 294, 321–326


Scientific Life: TrendsTalk

Interview with Anurag A. Agrawal

Anurag Agrawal, born in Allentown, Pennsylvania (USA), received his BA (Biology) and MA (Conservation Biology) from the University of Pennsylvania, where he was inspired by Daniel Janzen, a pioneering evolutionary ecologist, and became intrigued with plant–animal interactions. He completed his PhD (Population Biology) with Rick Karban at the University of California, Davis, and held a Postdoctoral Fellowship at the University of Amsterdam before becoming an Assistant Professor of Botany at the University of Toronto. In 2004, he joined the Department of Ecology & Evolutionary Biology at Cornell, where he is currently a Professor. His research has been broad, embracing chemical ecology, quantitative genetics, phylogenetic analyses, community dynamics and the nascent field of community genetics. Making his work a hobby and some of his hobbies his work has made being a plant biologist and naturalist an immense pleasure. What turned you on to plant biology in the first place? It’s hard to know how this happened to me, but I suppose it was spending much of my childhood outdoors, my mother’s intense love of vegetable gardening, and then some key serendipitous moments, like stumbling into Dan Janzen’s introductory biology course at Penn. I am the type of person who can get interested (and obsessed) by a lot of things, so I feel lucky to have landed here! What paper influenced you most? Ehrlich and Raven 1964. Not because of its specific content, but because of the conceptual linkage between something so mechanistic (plant-produced secondary compounds and their defensive impacts on insects) and something so bigpicture and central to patterns of biodiversity (how new species are formed, generating clades of hyperdiverse plants and herbivores). Remarkably, we don’t know if their hypotheses were correct, but evolutionary chemical ecology has certainly come of age, and great strides are being made right now. What big questions interest you in the long term? To what extent can we generalize in plant science? Are there laws that regulate the ways in which plants respond to the environment, evolve and diversify? I am a huge believer in the integration of work on mechanisms in model organisms and the study of patterns across many species. For example, it is remarkable that what we know about highly conserved traits from some of our model organisms (e.g. Arabidopsis and tomato) indicate that they interact with the environment in divergent ways. I think we have to reconcile the highly conserved blueprint of most plants

with the diversity of how they actually behave. In part that means a move towards non-model-omics, but also a conscious decision to value the study of patterns in wild species studied in their natural environment. What is the best (professional) advice you have been given? Two things, one specific and one general. As an assistant professor, ‘do another thesis project’. It is about the same time frame, results in the same thing (a novel, cohesive and advanced body of work), and is a concrete goal when thrown into academia on our own. More generally, succeeding in science isn’t easy, but it shouldn’t be a mystery. ‘Do whatever you have to in order to learn the culture of being a scientist.’ And what advice would you give? Be sure to play to your strengths and continually work on your weaknesses. There isn’t a single formula for success in science, but again, it shouldn’t be a mystery, and your recipe will require self-study. Most scientists could improve one or two things that are rate limiting steps (e.g. writing faster), which could be a major improvement. Oh, and you must be prepared to accept a steady stream of criticism and rejection, but there is no limit to what we can accomplish through dreaming and taking risks. What is the biggest hindrance to science? Two things, one general and one specific. First, there is often difficulty in accepting change. A colleague once told me that one of the great things about a life in science is that we have the ability to change what we work on, our approach and philosophy, and what we find inspiring. I couldn’t agree more. Nonetheless, sometimes change is thrust upon us, in terms of funding streams, technology, what questions are hot or passe´, etc. A challenge for the academic industry is allowing creative freedom while maintaining an environment where accepting change is facilitated. Second, most scientists sit on too much unpublished data... although there are many reasons for this (negative results, other things more pressing, student left the lab without finishing the project); it is a bit of a tragedy for the work to be done but not be available in the commons. What has been your biggest mistake in research? Letting my own impatience get the best of me. Was it really a mistake? A few times, yes. Like most things, there are tradeoffs, and my impatience has occasionally been beneficial as well. This interview is done. 1360-1385/$ – see front matter doi:10.1016/j.tplants.2012.03.010 Trends in Plant Science xx (2012) 1

1

Scientific Life: TrendsTalk

Trends in Plant Science May 2012, Vol. 17, No. 5


Interview with Martin Heil

Martin Heil is professor and senior researcher at the CINVESTAV-Irapuato in Me´xico, leading the plant ecology laboratory. His group studies co-adaptations in mutualisms and induced defences of plants against pathogens and herbivores. What influenced your path into plant biology? My way into biology was predictable ever since my early years in kindergarden, in Hanau, Germany, when I already considered plants and animals way more interesting than cars and ball games. Therefore, I studied Biology, Geology and Philosophy at Wu¨rzburg University. So why then did it become plant biology? Well, I always felt more comfortable in a botanical garden than in a zoo, and plants do not run away when a curious scientist is approaching. However, I always saw my interest at the interface between plants, microorganisms and animals and therefore focused in my PhD field work, in Malaysia, on a protective ant–plant mutualism. I continued to work on this topic during my postdoc at CEFE/CNRS in Montpellier, France. I then joined Wilhelm Boland’s Department of Bioorganic Chemistry at the Max-Planck-Institute for Chemical Ecology in Jena, Germany as junior group leader. At the age of 36, I accepted a full professorship at University of Duisburg-Essen, where it took me less than three years to find out that for me research is more interesting than administrative duties. For this reason, in 2007, I moved to CINVESTAV-Irapuato in Me´xico, where I am currently leading the plant ecology laboratory. What was the driving force for you to move to Mexico for your research? The degree of freedom that I have in my work and my private life and the very low load of administrative duties. Moreover, the level of appreciation that I receive here for my research is by orders of magnitude higher than before, the authorities of the institute give me the impression that they do whatever they can to support my research, and as long as I maintain my productivity I can practically do whatever I want. Fund raising is comparably easy, because Mexico invests relatively more in science than for example the USA, at least when we divide national funds for science by the number of scientists that are competing for these funds. Another important aspect is that the Mexican system as a whole appreciates your scientific productivity, in terms of publications, citations and graduated students, which not only strongly influences the personal salary, but also the chance to get projects accepted. How did you decide on your current research topics? I have no major strategy at all, but rather follow a stochastic, phenomena-based approach. My decisions are 244

mainly guided by curiosity, that is, when I see an interesting phenomenon in the field or read about a strange observation in the literature, it might easily occur that I start a new project in order to find out what is going on. Has your work been affected by the genomics revolution? Recently, yes, because the new high-throughput sequencing approaches allow the investigation of ecologically relevant phenomena in non-model species under field conditions. For example, our ant-Acacias are ecologically extremely interesting plants that possess multiple adaptations to maintain the ant–plant mutualism. However, because they are taxonomically very distant from any model plant, classical genetic tools were not suitable to approach these phenomena at the genetic level. With the recently developed highthroughput sequencing techniques, we are now able to search for the genetic control mechanisms that underlie a functioning ant–plant mutualism. What would you be if you were not a biologist? My way into biology seemed predestined; at least, I cannot remember ever actively considering any alternatives. However, if I had not pursued an academic career, being a ballet dancer or opera singer (tenor) would have been attractive possibilities. In fact, I personally feel more like an artist than a classical scientist, a phenomenon that perhaps also underlies my ‘‘emotion-guided’’ approach in the selection of research projects. What is your favorite book? Difficult to mention only one. If four are allowed, the selection would be: ‘‘Der Zauberberg’’ by Thomas Mann, ‘‘La tia Julia y el escribidor’’ de Mario Vargas Llosa, ‘‘100 anõs de soledad’’ de Gabriel Garcıá Marquez and, well, to be honest... ‘‘Die Stadt der traümenden Bu¨cher’’ by Walter Moers. Do you have a scientific hero? Immanuel Kant and Werner Heisenberg, because I think that both had the most dramatic impact on our understanding of what science can achieve and of its limitations, and of the way in which we perceive our position in reality. Are scientific controversies helpful? Of course they are, as long as they are maintained free of personal attacks and avoid the abuse of political positions (including the position as a referee) to suppress the unwanted ideas. If we believe Thomas Kuhn, revolutions are the mechanism by which science proceeds. 1360-1385/$ – see front matter http://dx.doi.org/10.1016/j.tplants.2012.03.013 Trends in Plant Science, May 2012, Vol. 17, No. 5


Opinion


Plants as alternative hosts for Salmonella Adam Schikora1, Ana V. Garcia2 and Heribert Hirt2 1

Institute for Plant Pathology and Applied Zoology, Research Centre for BioSystems, Land Use and Nutrition, JL University Giessen, Heinrich-Buff-Ring 26-32, 35392 Giessen, Germany 2 URGV Plant Genomics, INRA/CNRS/University of Evry, 2 rue Gaston Cre´mieux, 91000 Evry, France

Recent findings show that many human pathogenic bacteria can use multiple host organisms. For example, Salmonella Typhimurium can use plants as alternative hosts to humans and other animals. These bacteria are able to adhere to plant surfaces and actively infect the interior of plants. Similarly to the infection of animal cells, S. Typhimurium suppresses plant defense responses by a type III secretion mechanism, indicating that these bacteria possess a dedicated multi-kingdom infection strategy, raising the question of host specificity. In addition, evidence is accumulating that the interaction of Salmonella with plants is an active process with different levels of specificity, because different Salmonella serovars show variations in pathogenicity, and different plant species reveal various levels of resistance towards these bacteria. Plant-originated salmonellosis Several reports indicated that bacteria, which are pathogenic to humans and other mammals, are able to infect plants. Salmonella enterica, Pseudomonas aeruginosa, Burkholderia cepacia, Erwinia spp., Staphylococcus aureus, Escherichia coli O157:H7 and Listeria monocytogenes infect animals and plants [1–5]. Amongst these pathogens, Salmonella bacteria are the major cause of food poisoning. These Gram-negative enteropathogenic bacteria can successfully colonize animals, humans and plants. Their genus is divided into two species: Salmonella bongori and Salmonella enterica, encompassing several hundred isolates, which are typically named after the place of origin [6]. The species S. enterica is additionally divided into seven subspecies, one of them, S. enterica subsp. enterica, is the major cause of salmonellosis in humans. The most common mode of infection is ingestion of contaminated food or water. Moreover, many reports have linked food poisoning with the consumption of Salmonella-contaminated raw vegetables and fruits (for review see [2,7]). Studies in various European countries revealed that in 2007, 0.3–2.3% of raw vegetables were infected with Salmonella bacteria [8]. In the USA, the proportion of raw food-associated salmonellosis outbreaks increased from 0.7% in the 1960s to 6% in the 1990s [9], and crossed 25% in recent years [10]. Most studies on Salmonella–plant interactions suggested an epiphytic lifestyle of Salmonella on plants. However, a growing body of evidence Corresponding author: Hirt, H. ([email protected]).

points to a directed process in which the bacteria infect various plants and use them as viable hosts (Table 1) [11–22]. The ability to infect and grow on such diverse hosts is a remarkable example of the lack of specificity seen in so many other microbes (Figure 1). Do plants serve as alternative hosts or are they part of the Salmonella life cycle? Adhesion is typically the first step of an infection by Salmonella. Diverse S. enterica serovars have been shown to adhere to plant surfaces, and many Salmonella serovars bind to plants significantly better than for instance the pathogenic E. coli strain O157:H7 [23]. Evidence suggests that Salmonella actively attach to plant tissues and only viable bacteria can successfully colonize plants [19]. In a screen of 6000 S. Newport mutants, 20 mutants were identified with lower attachment ability to Medicago sativa (alfalfa) sprouts [12]. Interestingly, some of the genes identified in this study code for the surface-exposed aggregative fimbria nucleator curli (agfB) and for the global stress regulator rpoS which regulates the production of curli, cellulose and other adhesins that are important also for animal pathogenicity. AgfD, which was also identified in this study, plays not only a central role in the ability to attach to plant surfaces [24], but also in the environmental fitness and the pathogenicity of the bacteria toward animals [25]. In addition, it was shown that yihO (involved in O-antigen capsule formation) and bcsA (coding for a cellulose synthase) are also important for adhesion to alfalfa sprouts [24], whereas cellulose and curli are involved in transmission of S. Typhimurium from water onto parsley (Petroselinum hortense) leaves [26]. In another study, two previously uncharacterized genes (STM0278 and STM0650) were characterized as important factors for the infection of alfalfa sprouts, due to their essential role in biofilm formation and swarming [11]. It is thus becoming evident that the genetic equipment of Salmonella, previously thought to be animal-infection specific, plays an important role in the infection of animals and plants alike. Surprisingly, a comparative study on the internal colonization in lettuce (Lactuca sativa) leaves by five S. enterica serovars (Dublin, Enteritidis, Montevideo, Newport and Typhimurium) indicated significant differences between the different serovars, indicating that distinct genetic backgrounds have an impact on the pathogenic behavior towards plants [16]. A similar study conducted on the

1360-1385/$ – see front matter ß 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.tplants.2012.03.007 Trends in Plant Science xx (2012) 1–5

1


Opinion


Table 1. Known interactions between Salmonella and plantsa Salmonella strain S. Anatum DMST 19600 S. enterica Dublin S. enterica, diverse serovars S. enterica, diverse serovars S. enterica, diverse serovars S. enterica, diverse serovars S. Newport S. Newport S. Newport, Enteriditis, mutants S. Thompson RM1987 S. Typhimurium S. Typhimurium S. Typhimurium 14028 S. Typhimurium 14028 S. S. S. S. S. S. S.

Typhimurium Typhimurium Typhimurium Typhimurium Typhimurium Typhimurium Typhimurium

14028 14028, 1344 DT104 MAE110, MAE119 SL1344 SL1344 14028

Infected plant Cabbage Lettuce Lettuce Arabidopsis Tomato, pepper Lettuce, cabbage Alfalfa Alfalfa Alfalfa Lettuce Barley (Hordeum vulgare) Potato (Solanum tuberosum) Tomato fruits Arabidopsis Arabidopsis Tobacco Lettuce Tomato Lettuce Diverse Arabidopsis

Main finding Temperature-dependent susceptibility to infection Colonization of lettuce and transcriptome of response to infection Different serovars vary in their colonizing capacity Strains from O-serogroup induce chlorosis and wilting in Arabidopsis Cultivar-dependent colonization, trichomes as infection point Serovar-dependent divergences in attachment to leaves Identification of two new genes required for attachment to plants Screen of 6000 mutants for their ability to attach to plant surface Cellulose and O-antigen capsule play role in the attachment to plants Increased infection was observed in elderly leaves Colonization of barley roots

Refs [46] [47] [16] [30] [13] [27] [11] [12] [24] [48] [49]

Attachment to plant surface is an active process

[19]

Screen for bacterial genes expressed upon plant infection Plants induce defense mechanisms after infection, bacteria internalized in plants cells Suppression of plant immune system is T3SS-dependent Wild type bacteria suppress plant defense reactions A passage via lettuce increased attachment capacity to epithelial cells Bacteria spread systemically and colonize non-infected leaves and fruits Internalization via stomata is light dependent and requires chemotaxis Internalization of bacteria varies between leafy vegetables Plant defense is required for resistance towards Salmonella

[18] [20] [21] [45] [50] [22] [17] [14] [15]

a

The majority of the studies focus on different serovars of S. enterica subspecies enterica interacting with Arabidopsis or plants traditionally associated with salmonellosis outbreaks such as lettuce, tomato and alfalfa. The list presented here summarizes the research on the interaction between Salmonella and the plant immune system, as well as the genetic requirement to infect plants. Due to length restrictions, it is impossible to cover comprehensively the broad literature of different plant-originated outbreaks and the anti-microbial activity of diverse plants.

serovars Braenderup, Negev, Newport, Tennessee and Thompson, likewise revealed differences between the tested serovars [27]. Interestingly, the authors pointed out a correlation between the capacity to produce biofilms and the attachment to leaves, with S. Thompson producing the strongest biofilms and showing the most efficient adhesion to lettuce leaves [27]. Salmonella can live inside plants In animals, Salmonella actively enter epithelial and other cell types in order to replicate and spread through the organism. The question whether Salmonella use a similar strategy to infect plants is therefore of great interest. Salmonella were found to form biofilm-like structures on the surface of roots, preferentially colonizing regions around emerging lateral roots and wounded tissues [15,20]. The formation of biofilms of Salmonella on leaves was also reported. Recently, three reports presented the possible entry points of bacteria to the inner layers of leaves [13,14,17] and it was postulated that trichomes are preferential colonization sites [13]. By contrast, it was shown that Salmonella use stomata as entry points in order to penetrate lettuce leaves [17]. Moreover, bacterial aggregation near stomata occurs only under light conditions when the stomata are open. Artificial opening of the stomata in the dark had no impact on the bacterial behavior, suggesting that bacteria are attracted to photosynthesis-dependent products. Previously, we showed that the GFP-marked S. Typhimurium 14028s bacteria can be observed inside root hairs at 3 h, and bacterial titers increased at 20 h after inoculation of Arabidopsis plants [20]. 2

Additional tests revealed that motility and the ability of chemotaxis are essential for Salmonella to colonize the interior of lettuce leaves [17]. In a follow-up report, the same group demonstrated that not all plants are equally susceptible (or resistant) to Salmonella internal infection. Using GFP-marked bacteria, the authors analyzed the internalization of the S. Typhimurium strain 1344 in many leafy vegetables and herbs [14]. In the same year, another study reported that S. Typhimurium strain MAE110 is able to translocate within tomato (Solanum lycopersicum) plants, infecting distal, non-infected leaves and fruits without visible symptoms and only slightly reducing plant growth [22]. Interestingly, while some plant species [e.g. arugula (Diplotaxis tenuifolia)], allowed Salmonella to internalize, some others (e.g. parsley), seemed to have effective means to prevent infection [14]. Studies on lettuce, cabbage (Brassica oleracea) and tomato demonstrated significant differences in the susceptibility to Salmonella infection [13,16], pointing to an important role of plant innate immunity in modulating the response to infection by these bacteria. By contrast, pathogenic bacteria often use type III secretion system (T3SS)-dependent injection of effector proteins in order to modulate host physiology and suppress the immune system. To answer the question whether Salmonella rely on T3SS for infection of plants, mutants in two Salmonella T3SS were tested for their performance on plants. Both of the T3SS mutants are unable to inject effector proteins into host cells and are therefore not virulent for animal hosts [28,29]. Although these T3SS mutant strains showed normal proliferation rates when grown in standard medium, their proliferation in


Opinion


H+ ROS T3SS

H+

PRRs

MAMPs

H+

MAPKKK MAPKK MAPK MAPK TF Defense related genes


Figure 1. Wild type Salmonella are able to attach to plant surfaces and infect plants via stomata openings or roots. Upon infection, Salmonella hinder the enhanced production of reactive oxygen species (ROS) and prevent pH changes in the apoplast. Moreover, Salmonella actively prevent the transcriptional activation of defenserelated genes. Abbreviations: MAMP, microbe-associated molecular pattern (yellow circles); MAPK, mitogen-activated protein kinase; PRR, pattern recognition receptor; T3SS, type III secretion system; TF, transcription factor; red circles represent Salmonella effectors; green circles represent products of defense-related genes.

Arabidopsis (Arabidopsis thaliana) plants was compromised, indicating that both SPI-1- and SPI-2-encoded type III secretion systems are needed for successful plant infection [21]. Plant responses to Salmonella infection Upon inoculation, Arabidopsis responds to Salmonella with a rapid induction of defense responses, including the activation of mitogen-activated protein kinases MPK3, MPK4 and MPK6 that is followed by the expression of a number of defense genes, such as PDF1.2 or the pathogenesis-related genes PR2 and PR4 [20]. Transcriptome analysis of Arabidopsis plants showed differential expression of about 250 and 1300 genes at 2 and 24 h after Salmonella infection, respectively. With the exception of 32 genes, the Salmonella-induced differentially expressed genes were also affected by inoculation with the non-pathogenic E. coli laboratory strain DH5a and the pathogenic Pseudomonas syringae strain DC3000 [21]. Among the genes that were induced by E. coli DH5a, S. Typhimurium 14028 and P. syringae DC3000, about 160 (including various WRKY and bZIP transcription factors as well as protein kinases and phosphatases) could be identified as a core set of Arabidopsis genes responsive to common bacterial exposure [21].

Towards identification of the plant Salmonella receptors A recent study examined the macroscopic symptoms of wilting and chlorosis in Arabidopsis plants after infiltration with different serovars of S. enterica subsp. enterica, as well as S. enterica subsp. arizonae and diarizonae [30]. Infiltration with S. Senftenberg and also with S. Cannstatt, Krefeld and Liverpool, all of which belong to the serogroup E4 (O: 1, 3, 19) possessing the O-antigen, resulted in rapid wilting and chlorosis. By contrast, infiltration with serovars lacking the O-antigen provoked no symptoms [30]. In addition, the authors stated that the response to Salmonella infiltration is independent of the most prominent and studied pattern recognition receptors, suggesting that specific receptors for Salmonella O-antigen could exist in Arabidopsis. Salmonella factors interacting with the plant immune system In humans, salmonellosis develops after the bacteria enter epithelial cells of the intestine [31]. Although a typical infection usually leads to a self-limiting gastroenteritis, Salmonella can cause systemic infections by invading spleen, liver and other organs in susceptible hosts. Studies of the infection mechanisms in animals have shown that 3


Opinion Salmonella actively remodel the host cell physiology and architecture, and suppress the host immune system by injecting a cocktail of effectors delivered by T3SS. A recently published literature survey revealed a standard list of 62 protein–protein interactions between 22 Salmonella proteins and numerous human proteins [32]. Salmonella enterica subsp. enterica has two distinct T3SSs, T3SS-1 and T3SS-2, encoded by the Salmonella Pathogenicity Islands (SPI) SPI-1 and SPI-2, respectively [33,34]. T3SS-1 secretes at least 16 proteins of which six were shown to interact with the host signaling cascades and the cytoskeleton. T3SS-2 secretes at least 19 S. entericaspecific effector proteins that are involved in survival and multiplication within the host cell [35,36]. The expression and the secretion of SPI-1 and SPI-2 encoded effectors are tightly regulated. Recently, the cytoplasmic SpaO–OrgA–OrgB complex was identified as the sorting platform for T3SS effectors that determines the appropriate hierarchy for protein secretion [37]. This complex enables the sequential delivery of translocases before the secretion of the actual effectors. The authors also described the role of specific chaperones in the recognition and loading of effectors into the sorting SpaO–OrgA–OrgB complex, and postulated that similar sorting platforms might exist in other bacteria [37]. Even though many reports suggest that the mechanisms used by Salmonella to infect animal and plant hosts could be similar, the role of Salmonella T3SS effectors during plant infections remains unclear. To date, 44 Salmonella effectors have been described to be injected into animal and human cells through one or both T3SSs (reviewed in [38]). Several of these effectors target MAPK cascades, which are important regulators of the immune response in animals and plants. SpvC from Salmonella spp. belongs to the OspF family initially identified in Shigella spp. OspF encodes a phosphothreonine lyase that dephosphorylates the pTXpY double phosphorylated activation loop in the ERK1/2 kinases [39–41]. Interestingly, P. syringae HopAI1 is a homolog of SpvC/OspF, and encodes a phosphothreonine lyase that dephosphorylates the threonine residue in the activation loop of activated MAPKs [42]. When expressed in Arabidopsis, HopAI1 directly interacts with MPK3 and MPK6, attenuating flg22-induced MAPK activation and downstream defense responses [40–42]. Besides OspF/SpvC/HopAI1, also the Pseudomonas effector HopPtoD2 has homologs in human pathogenic bacteria. HopPtoD2 is a tyrosine phosphatase which inhibits pathogen-triggered programmed cell death [43], while its homolog from Salmonella SptP, inhibits phosphorylation and membrane localization of Raf kinase and therefore the activation of ERK2 [44]. It is tempting to speculate that the biochemical features of these effectors are conserved between animal and plant hosts, providing Salmonella and other pathogenic bacteria with efficient tools for suppressing the host immune systems. A suppression of the defense responses was recently reported during the infection of tobacco (Nicotiana tabacum) plants with S. Typhimurium. In contrast to living Salmonella, dead or chloramphenicol-treated bacteria elicited an oxidative burst and pH changes in tobacco cells [45], indicating that Salmonella actively engages in the suppression of the 4


plant defense responses. Similar conclusions were reached when comparing the Arabidopsis responses against S. Typhimurium wild type and the T3SS mutants invA or prgH, which lack a functional T3SS [21,45]. Whereas Salmonella wild type and prgH mutants provoke changes in more than 1600 Arabidopsis genes after 24 h, a group of 649 genes is specifically induced by infection with the T3SS mutant. Many of these prgH-specific genes encode proteins related to pathogen responses and ubiquitin-mediated protein degradation. This group of genes also includes BAK1, BIK1, WRKY18 and WRKY33, EIN3, PR4 and PUB23, all of which are marker genes that are upregulated upon plant pathogen infections. The lower expression level of those genes upon infection with wild type Salmonella suggests that the T3SS mutant is unable to employ an effective immune suppression mechanism. These results suggest that Salmonella depend on the T3SS during plant infection and actively suppress immune responses. Concluding remarks Along with E. coli, Salmonella belong to the best-studied bacteria today. The growing number of human infections with pathogenic bacteria derived from vegetables and fruits raise the question of the host specificity mechanisms of these bacteria. Recent reports clearly demonstrate that Salmonella not only passively survive, but also actively infect plants. Moreover, infection of plants depends on the active suppression of the host immune responses by Salmonella. Further studies are clearly warranted to uncover the extent to which the factors and mechanisms employed by Salmonella to infect animals are also used against plants and will likely lead to a better understanding of the evolution of specificity. Acknowledgments The work of AVG and HH is supported from a grant of the ERANET Systems Biology project SHIPREC (Salmonella Host Interaction Project European Consortium). The authors would like to apologize to all colleagues whose work could not be cited because of space limitations.

References 1 Haapalainen, M. et al. (2009) Soluble plant cell signals induce the expression of the type III secretion system of Pseudomonas syringae and upregulate the production of pilus protein HrpA. Mol. Plant Microbe Interact. 22, 282–290 2 Holden, N. et al. (2009) Colonization outwith the colon: plants as an alternative environmental reservoir for human pathogenic enterobacteria. FEMS Microbiol. Rev. 33, 689–703 3 Plotnikova, J.M. et al. (2000) Pathogenesis of the human opportunistic pathogen Pseudomonas aeruginosa PA14 in Arabidopsis. Plant Physiol. 124, 1766–1774 4 Prithiviraj, B. et al. (2005) Staphylococcus aureus pathogenicity on Arabidopsis thaliana is mediated either by a direct effect of salicylic acid on the pathogen or by SA-dependent, NPR1-independent host responses. Plant J. 42, 417–432 5 Milillo, S.R. et al. (2008) Growth and persistence of Listeria monocytogenes isolates on the plant model Arabidopsis thaliana. Food Microbiol. 25, 698–704 6 Lan, R. et al. (2009) Population structure, origins and evolution of major Salmonella enterica clones. Infect. Genet. Evol. 9, 996–1005 7 Brandl, M.T. (2006) Fitness of human enteric pathogens on plants and implications for food safety. Annu. Rev. Phytopathol. 44, 367–392 8 Westrell, T. et al. (2009) Zoonotic infections in Europe in 2007: a summary of the EFSA-ECDC annual report. Euro Surveill. 14 9 Sivapalasingam, S. et al. (2004) Fresh produce: a growing cause of outbreaks of foodborne illness in the United States, 1973 through 1997. J. Food Prot. 67, 2342–2353


Opinion 10 Rangel, J.M. et al. (2005) Epidemiology of Escherichia coli O157:H7 outbreaks, United States, 1982–2002. Emerg. Infect. Dis. 11, 603–609 11 Barak, J.D. et al. (2009) Previously uncharacterized Salmonella enterica genes required for swarming play a role in seedling colonization. Microbiology 155, 3701–3709 12 Barak, J.D. et al. (2005) Salmonella enterica virulence genes are required for bacterial attachment to plant tissue. Appl. Environ. Microbiol. 71, 5685–5691 13 Barak, J.D. et al. (2011) Colonization of tomato plants by Salmonella enterica is cultivar dependent, and type 1 trichomes are preferred colonization sites. Appl. Environ. Microbiol. 77, 498–504 14 Golberg, D. et al. (2011) Salmonella Typhimurium internalization is variable in leafy vegetables and fresh herbs. Int. J. Food Microbiol. 145, 250–257 15 Iniguez, A.L. et al. (2005) Regulation of enteric endophytic bacterial colonization by plant defenses. Mol. Plant Microbe Interact. 18, 169–178 16 Klerks, M.M. et al. (2007) Differential interaction of Salmonella enterica serovars with lettuce cultivars and plant-microbe factors influencing the colonization efficiency. ISME J. 1, 620–631 17 Kroupitski, Y. et al. (2009) Internalization of Salmonella enterica in leaves is induced by light and involves chemotaxis and penetration through open stomata. Appl. Environ. Microbiol. 75, 6076–6086 18 Noel, J.T. et al. (2010) Specific responses of Salmonella enterica to tomato varieties and fruit ripeness identified by in vivo expression technology. PLoS ONE 5, e12406 19 Saggers, E.J. et al. (2008) Salmonella must be viable in order to attach to the surface of prepared vegetable tissues. J. Appl. Microbiol. 105, 1239–1245 20 Schikora, A. et al. (2008) The dark side of the salad: Salmonella Typhimurium overcomes the innate immune response of Arabidopsis thaliana and shows an endopathogenic lifestyle. PLoS ONE 3, e2279 21 Schikora, A. et al. (2011) Conservation of Salmonella infection mechanisms in plants and animals. PLoS ONE 6, e24112 22 Gu, G. et al. (2011) Internal colonization of Salmonella enterica serovar Typhimurium in tomato plants. PLoS ONE 6, e27340 23 Barak, J.D. et al. (2002) Differences in attachment of Salmonella enterica serovars and Escherichia coli O157:H7 to alfalfa sprouts. Appl. Environ. Microbiol. 68, 4758–4763 24 Barak, J.D. et al. (2007) The role of cellulose and O-antigen capsule in the colonization of plants by Salmonella enterica. Mol. Plant Microbe Interact. 20, 1083–1091 25 Gibson, D.L. et al. (2006) Salmonella produces an O-antigen capsule regulated by AgfD and important for environmental persistence. J. Bacteriol, 188, 7722–7730 26 Lapidot, A. and Yaron, S. (2009) Transfer of Salmonella enterica serovar Typhimurium from contaminated irrigation water to parsley is dependent on curli and cellulose, the biofilm matrix components. J. Food Prot. 72, 618–623 27 Patel, J. and Sharma, M. (2010) Differences in attachment of Salmonella enterica serovars to cabbage and lettuce leaves. Int. J. Food Microbiol. 139, 41–47 28 Behlau, I. and Miller, S.I. (1993) A PhoP-repressed gene promotes Salmonella Typhimurium invasion of epithelial cells. J. Bacteriol. 175, 4475–4484 29 Hensel, M. et al. (1997) Functional analysis of ssaJ and the ssaK/U operon, 13 genes encoding components of the type III secretion apparatus of Salmonella Pathogenicity Island 2. Mol. Microbiol. 24, 155–167


30 Berger, C.N. et al. (2011) Salmonella enterica strains belonging to O serogroup 1,3,19 induce chlorosis and wilting of Arabidopsis thaliana leaves. Environ. Microbiol. 13, 1299–1308 31 Patel, J.C. et al. (2005) The functional interface between Salmonella and its host cell: opportunities for therapeutic intervention. Trends Pharmacol. Sci. 26, 564–570 32 Schleker, S. et al. (2011) The current Salmonella–host interactome. Proteomics Clin. Appl. 6, 117–133 33 Collazo, C.M. and Galan, J.E. (1997) The invasion-associated type-III protein secretion system in Salmonella–a review. Gene 192, 51–59 34 Hensel, M. (2000) Salmonella pathogenicity island 2. Mol. Microbiol. 36, 1015–1023 35 Kuhle, V. and Hensel, M. (2004) Cellular microbiology of intracellular Salmonella enterica: functions of the type III secretion system encoded by Salmonella pathogenicity island 2. Cell. Mol. Life Sci. 61, 2812–2826 36 Waterman, S.R. and Holden, D.W. (2003) Functions and effectors of the Salmonella pathogenicity island 2 type III secretion system. Cell. Microbiol. 5, 501–511 37 Lara-Tejero, M. et al. (2011) A sorting platform determines the order of protein secretion in bacterial type III systems. Science 331, 1188–1191 38 Heffron, F. et al. (2011) Salmonella-secreted virulence factors. In Salmonella. From Genome to Function (Porwollik, S., ed), pp. 187– 223, Caister Academic Press 39 Mazurkiewicz, P. et al. (2008) SpvC is a Salmonella effector with phosphothreonine lyase activity on host mitogen-activated protein kinases. Mol. Microbiol. 67, 1371–1383 40 Arbibe, L. et al. (2007) An injected bacterial effector targets chromatin access for transcription factor NF-kappaB to alter transcription of host genes involved in immune responses. Nat. Immunol. 8, 47–56 41 Li, H. et al. (2007) The phosphothreonine lyase activity of a bacterial type III effector family. Science 315, 1000–1003 42 Zhang, J. et al. (2007) A Pseudomonas syringae effector inactivates MAPKs to suppress PAMP-induced immunity in plants. Cell Host Microbe 1, 175–185 43 Espinosa, A. et al. (2003) The Pseudomonas syringae type III-secreted protein HopPtoD2 possesses protein tyrosine phosphatase activity and suppresses programmed cell death in plants. Mol. Microbiol. 49, 377–387 44 Lin, S.L. et al. (2003) SptP, a Salmonella typhimurium type IIIsecreted protein, inhibits the mitogen-activated protein kinase pathway by inhibiting Raf activation. Cell. Microbiol. 5, 267–275 45 Shirron, N. and Yaron, S. (2011) Active suppression of early immune response in tobacco by the human pathogen Salmonella Typhimurium. PLoS ONE 6, e18855 46 Hawaree, N. et al. (2009) Effects of drying temperature and surface characteristics of vegetable on the survival of Salmonella. J. Food Sci. 74, E16–E22 47 Klerks, M.M. et al. (2007) Physiological and molecular responses of Lactuca sativa to colonization by Salmonella enterica serovar Dublin. Appl. Environ. Microbiol. 73, 4905–4914 48 Brandl, M.T. and Amundson, R. (2008) Leaf age as a risk factor in contamination of lettuce with Escherichia coli O157:H7 and Salmonella enterica. Appl. Environ. Microbiol. 74, 2298–2306 49 Kutter, S. et al. (2006) Colonization of barley (Hordeum vulgare) with Salmonella enterica and Listeria spp. FEMS Microbiol. Ecol. 56, 262–271 50 Oliveira, M. et al. (2011) Pathogenic potential of Salmonella Typhimurium DT104 following sequential passage through soil, packaged fresh-cut lettuce and a model gastrointestinal tract. Int. J. Food Microbiol. 148, 149–155

5

Review


Role of phytohormones in insect-specific plant reactions Matthias Erb1, Stefan Meldau2 and Gregg A. Howe3 1

Root–Herbivore Interactions Group, Max Planck Institute for Chemical Ecology, Hans-Kno¨ll-Str. 8, 07745 Jena, Germany Department of Molecular Ecology, Max Planck Institute for Chemical Ecology, Hans-Kno¨ll-Str. 8, 07745 Jena, Germany 3 Biochemistry & Molecular Biology, 122 Plant Biology Lab, Michigan State University, East Lansing, MI 48824-1219, USA 2

The capacity to perceive and respond is integral to biological immune systems, but to what extent can plants specifically recognize and respond to insects? Recent findings suggest that plants possess surveillance systems that are able to detect general patterns of cellular damage as well as highly specific herbivoreassociated cues. The jasmonate (JA) pathway has emerged as the major signaling cassette that integrates information perceived at the plant–insect interface into broad-spectrum defense responses. Specificity can be achieved via JA-independent processes and spatio-temporal changes of JA-modulating hormones, including ethylene (ET), salicylic acid (SA), abscisic acid (ABA), auxin, cytokinins (CK), brassinosteroids (BR) and gibberellins (GB). The identification of receptors and ligands and an integrative view of hormone-mediated response systems are crucial to understand specificity in plant immunity to herbivores. Know your enemy: a golden rule of plant defense? ‘If you know your enemies and know yourself, you can win a hundred battles without a single loss’, states Sun Tzu in his ancient military treatise The Art of War. Plants, as primary producers of organic matter in terrestrial ecosystems, must continuously resist a multitude of attackers and, unlike the armies of Sun Tzu, do not have the option of retreating to safe ground. Have plants nevertheless evolved the capacity to ‘know’ the attacking enemies and adjust their defenses accordingly? In this review, we use the paradigm of molecular specificity in plant–pathogen interactions as a framework to discuss potential mechanisms by which plants specifically recognize and respond to insect herbivores. Plants recognize herbivores via mechanical and chemical cues An appropriate defense response to a biotic threat requires initial recognition. Pathogens are recognized when conserved patterns of microbial molecules, called microbe- or pathogen-associated molecular patterns (MAMPs or PAMPs), are detected by pattern recognition receptors (PRRs) on the surface of the host plant cell, leading to PAMP-triggered immunity (PTI; Figure 1). Damage-associated molecular patterns (DAMPs), which are endogenous molecules that are produced by the plant after infection, Corresponding author: Erb, M. ([email protected]).

250

are also recognized by PRRs to trigger defensive reactions [1]. Pathogens can evade this innate immune response through the action of effector proteins that, upon delivery into the host cell, suppress PTI. Some plant genotypes Glossary Appropriate response: a phenotypical change following herbivory that provides a benefit to the plant. This benefit can be realized either by increasing resistance and fending off the attacker, or by changing the primary metabolism to enable a more effective regrowth after attack. Appropriate responses are not necessarily based on specific recognition and specific metabolic changes, as plants can use general mechanisms to defend themselves against a variety of attackers. Different chewing herbivores are likely to be susceptible to the same defensive mechanisms. By contrast, phloem feeders might require different measures of protection because they only feed on specialized cells. From an adaptive point of view, truly specific responses can be expected to be appropriate. Direct crosstalk: a phenomenon in which two or more hormone pathways either share a common signaling component; for example, the use of corepressor TOPLESS (TPL) by both the JA and auxin pathways, or contain components that physically interact to modify the signal output (e.g. the JAZ–DELLA interaction). The biological significance of direct crosstalk in shaping the outcome of plant– insect interactions remains to be demonstrated. Hormone crosstalk: a phenomenon in which a signal transmitted through one hormone pathway stimulates or represses signal output (e.g. a physiological or defense-related response) from another signaling pathway. Interactions between the hormone signals can be direct or indirect (see direct and indirect crosstalk). Indirect crosstalk: a common phenomenon in which two or more hormone pathways are integrated at the hormone response gene–network level rather than at the upstream level of signal transduction. One example is the JAinduced expression of NtPYL4, which affects the ability of ABA to regulate alkaloid production in tobacco. Specificity of recognition: the extent to which a plant can discriminate the presence of and/or attack by different herbivores. Specific recognition of arthropod herbivores can occur at different levels, ranging from phyla (i.e. distinct detection of arthropods compared with vertebrates) to species (i.e. distinct detection of two different herbivore species). Little is known about the molecular mechanisms underlying plant recognition of herbivores; however, generalized examples based on the PTI/ETI paradigm are informative. For example, a high degree of specificity in recognition could be achieved by R gene products that evolved to recognize effector molecules in adapted insects. Low-level specificity might involve the action of mechanosensors that detect insect movement on the leaf surface. Receptor-mediated recognition of HAMPs and/or DAMPs produced at the site of insect feeding is expected to provide an intermediate level of specificity because these signals are common in plant interactions with multiple insect species. Specificity of response: the extent to which plant physiological and/or metabolic changes elicited by the specific perception of a given herbivore are distinct from changes elicited by the perception of other attackers. Unlike the adaptive immune system in animals, which creates an immunological memory of a specific invading pathogen, the recognition of many insects (e.g. chewing herbivores) is channeled into a general defense response. Many measured differences in responses are not based on specific perception but are likely to be artifacts of secondary stress factors. These responses are referred to as ‘distinct’ or ‘different’ but not ‘specific’. Examples of specific responses are the different phenotypical changes triggered by different putative aphid receptors [122].

1360-1385/$ – see front matter ß 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.tplants.2012.01.003 Trends in Plant Science, May 2012, Vol. 17, No. 5

[(Figure_1)TD$IG]

Review


PAMPs

HAMPS

Eggs

DAMPs

5

6 DAMPS

1

PRRs

4 Wounding 2

3

R

WIR

8

ETI

HTI

R

ETI

PTI

Plant leaf

HTI

7

Effectors

Phytohormone signaling and induced resistance TRENDS in Plant Science

Figure 1. Molecular recognition of pathogens and herbivores by plants. 1. Microbe-, pathogen- and damage-associated molecular patterns (MAMPs, PAMPs and DAMPs) are recognized by pattern recognition receptors (PRRs) and lead to PAMP-triggered immunity (PTI). 2. Pathogen effectors suppress PTI. 3. Resistance gene products recognize effectors and lead to effector-triggered immunity (ETI). 4. Oviposition-associated compounds are recognized by unknown receptors and trigger defensive responses. 5. Putative herbivore-associated molecular patterns (HAMPs) are recognized by receptors and lead to herbivore-triggered immunity (HTI). 6. Wounding leads to the release of DAMPs and to wound-induced resistance (WIR). 7. Effector-like molecules from insects can suppress HTI and WIR. Uncharacterized elements are indicated by broken lines.

again contain disease resistance (R) proteins that specifically recognize pathogen effectors, resulting in effectortriggered immunity (ETI) [2]. Although the PTI/ETI model is sometimes regarded as an oversimplification [3], the molecular identification of the involved ligands and receptors has enabled conclusions to be drawn about the specificity of recognition in plant–pathogen interactions: in general, PTI is based on non-specific recognition of common microbial molecules, whereas ETI is triggered by highly pathogen-specific compounds [4]. To what extent can the PTI/ETI model inform research aimed at elucidating the specificity of recognition in plant– herbivore interactions? In comparison to pathogens, insects are highly complex multicellular organisms with various lifestyles and behavioral patterns. Cues emanating from these patterns may be used by the plant to recognize the threat of herbivory and to mount appropriate defensive responses [5] (Figure 1). The first contact with the herbivore often occurs when the tarsi of an arriving insect touch the leaf surface. Landing and walking on a plant will exert pressure, break trichomes and deposit chemicals from tarsal pads on the leaf [5]. Plants have evolved mechanisms to sense pressure. The Venus fly trap (Dionaea muscipula), for example, closes immediately when its sensory hairs are stimulated by insects [6]. Non-carnivorous plants are also highly sensitive to touch [7]. In at least some cases, mechanostimulation by repeated touching is sufficient to induce the accumulation of jasmonic acid (JA) [8], the precursor of the defense hormone jasmonoyl-L-isoleucine (JA-Ile). Breaking of tomato (Solanum lycopersicum) leaf trichomes by adult moths or caterpillars induces hydrogen peroxide (H2O2) formation and expression of defensive proteinase inhibitors [9]. To date, there is no indication that this type of ‘early warning’ response is specific for particular insect

species, and the observed effects may be mostly related to DAMP-like effects (see below). Oviposition represents another opportunity for plants to detect insect herbivores. The formation of necrotic zones following egg deposition has been observed in black mustard (Brassica nigra) and certain potato (Solanum spp.) clones [10,11]. In pea (Pisum sativa) plants, long-chain alpha,omega-diols (bruchins) deposited during oviposition by pea weevils (Bruchus pisorum) on pea pods trigger the formation of undifferentiated cells beneath the eggs, which increase plant resistance by hindering the larvae when they try to burrow into the pod [12]. Oviposition can be accompanied by wounding, and in the interaction between the elm leafbeetle (Xanthogaleruca luteola) and the field elm (Ulmus minor), for example, oviduct secretions induce defenses only when they are released into oviposition wound sites [13]. Overall, some oviposition-associated cues seem to act as MAMP-like molecular patterns that can be used by plants to recognize and predict herbivore attack. Consistent with the PTI/ETI framework, oviposition effectors may be produced by herbivores to suppress the plant immune response (Box 1). Taken together, these findings suggest that oviposition events trigger plant defense reactions in an insectand potentially even species-specific manner. Herbivory disrupts the integrity of plant tissue, and many plant defense responses can be triggered by mechanical wounding alone [14–16], leading to wound-induced resistance (WIR). Extensive studies of the wound response in model plants such as tomato and Arabidopsis (Arabidopsis thaliana) have identified plant-derived compounds that trigger anti-insect defense responses. Such compounds are potentially recognized by PRRs and, thus, can be defined conceptually as DAMPs [17]. Among the DAMPs shown to activate anti-insect defenses in tomato are cell wall-derived oligosaccharides and the peptide 251

Review Box 1. Herbivore effectors Just as plants recognize a variety of herbivore-derived cues, there is evidence that herbivores can use effector molecules to suppress plant defenses: Oviposition fluids trigger the SA pathway in Arabidopsis, which increases the growth of Egyptian cotton leafworm (Spodoptera littoralis) larvae [58,123]. Oviposition can also suppress herbivoreinduced plant volatiles in maize [124]. During feeding, insects secrete effector-like compounds to suppress plant immunity. The best-known example is glucose oxidase produced by the salivary glands of various lepidopteran insects [26]. Many aphids also produce effector-like compounds [125], and other examples of herbivore-mediated suppression of plant defenses via as yet unknown mechanisms have also been documented [36,72]. Herbivores can produce plant hormones or hormone mimics to manipulate the host defense responses [126]. Insect herbivores are hosts to microorganisms (e.g. endosymbionts) and surface-dwelling parasites that produce compounds that potentially interfere or otherwise affect plant immunity [36]. For example, a recent study suggests that Wolbachia endosymbionts suppress the induction of maize genes involved in defense against the Western corn rootworm (Diabrotica virgifera), which feeds on roots [127]. Bacterial symbionts are also involved in the production of cytokinins that are secreted by the larvae of a leafminer moth (Phyllonorycter blancardella) to inhibit leaf senescence to maintain a source of food for the larvae [128]. The PTI/ETI paradigm indicates that plants have evolved various ways to recognize and respond defensively to pathogen effectors. Have plants also acquired the capacity to recognize insect effectors? Although there is evidence to suggest that this is indeed the case for hymenopterans [25], the ligands and receptors that constitute this form of recognition have yet to be identified and characterized [125]. Future research focusing on the identification of insect effectors and their mechanism of action is likely to mark a new phase in plant– insect interaction research.

signal systemin [18]. These studies lend support to the idea that many plant defense responses against herbivores are mediated by recognition of the ‘damaged self’ of the plant [17,19]. Analogous to danger signal models in the vertebrate immune system, when plant tissue suffers mechanical damage, this is likely to disrupt intracellular compartmentalization in ways that lead to the production of molecules that trigger general plant immune responses. It is becoming clear that a second layer of perception in addition to WIR can enable plants to detect herbivores more specifically: plants seem to recognize compounds that are released by herbivores during feeding. Extensive genetic analysis of the Hessian fly (Mayetiola destructor)– wheat (Triticum spp.) interaction [20,21] and the cloning of receptor-like R genes have demonstrated that there is a high degree of specificity of perception in this case [22–24]. The recognition systems for hemipteran and dipteran parasites, together with the identification of possible effectors (Box 1), appear to conform to the general PTI/ETI theory [25]. Less is known about the mechanisms by which plants perceive chewing insects, such as beetles and caterpillars, which constitute the vast majority of insect herbivore species. Numerous studies have shown that insect oral secretions, when applied to artificial wound sites, amplify the wound response of the plant [26–29]. Identified elicitors include fatty acid-amino acid conjugates (FACs), sulfur-containing fatty acids (caeliferins), peptides from digested plant proteins, and lipases [30–34]. Based on their 252


eliciting activity, at least the insect-derived compounds can be conceptually classified as herbivore-associated molecular patterns (HAMPs), which presumably are recognized by PRRs at the cell surface [35,36] to trigger HAMPinduced immunity (HTI). Even though HAMP receptors have not been identified, several trends have emerged concerning the specificity of elicitor-mediated recognition of chewing herbivores by plants. First, herbivore-derived elicitors boost the amplitude of wound-induced defense responses [30,31,33,34,37]. Second, different herbivore species produce qualitatively and quantitatively different elicitor combinations [38,39]. Third, the activity of the different known elicitors varies between plant species [37]. Taken together, these observations suggest the potential for elicitor-mediated, species-specific recognition of chewing insects by plants. Plant hormone regulatory networks integrate different herbivore recognition cues Following the recognition of an attacker, plants use different signaling cascades to reprogram their phenotype. Extended PTI/ETI models in plant–pathogen interactions suggest that, although the recognition of pathogens can be very specific, plants have a ‘common downstream signaling machinery’ [40] that is activated upon recognition of many different attackers. To what extent is this paradigm valid for plant–insect interactions? The JA signaling cascade, including the wound hormone JA-Ile, is widely considered to be a master regulator of plant resistance to arthropod herbivores as well as various pathogens [15,17,41–45], and JAs may therefore represent the core signaling pathway for activating resistance to insects. Disruption of plant tissue integrity during insect feeding triggers the production of JA-Ile and the activation of a well-defined signal transduction chain, leading to transcriptional activation of defense responses. Thus, it is the defining feature of most, if not all herbivores, namely the need to obtain nutrition from plant tissues, which betrays the presence of the attacker to the host. A major unresolved question is the extent to which herbivore-induced production of JA-Ile is promoted by signals originating from the signals of the plant (i.e. DAMPs, ‘self’) versus those from the herbivores (i.e. HAMPs, ‘non-self’) [19]. Mechanical wounding is sufficient to trigger robust local and systemic increases in JA-Ile levels within minutes of leaf injury, which indicates that herbivore-associated cues are not strictly required to activate the response [17,46,47]. However, the severity of crushingtype wounds typically used in these studies may bypass a requirement for HAMPs in the elicitation of herbivoreinduced responses. Research aimed at identifying herbivore-derived elicitors has therefore relied on the application of insect oral secretions to wound sites created by mild wounding regimens that do not elicit a strong defense response in the absence of oral secretion [35, 36,48–50]. Importantly, defense responses are attenuated in leaf-feeding lepidopteran herbivores that lack known elicitors in their oral secretions, lending support to the concept that wound-induced responses at an ecologically relevant intensity are potentiated by recognition factors [51].

Review


Hormone accumulation

[(Figure_2)TD$IG]

(a)

(b)

(c)

(d)

Time


Figure 2. Perception triggers herbivore- and tissue-specific hormonal network responses. (a–d) Conceptual kinetics of three hypothetical phytohormones are shown by solid and broken, black and gray lines. (a–c) with a white background represent the same tissue, whereas (d) with a darker background represents a different tissue or tissue age. (a,b) Different herbivores can elicit different hormonal responses. (c,d) Hormonal responses to the same herbivore can show tissue- or age-specific differences. A hypothetical hormone-responsive transcriptional network is then triggered by the different hormones. This network is represented here by specific transcripts (black circles), which differ in their expression intensity (different sizes of the circles) and interact in space and time (unbroken black lines represent strong interactions and broken black lines represent weak interactions). Gray ellipses denote specific groups of transcripts that are functionally related. The integration of spatiotemporal changes of hormone signaling into the downstream transcriptional network can lead to herbivore-specific plant responses.

We argue here that the JA pathway represents a conserved core-signaling machinery that is activated by both non-specific and specific recognition patterns following herbivore attack. However, how do plants fine tune their defense machinery to mount appropriate herbivore-specific responses beyond JAs? We propose two potential answers to this question. First, plants may use JA-independent, parallel pathways to create distinct response patterns. Second, specificity may be mediated through the activation of spatio-temporal modulators of the JA response (Figure 2). Evidence for the first concept comes from studies on the recognition and response system of tomato to the potato aphid Macrosiphum euphorbiae. Mi-1, a putative receptor, triggers SA-mediated signaling [52] and resistance independently of the JA pathway [53]. Plant recognition of, and response to, many other hemipterans seems to follow a similar pattern [54], which suggests that plants use JAindependent hormone response pathways to achieve specific resistance against phloem feeders. However, most herbivores inflict much greater cell damage than do phloem feeders, and will activate JA signaling and resistance. In this case, specificity may be achieved via cross-talk with other hormones (Figure 3). Indeed, JAinduced changes in gene expression typically depend on the context in which the hormone is perceived [55,56]. The best-studied hormones that alter JA-mediated defense responses and herbivore resistance are SA and ET. In general, SA antagonizes JA-induced resistance, whereas ET can have both positive and negative effects. For ET, some of the transcriptional responses that are modulated

by crosstalk with JA were recently shown to be mediated by the ET-stabilized transcriptional regulator EIN3 [57]. Several studies have shown that SA and ET are specifically modulated by different herbivore elicitors [29,37] and may thereby provide a degree of hormone-mediated specificity. A striking example of how herbivore-induced SA signaling can modulate JA-dependent defenses comes from research on A. thaliana: Oviposition by the cabbage butterfly (Pieris brassicae) induces SA accumulation and reduces the induction of JA-responsive genes, leading to reduce plant resistance against S. littoralis [58]. SA–JA–ET crosstalk has been reviewed in detail elsewhere [59,60] (also see other reviews in this special issue). However, ABA, auxins, GB, CK and BR have received less attention as potential factors that modulate herbivore resistance. The following discussion highlights examples from the recent literature that indicate that these hormones also play an important role in mediating specificity in herbivory-induced defense responses. Abscisic acid ABA levels in maize (Zea mays) are increased during attack by the specialist root herbivore western corn rootworm (Diabrotica virgifera virgifera), but not by mechanical wounding alone [27,61], and in Arabidopsis after induction with wounding and the oral secretions of the desert locust (Schistocerca gregaria), a generalist herbivore [34]. ABA levels also increased in a goldenrod species (Solidago altissima) after induction by the tobacco budworm (Heliothis virescens) caterpillar, but not by the gall-inducing 253

Review

[(Figure_3)TD$IG]


Secondary stress effects

Receptor-mediated perception

HAMPs

Oxidative stress

Wounding

DAMPS

Perception

SA Hormonal regulation

ABA Auxin

Desiccation

pH

CKs JA-Ile

GAs

SA

JAZ

ABA

DELLAs

BRs

CKs

SCFCOI1

EIN3

ET

MYB MYC

Phenotypical changes

JA-dependent responses Specific responses

Non-specific responses TRENDS in Plant Science

Figure 3. The jasmonate (JA) core pathway and its modulating factors. A conceptual, non-exhaustive overview is presented. General and specific herbivore-associated patterns, including herbivore-associated molecular patterns (HAMPs), damage-associated molecular patterns (DAMPs) and wounding, activate the JA pathway (blue area). Increased accumulation of jasmonoyl-L-isoleucine (JA-Ile) promotes the interaction of JAZ proteins with the SCF ubiquitin ligase SCFCOI1. Ubiquitin-dependent degradation of JAZs by the 26S proteasome releases transcription factors from their JAZ-bound repressed state, thereby activating the expression of transcriptional regulons that promote defense and inhibit vegetative growth. JA-independent hormonal pathways are also induced (purple area), and several hormones, including salicylic acid (SA), ethylene (ET), auxin, gibberellins (GA), cytokinins (CK) and brassinosteroids (BR) modulate JA metabolism and signaling (light-blue area). Herbivory also leads to oxidative stress, changes in intracellular pH and desiccation, which modulate the JA pathway either directly or indirectly through other hormones. Together, this leads to complex phenotypic changes that comprise both specific and general responses, the majority of which can be linked back to the JA pathway.

caterpillar Gnorimoschema gallaesolidaginis [62]. ABA synthesis and signaling affect herbivore-induced transcript levels and JA biosynthesis in Arabidopsis [63,64], JA-inducible defense responses in maize [27] and resistance to herbivores in tomato [65]. ABA and JA synergistically induce MYC2-dependent gene expression during wound responses. MYC2 encodes a nuclear localized basic helix-loop-helix– type transcription factor that acts as both activator and repressor of JA-mediated gene expression and serves as an integration point between ABA and JA signaling [66–68]. In tobacco (Nicotiana tabacum), JA also regulates the expression of NtPYL4, a gene belonging to the PYR/PYL/RCAR family, which encodes an ABA receptor protein, thereby affecting ABA-induced levels of root alkaloids [69]. The same study demonstrated that AtPYL4 and AtPYL5 mutants in Arabidopsis are more sensitive to JA-induced growth inhibition and less sensitive to JA-induced anthocyanin accumulation. Although the molecular mechanisms behind ABA–JA crosstalk are still elusive, recent findings suggest that both pathways share similar regulatory proteins. The co-repressor TOPLESS (TPL) interacts with ethylene-responsive element binding factor-associated amphiphilic repression (EAR)-motif proteins to repress transcription of genes involved in several hormone pathways [70]. The EAR-motif protein Novel Interactor of JAZ (NINJA) 254

connects TPL to the JAZ complex, thereby mediating repression of genes demarcated by JAZ-bound transcription factors, such as MYC2. TPL also interacts with NINJArelated proteins that are part of a complex that mediates ABA-induced degradation of negative transcriptional regulators [71]. Taken together, these findings indicate that ABA and JA are tightly interconnected and that regulation of ABA levels in response to herbivory can modulate JAdriven defense responses (Figure 3). However, because ABA is also an important signal for responses to desiccation, which is an effect that accompanies herbivore attack in many cases, it remains to be determined to what extent this stress hormone is involved in recognition-mediated responses to insect feeding. The application of Egyptian cotton leafworm (Spodoptera littoralis) regurgitant to Arabidopsis can reduce wound-induced stomatal closure and water loss [72], whereas S. gregaria oral secretions induce ABA levels [34], which suggests specific elicitor-mediated regulation of this hormone. Auxin Levels of the auxin indole-3-acetic acid (IAA) are elevated in plants attacked by gall-feeding insects [62,73]. By contrast, IAA levels in the leaves of a species of wild tobacco (Nicotiana attenuata) are reduced within three days after

Review simulated herbivory [74]. It is known that plant resistance to pathogens can be modulated through changes in auxin sensitivity. For example, the perception of the bacterial elicitor flagellin decreases auxin sensitivity, thereby elevating resistance to Pseudomonas syringae [75]. Concomitantly, P. syringae suppresses host defense by promoting auxin production via delivery of effectors into the plant cell [76,77]. Treatments with synthetic auxin directly suppress SA-induced defense responses [78], which can be linked to SA-mediated resistance to phloem-feeding insects [52]. Whether insects that are negatively affected by SA-mediated defenses can alter auxin homeostasis or signaling to suppress host defense is not known. In addition to regulation of SA signaling, studies have suggested that there is an intimate molecular interplay between auxin and JA signaling: auxin formation in Arabidopsis roots is enhanced by JA-mediated induction of genes involved in auxin biosynthesis and transport [68,79]. In N. attenuata leaves, JA negatively regulates woundinduced decreases in auxin content [74], demonstrating that the effects of JA on auxin biosynthesis are tissue specific and possibly also species specific [55]. Importantly, these data suggest that herbivore-induced JA levels affect auxin homeostasis. Conversely, there is evidence to indicate that auxin enhances JA biosynthesis and signaling [80–82]. JAZ1 and MYC2 are coregulated by auxin and JA, demonstrating the potential for crosstalk between both hormones [81,82]. Analogous to ABA signaling, EAR-motif-containing AUX/IAA proteins, which are negative regulators of auxin-induced responses, also interact with TPL [71], suggesting that TPL acts as an integrator of multiple hormone pathways. Another protein that links auxin and JA responses is suppressor of G-two allele of SKP1 (SGT1), which connects chaperone-mediated protein assembly and ubiquitin-mediated protein degradation. SGT1 mutants of Arabidopsis are compromised in their sensitivity to both auxin and JA [83], and silencing SGT1 in N. attenuata attenuates JA levels, defense metabolite accumulation, and resistance to the tobacco hornworm (Manduca sexta) caterpillar [84]. Auxin can also regulate plant defense responses independently of SA and JA [85,86]. These findings demonstrate that auxin is a potent modifier of herbivore-relevant defense responses and indicate that plants may modulate auxin levels to mediate attacker specificity. Gibberellins Studies with plants altered in GB signaling have suggested a role for GB in herbivore-induced defense responses. DELLA proteins are negative transcriptional regulators of gibberellic acid (GA)-induced gene expression and are considered to play key roles in integrating plant responses to diverse developmental and environmental stimuli [87]. Remarkably, GAs affect JA signaling through competitive binding of DELLAs to JAZ proteins, thereby preventing JAZ–MYC2 interaction and promoting MYC2-induced transcriptional responses [88]. GA perception leads to degradation of DELLAs, which ultimately leads to inhibition of MYC2 and diminished JA responses. Accordingly, alteration of DELLA levels affects JA biosynthesis and signaling [89,90].


Cytokinins In N. attenuata, CK-related transcripts are among the genes that are most strongly regulated by FAC elicitors [91,92], suggesting that CK has a role in the hormonal regulatory network. In addition, gall-forming insects and possibly some leaf miners modulate plant CK levels, presumably to maintain the sink status of the infected tissues [62,73,93]. Isopentenyltransferases (IPT) represent the rate-limiting step in CK biosynthesis, and IPT overexpression increases resistance of common tobacco to M. sexta [94]. Several lines of evidence also support an important role for CK in the activation of JA biosynthesis. Transgenic tobacco (N. tabacum cv. Xanthi nc) plants that overexpress a small GTP-binding protein accumulate high levels of CK, resulting in increased rates of JA production after wounding, a response that can be mimicked by long-term CK treatments [95]. Furthermore, CK treatments of hybrid poplar (Populus sp.) leaves increase the wound-induced JA burst and the expression of genes involved in JA biosynthesis [96]. The same study also shows that wounding and CK treatments of sink but not source leaves impairs gypsy moth (Lymantria dispar) larval performance, suggesting that CK-mediated resistance to insects depends on leaf ontogeny. CK levels in leaves are thought to be regulated by ontogenic constraints because the hormone accumulates to high levels in younger leaves, whereas reduced CK levels promote leaf senescence [97,98]. Because CKs modulate herbivore-induced defenses, the CK status of a given tissue might determine the intensity of the defense response of that particular tissue after perception of herbivory and, thus, contribute to tissuespecific responses in herbivory-induced signaling [99] (Figure 2). Brassinosteroids Recent findings have also suggested important roles for BRs in herbivore resistance. BRs antagonize JA-mediated trichome density and defense metabolite accumulation in tomato [100]. BRs are also known to repress JA-governed inhibition of root growth [101]. BR are perceived by BR insensitive 1 (BRI1), a leucine-rich repeat receptor-like kinase [102,103]. BRI1-associated kinase 1 (BAK1) interacts with BRI1 and plays an essential role in BR signaling [104]. Apart from BR signaling, BAK1 also interacts with the flagellin receptor FLS2 and is required for multiple MAMP-elicited responses [105]. Silencing BAK1 in N. attenuata reduces wound- and herbivory-induced JA and JA-Ile levels and JA-induced trypsin proteinase inhibitor (TPI) activity [106]. Whether BAK regulates HAMP or DAMP perception to modulate JA levels, or whether the effects in BAK1-silenced plants are the result of changes in BR perception, requires further analysis [106]. Taken together, these examples illustrate the many possibilities that plants have to modulate the JA pathway to achieve specific responses. However, apart from JA/SA crosstalk (see Whitham and colleagues in this special issue), clear examples of specific induction of JA-modulators following herbivore recognition remain scarce. The identification of herbivore receptors followed by in-depth analysis of their downstream targets should help to fill 255

Review this knowledge gap. Another open question is whether the JA pathway and its modulating factors function in a similar manner in different plant species. Interspecific comparative approaches would be important to be construct generalized hormonal networks that mediate specificity. Do recognition-induced plant hormone networks trigger specific and appropriate defense responses? The recognition systems used by plants to perceive herbivore attack are integrated with hormone response pathways that reprogram the plant. However, what evidence is there that the resulting responses are specific and appropriate (see Glossary) for defense against the attacking herbivore? Again, for hemipterans, compelling examples of gene-for-gene resistance link specific recognition to both specific and appropriate responses [54]. In the case of chewing herbivores, many studies have also demonstrated the differential responses of plants to different insect species [107–110], but to date evidence that these responses provide specific resistance is rare. Given the complexity of host transcriptional responses to herbivory, some of these differences may be attributed to a variety of experimental factors that influence the response. Plant growth conditions, plant and insect developmental stage, herbivore density and treatment duration are among the parameters that are expected to have major effects on transcript profiles. Also, in some cases, different insects from the same feeding guild elicit similar or converging responses via the general JA-signaling cassette [111– 114]. The question thus arises whether, from an adaptive point of view, plants benefit from tailoring their response to different chewing herbivores, or whether a generalized response following recognition is the most pertinent strategy? There is evidence to indicate that, among the multitude of defenses induced by one herbivore species, some responses target specific types of herbivore, even within the same feeding guild. One example comes from work on the JA-regulated defensive enzyme threonine deaminase (TD2), which degrades the essential amino acid Thr in the lepidopteran gut [115]. Although TD2 expression in tomato leaves is induced in response to attack by both beet armyworm (Spodoptera exigua) and cabbage looper (Trichoplusia ni) caterpillars and by the Colorado potato beetle (Leptinotarsa decemlineata), the defensive activity of the enzyme is only activated in the gut of lepidopteran herbivores, not in the gut of the coleopteran herbivore [116]. Levels of the JA-regulated non-protein amino acid Nd-acetylornithine in Arabidopsis increase in response to feeding by larvae of the small white (Pieris rapae) and the diamondback moth (Plutella xylostella) as well as the green peach aphid (Myzus persicae), but this defense appears to target only the aphids [45]. Also, distinct patterns of volatile compounds have been shown to be induced different herbivores, leading to specific attraction of natural enemies [117,118] (see McCormick and colleagues in this special issue). However, also in this case, it remains open whether the differences in the reactions of the plants are based on specific recognition patterns. 256


Overall, generalist and specialist herbivores might be susceptible to different types of defense, and plants may benefit from detecting highly adapted herbivores and adjusting their regulation of quantitative and qualitative direct and indirect defenses. Specialist herbivores in particular may have found ways to suppress plant defenses (Box 1) or to circumvent them via behavioral adaptations, which, in analogy to the PTI/ETI model in plant–pathogen interactions, may have led to the counter-evolution of specifically adapted defense mechanisms in plants [119,120]. Until today, few mechanistic examples of plant-counter adaptations to specialists are known (but see other articles in this special issue), and further research is required to disentangle whether differential responses of plants to chewing herbivores are truly specific, and whether plants have evolved to tailor their response to different chewing attackers. Conclusions and future directions: piecing together the recognition–response puzzle A recurring theme in all spheres of plant–herbivore biology is the ability of each player to perceive and respond to cues generated by the other; this exchange of information provides an excellent focal point for elucidating the basic chemical and molecular principles of plant–herbivore interactions. Understanding the mechanisms behind perception and response may also inform studies about their evolutionary history. In contrast to well-established models describing the evolution of plant–pathogen interactions [121], our understanding of how molecular recognition and response systems shape plant–herbivore relationships is still in its infancy, and many important questions remain unanswered (Box 2). Nevertheless, the literature supports several general conclusions about specificity in plant–herbivore interactions. First, plants perceive different arthropods by integrating various environmental cues, ranging from mechanostimulation by insects walking on plant surfaces to contact with salivary components during feeding. Second, the perception of herbivores triggers regulatory responses that include different phytohormones, with the JA pathway playing a dominant role in host resistance. Third, although JA signaling is highly conserved, it is becoming increasingly clear that multiple hormone response pathways interact to translate initial perception events into appropriate responses that increase plant fitness in the presence of hostile aggressors.

Box 2. Outstanding questions Which receptors are involved in plant perception of herbivores? How prevalent are herbivore effectors and how do they act to suppress the plant immune system? Are herbivore effectors recognized by plant R genes in accordance with the ETI/PTI model in plant–pathogen interactions? How do plants integrate information derived from multiple herbivore- and plant-derived cues? Which JA-independent processes mediate specific plant responses to herbivore attack? What is the precise role of growth hormones (GBs, CKs, auxin and BRs) in modulating plant immunity to herbivores? Is the recognition of a specific chewing herbivore translated into a distinct defense response?

Review Acknowledgments We thank Martin Heil and Anurag Agrawal for the invitation to contribute to this special issue. Georg Jander and Ian Baldwin provided helpful comments on an earlier version of this manuscript. This work is supported by a Swiss National Science Foundation Fellowship to ME (PBNEP3-134930). Plant–insect interaction research in the GAH laboratory is supported by grants from the National Institutes of Health (R01GM57795), the Chemical Sciences, Geosciences and Biosciences Division, Office of Basic Energy Sciences, Office of Science, US Department of Energy (grant DE-FG02-91ER20021) and the US Department of Agriculture (2007-35604-17791).

References 1 Boller, T. and Felix, G. (2009) A renaissance of elicitors: perception of microbe-associated molecular patterns and danger signals by patternrecognition receptors. Annu. Rev. Plant Biol. 60, 379–406 2 Jones, J.D.G. and Dangl, J.L. (2006) The plant immune system. Nature 444, 323–329 3 Thomma, B.P.H.J. et al. (2011) Of PAMPs and effectors: the blurred PTI-ETI dichotomy. Plant Cell Online 23, 4–15 4 Dodds, P.N. et al. (2006) Direct protein interaction underlies gene-forgene specificity and coevolution of the flax resistance genes and flax rust avirulence genes. Proc. Natl. Acad. Sci. U.S.A. 103, 8888–8893 5 Hilker, M. and Meiners, T. (2010) How do plants ‘notice’ attack by herbivorous arthropods? Biol. Rev. 85, 267–280 6 Forterre, Y. et al. (2005) How the Venus flytrap snaps. Nature 433, 421–425 7 Braam, J. (2005) In touch: plant responses to mechanical stimuli. New Phytol. 165, 373–389 8 Tretner, C. et al. (2008) Mechanostimulation of Medicago truncatula leads to enhanced levels of jasmonic acid. J. Exp. Bot. 59, 2847–2856 9 Peiffer, M. et al. (2009) Plants on early alert: glandular trichomes as sensors for insect herbivores. New Phytol. 184, 644–656 10 Shapiro, A.M. and DeVay, J.E. (1987) Hypersensitivity reaction of Brassica nigra L. (Cruciferae) kills eggs of Pieris butterflies (Lepidoptera: Pieridae). Oecologia 71, 631–632 11 Balbyshev, N.F. and Lorenzen, J.H. (1997) Hypersensitivity and egg drop: a novel mechanism of host plant resistance to Colorado potato beetle (Coleoptera: Chrysomelidae). J. Econ. Entomol. 90, 652–657 12 Doss, R.P. et al. (2000) Bruchins: insect-derived plant regulators that stimulate neoplasm formation. Proc. Natl. Acad. Sci. U.S.A. 97, 6218– 6223 13 Meiners, T. and Hilker, M. (2000) Induction of plant synomones by oviposition of a phytophagous insect. J. Chem. Ecol. 26, 221–232 14 Mitho¨fer, A. et al. (2005) Effects of feeding Spodoptera littoralis on lima bean leaves. II. Continuous mechanical wounding resembling insect feeding is sufficient to elicit herbivory-related volatile emission. Plant Physiol. 137, 1160–1168 15 Howe, G. and Jander, G. (2008) Plant immunity to insect herbivores. Annu. Rev. Plant Biol. 59, 41–66 16 Bricchi, I. et al. (2010) Robotic mechanical wounding (MecWorm) versus herbivore-induced responses: early signaling and volatile emission in Lima bean (Phaseolus lunatus L.). Planta 232, 719–729 17 Koo, A.J.K. and Howe, G.A. (2009) The wound hormone jasmonate. Phytochemistry 70, 1571–1580 18 Ryan, C.A. (2000) The systemin signaling pathway: differential activation of plant defensive genes. Biochim. Biophys. Acta 1477, 112–121 19 Heil, M. (2009) Damaged-self recognition in plant herbivore defence. Trends Plant Sci. 14, 356–363 20 Sardesai, N. et al. (2005) Identification and mapping of H32, a new wheat gene conferring resistance to Hessian fly. Theor. Appl. Genet. 111, 1167–1173 21 Harris, M.O et al. (2003) Grasses and gall midges: plant defense and insect adaptation. Annu. Rev. Entomol. 48, 549–577 22 Rossi, M. et al. (1998) The nematode resistance gene Mi of tomato confers resistance against the potato aphid. Proc. Natl. Acad. Sci. U.S.A. 95, 9750–9754 23 Boissot, N. et al. (2010) Mapping and validation of QTLs for resistance to aphids and whiteflies in melon. Theor. Appl. Genet. 121, 9–20 24 De Vos, M. and Jander, G. (2009) Myzus persicae (green peach aphid) salivary components induce defence responses in Arabidopsis thaliana. Plant Cell Environ. 32, 1548–1560


25 Hogenhout, S.A. and Bos, J.I.B. (2011) Effector proteins that modulate plant–insect interactions. Curr. Opin. Plant Biol. 14, 422–428 26 Musser, R.O. et al. (2002) Herbivory: caterpillar saliva beats plant defences – a new weapon emerges in the evolutionary arms race between plants and herbivores. Nature 416, 599–600 27 Erb, M. et al. (2009) Signal signature of aboveground-induced resistance upon belowground herbivory in maize. Plant J. 59, 292–302 28 Turlings, T.C.J. et al. (1993) An elicitor in caterpillar oral secretions that induces corn seedlings to emit chemical signals attractive to parasitic wasps. J. Chem. Ecol. 19, 411–425 29 Diezel, C. et al. (2009) Different lepidopteran elicitors account for cross-talk in herbivory-induced phytohormone signaling. Plant Physiol. 150, 1576–1586 30 Alborn, H.T. et al. (2007) Disulfooxy fatty acids from the American bird grasshopper Schistocerca americana, elicitors of plant volatiles. Proc. Natl. Acad. Sci. U.S.A. 104, 12976–12981 31 Alborn, H.T. et al. (1997) An elicitor of plant volatiles from beet armyworm oral secretion. Science 276, 945–949 32 Schmelz, E.A. et al. (2006) Fragments of ATP synthase mediate plant perception of insect attack. Proc. Natl. Acad. Sci. U.S.A. 103, 8894– 8899 33 Halitschke, R. et al. (2001) Molecular interactions between the specialist herbivore Manduca sexta (Lepidoptera, Sphingidae) and its natural host Nicotiana attenuata. III. Fatty acid-amino acid conjugates in herbivore oral secretions are necessary and sufficient for herbivore-specific plant responses. Plant Physiol. 125, 711–717 34 Scha¨fer, M. et al. (2011) Lipase activity in insect oral secretions mediates defense responses in Arabidopsis. Plant Physiol. 156, 1520–1534 35 Mitho¨fer, A. and Boland, W. (2008) Recognition of herbivoryassociated molecular patterns. Plant Physiol. 146, 825–831 36 Felton, G.W. and Tumlinson, J.H. (2008) Plant–insect dialogs: complex interactions at the plant–insect interface. Curr. Opin. Plant Biol. 11, 457–463 37 Schmelz, E.A. et al. (2009) Phytohormone-based activity mapping of insect herbivore-produced elicitors. Proc. Natl. Acad. Sci. U.S.A. 106, 653–657 38 Yoshinaga, N. et al. (2010) Fatty acid-amino acid conjugates diversification in lepidopteran caterpillars. J. Chem. Ecol. 36, 319– 325 39 Mori, N. et al. (2001) Enzymatic decomposition of elicitors of plant volatiles in Heliothis virescens and Helicoverpa zea. J. Insect Physiol. 47, 749–757 40 Katagiri, F. and Tsuda, K. (2010) Understanding the plant immune system. Mol. Plant Microbe Interact. 23, 1531–1536 41 Stout, M.J. et al. (2005) Plant-mediated interactions between pathogenic microorganisms and herbivorous arthropods. Annu. Rev. Entomol. 51, 663–689 42 Bostock, R.M. (2005) Signal crosstalk and induced resistance: straddling the line between cost and benefit. Annu. Rev. Phytopathol. 43, 545–580 43 Glazebrook, J. (2005) Contrasting mechanisms of defense against biotrophic and necrotrophic pathogens. Annu. Rev. Phytopathol. 43, 205–227 44 Farmer, E.E. and Dubugnon, L. (2009) Detritivorous crustaceans become herbivores on jasmonate-deficient plants. Proc. Natl. Acad. Sci. U.S.A. 106, 935–940 45 Adio, A. et al. (2011) Biosynthesis and defensive function of Nacetylornithine, a jasmonate-induced Arabidopsis metabolite. Plant Cell 23, 3303–3318 46 Chung, H.S. et al. (2008) Regulation and function of Arabidopsis JASMONATE ZIM-domain genes in response to wounding and herbivory. Plant Physiol. 146, 952–964 47 Glauser, G. et al. (2008) Spatial and temporal dynamics of jasmonate synthesis and accumulation in Arabidopsis in response to wounding. J. Biol. Chem. 283, 16400–16407 48 Walling, L.L. (2009) Adaptive defense responses to pathogens and insects. In Plant Innate Immunity (Van Loon, L.C., ed.), pp. 551–612, Elsevier 49 Kessler, A. and Baldwin, I.T. (2002) Plant responses to insect herbivory: the emerging molecular analysis. Annu. Rev. Plant Biol. 53, 299–328 257

Review 50 Bonaventure, G. et al. (2011) Herbivore-associated elicitors: FAC signaling and metabolism. Trends Plant Sci. 16, 294–299 51 De Moraes, C.M. and Mescher, M.C. (2004) Biochemical crypsis in the avoidance of natural enemies by an insect herbivore. Proc. Natl. Acad. Sci. U.S.A. 101, 8993–8997 52 Li, Q. et al. (2006) Mi-1-mediated aphid resistance involves salicylic acid and mitogen-activated protein kinase signaling cascades. Mol. Plant Microbe Interact. 19, 655–664 53 Bhattarai, K.K. et al. (2007) Coi1-dependent signaling pathway is not required for Mi-1-mediated potato aphid resistance. Mol. Plant Microbe Interact. 20, 276–282 54 Dogimont, C. et al. (2010) Host plant resistance to aphids in cultivated crops: genetic and molecular bases, and interactions with aphid populations. C. R. Biol. 333, 566–573 55 Pauwels, L. et al. (2009) Jasmonate-inducible gene: what does it mean? Trends Plant Sci. 14, 87–91 56 Erb, M. and Glauser, G. (2010) Family business: multiple members of major phytohormone classes orchestrate plant stress responses. Chemistry 16, 10280–10289 57 Zhu, Z.Q. et al. (2011) Derepression of ethylene-stabilized transcription factors (EIN3/EIL1) mediates jasmonate and ethylene signaling synergy in Arabidopsis. Proc. Natl. Acad. Sci. U.S.A. 108, 12539–12544 58 Bruessow, F. et al. (2010) Insect eggs suppress plant defence against chewing herbivores. Plant J. 62, 876–885 59 Robert-Seilaniantz, A. et al. (2011) Hormone crosstalk in plant disease and defense: more than just jasmonate-salicylate antagonism. Annu. Rev. Phytopathol. 49, 317–343 60 Pieterse, C.M.J. et al. (2009) Networking by small-molecule hormones in plant immunity. Nat. Chem. Biol. 5, 308–316 61 Erb, M. et al. (2011) The role of abscisic acid and water stress in root herbivore-induced leaf resistance. New Phytol. 189, 308–320 62 Tooker, J.F. and De Moraes, C.M. (2011) Feeding by a gall-inducing caterpillar species alters levels of indole-3-acetic and abscisic acid in Solidago altissima (Asteraceae) stems. Arthropod Plant Interact. 5, 115–124 63 Adie, B. et al. (2007) Modulation of plant defenses by ethylene. J. Plant Growth Regul. 26, 160–177 64 Bodenhausen, N. and Reymond, P. (2007) Signaling pathways controlling induced resistance to insect herbivores in Arabidopsis. Mol. Plant Microbe Interact. 20, 1406–1420 65 Thaler, J.S. and Bostock, R.M. (2004) Interactions between abscisicacid-mediated responses and plant resistance to pathogens and insects. Ecology 85, 48–58 66 Lorenzo, O. et al. (2004) Jasmonate-insensitive1 encodes a MYC transcription factor essential to discriminate between different jasmonate-regulated defense responses in Arabidopsis. Plant Cell 16, 1938–1950 67 Anderson, J.P. et al. (2004) Antagonistic interaction between abscisic acid and jasmonate-ethylene signaling pathways modulates defense gene expression and disease resistance in Arabidopsis. Plant Cell 16, 3460–3479 68 Dombrecht, B. et al. (2007) MYC2 differentially modulates diverse jasmonate-dependent functions in Arabidopsis. Plant Cell 19, 2225– 2245 69 Lackman, P. et al. (2011) Jasmonate signaling involves the abscisic acid receptor PYL4 to regulate metabolic reprogramming in Arabidopsis and tobacco. Proc. Natl. Acad. Sci. U.S.A. 108, 5891–5896 70 Szemenyei, H. et al. (2008) TOPLESS mediates auxin-dependent transcriptional repression during Arabidopsis embryogenesis. Science 319, 1384–1386 71 Pauwels, L. et al. (2010) NINJA connects the co-repressor TOPLESS to jasmonate signalling. Nature 464, 788–791 72 Consales, F. et al. (2011) Insect oral secretions suppress woundinduced responses in Arabidopsis. J. Exp. Bot. 63, 727–737 73 Mapes, C.C. and Davies, P.J. (2001) Indole-3-acetic acid and ball gall development on Solidago altissima. New Phytol. 151, 195–202 74 Onkokesung, N. et al. (2010) Jasmonic acid and ethylene modulate local responses to wounding and simulated herbivory in Nicotiana attenuata leaves. Plant Physiol. 153, 785–798 75 Navarro, L. et al. (2006) A plant miRNA contributes to antibacterial resistance by repressing auxin signaling. Science 312, 436–439

258


76 Chen, I.C. et al. (2007) Glutathione S-transferase interacting with farred insensitive 219 is involved in phytochrome A-mediated signaling in Arabidopsis. Plant Physiol. 143, 1189–1202 77 Uppalapati, S.R. et al. (2005) The phytotoxin coronatine and methyl jasmonate impact multiple phytohormone pathways in tomato. Plant J. 42, 201–217 78 Wang, L. et al. (2007) Independently silencing two JAR family members impairs levels of trypsin proteinase inhibitors but not nicotine. Planta 226, 159–167 79 Sun, J.Q. et al. (2009) Arabidopsis ASA1 is important for jasmonatemediated regulation of auxin biosynthesis and transport during lateral root formation. Plant Cell 21, 1495–1511 80 Nagpal, P. et al. (2005) Auxin response factors ARF6 and ARF8 promote jasmonic acid production and flower maturation. Development 132, 4107–4118 81 Grunewald, W. et al. (2009) Expression of the Arabidopsis jasmonate signalling repressor JAZ1/TIFY10A is stimulated by auxin. EMBO Rep. 10, 923–928 82 Tiryaki, I. and Staswick, P.E. (2002) An Arabidopsis mutant defective in jasmonate response is allelic to the auxin-signaling mutant axr1. Plant Physiol. 130, 887–894 83 Gray, W.M. et al. (2003) Arabidopsis SGT1b is required for SCF(TIR1)-mediated auxin response. Plant Cell 15, 1310–1319 84 Meldau, S. et al. (2011) SGT1 regulates wounding- and herbivoryinduced jasmonic acid accumulation and Nicotiana attenuata’s resistance to the specialist lepidopteran herbivore Manduca sexta. New Phytol. 189, 1143–1156 85 Ding, X.H. et al. (2008) Activation of the indole-3-acetic acid-amido synthetase GH3-8 suppresses expansin expression and promotes salicylate- and jasmonate-independent basal immunity in rice. Plant Cell 20, 228–240 86 Llorente, F. et al. (2008) Repression of the auxin response pathway increases Arabidopsis susceptibility to necrotrophic fungi. Mol. Plant 1, 496–509 87 Grant, M.R. and Jones, J.D.G. (2009) Hormone (dis)harmony moulds plant health and disease. Science 324, 750–752 88 Hou, X. et al. (2010) DELLAs modulate jasmonate signaling via competitive binding to JAZs. Dev. Cell 19, 884–894 89 Cheng, H. et al. (2009) Gibberellin acts through jasmonate to control the expression of MYB21, MYB24, and MYB57 to promote stamen filament growth in Arabidopsis. PLoS Genet. 5, e1000440 90 Navarro, L. et al. (2008) DELLAs control plant immune responses by modulating the balance of jasmonic acid and salicylic acid signaling. Curr. Biol. 18, 650–655 91 Hui, D.Q. et al. (2003) Molecular interactions between the specialist herbivore Manduca sexta (Lepidoptera, Sphingidae) and its natural host Nicotiana attenuata: V. Microarray analysis and further characterization of large-scale changes in herbivore-induced mRNAs. Plant Physiol. 131, 1877–1893 92 Gilardoni, P. et al. (2010) SuperSAGE analysis of the Nicotiana attenuata transcriptome after fatty acid-amino acid elicitation (FAC): identification of early mediators of insect responses. BMC Plant Biol. 10, 66 93 Giron, D. et al. (2007) Cytokinin-mediated leaf manipulation by a leafminer caterpillar. Biol. Lett. 3, 340–343 94 Smigocki, A. et al. (1993) Cytokinin-mediated insect resistance in Nicotiana plants transformed with the IPT gene. Plant Mol. Biol. 23, 325–335 95 Sano, H. et al. (1996) Regulation by cytokinins of endogenous levels of jasmonic and salicylic acids in mechanically wounded tobacco plants. Plant Cell Physiol. 37, 762–769 96 Dervinis, C. et al. (2010) Cytokinin primes plant responses to wounding and reduces insect performance. J. Plant Growth Regul. 29, 289–296 97 Boonman, A. et al. (2007) Cytokinin import rate as a signal for photosynthetic acclimation to canopy light gradients. Plant Physiol. 143, 1841–1852 98 Ori, N. et al. (1999) Leaf senescence is delayed in tobacco plants expressing the maize homeobox gene knotted1 under the control of a senescence-activated promoter. Plant Cell 11, 1073–1080 99 Ballare, C.L. (2011) Jasmonate-induced defenses: a tale of intelligence, collaborators and rascals. Trends Plant Sci. 16, 249–257

Review 100 Campos, M.L. et al. (2009) Brassinosteroids interact negatively with jasmonates in the formation of anti-herbivory traits in tomato. J. Exp. Bot 60, 4346–4360 101 Ren, C. et al. (2009) A leaky mutation in DWARF4 reveals an antagonistic role of brassinosteroid in the inhibition of root growth by jasmonate in Arabidopsis. Plant Physiol. 151, 1412–1420 102 Li, J.M. and Chory, J. (1997) A putative leucine-rich repeat receptor kinase involved in brassinosteroid signal transduction. Cell 90, 929– 938 103 Kinoshita, T. et al. (2005) Binding of brassinosteroids to the extracellular domain of plant receptor kinase BRI1. Nature 433, 167–171 104 Vert, G. (2008) Plant signaling: brassinosteroids, immunity and effectors Are BAK! Curr. Biol. 18, R963–R965 105 Heese, A. et al. (2007) The receptor-like kinase SERK3/BAK1 is a central regulator of innate immunity in plants. Proc. Natl. Acad. Sci. U.S.A. 104, 12217–12222 106 Yang, D.H. et al. (2011) BAK1 regulates the accumulation of jasmonic acid and the levels of trypsin proteinase inhibitors in Nicotiana attenuata’s responses to herbivory. J. Exp. Bot. 62, 641–652 107 Reymond, P. et al. (2000) Differential gene expression in response to mechanical wounding and insect feeding in Arabidopsis. Plant Cell 12, 707–719 108 Walling, L.L. (2000) The myriad plant responses to herbivores. J. Plant Growth Regul. 19, 195–216 109 Thompson, G. and Goggin, F. (2006) Transcriptomics and functional genomics of plant defence induction by phloem-feeding insects. J. Exp. Bot. 57, 755–766 110 Ralph, S.G. et al. (2006) Conifer defence against insects: microarray gene expression profiling of Sitka spruce (Picea sitchensis) induced by mechanical wounding or feeding by spruce budworms (Choristoneura occidentalis) or white pine weevils (Pissodes strobi) reveals large-scale changes of the host transcriptome. Plant Cell Environ. 29, 1545–1570 111 Reymond, P. et al. (2004) A conserved transcript pattern in response to a specialist and a generalist herbivore. Plant Cell 16, 3132–3147 112 Voelckel, C. and Baldwin, I.T. (2004) Herbivore-induced plant vaccination. Part II. Array-studies reveal the transience of herbivore-specific transcriptional imprints and a distinct imprint from stress combinations. Plant J. 38, 650–663 113 Bidart-Bouzat, M.G. and Kliebenstein, D.J. (2011) An ecological genomic approach challenging the paradigm of differential plant responses to specialist versus generalist insect herbivores. Oecologia 167, 677–689


114 Vogel, H. et al. (2007) Different transcript patterns in response to specialist and generalist herbivores in the wild Arabidopsis relative Boechera divaricarpa. PLoS ONE 2, e1081 115 Chen, H. et al. (2007) Stability of plant defense proteins in the gut of insect herbivores. Plant Physiol. 143, 1954–1967 116 Gonzales-Vigil, E. et al. (2011) Adaptive evolution of threonine deaminase in plant defense against insect herbivores. Proc. Natl. Acad. Sci. U.S.A. 108, 5897–5902 117 Shiojiri, K. et al. (2010) Herbivore-specific, density-dependent induction of plant volatiles: honest or ‘cry wolf’ signals? PLoS ONE 5, e12161 118 Erb, M. et al. (2010) A tritrophic signal that attracts parasitoids to host-damaged plants withstands disruption by non–host herbivores. BMC Plant Biol. 10, 247 119 Becerra, J.X. et al. (2009) Macroevolutionary chemical escalation in an ancient plant–herbivore arms race. Proc. Natl. Acad. Sci. U.S.A. 106, 18062–18066 120 Agrawal, A.A. et al. (2008) Evolution of latex and its constituent defensive chemistry in milkweeds (Asclepias): a phylogenetic test of plant defense escalation. Entomologia Experimentalis et Applicata 128, 126–138 121 Jones, J.D. and Dangl, J.L. (2006) The plant immune system. Nature 444, 323–329 122 Guo, S. et al. (2009) Two independent resistance genes in the Medicago truncatula cultivar Jester confer resistance to two different aphid species of the genus Acyrthosiphon. Plant Signal. Behav. 4, 328–331 123 Little, D. et al. (2007) Oviposition by pierid butterflies triggers defense responses in Arabidopsis. Plant Physiol. 143, 784–800 124 Penãflor, M.F.G.V. et al. (2011) Oviposition by a moth suppresses constitutive and herbivore-induced plant volatiles in maize. Planta 234, 207–215 125 Bos, J.I.B. et al. (2010) A functional genomics approach identifies candidate effectors from the aphid species Myzus persicae (green peach aphid). PLoS Genet. 6, e1001216 126 Schultz, J.C. (2002) Shared signals and the potential for phylogenetic espionage between plants and animals. Integr. Comp. Biol. 42, 454– 462 127 Barr, K.L et al. (2010) Microbial symbionts in insects influence downregulation of defense genes in maize. PLoS ONE 5, e11339 128 Kaiser, W. et al. (2010) Plant green-island phenotype induced by leafminers is mediated by bacterial symbionts. Proc. R. Soc. B: Biol. Sci. 277, 2311–2319

259


Review


Evolution of jasmonate and salicylate signal crosstalk Jennifer S. Thaler1, Parris T. Humphrey2 and Noah K. Whiteman2 1 2

Department of Entomology and Department of Ecology and Evolutionary Biology, Cornell University, Ithaca, NY, USA Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, AZ, USA

The evolution of land plants approximately 470 million years ago created a new adaptive zone for natural enemies (attackers) of plants. In response to attack, plants evolved highly effective, inducible defense systems. Two plant hormones modulating inducible defenses are salicylic acid (SA) and jasmonic acid (JA). Current thinking is that SA induces resistance against biotrophic pathogens and some phloem feeding insects and JA induces resistance against necrotrophic pathogens, some phloem feeding insects and chewing herbivores. Signaling crosstalk between SA and JA commonly manifests as a reciprocal antagonism and may be adaptive, but this remains speculative. We examine evidence for and against adaptive explanations for antagonistic crosstalk, trace its phylogenetic origins and provide a hypothesistesting framework for future research on the adaptive significance of SA–JA crosstalk. Attack, hormonal signaling and plant defense Sessile organisms, such as terrestrial green plants, are subject to pervasive attack by diverse attackers. These attackers include microbial pathogens (e.g. viruses, bacteria and fungi), macroscopic herbivores and parasites (e.g. parasitic plants and arthropods) and browsing herbivores (e.g. ungulates). The vast majority of attackers are relatively specialized in terms of the number of host species that they utilize (specialists), and a minority are less restricted in host range (generalists) [1,2]. Over the past 470 million years [3], plants have evolved effective inducible defense systems [4] to cope with attack by these diverse and abundant enemies. Yet, the specific match between particular attackers and plant defense traits, and whether attackers have the upper hand in these interactions, is poorly understood [5]. The specificity of plant–attacker interactions, from both sides of the equation, has important implications for understanding the evolution of resistance in plants and the evolution of virulence in the enemies [6]. Plants have to balance the costs and potential benefits of investing in defense in an environment where enemy attack is variable. On the one hand, defenses are costly to produce; in the absence of enemies, deploying defenses reduces plant fitness [7]. Because they are costly to produce, natural selection is presumed to favor the evolution of inducibility, meaning that these defenses are only produced in the Corresponding author: Whiteman, N.K. ([email protected]).

presence of attack. On the other hand, having an immediate impact on an attacker could be paramount to deterring further attacks. Plants generally strike a balance and maintain constitutive and inducible defenses. However, individual plants are likely to be attacked by more than one organism. Microbial pathogens, which are typically endophagous and single-celled, require vastly different defenses than macroscopic herbivores, which may even move among plant individuals while feeding. Among herbivores, different defenses are required for different guilds. For example, defense traits that are effective against aphids, which feed on plant phloem, are distinct from those that are effective against caterpillars, which typically defoliate plants [8]. Characterization of the specificity of the plant response is a focus of intense research among ecologists and plant scientists [5,9,10]. Of particular interest in this review is whether adaptive tailoring of the response occurs, or if tailoring is a byproduct of manipulation by enemies. Despite the caveats discussed above, the inducible plant defense system can be generally divided into two branches – one effective primarily against biotrophic (feeding on living tissue) pathogens and one against herbivores and necrotrophic (feeding on dead tissue) pathogens [11]. Inducible defenses are incredibly diverse and include morphological structures such as trichomes, fast-killing toxins such as alkaloids, digestibility reducers such as proteinase inhibitors and indirect defenses such as extrafloral nectaries and plant volatiles that can recruit other insects that deter herbivores [1,12–14]. Several plant hormones regulate the production of downstream resistance molecules in each branch. The SA pathway is primarily induced by and effective in mediating resistance against biotrophic pathogens and the JA pathway is primarily induced by and effective in mediating resistance against herbivores and necrotrophic pathogens [9]. This is an overly simplistic view of the complex repertoire of plant hormones that probably play a role in mediating inducible defenses, including abscisic acid (ABA), auxin, brassinosteroids, cytokinins, ethylene (ET) and gibberellic acid [15]. Interestingly, evidence from several distantly related plant species suggests that there can be evolutionarily conserved SA- and JA-signaling crosstalk resulting in reciprocal antagonism between the SA and JA signaling pathways [9]. The adaptive significance of this crosstalk, if any, is the focus of this review. The dynamics and genetic bases of SA–JA crosstalk, including the reciprocal antagonism often observed as a

1360-1385/$ – see front matter ß 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.tplants.2012.02.010 Trends in Plant Science xx (2012) 1–11

1


Review result, has mainly been dissected in the model plant Arabidopsis thaliana (Arabidopsis) [16–18]. The genetic basis of the reciprocal antagonism is extremely complex and an overview is presented below, in the context of the evolution of each of the major genetic players. Here we focus on SA and JA; however, ET is a critical third player from the perspective of understanding how plants prioritize and tailor their responses to diverse attackers and a brief focus on its role in mediating crosstalk is warranted. SA is typically prioritized over JA in Arabidopsis [19]. However, plants use ET to fine tune defenses by prioritizing JA induction over SA in response to multiple attackers [20]. ET also modifies the effect of a key protein (NPR1; NONEXPRESSOR OF PATHOGENESIS-RELATED GENES 1) involved in SA suppression of JA. In Arabidopsis, NPR1 is necessary for expression of SA-responsive genes and for repression of JA by SA. However, when ET is present, NPR1 function is no longer required for SA suppression of JA [20,21], suggesting that ET signaling acts to suppress JA in the presence of SA by bypassing NPR1. Many other plant hormones are also important in mediating the crosstalk, but the genetic bases of this crosstalk are less well studied. Recent approaches that examine genetic interaction networks in Arabidopsis have been fruitful for identifying candidate loci to be studied in detail for their potential role in defense signaling crosstalk [22]. The SA–JA crosstalk that often results in reciprocal antagonism between these two pathways has been interpreted as being an adaptive plant strategy, representing a cost-saving measure given that phenotypically different enemies are susceptible to distinct defense strategies. However, specific defenses that induce resistance to one attacker may render the plant more susceptible to another if alternative defenses are repressed by crosstalk [23]. We first focus on the phylogenetic distribution of crosstalk, candidate loci underlying crosstalk and the nature of the evidence used to assay for crosstalk. We then evaluate adaptive and nonadaptive evidence for the SA–JA reciprocal antagonism and illuminate a research path that integrates phylogenetic, genetic and ecological approaches towards testing explicit hypotheses on the origins and adaptive value of signal crosstalk. We end with a discussion of SA–JA signal interactions as a mechanism that generates specificity in plant–attacker interactions. Distribution of SA–JA reciprocal antagonism Although the SA–JA antagonism is clearly present in many plant species, an open question is whether there is a common genetic basis to this crosstalk and if so, whether the trait is conserved across all plants. Similarly, although it can be a reciprocal antagonism, the strength of the downregulation from each side of the SA and JA equation is not identical and may not be antagonistic across plants. We searched for all studies that tested for antagonisms in SA–JA signaling (Table 1). A paper was included as presenting evidence for SA–JA antagonism if there was a genetic or biochemical measure widely believed to be regulated by the jasmonate and salicylate pathways, or if one pathway was genetically manipulated and a response was measured in the other. Our survey included papers that 2


measured JA, SA or their derivatives, gene expression or chemical end-products known to be regulated by one of the pathways. In some studies, one pathway was elicited and had direct effects on the other pathway. In other studies, SA–JA antagonism was seen when induction of one pathway reduced the response to elicitation of the other pathway. We did not include studies that only found antagonisms in resistance to bioassay organisms if there was not evidence that the antagonism was due to SA–JA crosstalk. From the well-studied systems including Arabidopsis, tomato (Solanum lycopersicum) and tobacco (Nicotiana spp.), a subset of studies is included to highlight the ecological conditions under which antagonism can occur. There are systems that show conditionality in the antagonism and these were scored as having SA–JA antagonism for the purposes of Table 1 and are discussed in the text. It is important to point out that although there are a growing number of studies using biological inducers, most of the evidence for antagonism is based on treating plants with exogenous SA and JA, either singly or in combination. In most cases, there have not been studies that test whether there is a common genetic basis or a correlated gene expression phenotype that underlies the gross SA–JA antagonism reported across plants. Therefore, the results of this survey and our inferences on trait evolution need to be interpreted with caution because of the inherent limitations of screening for SA–JA antagonism using chemical elicitors and the lack of direct evidence for a common mechanism. The evolutionary interpretations below and our interpretations are hypotheses to be tested. The pathways that produce both hormones at the center of this story have ancient origins. SA is produced downstream of isochorismate synthase (ICS), which occurs in many green and red algae as well as in bacteria, and may have a plastid origin in plants [24]. By contrast, jasmonates are end-products of the ancient octadecanoid (C18) oxylipin pathway. Oxylipins are bioactive lipid derivatives that are used as signaling molecules in plants, animals, fungi [25], as well as in several marine algae species [26]. An allene oxide synthase (AOS) homolog (the second enzyme in the octadecanoid biosynthetic pathway) has been discovered in the moss Physcomitrella patens [27,28], and distant structural homologs to AOS have been putatively identified in three metazoan lineages [29]. The specific compounds JA and methyl JA also have been detected in P. patens [27,30–32], as well as in ferns [33], suggesting that JA production arose at least in the common ancestor of mosses, ferns and seed plants (Figure 1). Despite the ancient origins of each hormone, the antagonism between SA and JA may have more recent origins. SA– JA antagonism has been reported in a total of 17 plant species, including 11 crop plants and six wild species (Table 1). Ancestral state reconstruction [34] indicates that SA–JA antagonism evolved at least at the base of angiosperms, but possibly before the split of gymnosperms and angiosperms (Figure 1; using data from Table 1). The presence of orthologs of genes known to be involved in the SA–JA antagonism including NPR1, WRKY70 (WRKY DNA-binding protein 70), GRX480 (Glutaredoxin 480), ERF1 (ETHYLENE RESPONSE FACTOR 1), MYC2 (JASMONATE INSENSITIVE 1, JIN1), ORA59 (OCTADECANOID-RESPONSIVE

Method of JA elicitation

Arabidopsis thaliana

Method of SA elicitation Pieris brassicae eggs/egg extracts SA


SA


Hyaloperonospora parasitica Pseudomonas syringae


Arabidopsis thaliana Arabidopsis thaliana

Arabidopsis thaliana Arabidopsis thaliana Solanum lycopersicum (tomato) Solanum lycopersicum Solanum lycopersicum Solanum lycopersicum

Nicotiana tabacum (tobacco) Nicotiana tabacum

Nicotiana attenuata

Hordeum vulgare (barley)

–

–

Pathogens: Alternaria brassicola, Botrytis cinerea, insects: Frankliniella occidentalis, Pieris rapae MeJA

–

–

PDF1.2 expression decreases

–

[59]

MeJA

–

–

[10]

Mutant plants with elevated or suppressed SA SA Cucumber mosaic virus SA

–

–

–

MeJA

– – –

Genome wide effects JA inducible transcripts decrease Proteinase inhibitors decrease

Decreased resistance to Trichoplusia ni Trichoplusia ni resistance decreased as SA expression increased – – –

BTH

JA

PR4 (PATHOGENESIS RELATED 4) transcripts downregulated

Oxidative enzymes decrease

[53,55]

Botrytis cinerea

–

SA induced

Parasitic plant Cuscuta pentagona and SA deficient plants BTH

–

SA induced

Proteinase inhibitor transcripts decrease JA and herbivore induced plant volatiles decrease

JA

–

Polyphenol oxidase activity decrease

Decreased resistance to Spodoptera exigua and Trichoplusia ni B. cinerea disease increased Spodoptera exigua performance not affected Spodoptera exigua performance not affected

Mechanical damage

–

–

Tobacco mosaic virus inoculation Genetically reduced SA production –

–

–

Increased JA correlates with decreased SA JA and nicotine decrease

–

–

Fatty acid–amino acid conjugates from Spodoptera exigua oral secretion –

SA

JA and systemin

JA pathway inducibility measurement Ten insect-induced JA regulated transcripts decrease Peroxidase, polyphenol oxidase, chitinase, glucosinolates decrease PDF 1.2 (PLANT DEFENSIN 1.2) decreases

Bioassay result

Refs

Decreased resistance to Spodoptera littoralis Decreased resistance to Spodoptera exigua

[44]

–

[59]

[84]

[65]

[85] [45] [86]

[47] [48]

[52]

[70] [46]

JA, nicotine, polyphenol oxidase increase

Decreased resistance to Manduca sexta Increased resistance to Heliothis virescens

SA decreases

–

–

[80]

–

13-Hydroxyoctadecatri(di)enoic (JA suppressor) increase

–

[88]

[87]

3


Solanum lycopersicum cv. cerasiforme (wild tomato) Oryza sativa (rice)

–

SA pathway inducibility measurement SA induced


Plant species

Review

Table 1. Evidence for SA–JA antagonism across plant species

Method of SA elicitation BTH

Method of JA elicitation

Pisum sativum (pea)

SA

Phaseolus lunatus (lima bean) Gossypium hirsutum (cotton)

Cucumis sativus (cucumber)

Sorghum bicolor (sorghum) Ginkgo biloba Brassica carinata (Ethiopian mustard) Brassica nigra (black mustard) Brassica oleracea (cabbage) Brassica napus (oilseed rape) Asclepias tuberosa (butterfly milkweed)

SA pathway inducibility measurement Reduced chitinase levels on dual-elicited plants

JA pathway inducibility measurement –

Wounding, JA

–

Whitefly, SA

JA

–

JA, polyphenol oxidase downregulated JA, volatiles

Phenacoccus solenopsis (mealy bugs) SA

–

SA-induced volatiles and upregulation of SA-dependent transcripts

Gossypol and other transcripts downregulated

MeJA

Some SA transcripts downregulated

Transgenic suppression of SA Sclerotinia sclerotiorum (white mold) SA applied to roots SA applied to roots SA

–

–

JA

Bioassay result

Refs

Colletotrichum orbiculare disease severity lower on dual elicited plants –

[58]

[51]

Predatory mite attraction reduced –

[49]

some JA transcripts downregulated

–

[56]

JA, OPDA levels decrease

–

[72]

–

[89]

–

[66]

[50]

Sclerotinia sclerotiorum

JA transcripts upregulated after SA transcripts downregulated

–

–

SA transcripts upregulated only after JA transcripts are downregulated JA downregulated in roots

–

–

JA downregulated in roots

–

[66]

Mechanical wounding, Methyl jasmonate

–

–

[90]

Danaus plexippus (monarch) herbivory

SA decreases

Myrosinase-associated protein downregulated in dual-elicited plants JA upregulated

–

A.A. Agrawal, unpublished

Abbreviations: BTH, benzothiadiazole; JA, jasmonic acid; SA, salicylic acid.


Plant species

Review

4

Table 1 (Continued )



Review

E SA vid –J enc A e an for ta go ni

sm


Last common ancestor of land plants probably produced SA and JA. Bryophytes

? Physcomitrella patens ? Selaginella moellendorffii ? Ferns

Lycophytes Pteridophytes Orthologs of Arabidopsis genes important in SA-JA antagonism found in all available land plant genomes (see Table 2).

Ginkgo biloba

Gymnosperms

0.50

Picea abies Sorghum bicolor 0.50

Monocots 0.59

0.84 0.99

Zea mays Oryza sativa Hordeum vulgaris

Earliest node with support for origin of SA-JA antagonism.

0.99

Angiosperms

0.62

Pisum sativum

0.98

Rosids

Cucumis sativus 0.86 0.96

Eudicots

Gossypium hirsutum

Asterids NPR-1 modulates SA–JA antagonism 0.## Probability of SA–JA antagonism as ancestral state

Arabidopsis thaliana Brassica spp. **

0.95 0.70 0.50

Key:

Phaseolus lunatus

Asclepias tuberosa Asclepias exaltata

0.83

Solanum lycopersicum 0.98 0.99

Nicotiana tabacum Nicotiana attenuata

SA–JA antagonism present SA–JA antagonism absent

?

No data TRENDS in Plant Science

Figure 1. Phylogeny of green plants showing putative and reconstructed ancestral states for key aspects of SA–JA antagonism. Topology is based on that published on the Angiosperm Phylogeny Website [96]. Antagonism between SA and JA signaling has only been investigated in seed plants, and only sparsely among gymnosperms (see Table 1 for details and references). The ancestral state of the antagonism was inferred using the ace function in the R library APE [34] using maximum likelihood with branch lengths set to 1. Node labels are probabilities (between 0 and 1) of trait presence given equal gain/loss transition probabilities. The antagonism was probably present in the ancestor of all angiosperms, and in the ancestor of all seed plants, but whether the antagonism is present in the gymnosperms is equivocal given poor taxon sampling. To our knowledge, there are no data addressing the existence of SA–JA antagonism in sister taxa of seed plants. This is despite the occurrence of close genetic orthologs of many genes known to affect the antagonism in angiosperms (Table 2). BLASTs of Arabidopsis thaliana genes (Table 2) were conducted using blastp searches against the following taxa: Physcomitrella patens (NCBI taxon id: 3218), Selaginella moellendorffii (taxon id: 88036), Sorghum bicolor (taxon id: 4558), Zea mays (taxon id: 4577), Oryza sativa var. Japonica (taxon id: 39947), Solanum lycopersicum (taxon id: 4081), as well as an expressed sequence tag database for the fern Pteridium aquilinum. For all genes, a hit was found to all taxa (except P. aquilinum) with an e-value < e–10. Hits (e-value < e–6) to P. aquilinum were found for AtGRX480 and AtMPK4 using blastn searches against the nonhuman, nonrodent EST database; additional genes in this fern were possibly missed due to low coverage of the P. aquilinum transcriptome. Because of extensive gene and genome duplications across plants, BLAST results convey conservation of gene families, members of which were inherited by the ancestor of all land plants, although the vast majority of hits from the taxa above represent reciprocal best blast hits back to the Arabidopsis thaliana genes used as queries. Thus, in principle, the genetic machinery underpinning the SA–JA antagonism was available early on in the evolution of land plants. This in itself is not evidence of SA–JA antagonism. An NPR1 ortholog in Oryza sativa modulates the SA–JA antagonism, which is similar to NPR1 in Arabidopsis, suggesting this aspect of the antagonism may have been present before the split between monocots and eudicots. More extensive taxon sampling is required before evaluating the evolution of this function for NPR1 across plants. **Four Brassica species (B. carinata, B. nigra, B. oleracea and B. napus) all exhibit SA–JA antagonism (Table 1).

ARABIDOPSIS AP2/ERF 59), JAZ1-JAZ3 (JASMONATE ZIM-DOMAIN) are predicted, based on reciprocal best blastp searches using Arabidopsis proteins as subjects (Table 2), to have been present in the first land plants, after this lineage split with green algae. This suggests that many regulatory features of SA–JA crosstalk have diverse and potentially ancient roles in the cell. An ortholog of the canonical crosstalk regulator NPR1 was probably present in the ancestor of all land plants, indicating that the potential for this gene to mediate SA–JA antagonism exists in all

species in which the antagonism has been found (Table 1, Figure 1). NPR1 exhibits unique roles in SA–JA crosstalk in different extant plant species. Unlike in Arabidopsis, tobacco (Nicotiana attenuata) NPR1 acts as a negative regulator of signal crosstalk in the presence of herbivory by preventing SA from suppressing JA-responsive defenses [35]. In this study, herbivory induced SA and JA, as well as NPR1 gene expression [35]. This functioned to prevent SA from repressing JA defenses against the herbivore, thus prioritizing JA over SA. 5

AT gene symbol GRX480

Arabidopsis thaliana NP_174170.1

Solanum lycopersicum NP_001233988.1

Sorghum bicolor XP_002440249.1

Oryza sativa

Zea Mays

Physcomitrella patens XP_001770429.1

NP_001043812.1

ERF1

NP_188965.1

NP_001234695.1

XP_002463464.1

NP_001051973.1

NP_001170395.1

XP_002967934.1

XP_001779786.1

MYC2

NP_174541.1

AAF04917.1

XP_002467448.1

NP_001065478.1

AAD15818.1

XP_002987548.1

XP_001754025.1

NPR1

NP_176610.1

NP_001234558.1

XP_002455011.1

NP_001042286.1

NP_001152107.1

XP_002992598.1

XP_001778211.1

ORA59

NP_172106.1

NP_001234695.1

XP_002461637.1

NP_001051973.1

NP_001170395.1

XP_002966804.1

XP_001779786.1

WRKY70

NP_191199.1

NP_001234530.1

XP_002441930.1

NP_001055192.1

NP_001147748.1

XP_002961829.1

XP_001778254.1

MPK4

NP_192046.1

NP_001234660.1

XP_002467591.1

NP_001061028.2

NP_001105239.1

XP_002976336.1

XP_001763232.1

JAZ1

NP_564075.1

NP_001234883.1

XP_002465159.1

NP_001060268.1

NP_001150658.1

XP_002984538.1

XP_001785097.1

JAZ2

NP_565096.1

NP_001234883.1

XP_002461012.1

NP_001050322.1

NP_001148852.1

XP_002984538.1

XP_001785091.1

JAZ3

NP_566590.1

NP_001234373.1

XP_002462352.1

NP_001063121.1

NP_001141029.1

XP_002975031.1

XP_001754769.1

Role in JA–SA crosstalk in Arabidopsis This SA- and NPR1-induced glutaredoxin represses JA-responsive PDF1.2 in a TGA-transcription factor-dependent manner. This ET- and JA-responsive factor suppresses MYC2-dependent JA responses. ERF1 is suppressed by NPR1. This JA-induced transcription factor is inhibited by ET/JA-mediated ERF1 expression. SA suppression of JA in Arabidopsis is NPR1-dependent. This ET- and JA-responsive transcription factor is necessary for preventing SA suppression of JA in the presence of ET. NPR1-mediated suppression of JA is controlled by WRKY70 and downstream TGA transcription factors. This MAP kinase is a negative regulator of SA and a positive regulator of JA by suppressing the SA activators/JA repressors PAD4 and EDS1. JAZ proteins mediate JA crosstalk with a variety of other pathways, including SA, ET, Auxin and Gibberellin. In the absence of JA, JAZ proteins repress the JA-responsive TFs EIN3/EIL1, which suppresses SA synthesis through effects on ICS (SID2). JAZ3 (JAI3) was the first JAZ protein identified to repress the JA-responsive transcription factor, MYC2.

Refs [91]

[91,92]

[92]

[93] [20]

[94]

[95]

[39]

[39]

[39]


NP_001147414.1

Selaginella moellendorffii XP_002988222.1


Review

6

Table 2. Phylogenetic distribution of orthologs to Arabidopsis thaliana genes important in the JA–SA antagonism


Review Functional data, based on gene expression and/or other studies, show that NPR1 modulates SA–JA antagonism in rice, Arabidopsis and tomato, suggesting that this subfunction for NPR1 may have ancient origins in the common ancestor of monocots and eudicots. Another Arabidopsis gene involved in crosstalk, WRKY70, is present as a ortholog in rice (Oryza sativa WRKY13). These orthologs positively regulate SA-induced and negatively regulate JAinduced responses [36]. Although OsWRKY13 is not a one-to-one ortholog of WRKY70 in Arabidopsis, the central role of a WRKY transcription factor in modulating between SA- and JA-dependent responses is important in both species. In addition, a microarray analysis on OsWRKY13-overexpression rice lines found suites of SAand JA-regulated genes displaying reciprocal antagonism [37]. These are patterns similar to those found in Arabidopsis. Among their many functions, JAZ proteins repress the JA-responsive ethylene-signaling genes EIN3 (ETHYLENE INSENSITIVE 3)/EIL1 (ETHYLENE INSENSITIVE 3 LIKE1), which when expressed lead to suppression of SA synthesis [38]. Upon activation of the JA receptor COI1 (CORONATINE INSTENSITIVE 1), JAZ repressor proteins are degraded, allowing for the activation of JAresponsive signaling cascades [39]. JAZ-mediated repression and derepression appears to be important in mediating not only SA–JA crosstalk but also JA–ET, JA– Gibberellin and JA–auxin signal interactions [39]. This suggests that signal crosstalk may be a fundamental attribute of plant genetic networks [22,40] and may be commonly achieved through JAZ-mediated repression [39]. Orthologs of JAZ proteins identified in Arabidopsis have been discovered in P. patens and other early diverging land plants [41] (and this study). Together, the role of NPR1, WRKY and JAZ genes in regulating SA–JA and SA–JA–ET crosstalk from rice to eudicots suggests a generally conserved core genetic architecture to defense signaling in flowering plants. Nonetheless, the presence/absence of these genes is not sufficient evidence of any SA–JA antagonism. Although the antagonism frequently occurs, it also appears to be absent in several lineages: Picea abies (Pinaceae) (J. Arnerup, PhD thesis, Swedish University of Agricultural Sciences, 2011), Zea mays (Poaceae) [42] and Asclepias exaltata (Apocynaceae) (Table 1, Figure 1). For two closely related milkweed species studied in the same experiment, Asclepias tuberosa showed the antagonism whereas A. exaltata did not (A.A. Agrawal, personal communication). Antagonisms are common when chemical elicitors are the inducing agents [43] and when one pathway is genetically suppressed (ginkgo, Arabidopsis, tomato, tobacco) (Table 1). There is also widespread evidence that an antagonism occurs following induction by a biological agent. An extensive range of inducers in Arabidopsis has been investigated and the antagonism has been found following infection by bacteria, virus and fungi, leaf damage by thrips (Thysanoptera) and lepidopteran larvae, and oviposition of lepidopteran eggs [16,44,10,45]. Virus infection reduces induction of the JA pathway in tobacco [46]. In tomato, the antagonism occurs following infection by a parasitic plant and a fungus [47,48]. In lima bean (Phaseolus lunatus) and cotton plants (Gossypium hirsutum),


SA induction by whiteflies and mealybugs decreased sensitivity to JA [49,50]. Finally, in milkweed plants (A. tuberosa) monarch caterpillar feeding increased JA levels and decreased SA (A.A. Agrawal, personal communication). Does antagonism at the level of gene expression or hormone levels translate into a change in actual resistance level? In a small subset of these examples antagonism is inferred based on monitoring readouts of an end-product such as gossypol levels in cotton [50], polyphenol oxidase activity in pea (Pisum sativum) [51] and volatiles in cultivated tomato (Solanum lycopersicum) [48]. There are few examples that also include a bioassay to test for an antagonism, and when they do, an antagonism at the level of gene expression sometimes resulted in reduced resistance and sometimes it did not. Specifically, in Arabidopsis, cultivated tomato and tobacco, the antagonism has been shown to decrease resistance to a future attacker, yet in wild tomato there was no effect [52]. In the cases where the antagonism occurred following a biological inducer and resulted in decreased resistance, the inducing agent was usually a generalist attacker (whiteflies, aphids, Pseudomonas syringae; Table 1). Very few papers examined SA– JA antagonism in a field setting [53,35] and we found no study that measured the consequences of the antagonism for plant fitness. Although SA induction frequently suppresses JA induction, and plants have long been hypothesized to prioritize SA over JA induction, there are seven species in which JA responses were associated with the suppression of SA induction [54–58]. The sequence in which SA and JA are added exogenously in experiments influences the strength of the reciprocal antagonism [20], and the timing [59] or dosage [60] of hormone application is important for realization of the antagonism [59]. In some cases, SA and JA pathways are each upregulated by one attacker species, but their induction is not simultaneous. For example, following infection by Fusarium spp., a hemibiotrophic fungal pathogen, both the SA and JA pathways are induced after infection, but SA is important in establishing resistance early on, and JA is important in facilitating resistance during later time points [61]. Thus, although both the SA and JA pathways are induced by the same pathogen, the responses are temporally disconnected. Screens across genotypes of Arabidopsis revealed variation in priming of the SA and JA pathways that manifested as coinduction of SA and JA when a fungal species was used as the inducer [62]. However, these genotype-specific effects were only in the context of actual pathogen attack and were not observed when hormones were applied to plants [59]. All of this work points to the fact that the antagonism is highly context-dependent, both in terms of what is used to elicit SA and JA, the timing of the elicitation, and possibly with respect to genetic variation underlying the antagonism. The suppression of SA by JA is either triggered by a biological inducer (Arabidopsis, milkweed and Brassica), or follows after chemical or genetic manipulation of the SA pathway (tomato, millet, tobacco, cucumber). For example, the jasmonate mimic coronatine produced by Pseudomonas syringae activates the JA pathway and suppresses the SA pathway in Arabidopsis [63]. 7


Review Within-plant factors including the timing, concentration and location of induction influence whether crosstalk and an antagonism occurs. Effects of timing have been shown with elicitor studies that temporally manipulate the sequence of application, and with studies showing that endogenous JA and SA levels change inversely with each other [59,64]. Most studies only test for local but not systemic interactions. However, systemic antagonism in Arabidopsis is induced by Pseudomonas infection [10,65], and insect eggs and egg extracts only induced antagonism locally [44]. Root SA elicitation decreased JA inducibility within the root but did not reduce JA inducibility in shoot tissues [66]. We know little about how intensity of induction [55,67,68] and factors such as plant sectoriality and phenology influence signal antagonism. Because these aspects of plant form and growth influence hormone induction per se [69], they will probably influence the interaction between hormonal pathways. Adaptive and nonadaptive hypotheses for the antagonism Is the SA–JA antagonism an artifact of complex signaling? Plants have a limited number of hormone signal molecules, which by chance may sometimes interact to affect gene expression positively or negatively. In this scenario, different environmental conditions such as the location and timing of attacked generate specificity in the antagonism. Although this is possible, the existence of conserved genes (e.g. NPR1), conserved across several distantly related plant taxa that regulate SA–JA interactions in diverse taxa (e.g. rice, tobacco, Arabidopsis) makes this hypothesis unlikely. Is the SA–JA antagonism an ancient constraint found in plants and animals? Lipid-derived, jasmonate-like animal hormones such as prostaglandins are inhibited in animals by aspirin (i.e. acetylsalicylic acid). Because a similar antagonism is also widespread in plants (Figure 1), it may represent an ancient evolutionary constraint [70]. In addition, several genes that underpin crosstalk regulation in Arabidopsis have close homologs in the moss P. patens and the lycophyte Selaginella moellendorffii, in addition to several angiosperms (Table 2, Figure 1). This indicates that the genetic machinery to express and regulate crosstalk is widely conserved to this day and was probably ancestral to all land plants. However, gene presence/absence does not imply functional conservation. Because there is variation in whether the antagonism is expressed even between closely related taxa (Figure 1), expression of SA–JA antagonism is not an unbreakable constraint. Is SA–JA antagonism due to resource allocation costs of induction? There are fitness costs associated with the induction of SA and JA defenses in the absence of a natural enemy attack [71]. Thus, the SA–JA antagonism could be viewed as either a limitation of or adaptation to a resource-limited environment. There are at least two scenarios to consider whereby JA and SA pathways either regulate different 8


defense products or the same defense products. When crosstalk limits production of a product, antagonisms in the induction of each may prevent simultaneous induction [72]. When the crosstalk limits production of a product regulated by only one pathway, signal crosstalk can be a means of maintaining production of one product instead of another. The effect of elicitor concentration and exposure time on whether the antagonism is found supports this hypothesis [43]. Resource allocation costs are probably partially responsible for shaping the patterns of induction following attack, but several lines of evidence suggest they are not likely to be the only factor. The SA and JA pathways do not utilize the same precursors or components for their signal transduction pathways, which makes specific resource limitation less likely to be the explanation at that level. However, more importantly, strict competition for precursors, such as amino acids, should result in downregulation of many plant functions, not only particular JA or SA regulated genes. Therefore, costs alone do not explain the apparent specificity in the antagonism: decreased inducibility of the jasmonate pathway following light limitation is due to specific hormonal modulation [73], not simply reduced resource availability. Similarly, decreasing nitrogen availability actually increased the expression of the jasmonate pathway due to altered interactions between jasmonate and ethylene. Thus, decreasing nutrient levels can even increase defense expression, evidence against strict resource mediated SA–JA crosstalk [74]. Is SA–JA antagonism a means for the plant to adaptively tailor its responses to different enemies and also a target for manipulation by enemies? Downstream defenses that are modulated by the SA and JA pathways affect pathogens and herbivores, and each attacker may be affected by a different subset of these defense products. Thus, the adaptive tailoring hypothesis predicts that the plant should induce the components of each pathway that are most effective against the current attacker. This implies some degree of specificity on the plant’s part – if the plant is tailoring its defense response adaptively then different enemies must be recognized as distinct by the plant [5,75]. Many of the patterns described above almost make specificity axiomatic, such as the general asymmetry of SA and JA suppression, the important role of other hormones, the effects of the pattern of damage on the expression of crosstalk, and the effect of other enemies present on the plant [76]. A major unanswered question is whether crosstalk is adaptive for the plant [42]. If crosstalk tailors the plant’s response to a particular attacker this specificity should increase the plant’s resistance to that attacker. However, the selective advantage of manipulating crosstalk from the perspective of a particular attacker must be high. Thus, the specificity of response is a complex phenotype mediated by plant and attacker. The elicitors present in, for example, the saliva or accessory gland secretions from a particular herbivore species that is attacking a plant often determine the specificity of these responses in the plant [10,77]. Manipulation of hormonally regulated pathways may be a mechanism by which enemies can suppress induced


Review defenses in biochemically divergent plants [78]. Given that the SA–JA antagonism appears to be phylogenetically widespread and ancient, this method of manipulating the host plant has been available for a long time and may work against a wide diversity of plants [78]. Consistent with this hypothesis, generalist enemies have been found to induce SA–JA crosstalk in a way that benefits them [10,79–81]. Thus, experiments are required to understand who benefits (the plant, the plant’s attacker or neither?) and yet few studies explicitly connect the plant’s specific response to an effect on plant resistance [82] or performance. We propose that testing the adaptive value of specificity will require experiments that incorporate a ‘neutral’ inducer such as mechanical damage or pure hormonal application as controls. The effects of ‘neutral’ induction and induction in response to the biological organism can then Box 1. What experiments are needed to effectively test the adaptive significance of SA–JA antagonism? Constraints hypothesis (i) Do simple phylogenetic constraints explain SA–JA antagonism? Analysis of phylogenetic distribution of the SA–JA antagonism using common elicitors would illuminate repeated losses and gains. SA–JA antagonism is widespread across plants, but evidence is missing from early diverging lineages. (ii) Do the same pathways exist across plants for modulating the antagonism? This can be tested by measuring patterns of gene expression in candidate SA, JA and crosstalk modulator loci in dual elicitation experiments across plant diversity in a common environment [9]. If orthologous loci show common patterns of expression during dual elicitation, it is unlikely to be adaptive tailoring and more likely to be a constraint. Resource allocation costs of induction hypothesis (i) Resource limitation. Isotope tracer studies measuring flux of resources and precursors between the pathways would directly demonstrate resource diversion [43,83]. Resource limitation could also be tested by manipulating resource availability and addressing if SA–JA antagonism is weaker in resource-rich conditions. (ii) Cost of single versus dual elicitation. Is inducing both pathways more costly than inducing one? Plant fitness should be measured following single and dual elicitation.


be compared. For specificity to be adaptive for the plant, the plant’s response to the neutral and biological inducer must differ and this tailoring must benefit the plant. For example, evidence was found against adaptive specificity when chemical elicitation caused a similar pattern of crosstalk as biological induction [59]. If the attacker benefits, it may be manipulating the plant to its benefit. Alternatively, the response may not be adaptive for the plant or the attacker. Consistent with the adaptive tailoring hypothesis is that there is extensive variation in patterns of induction across plant diversity, which is a prerequisite for, but not yet evidence of, adaptation. In summary, critical data on the consequences of SA–JA antagonism for plants in the field are too scant to address this adaptive tailoring hypothesis at present. A prospectus on future experiments Our most important conclusion is that in order to test the various hypotheses proposed above: (i) measurements of the SA–JA reciprocal antagonism in the form of gene expression and biochemical activity must be coupled with pathogen and herbivore bioassays and simultaneous measurements of plant fitness, and (ii) that these experiments must be conducted in ecologically relevant settings and across plant diversity. From an evolutionary perspective, future experiments should attempt to test if the SA–JA antagonism arose in a reciprocal manner or sequentially with unidirectional antagonisms arising separately. Future genome sequencing of plant species where there is no evidence for the antagonism could reveal if, and perhaps how, SA–JA antagonism was lost or if there are other conditions under which the antagonism is expressed. Researchers should focus on understanding if indeed SA–JA reciprocal antagonism arose once and if there is a common genetic basis to this phenomenon across the plants in which it occurs. Specific recommendations are given in Box 1. Acknowledgments We thank Martin Heil, Anurag Agrawal, the Cornell Plant-Interactions Group and members of the Whiteman Laboratory at the University of Arizona for comments. We also thank Will Petry (UC-Irvine) for comments on the manuscript.

References Adaptive tailoring hypothesis (i) Biological elicitors result in varied expression levels and patterns of loci involved in the antagonism relative to chemical elicitors or mechanical elicitation. Greater variance and distinct patterns in the antagonism across attacker species would support the tailoring hypothesis. If chemical and biological inducers show similar patterns, this would not support adaptive tailoring. Higher resistance and plant performance following biological induction of SA–JA antagonism compared with chemical elicitation would be evidence for adaptive tailoring. (ii) Is genetic variation in the antagonism adaptive? Genotypes varying in the antagonism could be placed into environments varying in attacker composition. In environments with one attacker, the antagonism is more likely to benefit the plant compared with environments with multiple enemies (unless the antagonism is induced to the benefit of the attacker). Artificial selection experiments and forcing induction of the alternative pathway could reveal how natural selection shapes the antagonism.

1 Frankel, G. (1959) The raison d’etre of secondary plant substances. Science 129, 1466–1470 2 Jaenike, J. (1990) Host specialization in phytophagous insects. Annu. Rev. Ecol. Syst. 21, 243–273 3 Rubinstein, C.V. et al. (2010) Early Middle Ordovician evidence for land plants in Argentina (eastern Gondwana). New Phytol. 188, 365–369 4 Ausubel, F.M. (2005) Are innate immune signaling pathways in plants and animals conserved? Nat. Immunol. 6, 973–979 5 Erb, M. et al. Role of phytohormones in insect-specific plant reactions. Trends Plant Sci. 5, doi: 10.1016/j.tplants.2012.01.003. 6 Lambrechts, L. (2010) Dissecting the genetic architecture of host– pathogen specificity. PLoS Pathog. 6, e1001019 7 Heil, M. and Baldwin, I.T. (2002) Fitness costs of induced resistance: emerging experimental support for a slippery concept. Trends Plant Sci. 7, 61–67 8 De Vos, M. et al. (2005) Signal signature and transcriptome changes of Arabidopsis during pathogen and insect attack. Mol. Plant Microbe Interact. 18, 923–937 9 Glazebrook, J. (2005) Contrasting mechanisms of defense against biotrophic and necrotphic pathogens. Ann. Rev. Phytopathol. 43, 205–227 9


Review 10 Cui, J. et al. (2005) Pseudomonas syringae manipulates systemic plant defenses against pathogens and herbivores. Proc. Natl. Acad. Sci. U.S.A. 102, 1791–1796 11 Stout, M.J. et al. (2006) Plant-mediated interactions between pathogenic microorganisms and herbivorous arthropods. Annu. Rev. Entomol. 51, 663–689 12 Karban, R. and Baldwin, I.T. (1997) Induced Responses to Herbivory, The University of Chicago Press 13 Green, T.R. and Ryan, C.A. (1972) Wound-induced proteinase inhibitor in plant leaves: a possible defense mechanism against insects. Science 175, 776–777 14 Heil, M. et al. (2001) Extrafloral nectar production of the ant-associated plant, Macaranga tanarius, is an induced, indirect, defensive response elicited by jasmonic acid. Proc. Natl. Acad. Sci. U.S.A. 98, 1083–1088 15 Robert-Seilaniantz, A. et al. (2011) Hormone crosstalk in plant disease and defense: more than just JASMONATE-SALICYLATE antagonism. Annu. Rev. Phytopathol. 49, 317–343 16 Koornneef, A. and Pieterse, C.M.J. (2008) Cross talk in defense signaling. Plant Physiol. 146, 839–844 17 Verhage, A. et al. (2010) Plant immunity: it’s the hormones talking, but what do they say? Plant Physiol. 154, 536–540 18 Pieterse, C.M. et al. (2009) Networking by small-molecule hormones in plant immunity. Nat. Chem. Biol. 5, 308–316 19 Leon-Reyes, A. et al. (2010) Salicylate-mediated suppression of jasmonate-responsive gene expression in Arabidopsis is targeted downstream of the jasmonate biosynthesis pathway. Planta 232, 1423–1432 20 Leon-Reyes, A. et al. (2010) Ethylene signaling renders the jasmonate response of Arabidopsis insensitive to future suppression by salicylic Acid. Mol. Plant Microbe Interact. 23, 187–197 21 Leon-Reyes, A. et al. (2009) Ethylene modulates the role of NONEXPRESSOR OF PATHOGENESIS-RELATED GENES1 in cross talk between salicylate and jasmonate signaling. Plant Physiol. 149, 1797–1809 22 Arabidopsis Interactome Mapping Consortium (2011) Evidence for network evolution in an Arabidopsis interactome map. Science 333, 601–607 23 Beckers, G.J.M. and Spoel, S.H. (2006) Fine-tuning plant defence signalling: salicylate versus jasmonate. Plant Biol. 8, 1–10 24 Wildermuth, M.C. et al. (2001) Isochorismate synthase is required to synthesize salicylic acid for plant defence. Nature 414, 562–565 25 Brodhun, F. and Feussner, I. (2011) Oxylipins in fungi. FEBS J. 278, 1047–1063 26 Gerwick, W.H. (1994) Structure and biosynthesis of marine algal oxylipins. Biochim. Biophys. Acta 1211, 243–255 27 Bandara, P.K.G.S.S. et al. (2009) Cloning and functional analysis of an allene oxide synthase in Physcomitrella patens. Biosci. Biotechnol. Biochem. 73, 2356–2359 28 Oliver, J. et al. (2009) Pythium infection activates conserved plant defense responses in mosses. Planta 230, 569–579 29 Lee, D.S. et al. (2008) Structural insights into the evolutionary paths of oxylipin biosynthetic enzymes. Nature 455, 363–368 30 Anterola, A. et al. (2009) Physcomitrella patens has lipoxygenases for both eicosanoid and octadecanoid pathways. Phytochemistry 70, 40–52 31 Hashimoto, T. et al. (2011) Cloning and characterization of an allene oxide cyclase, PpAOC3, in Physcomitrella patens. Plant Growth Regul. 65, 239–245 32 Stumpe, M. et al. (2010) The moss Physcomitrella patens contains cyclopentenones but no jasmonates: mutations in allene oxide cyclase lead to reduced fertility and altered sporophyte morphology. New Phytol. 188, 740–749 33 Dathe, W. et al. (1989) Occurrence of jasmonic acid, related compounds and abscisic acid in fertile and sterile fronds of three Equisetum species. Biochem. Physiol. Pflanzen 185, 83–92 34 Paradis, E. et al. (2004) APE: analyses of phylogenetics and evolution in R language. Bioinformatics 20, 289–290 35 Rayapuram, C. and Baldwin, I.T. (2007) Increased SA in NPR1silenced plants antagonizes JA and JA-dependent direct and indirect defenses in herbivore-attacked Nicotiana attenuata in nature. Plant J. 52, 700–715 36 Qiu, D. et al. (2007) OsWRKY13 mediates rice disease resistance by regulating defense-related genes in salicylate- and jasmonatedependent signaling. Mol. Plant Microbe Interact. 20, 492–499 10


37 Qiu, D. et al. (2008) Rice gene network inferred from expression profiling of plants overexpressing OsWRKY13, a positive regulator of disease resistance. Mol. Plant 1, 538–551 38 Chen, H. et al. (2009) ETHYLENE INSENSITIVE3 and ETHYLENE INSENSITIVE3-LIKE1 repress SALICYLIC ACID INDUCTION DEFICIENT2 expression to negatively regulate plant innate immunity in Arabidopsis. Plant Cell 21, 2527–2540 39 Kazan, K. and Manners, J.M. (2012) JAZ repressors and the orchestration of phytohormone crosstalk. Trends Plant Sci. 17, 22–31 40 Sato, M. et al. (2010) Network modeling reveals prevalent negative regulatory relationships between signaling sectors in Arabidopsis immune signaling. PLoS Pathog. 6, e1001011 41 Katsir, L. et al. (2008) Jasmonate signaling: a conserved mechanism of hormone sensing. Curr. Opin. Plant Biol. 11, 428–435 42 Engelberth, J. et al. (2011) Low concentrations of salicylic acid stimulate insect elicitor responses in Zea mays seedlings. J. Chem. Ecol. 37, 263–266 43 Heil, M. and Bostock, R.M. (2002) Induced systemic resistance (ISR) against pathogens in the context of induced plant defences. Ann. Bot. 89, 503–512 44 Bruessow, F. et al. (2010) Insect eggs suppress plant defence against chewing herbivores. Plant J. 62, 876–885 45 Lewsey, M.G. et al. (2010) Disruption of two defensive signaling pathways by a viral RNA silencing suppressor. Mol. Plant Microbe Interact. 23, 835–845 46 Preston, C.A. et al. (1999) Tobacco mosaic virus inoculation inhibits wound-induced jasmonic acid-mediated responses within but not between plants. Planta 209, 87–95 47 El Oirdi, M. et al. (2011) Botrytis cinerea manipulates the antagonistic effects between immune pathways to promote disease development in tomato. Plant Cell 23, 2405–2421 48 Runyon, J.B. et al. (2008) Parasitism by Cuscuta pentagona attenuates host plant defenses against insect herbivores. Plant Physiol. 146, 987–995 49 Zhang, P-J. et al. (2009) Whiteflies interfere with indirect plant defense against spider mites in Lima bean. Proc. Natl. Acad. Sci. U.S.A. 106, 21202–21207 50 Zhang, P. et al. (2011) Suppression of jasmonic acid-dependent defense in cotton plant by the mealybug Phenacoccus solenopsis. PLoS ONE 6, e22378 51 Yang, H.R. et al. (2011) Effect of salicylic acid on jasmonic acid-related defense response of pea seedlings to wounding. Sci. Horticult. 128, 166– 173 52 Thaler, J. et al. (2002) Cross-talk between jasmonate and salicylate plant defense pathways: effects on several plant parasites. Oecologia 131, 227–235 53 Thaler, J.S. et al. (1999) Trade-offs in plant defense against pathogens and herbivores: a field demonstration of chemical elicitors of induced resistance. J. Chem. Ecol. 25, 1597–1609 54 Seo, S. et al. (1997) Jasmonic acid in wound signal transduction pathways. Physiol. Plant. 101, 740–745 55 Thaler, J.S. et al. (2002) Antagonism between jasmonate- and salicylate-mediated induced plant resistance: effects of concentration and timing of elicitation on defense-related proteins, herbivore, and pathogen performance in tomato. J. Chem. Ecol. 28, 1131–1159 56 Salzman, R.A. et al. (2005) Transcriptional profiling of sorghum induced by methyl jasmonate, salicylic acid, and aminocyclopropane carboxylic acid reveals cooperative regulation and novel gene responses. Plant Physiol. 138, 352–368 57 Kachroo, P. et al. (2001) A fatty acid desaturase modulates the activation of defense signaling pathways in plants. Proc. Natl. Acad. Sci. U.S.A. 98, 9448–9453 58 Liu, C. et al. (2008) Antagonism between acibenzolar-S-methyl-induced systemic acquired resistance and jasmonic acid-induced systemic acquired susceptibility to Colletotrichum orbiculare infection in cucumber. Physiol. Mol. Plant Pathol. 72, 141–145 59 Koornneef, A. et al. (2008) Kinetics of salicylate-mediated suppression of jasmonate signaling reveal a role for redox modulation. Plant Physiol. 147, 1358–1368 60 Bostock, R.M. (1999) Signal conflicts and synergies in induced resistances to multiple attackers. Physiol. Mol. Plant Pathol. 55, 99–109 61 Ding, L. et al. (2011) Resistance to hemi-biotrophic F. graminearum infection is associated with coordinated and ordered expression of diverse defense signaling pathways. PLoS ONE 6, e19008


Review 62 Ahmad, S. et al. (2011) Genetic dissection of basal defence responsiveness in accessions of Arabidopsis thaliana. Plant Cell Environ. 34, 1191–1206 63 Brooks, D.M. et al. (2005) The Pseudomonas syringae phytotoxin coronatine promotes virulence by overcoming salicylic aciddependent defences in Arabidopsis thaliana. Mol. Plant Pathol. 6, 629–639 64 Luo, Y. et al. (2011) Application of jasmonic acid followed by salicylic acid inhibits cucumber mosaic virus replication. Plant Pathol. J. 27, 53–58 65 Cui, J. et al. (2002) Signals involved in Arabidopsis resistance to Trichoplusia ni caterpillars induced by virulent and avirulent strains of the phytopathogen Pseudomonas syringae. Plant Physiol. 129, 551–564 66 van Dam, N.M. et al. (2004) Interactions between aboveground and belowground induction of glucosinolates in two wild Brassica species. New Phytol. 161, 801–810 67 van Wees, S.C.M. et al. (2000) Enhancement of induced disease resistance by simultaneous activation of salicylate- and jasmonatedependent defense pathways in Arabidopsis thaliana. Proc. Natl. Acad. Sci. U.S.A. 97, 8711–8716 68 Mur, L.A.J. et al. (2006) The outcomes of concentration-specific interactions between salicylate and jasmonate signaling include synergy, antagonism, and oxidative stress leading to cell death. Plant Physiol. 140, 249–262 69 Bledsoe, T.M. and Orians, C.M. (2006) Vascular pathways constrain C13 accumulation in large root sinks of Lycopersicon esculentum (Solanaceae). Am. J. Bot. 93, 884–890 70 Lee, A. et al. (2004) Inverse correlation between jasmonic acid and salicylic acid during early wound response in rice. Biochem. Biophys. Res. Commun. 318, 734–738 71 Lou, Y.G and Baldwin, I.T. (2004) Nitrogen supply influences herbivore-induced direct and indirect defenses and transcriptional responses to Nicotiana attenuata. Plant Physiol. 135, 496–506 72 Xu, M. et al. (2009) Complementary action of jasmonic acid on salicylic acid in mediating fungal elicitor-induced flavonol glycoside accumulation of Ginkgo biloba cells. Plant Cell Environ. 32, 960–967 73 Moreno, J.E. et al. (2009) Ecological modulation of plant defense via phytochrome control of jasmonate sensitivity. Proc. Natl. Acad. Sci. U.S.A. 106, 4935–4940 74 Schmelz, E.A. et al. (2003) Nitrogen deficiency increases volicitininduced volatile emission, jasmonic acid accumulation, and ethylene sensitivity in maize. Plant Physiol. 133, 295–306 75 Ali, J.G. and Agrawal, A.A. (2012) Specialist versus generalist insect herbivores and plant defense. Trends Plant Sci. 5, DOI: 10.1016/ j.tplants.2012.02.006 76 Rodriguez-Saona, C.R. et al. (2010) Molecular, biochemical, and organismal analyses of tomato plants simultaneously attacked by herbivores from two feeding guilds. J. Chem. Ecol. 36, 1043–1057 77 Uppalapati, S.R. et al. (2007) The phytotoxin coronatine contributes to pathogen fitness and is required for suppression of salicylic acid accumulation in tomato inoculated with Pseudomonas syringae pv. tomato DC3000. Mol. Plant Microbe Interact. 20, 955–965


78 Li, X. et al. (2002) Jasmonate and salicylate induce expression of herbivore cytochrome P450 genes. Nature 419, 712–715 79 Weech, M.H. et al. (2008) Caterpillar saliva interferes with induced Arabidopsis thaliana defence responses via the systemic acquired resistance pathway. J. Exp. Bot. 59, 2437–2448 80 Diezel, C. et al. (2009) Different lepidopteran elicitors account for crosstalk in herbivory-induced phytohormone signaling. Plant Physiol. 150, 1576–1586 81 Consales, F. et al. (2012) Insect oral secretions suppress woundinduced responses in Arabidopsis. J. Exp. Bot. 63, 727–737 82 Thaler, J.S. et al. (2010) Salicylate-mediated interactions between pathogens and herbivores. Ecology 91, 1075–1082 83 Baldwin, I.T. and Hamilton, W. (2000) Jasmonate-induced responses of Nicotiana sylvestris results in fitness costs due to impaired competitive ability for nitrogen. J. Chem. Ecol. 26, 915–952 84 Cipollini, D. et al. (2004) Salicylic acid inhibits jasmonic acid-induced resistance of Arabidopsis thaliana to Spodoptera exigua. Mol. Ecol. 13, 1643–1653 85 Schenk, P.M. et al. (2000) Coordinated plant defense responses in Arabidopsis revealed by microarray analysis. Proc. Natl. Acad. Sci. U.S.A. 97, 11655–11660 86 Doares, S.H. et al. (1995) Salicylic acid inhibits synthesis of proteinase inhibitors in tomato leaves induced by systemin and sasmonic acid. Plant Physiol. 108, 1741–1746 87 Felton, G.W. et al. (1999) Inverse relationship between systemic resistance of plants to microorganisms and to insect herbivory. Curr. Biol. 9, 317–320 88 Weichert, H. et al. (1999) Metabolic profiling of oxylipins upon salicylate treatment in barley leaves – preferential induction of the reductase pathway by salicylate. FEBS Lett. 464, 133–137 89 Yang, B. et al. (2010) Characterization of defense signaling pathways of Brassica napus and Brassica carinata in response to Sclerotinia sclerotiorum challenge. Plant Mol. Biol. Rep. 28, 253–263 90 Taipalensuu, J. et al. (1997) Regulation of the wound-induced myrosinase-associated protein transcript in Brassica napus plants. Eur. J. Biochem. 247, 963–971 91 Ndamukong, I. et al. (2007) SA-inducible Arabidopsis glutaredoxin interacts with TGA factors and suppresses JA-responsive PDF1.2 transcription. Plant J. 50, 128–139 92 Lorenzo, O. et al. (2004) JASMONATE-INSENSITIVE1 encodes a MYC transcription factor essential to discriminate between different jasmonate-regulated defense responses in Arabidopsis. Plant Cell 16, 1938–1950 93 Spoel, S.H. et al. (2003) NPR1 modulates cross-talk between salicylateand jasmonate-dependent defense pathways through a novel function in the cytosol. Plant Cell 15, 760–770 94 Li, J. et al. (2004) The WRKY70 transcription factor: a node of convergence for jasmonate-mediated and salicylate-mediated signals in plant defense. Plant Cell 16, 319–331 95 Brodersen, P. et al. (2006) Arabidopsis MAP kinase 4 regulates salicylic acid- and jasmonic acid/ethylene-dependent responses via EDS1 and PAD4. Plant J. 47, 532–546 96 Stevens, P.F. (2001 onwards) Angiosperm Phylogeny Website, Version 9 (June 2008); http://www.mobot.org/MOBOT/research/APweb/)

11

Review


Community specificity: life and afterlife effects of genes Thomas G. Whitham, Catherine A. Gehring, Louis J. Lamit, Todd Wojtowicz, Luke M. Evans, Arthur R. Keith and David Solance Smith Department of Biological Sciences and Merriam-Powell Center for Environmental Research, Northern Arizona University, Flagstaff, AZ 86011, USA

Community-level genetic specificity results when individual genotypes or populations of the same species support different communities. Our review of the literature shows that genetic specificity exhibits both life and afterlife effects; it is a widespread phenomenon occurring in diverse taxonomic groups, aquatic to terrestrial ecosystems, and species-poor to species-rich systems. Such specificity affects species interactions, evolution, ecosystem processes and leads to community feedbacks on the performance of the individuals expressing the traits. Thus, genetic specificity by communities appears to be fundamentally important, suggesting that specificity is a major driver of the biodiversity and stability of the world’s ecosystems. Genetic specificity by communities Specificity is often defined as the number of different host species with which a plant enemy or mutualist associates. Researchers in diverse fields have expanded this definition to include factors such as phylogenetic relationships among hosts [1,2]. We propose that specificity should be broadened in two additional ways. First, we provide evidence that specificity by plant associates such as pathogens, herbivores or mutualists frequently occurs below the species level. Studies of this genetic specificity are important for understanding the process of speciation [3]. Second, we show that entire communities of organisms can exhibit specificity for plants below the species level, in which different genotypes of plants support different communities (see Glossary). In turn, these communities can feed back to affect the performance of the individual genotypes with which they interact. Studies of specificity at this level demonstrate novel links between ecology and evolution. Figure 1 shows an example of specificity in which a diverse community of arthropods (103 species from 12 orders) exhibit specificity for individual tree genotypes [4]. Replicate clones of the same genotypes of narrowleaf cottonwood (Populus angustifolia) showed significant broad-sense heritability of the community phenotype. Similarly, repeated censuses across years showed high repeatability, demonstrating that the colonizing communities of arthropods responded similarly each year to individual tree genotypes. Specificity among species has long been Corresponding author: Whitham, T.G. ([email protected]).

studied in the context of coevolution. At the community level, coevolutionary dynamics undoubtedly play a role; however, we predict that the effects of plant genetics on the associated community is highly asymmetric [5] in which only a few species are coevolved and the genetically based interactions of a few species (e.g. plant–enemy interactions) are likely to define much of the community. Many associated species may have no individual feedbacks to the plant, whereas the whole community of arthropods or soil microbes, acting together, does affect the fitness of individual genotypes. Although several reviews have documented the extended community and ecosystem phenotypes of individual plant genotypes [6,7], this review emphasizes the breadth of model and non-model systems that have demonstrated genetic effects at the community level and is the first to place these within the context of specificity. Community specificity is largely an outgrowth of community heritability, which is the tendency for related individuals to support similar community members and ecosystem processes [6]. Thus, with low heritability, community specificity should be weak, but as heritability increases, specificity should also increase. Our review examines: (i) how subspecific levels of plant genetic variation (populations, genotypes) differentially affect community structure, ecosystem processes, diversity and stability across a range of organisms, Glossary Afterlife effects: the phenotypic effects of a plant that extends beyond the life of the individual plant or plant part such as leaf litter. Broad-sense heritability: the contribution of all genetic factors (additive, dominant, epistatic) to the total variance in phenotype. H2 is the broad-sense heritability of a traditional phenotype and H2C is the broad-sense heritability of a community or ecosystem phenotype [35]. Community and ecosystem phenotypes: the effects of genes at levels higher than the population [6]. Community genetics: the study of the genetic interactions that occur between species and their abiotic environment in complex communities [6]. Community specificity: the genetically based tendency for individual populations or individual genotypes within a species to support different communities of organisms and ecosystem processes. Community stability: the similarity in the community composition of associated species across years for individual plant genotypes or populations [4]. Deme: genetically distinct populations that form despite close proximity to one another [3]. Foundation species: ‘a single species that defines much of the structure of a community by creating locally stable conditions for other species. . .’ [72]. Other terms such as keystone, ecosystem engineers or dominant species have similar meanings and overlap in their definitions.

1360-1385/$ – see front matter ß 2012 Published by Elsevier Ltd. doi:10.1016/j.tplants.2012.01.005 Trends in Plant Science, May 2012, Vol. 17, No. 5

271

Review

[(Figure_1)TD$IG]


Yr3 Yr3 Yr1 Yr1

Yr3 Yr1

Yr3 Yr2

Yr2

Yr1 Tree genotype

Yr2

Key: Yr2

1000 HE-8 1020 1008

More stable community

Less stable community TRENDS in Plant Science

Figure 1. Non-metric multidimensional scaling (NMDS) analysis showing replicated genotypes of narrowleaf cottonwood (Populus angustifolia) vary in arthropod community composition and stability. Points represent the mean arthropod communities on replicate clones of an individual tree genotype and bars represent 95% confidence limits. These censuses were repeated each year for 3 years, so the mean community of each tree genotype is represented three times, once for each year (circles encompass the three years of each genotype). Because the within genotype variance is much less than the between genotype variance, individual genotypes differed significantly in arthropod community composition (Analysis of similarity (ANOSIM) R for community composition = 0.21, P < 0.0001). Furthermore, because the communities of some genotypes changed significantly less over the 3 years of study than other genotypes (see size of encompassing circles), there is also a significant genetic component to community stability (Restricted estimated maximum likelihood (REML) for community stability P < 0.0001). Replicate clones also showed significant broad-sense heritability of the community phenotype (species composition; H2C = 0.65). Images of the 11 most common arthropods from the study are shown clockwise from bottom left: Anthocoris antevolens, Chrysopa sp. 1, Harmonia axyridis, Boisea trivitatta, Tortricidae sp. 2, Listrus sp. 1, Ceratopogonidae sp. 1, Dasyhelia sp. 1, Thecabius populomonilus, Tortricidae sp. 1, Trombiculidae sp. 1, Gypona sp. 1, Pemphigus betae. We know that at least one member of the community, the bud-galling mite, Aceria parapopuli has evolved and adapted to individual tree genotypes indicating that specificity has both ecological and evolutionary implications. Photos courtesy of A. Keith, J. Dombroskie and K. Matz. Reproduced, with permission, from [4].

both in the context of the living organism and effects that endure after the organism has died; (ii) the effect of specificity on the evolution of dependent organisms and how genetically based species interactions among relatively few species can define a larger community; (iii) how the community can feed back to affect the genotypes expressing specific community and ecosystem phenotypes; and (iv) the key postulates that are necessary to demonstrate that specific genes are responsible for community and ecosystem phenotypes. Because community and ecosystem ecology have been largely genetics free, while molecular ecology has been largely community and ecosystem free, the demonstration of genetic specificity across these levels represents an important merger of disciplines. Community specificity to plant genotype Community genetics studies have revealed a diverse suite of plant-associated communities that exhibit specificity below the level of plant species. Initially, studies that focused on specificity of communities had a strong emphasis on arthropod herbivores, reflecting the field’s roots in plant–enemy interactions. However, this research has expanded to 272

include organisms such as fungal endophytes [8], mycorrhizal fungi [9], epiphytic and terrestrial plants [10,11], soil microbes [12] and terrestrial invertebrates [13]. Table 1 highlights the diversity of study systems where community specificity has been examined. Because studies based on whole communities within a genetics context are rare, to reflect more complex communities only studies with five or more species were included. Of the 75 communities shown in Table 1, 85% responded to genetic variation in focal plant species from 28 genera within 15 plant families, including angiosperms and gymnosperms. Aboveground arthropod and plant communities seemed particularly responsive, with 93.5% and 88.9%, respectively, showing a significant effect, whereas litter/soil invertebrates and microbial communities responded to plant genetics approximately 75% of the time. In addition to their phylogenetic diversity, organisms exhibiting community specificity represent a range of relationships with the focal plant, including mutualism, parasitism, commensalism, facilitation and competition. Evidence of community specificity comes from ecosystems around the world. Genetic variation in plants as varied as neotropical canopy trees, Tasmanian eucalypts,

Review


Table 1. Plant genera within which at least one focal plant species has been examined for community specificitya Focal plant taxa Growth Genetic scale form Asteraceae Shrub Intraspecific Artemisia c hybridization; population; subspecies Shrub Genetically based Baccharis phenotype Forb e Genotype Borrichia Subspecies Chrysothamnus Shrub

Biome and habitat

Study region

Study Communities responding method b

Refs

Temperate semi-arid shrubland

North America

CG; F

[73–75]

Temperate coastal dune Temperate coastal Temperate conifer forest Subspecies Temperate grassland and savannah Genotype Temperate coastal Genotype; ploidy; Temperate old-field f half-sibs

North America

CG; F

North America North America

CG F

Herbivores (stem, bud and leaf arthropods, deer); soil bacteriad; fungal endophytes (root, shoot and leaf) d Leaf, bud and stem arthropods; herbaceous and woody plants Leaf and stem arthropods Leaf, stem and bud arthropods

North America

CG

Leaf, stem and seed arthropods [79]


CG CG; F

Leaf and stem arthropods d Herbaceous plants; leaf, bud and stem arthropods; litter arthropods; arthropod pollinators

[77] [16,25,33, 80,81]

North America

CG

Leaf, flower and stem arthropods

[82]

Fungal leaf endophytes; leaf and stem herbivores (arthropods and small mammals)

[8,83]

Mycorrhizal fungi; soil bacteria; soil fungi Herbaceous plants; mycorrhizal fungi; soil bacteria; soil fungi; foliar and inflorescence herbivores (arthropods and mollusks)

[84]

Helianthus

Forb

Iva Solidago c

Shrub Forb

Apocynaceae Asclepias

Forb

Full-sibs

Temperate old-field

Tree; shrub

Genotype; half-sibs

Subarctic birch forest; Northern Europe CG; L boreal forest

Brassicaceae Alliaria

Forb

Brassica

Forb

Genetically based Temperate forest phenotype Genetically based Temperate grassland/ phenotype old-field; temperate coastal cliffs

Betulaceae Betula

Fabaceae Acacia c

North America

G

North America; Europe

CG; F; G

[17,76] [77] [78]

[14,85,86]

Tree; shrub

Population

Arid tropical savanna; Western Africa; temperate forest Australia

CG; G

Soil nematodes; foliar arthropods d

[87,88]

Tree

Population

South America

CG

Foliar arthropods

[89]

Tree; shrub

Genotype; half-sibs; genetically based phenotype

Temperate deciduous forest Temperate forest

North America; Japan; Europe

CG; F

Herbivores (leaf and stem arthropods, deer); litter microarthropodsd; soil bacteria d

[21,90–92]

Moraceae Brosimum

Tree

Genotype

Tropical rainforest

Central America

F

Litter and trunk arthropods, epiphytic plants

[10]

Myrtaceae Eucalyptus c

Tree

Sib-families; population

Temperate forest

Australia

CG

[13,15, 24,93]

Metrosideros c

Tree

Genetically based Tropical rainforest phenotype

Polynesia

CG

Trunk, litter and foliar arthropods; foliar pathogens; decomposer macro fungi d Foliar arthropods

Onagraceae Oenothera

Forb e

Full-sibs

Temperate meadow and old-field

North America

CG

Leaf and inflorescence arthropods; herbaceous plants d

[19,55,95]

Pinaceae Picea c

Tree

Genotype; genetically based phenotype

Boreal forest

Europe

CG

[9,96,97]

Pinus c

Tree

Population; genetically based phenotype

Temperate semi-arid woodland; temperate forest; boreal forest

North America; Europe

CG; F

Mycorrhizal fungi; soil microbes (bacteria and fungi); needle endophytic fungid; litter decay fungid; understory plants Mycorrhizal fungi; litter arthropods; soil bacteria; soil fungi; understory plants

Fagaceae Nothofagus c Quercus c

[94]

[12,31, 98,99]

273

Review


Table 1 (Continued ) Focal plant taxa Growth Genetic scale form Piperaceae Shrub Genotype Piper Poaceae Grass Population Ammophila

Biome and habitat

Study region

Study Communities responding method b

Refs

Tropical rainforest

Central America

CG

Leaf arthropods

[18]

Temperate coastal dune Temperate meadow Temperate coastal wetland

Europe

CG

[100]


CG CG

Root nematodes; foliar arthropods and mollusks Foliar arthropods Herbaceous plants

[101] [102]

North America

F

Herbaceous plants

[11]

[4,22,23, 26,35, 36,103]

[104]

Festuca Spartina c

Grass Grass

Genotype Genotype

Rosaceae Geum c

Forb

Genetically based Temperate alpine phenotype tundra

Salicaceae Populus c

Tree

Genotype; genetically based phenotype

Temperate forest; temperate riparian forest

North America, Europe

CG; F

Tree, shrub

Half-sibs

Temperate swamp

North America

CG

Leaf, bud and stem arthropods; root fungal endophytes; soil bacteria; soil fungi; mycorrhizal fungi; aquatic invertebratesd; aquatic fungi Leaf and stem arthropods

Solanaceae Datura

Forb

North America

CG

Foliar arthropods

[105]

Solanum

Forb

Genetically based Temperate semi-arid phenotype canyons and dry stream channels Genotype; Temperate old-field population

North America

CG

Herbivores (stem, flower, leaf and fruit arthropods, voles)

[20]

Salix c

a Studies focus on non-agricultural systems where the response of at least five community members were measured. Additionally, studies were chosen to represent the diversity of plant genera that have been examined. b

Study method codes: CG = common garden, F = field, L = laboratory, G = greenhouse or nursery (communities in soils only).

c

Taxa where status as a foundation species is established.

d

A community that has not shown a significant response in any study for a particular focal plant taxa.

e

Broad-leaved, non-woody flowering plants as distinguished from grasses and sedges.

f

Abandoned agricultural fields in various stages of recovery.

coastal dune shrubs, boreal conifers, alpine cushions and old-field (i.e. abandoned agricultural fields) forbs influences associated communities (Table 1). The scale of focal plant genetics varies among studies, including subspecies, populations, genotypes and sib-families. Although community specificity to each of these genetic levels has different ecological and evolutionary implications, these studies emphasize that genetic effects on diverse communities are common. In some systems, quantitative traits of the focal plant, including chemistry [14], bark characteristics [15], productivity [16] and architecture [17], are identified as potential mechanisms linking communities to plant genetics. Focal plants examined in this research are often foundation species (Table 1). Because of their strong influences on communities and ecosystems, genetic variation in the traits of foundation species is most likely to have a large impact on associated communities [6]. The extended effects of genes on community specificity also occur within plant species that do not define habitats. In a hyperdiverse tropical forest, the communities of trunk and litter arthropods, as well as epiphytic bromeliads and orchids, were more similar on genetically similar breadnut trees (Brosimum alicastrum) relative to genetically dissimilar individuals [10]. In addition, in a Costa Rican rainforest, communities of arthropod herbivores were found to 274

vary among genotypes of the understory shrub Piper arieianum [18]. Even non-foundation species such as evening primrose (Oenothera biennis) and horsenettle (Solanum carolinense) exhibit genetically variable effects on associated arthropod communities [19,20]. These examples illustrate two important points: intraspecific genetic variation in non-foundation plant species can exhibit community specificity, and these effects can occur in diverse systems, including species-rich rainforests. Specificity of afterlife effects Some of the longest lasting effects of plant genetic identity on communities and processes may occur after the death of either the whole plant or its components (e.g. litter, Table 1). All seven plant genera and families referenced in this section showed some afterlife response by biota or processes to subspecific genetic variation in plants. The afterlife effects of genetically based differences in leaf litter properties, in particular chemistry, are well documented [17,21–23], although the afterlife effects of bark and dead wood have also been examined [15,24]. Subspecific genetic variation in plants can influence soil and litter microbial biomass, microbial biomass nitrogen, microbial communities and litter arthropod communities [10,13,22,23,25,26]; however, soil and litter biota are not always sensitive to the

Review afterlife effects of fine-scale plant genetic variation [21]. These biotic responses may influence decomposition and nutrient dynamics, extending the effects of specificity to ecosystem processes [21,22]. The response of soil organisms and soil processes to genetically based differences in litter expands the temporal footprint of plant genetic identity beyond that of living tissue. Researchers have documented afterlife effects of litter lasting for days to years [21,22]. Genotypic variation in turkey oak (Quercus laevis) litter chemistry altered soil and litter nutrient dynamics during an 18–36 month sampling period [21]. We hypothesize that afterlife effects may last longer than have been reported to date, particularly if recalcitrant compounds such as lignin and polyphenols vary by genotype or population, as is often the case [21– 23], or if soil processes are influenced by repeated litter deposition by long-lived plants. Genetic specificity by plant enemies can also result in afterlife effects in litter and soil. Genetically based susceptibility to herbivores [3,27,28] results in damage to living leaves that translates to altered litter chemistry, which then influences litter and soil nutrient dynamics, soil microbial communities, microarthropod abundance and decomposition rates [29–32]. For example, the abundance of galls made by the rosette gall midge (Rhopalomyia solidaginis) on tall goldenrod (Solidago altissima) is dependent upon plant genotype and ploidy [27,33]. Leaf litter associated with galls contains higher initial carbon concentration, exhibits shortterm lower mass loss and retains more nitrogen than litter not associated with galls [32]. These studies provide evidence that life and afterlife effects of genetic specificity do not operate independently of each other. They can be linked by other organisms, and plant enemies may be particularly important in this context. Specificity of species interactions The community phenotypes described above result from interactions among species that can lead to greater specificity. It is important to distinguish between ecological [34] and genetic interactions [35] among species in which ecologists and geneticists define interactions differently for their purposes. In the ecological sense, two general types of indirect interactions have been described: interaction chains and interaction modifications [34]. These ‘indirect ecological interactions’ are distinct from ‘indirect genetic interactions’ in that they require more than two species, and do not consider the particular effect plant genotype may have on affected species. Yet indirect ecological interactions [34] can be linked to indirect genetic interactions [35] as follows. First, as an example of an interaction chain, plant genotypes can influence communities through another species [14]. With cottonwood (P. angustifolia), genetically based resistance and susceptibility to the galling aphid (Pemphigus betae) differentially shape the surrounding community and ecosystem [4,28,30,36]. Using experiments with naturally occurring tree genotypes, aphidsusceptible trees supported a different community of foliar arthropods, different rates of litter decomposition and more stable communities through time than resistant trees. Furthermore, the experimental removal of aphids from susceptible trees showed that the genotypic effects on


the community were mediated by the aphids [36]. This example shows that plant–enemy interactions, specific to particular plant genotypes, can shape a much larger community. In particular, it shows the importance that plant genotype, environment and selection in a community context can have on community organization. Second, in what has been termed interaction modification [34], plant genotype may mediate the interaction between two other species. For example, interactions between aphids and tending ants on common milkweed (Asclepias syriaca) ranged from antagonistic to mutualistic depending on plant genotype [37]. Thus, not only did the quantity of species change but the type of species interactions changed across plant genotypes, a result that has also been observed in other systems [19,28]. Here too, plant genotype, environment and selection in a community context significantly define community organization. In contrast to the ecological requirement that indirect effects are mediated by a third species, for geneticists, two classes of indirect genetic effects are identified. Indirect genetic effects (IGEs) refer to the influence that an individual genotype has on the expression of phenotype in other individuals within the same species, whereas interspecific indirect genetic effects (IIGEs) refer to the influence that an individual genotype in one species has on phenotypic expression among individuals in another species, with consequent effects on fitness [35]. Thus, in an evolutionary sense, IIGE theory shows that genetic variation in one species can influence the fitness and distribution of other species. Again, using the cottonwood (P. angustifolia), species interaction models showed that plant genotype can affect the fitness of other community members and that the accumulation of these fitness effects on multiple species can shape the unique communities that individual plant genotypes support [35]. This represents an important step towards incorporating genetics into community analyses and better understanding the genetic basis of ‘indirect effects’ in ecology. Although the extended consequences of these differing interactions on the rest of the community have not yet been shown in the wild, it is noteworthy that simply quantifying the abundances of species in a community is not sufficient to characterize the differing community dynamics among plant genotypes. Furthermore, most studies have focused on just a few species; considering the community in its entirety could reveal a larger and more realistic suite of genetically based interactions. For example, network theory may be a powerful tool to explore genetically based interactions in whole communities [38]. Network analysis revealed that communities were generally resilient to random elimination of species, but communities collapsed with the removal of well-connected species [39], suggesting that community stability is tied to the fate of a few, wellconnected foundation species. Specificity affects the evolution of dependent organisms The community context of the evolution between plants and their enemies can shape specificity as well as be a product of it. Specialization occurs in microbes, arthropods and vertebrates to plant populations and genotypes [3,40–42], 275

Review which can be influenced by the community context [43]. This community-driven evolution and specificity at the population level is evident in the interactions between lodgepole pine (Pinus contorta) and its seed predators. Red squirrels (Tamiasciurus hudsonicus) are dominant seed predators and independently drive the evolution of cone shape; however, in their absence crossbill (Loxia curvirostra complex) bill size and cone shape show evidence of a coevolutionary arms race [41]. Where squirrels are absent, a moth seed predator (Eucosma recissoriana) also influences cone evolution, underscoring the importance of a community context in specialization [44]. Selection on cones from squirrels influences serotiny (i.e. cones that release their seeds after fire) so extensively that it may shape fire dependency of seedling establishment and stand dynamics, resulting in community and ecosystem consequences [45]. Community context can also influence the evolution of dependent species in response to genetic variation at a finer scale than populations. The adaptive deme hypothesis [3] posits that individual host plants within a single population represent heterogeneous environments, which can lead to the formation of adapted demes of dependent herbivores and microbes. Reciprocal transfer experiments have demonstrated adaptation to individual plant genotypes across a broad range of arthropod and microbial taxa [3,46–48]. Where genetic interactions between species are affected by a third species [49], they indicate that community context drives genetic specificity [50]. When these interactions involve foundation species, for example, narrowleaf cottonwood (P. angustifolia) and the galling aphid (P. betae) [4,28], adaptive deme formation may influence the specificity of entire communities at the subspecies level. The community context in which differentiation of hostassociated herbivore lineages occurs may lead to differentiation of taxa outside the direct interaction between plants and their enemies. For example, parasitoids of two herbivores, each consisting of genetically divergent lineages on different Solidago species, have themselves differentiated in response to herbivore evolution, demonstrating evolutionary consequences at the community level driven by plant genetic variation [51]. Similar interactions are likely in response to herbivore specificity at finer levels of plant genetic variation, and a test of this would represent a major step forward in community genetics. Although most studies have examined species pairs when investigating the evolutionary consequences of plant genetic variation, it is clear that the community context of evolution is a frontier of community genetics research. Specificity affects community diversity and stability Specificity to genotypes can extend beyond community and ecosystem phenotypes to affect biodiversity and community stability. Often considered emergent properties, community diversity and stability can result from the sum of communities over all individual host genotypes (i.e. additive) or as a result of complex interactions of multiple communities and many genotypes that cannot be predicted by simple summation (i.e. non-additive) [52]. A clearer understanding of how genetically based traits may promote biodiversity has conservation value and could help prevent the loss of biodiversity. 276


Plant genetic diversity at both the individual and population levels can alter associated species diversity. Although some studies have found negative or no effects of plant genotypic diversity on associated community biodiversity [25,53,54], a greater number show that increased plant genetic diversity positively affects species richness and diversity (Table 1). With these patterns becoming more apparent, research is now focusing on potential mechanisms. For example, in tall goldenrod (S. altissima), increases in genotypic diversity led to greater aboveground net primary productivity, resulting in non-additive increases in arthropod community diversity [55]. A similar focus on the potential mechanisms by which genetic diversity can affect associated communities is necessary to predict when effects will be absent, additive or non-additive. Community specificity can also result in variation in community stability among genotypes and populations. A recent study showed that individual plant genotypes differed significantly in the stability of their associated arthropod communities across multiple years of study. They also showed that stability could be considered a heritable plant trait and that differences were likely to be the result of an indirect interaction with the galling aphid (P. betae), a foundation herbivore [4]. These results emphasize that although the composition of a large community can be structured by host genetic differences, there may be one or a few species that respond to those differences, which in turn define much of the remaining community. Genotypes of the biennial evening primrose (O. biennis) also varied in arthropod community structure, but the effects of particular genotypes changed across years, potentially due to large changes in phenotype [56]. Expanded to the stand or patch scale, variation among plant genotypes in diversity and stability could provide an additional mechanism linking diversity and stability. Furthermore, just as plant diversity has been shown to affect the stability of plant communities [57], it seems probable that genotypic diversity within a species will affect the stability of associated communities such as arthropods across multiple scales. Although studies are limited, it appears that plant genetic differences influence biodiversity and stability [4,56,58]. They suggest that complex community properties may be understood by considering the potential specificity of entire communities to host genetic variation. Because natural selection acting on individual tree genotypes can affect the associated community properties of diversity and stability, the diversity–stability hypothesis itself [59] may be genetically based and subject to natural selection. The conservation consequences of specificity are important because the choice of genotypes used in restoration could stabilize or destabilize a community as well as determine overall biodiversity. Genetic specificity and community feedbacks The community and ecosystem phenotypes that result from genetic specificity may feed back to affect the fitness of the individuals generating those phenotypes, resulting in an eco-evolutionary feedback, or ‘bidirectional interaction that unifies ecology and evolution’ [7,60]. For example, genotypes of P. angustifolia differentially affect soil microbial

Review


communities and the nitrogen transformations they mediate [26,30]. A reciprocal transplant experiment showed that P. angustifolia seedlings survived twice as well and were larger when grown in their local (maternal) soils, providing evidence of a positive feedback [61]. Similarly, research in plant–pollinator systems has demonstrated selection by the pollinator community on flowering phenology [62] and a gene associated with such changes [63], suggests that a mechanistic understanding of these relationships is attainable.

Community–genotype feedbacks can be positive or negative, with fundamentally different consequences for community specificity. If feedbacks are positive between individual plant genotypes and their associated community, and of similar strength among plant genotypes, both genotypic and community diversity will be maintained. However, if some genotype–community associations have higher fitness than others, the former will be selected for, ultimately reducing genotypic variation and species diversity. For example, P. angustifolia seedling genotypes varied more

[(Figure_2)TD$IG] Natural variation in nicotine

(b) Relative expression

Nicotine (µg/g)

(a)

400 300 200 100 0

Experimental gene silencing 1

0.5

0 WT

Accession

(c)

pmt expression

(d)

Herbivore community

Pollinator community

(i) 1

Nicotine

H N N

CH3

µg/mg leaf

0.5

Anatabine

1

0.5

H N N

0

H

0 WT

WT Silenced

1

Nectar volume (µl) after exposure to pollinators

(i)

µg/mg leaf

Silenced

Silenced

0.75 0.5 0.25 0 WT

20

All herbivores

15 10 5 0 WT

Silenced

Relative repellence when nicotine added to nectar

Leaf area damaged (%)

(ii)

Silenced

(ii) 0 −10 −20 −30 −40 −50 Ha

s

h

ird

ot

m wk

i

b ng

m

ts An

m

Hu


Figure 2. Genetic variation in the nicotine defenses of tobacco (Nicotiana attenuata) affects entire communities. (a) Different tobacco accessions exhibit natural variation in the production of nicotine. Photograph of N. attenuata in the wild (courtesy of USDA-NRCS PLANTS Database). (b) Silencing the putrescine N-methyl transferase ( pmt) genes responsible for this variation in the laboratory leads to altered production of the pyridine alkaloids, nicotine and anatabine (photo of silenced pmt plant courtesy of Anke Steppuhn). (c) Silenced pmt plants produce lower foliar nicotine levels (i), leading to higher levels of herbivory by a large suite of herbivores (ii). Herbivore examples include, clockwise from top in (c) (ii), Manduca sexta (Clemson University – USDA Cooperative Extension Slide Series, Bugwood.org), Timeroptropis sp. (courtesy of David J. Ferguson), Diabrotica undecimpunctata (Creative Commons Attribution License), Epitrix hirtipennis (Clemson University – USDA Cooperative Extension Slide Series, Bugwood.org), Spodoptera exigua (courtesy of Anke Steppuhn). (d) (i) Nicotine also alters pollinator preference, with pmt silenced plants preferred over wild type because of the lower level of nicotine in the nectar. (d) (ii) Nicotine is a deterrent for common pollinators of N. attenuata, such as the hawkmoth, hummingbirds and ants, as demonstrated by the experimental addition of nicotine to the nectar of pmt silenced plants. Data and images modified, with permission, from [66–68].

277

Review than twofold in the survival advantage they gained by associating with their local versus foreign soil community [61], setting the stage for selection against the less favorable genotype–community combinations. By contrast, negative feedback is unlikely to result in community specificity, but instead may result in strong selection against some genotypes. The cushion plant (Geum rossii) has two distinct architectural phenotypes. Common garden studies indicate that there is a genetic basis to these architectures and they are associated with distinct associated plant community phenotypes in the field. The more species-rich plant community associated with cushions with an open architecture significantly reduces their fitness, potentially leading to directional selection against open cushions and the plant community associated with them [11]. The interactions between genotypes and their associated communities vary with environmental context, leading to complex spatial and temporal dynamics. In G. rossii, open architecture cushions persist because they are favored in locations where environmental disturbance is high, showing the importance of spatial variation to feedback dynamics [11]. Studies of garlic mustard (Alliaria petiolata) across its introduced range provide an example of the importance of temporal variation. Genotypes of A. petiolata vary in their production of a family of allelochemicals, the glucosinolates. Glucosinolates alter the soil microbial community, particularly the arbuscular mycorrhizal fungi upon which many native plants depend for resource uptake [64]. Early in the invasion of a site, genotypes with high glucosinolate production are favored because they make A. petiolata more competitive against native plants. However, later in the invasion when A. petiolata interacts mostly with conspecifics, genotypes that invest less in the costly production of glucosinolate are favored [64]. Variation in the selection pressures associated with plant–plant interactions through time results in altered geographic patterns of genotype–microbial community feedbacks. These studies, along with similar research in other plant–soil systems [23] suggest that community and ecosystem feedbacks can influence our understanding of the evolution of specificity and its ecological impacts. Postulates of genetic specificity at higher levels Four postulates (analogous to Koch’s postulates for demonstrating the causal relationship between a microbe and a disease) have been proposed for testing the hypothesis that specific genes have community and ecosystem phenotypes [65]. These include: (i) the demonstration that a target organism affects other community members and/or ecosystem processes; (ii) the demonstration of key traits in the target organisms that are heritable; (iii) the demonstration of genotypic variation in these traits that result in different communities and/or ecosystem processes; and (iv) the identification of target gene(s) or their expression to evaluate a community and ecosystem effect experimentally. The fourth postulate could involve the use of quantitative trait loci mapping, genome-wide and fine-scale association mapping, knockouts, knockins, or other technologies tailored to the practical and ethical concerns of a given study. 278


Although evidence in support of two or more of these postulates has been found in numerous systems, only a few studies have tested all four. Using transformed native tobacco (Nicotiana attenuata) with silenced putrescine Nmethyl transferase ( pmt) genes, the production of nicotine defenses was reduced by 95% relative to the wild type (Figure 2) [66,67]. When planted in their native habitat and exposed to their natural community of herbivores, transformed plants suffered three times greater defoliation than wild-type plants. The silencing of nicotine genes also affected native insect and avian pollinators as well as nectar robbers [68], supporting all four postulates. Similar connections between plant genes and diverse community members have been observed in native populations of Arabidopsis thaliana [69]. With the increased use of transgenic plants on a global scale, the fourth postulate will receive widespread testing. For example, genetically modified aspen (Populus tremula P. tremuloides) expressing Bacillus thuringiensis (Bt) toxins that affect insect herbivores have unintended effects on the non-target aquatic insect community [70]. Such multidisciplinary research and experimental confirmation is likely to become common and should allow us to expand specificity studies to new levels. Concluding remarks Several major findings and future directions have emerged. (i) Genetic specificity at the community and ecosystem level has been demonstrated in diverse taxonomic groups, aquatic to terrestrial ecosystems, and from species-poor to species-rich systems (Table 1). Future work should assess how common this type of specificity is in different environments and investigate the cause(s) of significant differences. (ii) The effects of plant genetics can extend after the death of living plant tissue, but the breadth of the temporal and spatial footprints of these effects remains unknown. (iii) The genetically based interactions of relatively few species (e.g. foundation species) appear to drive the structure of a much larger community. For most systems, particularly species-rich systems, we have yet to characterize the highly interactive species, particularly those that might be ‘hidden’ players such as microbes. Network analysis at the population and genotype levels could provide a powerful tool for identifying key interactions. (iv) Specificity in one species for individual host genotypes or populations can affect the evolution of other species, which in turn can affect a larger community. Importantly, these genetically based interactions can change across the landscape to result in a geographic mosaic of evolution. The community context of evolution remains a frontier of ecology and evolutionary biology. (v) Because different genotypes support different communities, even emergent properties such as community stability are, in part, defined by genetic specificity. Understanding the links between biodiversity and genetic diversity is an important challenge. (vi) Because communities can differentially affect the performance of the individual genotypes they are associated with, feedbacks provide a major mechanism for evolution. In combination, it appears that genetic specificity is a fundamental feature of most ecosystems that has important ecological, evolu-

Review tionary and applied consequences. For example, if individual genotypes or populations differ in their response to climate change [71], such interactions could alter community structure, biodiversity, stability and specificity. Because of community specificity, losses of plant genetic diversity, even in common species, may cascade to affect whole communities. Acknowledgments Our work was supported by NSF DEB-0425908, NSF DEB-0816675 and Science Foundation Arizona. A National Science Foundation Integrative Graduate Education and Research Traineeship grant provided support for L.J.L., D.S.S., L.M.E. and A.R.K. We apologize for not being able to reference more of the research by our colleagues owing to space and citation limitations. We thank Steve Shuster for his comments on the manuscript.

References 1 Gilbert, G.S. and Webb, C.O. (2007) Phylogenetic signal in plant pathogen-host range. Proc. Natl. Acad. Sci. U.S.A. 104, 4979–4983 2 Poulin, R. et al. (2011) Host specificity in phylogenetic and geographic space. Trends Parasitol. 27, 355–361 3 Mopper, S. (1996) Adaptive genetic structure in phytophagous insect populations. Trends Ecol. Evol. 11, 235–238 4 Keith, A.R. et al. (2010) A genetic basis to community repeatability and stability. Ecology 11, 3398–3406 5 Bascompte, J. et al. (2006) Asymmetric coevolutionary networks facilitate biodiversity maintenance. Science 312, 431–433 6 Whitham, T.G. et al. (2006) A framework for community and ecosystem genetics: from genes to ecosystems. Nat. Rev. Genet. 7, 510–523 7 Haloin, J.R. and Strauss, S.Y. (2008) Interplay between ecological communities and evolution: a review of feedbacks from microevolutionary to macroevolutionary scales. Ann. N. Y. Acad. Sci. 1133, 87–125 8 Elamo, P. et al. (1999) Birch family and environmental conditions affect endophytic fungi in leaves. Oecologia 118, 151–156 9 Korkama, T. et al. (2006) Ectomycorrhizal community structure varies among Norway spruce (Picea abies) clones. New Phytol. 171, 815–824 10 Zytynska, S.E. et al. (2011) Genetic variation in a tropical tree species influences the associated epiphytic plant and invertebrate communities in a complex forest ecosystem. Philos. Trans. R. Soc. Lond. B: Biol. Sci. 366, 1329–1336 11 Michalet, R. et al. (2011) Phenotypic variation in nurse traits and community feedbacks define an alpine community. Ecol. Lett. 14, 433–443 12 Kuske, C.R. et al. (2003) The pinyon rhizosphere, plant stress, and herbivory affect the abundance of microbial decomposers in soils. Microb. Ecol. 45, 340–352 13 Barbour, R.C. et al. (2009) A footprint of tree-genetics on the biota of the forest floor. Oikos 118, 1917–1923 14 Lankau, R.A. et al. (2011) Plant-soil feedbacks contribute to an intransitive competitive network that promotes both genetic and species diversity. J. Ecol. 99, 176–185 15 Barbour, R.C. et al. (2009) Biodiversity consequences of genetic variation in bark characteristics within a foundation tree species. Conserv. Biol. 23, 1146–1155 16 Crutsinger, G.M. et al. (2008) Intraspecific diversity and dominant genotypes resist plant invasions. Ecol. Lett. 11, 16–23 17 Crutsinger, G.M. et al. (2010) Genetic variation within a dominant shrub species determines plant species colonization in a coastal dune ecosystem. Ecology 91, 1237–1243 18 Marquis, R.J. (1990) Genotypic variation in leaf damage in Piper arieianum (Piperaceae) by a multispecies assemblage of herbivores. Evolution 44, 104–120 19 Johnson, M.T.J. (2008) Bottom-up effects of plant genotype on aphids, ants and predators. Ecology 89, 145–154 20 Wise, M.J. (2007) Evolutionary ecology of resistance to herbivory: an investigation of potential genetic constraints in the multiple-herbivore community of Solanum carolinense. New Phytol. 175, 773–784


21 Madritch, M.D. and Hunter, M.D. (2005) Phenotypic variation in oak litter influences short- and long-term nutrient cycling through litter chemistry. Soil Biol. Biochem. 37, 319–327 22 LeRoy, C.J. et al. (2007) Within-species variation in foliar chemistry influences leaf-litter decomposition in a Utah river. J. North Am. Benthol. Soc. 26, 426–438 23 Madritch, M.D. and Lindroth, R.L. (2011) Soil microbial communities adapt to genetic variation in leaf litter inputs. Oikos 120, 1696–1704 24 Barbour, R.C. et al. (2009) Relative importance of tree genetics and microhabitat on macrofungal biodiversity on coarse woody debris. Oecologia 160, 335–342 25 Crutsinger, G.M. et al. (2008) Disparate effects of plant genotype diversity on foliage and litter arthropod communities. Oecologia 158, 65–75 26 Schweitzer, J.A. et al. (2008) Plant–soil-microorganism interactions: heritable relationship between plant genotype and associated soil microorganisms. Ecology 89, 773–781 27 Crawford, K.M. et al. (2007) Host–plant genotypic diversity mediates the distribution of an ecosystem engineer. Ecology 88, 2114–2120 28 Smith, S.D. et al. (2011) A geographic mosaic of trophic interactions and selection: trees, aphids and birds. J. Evol. Biol. 24, 422– 429 29 Chapman, S.K. et al. (2003) Insect herbivory increases litter quality and decomposition: an extension of the acceleration hypothesis. Ecology 84, 2867–2876 30 Schweitzer, J.A. et al. (2005) The interactions of plant genotype and herbivory decelerate leaf litter decomposition and alter nutrient dynamics. Oikos 110, 133–145 31 Classen, A.T. et al. (2006) Impacts of herbivorous insects on decomposer communities during the early stages of primary succession in a semi-arid woodland. Soil Biol. Biochem. 38, 972–982 32 Crutsinger, G.M. et al. (2008) Galling by Rhopalomyia solidaginis alters Solidago altissima architecture and litter nutrient dynamics in an old-field ecosystem. Plant Soil 303, 95–103 33 Halverson, K. et al. (2008) Differential attack on diploid, tetraploid, and hexaploid Solidago altissima L. by five insect gallmakers. Oecologia 154, 755–761 34 Wooton, J.T. (1994) The nature and consequences of indirect effects in ecological communities. Annu. Rev. Ecol. Syst. 25, 443–466 35 Shuster, S.M. et al. (2006) Community heritability measures the evolutionary consequences of indirect genetic effects on community structure. Evolution 60, 991–1003 36 Dickson, L.L. and Whitham, T.G. (1996) Genetically-based plant resistance traits affect arthropods, fungi, and birds. Oecologia 106, 400–406 37 Mooney, K.A. and Agrawal, A.A. (2008) Plant genotype shapes ant– aphid interactions: implications for community structure and indirect plant defense. Am. Nat. 171, E195–E205 38 Bascompte, J. (2009) Disentangling the web of life. Science 325, 416–419 39 Sole´, R.V. and Montoya, J.M. (2001) Complexity and fragility in ecological networks. Proc. R. Soc. Lond. B 268, 2039–2045 40 Rausher, M.D. (2001) Co-evolution and plant resistance to natural enemies. Nature 411, 857–864 41 Benkman, C. et al. (2003) Reciprocal selection causes a coevolutionary arms race between crossbills and lodgepole pine. Am. Nat. 162, 182–194 42 Gilbert, G.S. and Parker, I.M. (2010) Rapid evolution in a plant– pathogen interaction and the consequences for introduced host species. Evol. Appl. 3, 144–156 43 Craig, T.P. et al. (2007) Geographic variation in the evolution and coevolution of a tritrophic interaction. Evolution 61, 1137–1152 44 Siepielski, A.M. and Benkman, C.W. (2004) Interactions among moths, crossbills, squirrels, and lodgepole pine in a geographic selection mosaic. Evolution 58, 95–101 45 Benkman, C.W. and Siepielski, A.M. (2004) A keystone selective agent? Pine squirrels and the frequency of serotiny in lodgepole pine. Ecology 85, 2082–2087 46 Allen, R.L. et al. (2004) Host–parasite coevolutionary conflict between Arabidopsis and downy mildew. Science 306, 1957–1960 279

Review 47 Capelle, J. and Neema, C. (2005) Local adaptation and population structure at a micro-geographical scale of a fungal parasite on its host plant. J. Evol. Biol. 18, 1445–1454 48 Evans, L.M. et al. (2008) Tree hybridization and genotypic variation drive cryptic speciation of a specialist mite herbivore. Evolution 62, 3027–3040 49 Tetard-Jones, C. et al. (2007) Genotype-by-genotype interactions modified by a third species in a plant–insect system. Am. Nat. 170, 492–499 50 Neuhauser, C. et al. (2003) Community genetics: expanding the synthesis of ecology and genetics. Ecology 84, 545–558 51 Stireman, J.O., III et al. (2006) Cascading host-associated genetic differentiation in parasitoids of phytophagous insects. Proc. R. Soc. Lond. B 273, 523–530 52 Hughes, A.R. et al. (2008) Ecological consequences of genetic diversity. Ecol. Lett. 11, 609–623 53 Kanaga, M.K. et al. (2009) Plant genotypic diversity and environmental stress interact to negatively affect arthropod community diversity. Arthropod: Plant Interact. 3, 249–258 54 Johnson, D. et al. (2010) Plant genotypic diversity does not beget root– fungal species diversity. Plant Soil 336, 107–111 55 Crutsinger, G.M. et al. (2006) Plant genotypic diversity predicts community structure and governs an ecosystem process. Science 313, 966–968 56 Johnson, M.T.J. and Agrawal, A.A. (2007) Covariation and composition of arthropod species across plant genotypes of evening primrose (Oenothera biennis). Oikos 116, 941–956 57 Weigelt, A. et al. (2008) Does biodiversity increase spatial stability in plant community biomass? Ecol. Lett. 11, 338–347 58 Whitlock, R. et al. (2007) The role of genotypic diversity in determining grassland community structure under constant environmental conditions. J. Ecol. 95, 895–907 59 Elton, C.S. (1958) The Ecology of Invasions by Animals and Plants, Methuen 60 Post, D.M. and Palkovacs, E.P. (2009) Eco-evolutionary feedbacks in community and ecosystem ecology: interactions between the ecological theatre and the evolutionary play. Philos. Trans. R. Soc. Lond. B 364, 1629–1640 61 Smith, D.S. et al. (2012) Soil-mediated local adaptation alters seedling survival and performance. Plant Soil 352, 243–251 62 Sandring, S. and Agren, J. (2009) Pollinator-mediated selection on floral display and flowering time in the perennial herb Arabidopsis lyrata. Evolution 63, 1292–1300 63 Thornsberry, J.M. et al. (2001) Dwarf8 polymorphisms associate with variation in flowering time. Nat. Gen. 28, 286–289 64 Lankau, R.A. et al. (2009) Evolutionary limits ameliorate the negative impact of an invasive plant. Proc. Natl. Acad. Sci. U.S.A. 106, 15362– 15367 65 Wymore, A.S. et al. (2011) Genes to ecosystems: exploring the frontiers of ecology with one of the smallest biological units. New Phytol. 191, 19–36 66 Steppuhn, A. et al. (2004) Nicotine’s defensive function in nature. PLoS Biol. 2, 1074–1080 67 Wu, J. et al. (2008) A comparison of two Nicotiana attenuata accessions reveals large differences in signalling induced by oral secretions of the specialist herbivore Manduca sexta. Plant Phys. 146, 927–939 68 Kessler, D. and Baldwin, I.T. (2006) Making sense of nectar scents: the effects of nectar secondary metabolites on floral visitors of Nicotiana attenuata. Plant J. 49, 840–854 69 Todesco, M. et al. (2010) Natural allelic variation underlying a major fitness trade-off in Arabidopsis thaliana. Nature 465, 632–636 70 Axelsson, E.P. et al. (2011) Leaf litter from insect-resistant transgenic trees causes changes in aquatic insect community composition. J. Appl. Ecol. 48, 1472–1479 71 Sthultz, C.M. et al. (2009) Deadly combination of genes and drought: increased mortality of herbivore-resistant trees in a foundation species. Global Change Biol. 15, 1949–1961 72 Ellison, A.M. et al. (2005) Loss of foundation species: consequences for the structure and dynamics of forested ecosystems. Front. Ecol. Environ. 3, 479–486

280


73 Graham, J.H. et al. (1995) Narrow hybrid zone between 2 subspecies of big sagebrush (Artemisia tridentata, Asteraceae). 2. Selection gradients and hybrid fitness. Am. J. Bot. 82, 709–716 74 Messina, F.J. et al. (2002) Trade-off between plant growth and defense? A comparison of sagebrush populations. Oecologia 131, 43–51 75 Miglia, K.J. et al. (2007) Genotype, soil type, and locale effects on reciprocal transplant vigor, endophyte growth, and microbial functional diversity of a narrow sagebrush hybrid zone in Salt Creek, Canyon, Utah. Am. J. Bot. 94, 425–436 76 Rudgers, J.A. and Whitney, K.D. (2006) Interactions between insect herbivores and a plant architectural dimorphism. J. Ecol. 94, 1249–1260 77 Stiling, P. and Rossi, A.M. (1995) Coastal insect herbivore communities are affected more by local environmental conditions than by plant genotype. Ecol. Entomol. 20, 184–190 78 Floate, K.D. et al. (1996) Distinguishing intrapopulational categories of plants by their insect faunas: galls on rabbitbrush. Oecologia 105, 221–229 79 Whitney, K.D. et al. (2006) Adaptive introgression of herbivore resistance traits in the weedy sunflower Helianthus annuus. Am. Nat. 167, 794–807 80 Maddox, G.D. and Root, R.B. (1987) Resistance to 16 diverse species of herbivorous insects within a population of goldenrod, Solidago altissima: genetic variation and heritability. Oecologia 72, 8–14 81 Genung, M.A. et al. (2010) Non-additive effects of genotypic diversity increase floral abundance and abundance of floral visitors. PLoS ONE 5, e8711 82 Agrawal, A.A. (2005) Natural selection on common milkweed (Asclepias syriaca) by a community of specialized insect herbivores. Evol. Ecol. Res. 7, 651–667 83 Tikkanen, O.P. et al. (2003) No negative correlation between growth and resistance to multiple herbivory in a deciduous tree, Betula pendula. Forest Ecol. Manag. 177, 587–592 84 Lankau, R.A. (2011) Intraspecific variation in allelochemistry determines an invasive species’ impact on soil microbial communities. Oecologia 165, 453–463 85 Lankau, R.A. and Strauss, S.Y. (2007) Mutual feedbacks maintain both genetic and species diversity in a plant community. Science 317, 1561–1563 86 Newton, E.L. et al. (2009) Glucosinolate polymorphism in wild cabbage (Brassica oleracea) influences the structure of herbivore communities. Oecologia 160, 63–76 87 Duponnois, R. et al. (2002) Influence of the controlled dual ectomycorrhizal and rhizobal symbiosis on the growth of Acacia mangium provenances, the indigenous symbiotic microflora and the structure of plant parasitic nematode communities. Geoderma 109, 85–102 88 Andrew, N.R. and Hughes, L. (2007) Potential host colonization by insect herbivores in a warmer climate: a transplant experiment. Global Change Biol. 13, 1539–1549 89 Garibaldi, L.A. et al. (2011) Environmental and genetic control of insect abundance and herbivory along a forest elevational gradient. Oecologia 167, 117–129 90 Ito, M. and Ozaki, K. (2005) Response of a gall wasp community to genetic variation in the host plant Quercus crispula: a test using halfsib families. Acta Oecol. 27, 17–24 91 Cha, D.H. et al. (2010) Red oak responses to nitrogen addition depend on herbivory type, tree family, and site. Forest Ecol. Manag. 259, 1930–1937 92 Tack, A.J.M. and Roslin, T. (2011) The relative importance of host– plant genetic diversity in structuring the associated herbivore community. Ecology 92, 1594–1604 93 Barbour, R.C. et al. (2009) A geographic mosaic of genetic variation within a foundation tree species and its community-level consequences. Ecology 90, 1762–1772 94 Gruner, D.S. et al. (2005) The effects of foliar pubescence and nutrient enrichment on arthropod communities of Metrosideros polymorpha (Myrtaceae). Ecol. Entomol. 30, 428–443 95 Johnson, M.T.J. et al. (2008) Environmental variation has stronger effects than plant genotype on competition among plant species. J. Ecol. 96, 947–955

Review


96 Korkama, T. et al. (2007) Do same-aged but different height Norway spruce (Picea abies) clones affect soil microbial community? Soil Biol. Biochem. 39, 2420–2423 97 Korkama-Rajala, T. et al. (2008) Decomposition and fungi of needle litter from slow- and fast-growing Norway spruce (Picea abies) clones. Microb. Ecol. 56, 76–89 98 Pakeman, R.J. et al. (2006) The extended phenotype of Scots pine Pinus sylvestris structures the understorey assemblage. Ecography 29, 451–457 99 Leski, T. et al. (2010) Ectomycorrhizal community structure of different genotypes of Scots pine under forest nursery conditions. Mycorrhiza 20, 473–481 100 Vandegehuchte, M.L. et al. (2011) Contrasting covariation of aboveand belowground invertebrate species across plant genotypes. J. Anim. Ecol. 80, 148–158

101 Faeth, S.H. and Shochat, E. (2010) Inherited microbial symbionts increase herbivore abundances and alter arthropod diversity on a native grass. Ecology 91, 1329–1343 102 Proffitt, C.E. et al. (2005) Spartina alterniflora genotype influences facilitation and suppression of high marsh species colonizing an early successional salt marsh. J. Ecol. 93, 404–416 103 Karlinski, L. et al. (2010) Relationship between genotype and soil environment during colonization of poplar roots by mycorrhizal and endophytic fungi. Mycorrhiza 20, 315–324 104 Roche, B.M. and Fritz, R.S. (1997) Genetics of resistance of Salix sericea to a diverse community of herbivores. Evolution 51, 1490–1498 105 Hare, J.D. and Elle, E. (2002) Variable impact of diverse insect herbivores on dimorphic Datura wrightii. Ecology 83, 2711– 2720

Plant Science Conferences in 2012 23rd International Conference on Arabidopsis research (ICAR) 3 – 7 July, 2012 Vienna, Austria http://www.icar2012.org/ Plant Biology 2012 20 – 24 July, 2012 Austin, USA http://www.aspb.org/meetings EPSO meeting 29 July – 3 August Freiburg, Germany http://www.plant-biology-congress2012.de/freiburg.html -------------------------------------------------------------------------------------------------------------------Suggest a conference Please use the form at http://www.cell.com/conferences/SuggestConference to suggest a conference for Cell Press the Conference Calendar.

281

Review


Unifying concepts and mechanisms in the specificity of plant–enemy interactions Luke G. Barrett1 and Martin Heil2 1 2

CSIRO Plant Industry, GPO Box 1600, Canberra ACT, 2601, Australia Departamento de Ingenierıá Gene´tica, CINVESTAV, Irapuato, Me´xico

Host ranges are commonly quantified to classify herbivores and plant pathogens as either generalists or specialists. Here, we summarize patterns and mechanisms in the interactions of plants with these enemies along different axes of specificity. We highlight the many dimensions within which plant enemies can specify and consider the underlying ecological, evolutionary and molecular mechanisms. Host resistance traits and enemy effectors emerge as central players determining host utilization and thus host range. Finally, we review approaches to studying the causes and consequences of variation in the specificity of plant–enemy interactions. Knowledge of the molecular mechanisms that determine host range is required to understand host shifts, and evolutionary transitions among specialist and generalist strategies, and to predict potential host ranges of pathogens and herbivores. The importance of specificity in plant–enemy interactions Herbivores and plant pathogens make use of only a subset of the plant species and organs to which they are exposed. Such specialization is ubiquitous in plant–enemy interactions (see Glossary) and can have important consequences for their ecological and evolutionary dynamics. In a broad sense, specialization to the many different niches represented by plant communities has facilitated the evolution of the enormous diversity of herbivorous animals and microbial pathogens [1,2]. In turn, the specialization of plant enemies can influence rates of encounter with hosts [3] and with competitors or members of the third trophic level [4–6], and the local coexistence of plant species [3,7,8]. Specialization is also important from a broad array of applied perspectives. In particular, questions concerning the potential host range of plant enemies become crucial in a world in which both plants and their enemies have highly increased mobility, mainly because of human activities [9]. Despite the general importance of specificity in plant– enemy interactions, clearly defining the term ‘specialization’ is surprisingly challenging, and understanding the causes and consequences of specialization on ecological or evolutionary timescales remains an even more difficult Corresponding author: Heil, M. ([email protected]).

282

task. In part, this is because specialization can evolve along multiple axes, often simultaneously [10]. Specialization is usually considered as the process of adaptation to a limited spectrum of potential resources, although evidence is accumulating that the adaptations required by generalists might be as complex as those required by specialists [3]. Research into the molecular mechanisms of host utilization by plant enemies and the ecology and evolution of specificity has progressed mostly independently. Likewise, plant–herbivore and plant–pathogen interactions have only rarely been subject to general synthesis [11]. This is despite many commonalities: herbivores and pathogens often exploit the same plant species or plant organs, must overcome the same defence mechanisms, have similar effects on plant fitness and share clear demographic similarities. In this review, we identify concepts and mechanisms of general importance to the evolution of specificity in interactions between plants and their enemies. We first Glossary Effector: a molecule secreted by a plant enemy to manipulate host resistance. Effectors are commonly polymorphic among strains of the same species of pathogen or herbivore. Effector-triggered immunity (ETI): a plant resistance response that is activated upon recognition of enemy effectors by NB-LRRs. Enemy (plant enemy): used here to denominate herbivores and plant pathogens; that is, animals and microorganisms that form the second trophic level. Host range (potential): the host species or organs that could be used by an enemy in the absence of all other (usually geographical, behavioural or temporal) barriers. Host range (realized): the current host range of a plant enemy. Microbe-associated molecular pattern (MAMP): synonym to PAMP. Nucleotide-binding and leucine-rich repeat protein (NB-LRR): plant resistance proteins that act as receptors for effector molecules. NB-LRRs are often polymorphic among races or populations of plants. Pathogen-associated molecular pattern (PAMP): phylogenetically conserved molecular motifs, such as chitin and flagellin, that are recognized by plants as indicators of attacking pathogens. Pattern recognition receptor (PRR): proteins serving the perception of PAMPs, usually conserved within a host species or larger taxonomic group. Specificity (geographic): differences in host ranges among populations of an enemy. Specificity (ontogenetic): specificity in host use among different developmental stages of the enemy. Specificity (phylogenetic): specificity concerning the phylogenetic distances among host species. Specificity (structural): specificity concerning different structures or organs of the host.


Review


highlight the many axes along which specificity can evolve and discuss the fitness benefits of different strategies. We then review the common ecological, evolutionary and molecular mechanisms that determine patterns of host specificity. Only a mechanistic understanding of the determinants of host range will identify the reasons behind host shifts, and transitions between specialist and generalist strategies, and enable researchers to predict the potential host ranges of geographically isolated enemies. Effectors emerge as a common molecular concept that determines the host spectra of pathogens and herbivores [12–15]. Specialist pathogens and many specialist herbivores have highly specific effectors that facilitate the exploitation of specific hosts, but which in turn are often recognized as ‘avirulence’ (avr) genes by resistant hosts [14,16]. By contrast, generalist enemies commonly have multiple or promiscuous effectors or digestive enzymes that successfully suppress or overcome resistance responses in many different hosts [14,17–22]. Thus, host resistance traits and enemy strategies to overcome these traits are central players in defining host range. Based on these observations, we question the general hypothesis that specialists are more adapted than generalists and suggest that generalists are better understood as ‘multi-host specialists’. We finish with concrete suggestions as to how next-generation sequencing techniques can be used to investigate natural host ranges of herbivores and plant pathogens and to understand the molecular mechanisms that explain why certain plant enemies utilize specific organs of specific hosts.

The multifaceted nature of specificity The specificity of the interactions between plants and their enemies can range from tightly coupled associations among species pairs, through to diffuse relationships among diverse communities of prospective partners (for recent reviews, see [3]). However, the number of host species that an enemy can exploit (Figure 1) is only one aspect of its specificity. Specificity can manifest in different ways, and simple similarities in the overall number of host species attacked can mask fundamental differences in the biology of the organisms involved (Box 1). Potential and realized host ranges The host species utilized by a plant enemy in nature (the realized host range) does not necessarily reflect the species that it could attack in principle (its potential host range). A modern history of repeated invasions by plant enemies attests to the importance of geographical barriers in limiting realized host ranges [9]. As a consequence of geographical and behavioral limitations, plant enemies in nature seldom utilize all potential hosts. For example, the recent arrival of a single genotype of the endemic Brazilian rust pathogen Puccinia psidii in Australia (to which a wide range of species in the family Myrtaceae are potential hosts) has added more than 100 species to the realized host range of this pathogen (http://www.outbreak.gov.au/ pests_diseases/pests_diseases_plant/myrtle-rust/national_ host_list.html). However, despite being a growing problem worldwide, researchers currently lack the ability to predict potential host spectra accurately. Pathogens in particular

[(Figure_1)TD$IG] Agonopterix alstroemeriana (1)

1

2

3

Phascolarctos Myzus persicae cinereus (2–3) (several hundred)

4

10

Spodoptera littoralis (>500)

100

Bemisia tabaci (>500)

500

Capra aegagrus hircus (>1000)

1000

Number of host species

Cephaloleia placida (1)

Tetraopes tetraophthalmus (1)

Cephaloleia belti (11)

Popilia japonica (>300) TRENDS in Plant Science

Figure 1. Host species number as the conventional concept of specialization. The number of host species (in parentheses after species names) that can be utilized by a plant enemy rank from a single one to over 1000. Neither feeding mode nor size or taxonomic position appear to be good predictors of the position of a plant enemy on this first axis of specialization. Images reproduced, with permission, from: Eric M. Coombs (Agonopterix alstroemeriana); W. Billen (Bemisia tabaci); Biologische Bundesanstalt fu¨r Land- und Forstwirtschaft (Spodoptera littoralis); S. Bauer (Myzus persicae); and David Cappaert (Popilia Japonica) (all published at http://www.bugwood.org under Creative Commons Attribution 3.0 License); Edith Freitag and Matthias Goeke (Capra aegagrus hircus; http://www.tiere-der-heimat.de); Marc E. Ellis at H2O pictures (http:// www.h2opictures.com) (Phascolarctos cinereus); [62] and Carlos Garcıá-Robledo (Cephaloleia beetles); and Anurag A. Agrawal (Tetraopes tetraophthalmus).

283

Review are being increasingly discovered in association with novel host species, and often cause either unfamiliar or no obvious symptoms [23]. Humanopathogenic bacteria, such as Salmonella, Escherichia coli or Klebsiella pneumoniae, can also develop in plants [17,24,25], and endophytic fungi that were isolated from surface-sterilized, symptom-free leaves of diverse hosts commonly comprise multiple strains of the plant-pathogenic genera Alternaria [26–28], Colletotrichum [26,28,29] and Fusarium [26,28,30]. Are such endophytes non-pathogenic relatives of common pathogens, or are they pathogenic only under certain conditions, and what are the consequences of these alternative life stages for disease establishment and spread? In the ‘Perspectives’ section, we discuss how next-generation sequencing can be applied to investigate realized host ranges of herbivores and pathogens and how research into the molecular mechanisms used by plant enemies to overcome the resistance of their hosts will help to understand and reliably predict potential host ranges. Axes of specificity As stated by Daniel H. Janzen [31], plant enemies do not simply ‘eat latin binomials’. Rather, they are adapted to exploit selected parts of selected organs of selected plants, and the evolutionary relationships among host species commonly affect the probability that a given plant species can be attacked by a particular enemy species. This poses the question ‘to what is the specialist specialized?’ Specialization may manifest along various axes (Box 1). First, it can vary throughout the development of the enemy or host. Larval and adult stages of many insect herbivores [32,33] and different pathogenic spore stages [34] often have only partially overlapping or even completely


separate host ranges (‘ontogenetic specificity’), and most plant enemies can utilize only defined host developmental stages or organs (‘structural specificity’). If the traits that make a host suitable for a particular enemy tend to be distributed among phylogenetically related hosts, then the ‘phylogenetic specificity’ of the enemy is high. In fact, the capacity of most herbivores and pathogens to exploit multiple hosts decreases with the phylogenetic distance among host species [14,35–39]. By contrast, few species are ‘true generalists’ that are capable of exploiting numerous completely unrelated host taxa. Examples include Phytophthora cinnamomi, which attacks more than 1000 plant species in numerous families, including Myrtaceae, Coniferales and Fagaceae [40]; the Japanese beetle, Popillia japonica, whose adults feed on the foliage, fruits and flowers of over 300 species of plants from at least 79 plant families (http://pubs.ext.vt.edu/2909/2909-1411/2909_ 1411_pdf.pdf), and classic ‘model’ generalists, such as whitefly (Bemisia tabaci) and Spodoptera littoralis. Finally, host ranges might differ according to specific environments or habitats (‘geographic specificity’) [41]. Importantly, specialization can vary almost independently along all the axes that we describe here. Thus, simply counting susceptible species limits one’s ability to understand the processes that drive the evolution of specialization in plant–enemy interactions. Generalist species as conglomerates of specialized genotypes The observation that different populations of plant enemies can attack different host species indicates the potential importance of within-species genetic structure to an understanding of the evolution of host range [10,41]. In

Box 1. How can one define a specialist? The conceptual part of the problem We identify an urgent need for obtaining standardized (and, hence, comparable) methods for the quantification of host ranges along different axes of specialization. Ideally, one would quantify how well every developmental stage of the enemy does on any of its potential host organs, or species. Useful data in this context would be growth rates, population densities of pathogens, time required to conclude certain developmental stages or, ultimately, fitness. As suggested in [36], specificity then can be quantified using classic diversity indices (e.g. Simpson and Shannon-Weaver indices) along at least four axes as ontogenetic, structural, phylogenetic and geographic specificity (Figure I). Their quantitative comparison is important when considering ecological and evolutionary consequences of specialization and will provide the basis for the understanding of underlying mechanisms. For the enemy, ontogenetic specificity denominates specificity at different developmental stages (Figure Ia) such as, for example, beetle larvae and adults that perform differently on the same plant species [32]. Structural specificity means the degree of specialization on a specific host organ, or developmental stage of the host (Figure Ib). For example, Peronospora downy mildews, which are solely restricted to infecting flowers [76], have a higher structural specificity than does the Japanese beetle, Popillia japonica, whose adults feed on foliage, fruits and flowers (http://pubs.ext.vt.edu/2909/2909-1411/ 2909_1411_pdf.pdf). Phylogenetic specificity (Figure Ic) expresses the general tendency of herbivores and plant pathogens to restrict their host range to related species and can be high even for enemies that utilize multiple host species. Examples are Puccinia psidii, which attacks a large number of species that all belong to the family Myrtaceae [112] and the beetle Cephaloleia belti, which attacks 15 host species but only from the Zingiberales [62]. By contrast, the 284

above-mentioned Peronospora downy mildews have been isolated from diverse plants belonging to the Orobanchaceae, Lamiaceae, Asteraceae and Dipsacaceae [76] and so showed higher structural than phylogenetic specialization. Finally, geographic specificity quantifies the differences among host uses of geographically disparate populations of a plant enemy (Figure Id) [41] and will be particularly strong for apparent generalist species that in fact represent groups of locally adapted cryptic specialists. Importantly, the consequences of these different levels of specificity for the realized and potential host range and geographic range of a species are very different, as can easily be illustrated in terms of set theory (Figure Ie). The host range of an individual comprises the host ranges of all of its ontogenetic stages, meaning that its niche represents the intersection of the niches of its life stages (Figure Iei). At least one host for every ontogenetic stage must be present at the same site and in the correct temporal order to allow an individual to express positive fitness. As stated in [32], ‘the breadth of environments in which a species can succeed is ultimately determined by the full pattern of its vital rates in each environment’. By contrast, the host range of a species is the sum of the host ranges of all of its geographic populations or genetic races (i.e. Figure Ieii). These aspects are important for invasion biology, for example, where an enemy that utilizes different hosts during its different developmental stages can invade only a region in which all hosts are present. By contrast, all individual populations or genetic lines of an enemy represent the source of potentially invasive founder individuals; therefore, species with very different host ranges in their different habitats, or genetic lines, have a higher level of releasing invasive progeny.

Review


[(Figure_I)TD$G]

Ontogenetic specificity

(a)

(b)

Structural specificity High

High

Low

Flower Young leaf Old leaf

Low

Fruit Shoot Root Host species

Ontogenetic stage 1 Ontogenetic stage 2

0 0 Enemy performance

Host organs

Phylogenetic specificity

(c)

High

(d) Low

Geographic specificity High

Low

0 0 Enemy performance (e)

(i)

Population 2

Population 3

Host species

Host range of individual

Host range of stage 1 (ii)

Population 1

Host range of stage 2

Host range of species

Host range of population 1

Host range of population 2 TRENDS in Plant Science

Figure I. Levels of specialization and their consequence for host ranges. Specialization can occur as (a) ontogenetic specificity, which describes specificity at different enemy developmental stages, (b) structural specificity towards certain host organs or developmental stages of the host, (c) phylogenetic specificity, and (d) geographic specificity, which describes different degrees of specialization among populations of the same species. Set theory (e) illustrates that the host range of an individual is formed by the intersection of the host ranges of all of its ontogenetic stages, whereas the overall host range of a species is represented by the union of the host ranges of all its populations.

particular, enemies recorded on a wide number of hosts (e.g. P. cinnamomi or P. japonica) should not be classified a priori as ‘mega-generalists’, because such species may exist as complexes of genetically discrete, host-specialized lineages. For example, a recent study reported a strong assortment of specific genotypes of the fungus Beauveria bassiana with specific hosts and other environmental conditions [42], and the generalist herbivore B. tabaci is also likely to represent a species complex [43]. In fact, many species with apparently wide host ranges exist as (often cryptic) complexes of closely related subspecies, host-associated lineages, locally adapted populations and individual specialists, all of which individually utilize a narrower host spectrum than does the species as a whole [41,44–47]. Such species can be termed ‘generalists’ only when considered as an entire, taxonomically defined unit [41] and might be

better described and understood as a complex of functionally disparate, but closely related, specialists [41,48]. This intraspecific variability in host plant utilization is likely to be critical to understanding the emergence of true specialists, because specialization, similar to all adaptive processes, requires genetic variation within populations upon which evolution can act (see the section below on biotic heterogeneity and the emergence of specialists and the ‘jack of all trades – master of none’ principle). Given such complexity, it is clear that categorically assigning enemies to generalist or specialist strategies is problematic. Rather, these terms need to be considered as end points along various continua of specificity. However, we argue that it remains important to differentiate among strategies, at least in a relative sense, because there are numerous ecological and evolutionary consequences of 285

Review


specialist versus generalist life-histories. Although the terms are undoubtedly relative, evolution and ecology also act at relative scales (e.g. the same absolute number of offspring means a higher fitness in one environment, and lower fitness in a second environment). Therefore, distinguishing among strategies enhances one’s ability to understand important processes and interpret patterns of specialization. Ecological and evolutionary patterns The ecological and evolutionary mechanisms that determine variation in host range mirror more general processes that drive the evolution of ecological specialization and the maintenance of biological diversity [10]. Here, we review the most widely cited hypotheses dealing with the phenomenon of specialization and how recent phylogenetic studies have challenged their universal applicability. Adaptive radiation, ecological fitting and enemy-free space The concept that is perhaps most widely cited to explain how the interactions of plants with their enemies enhance specialization is adaptive radiation, sensu Ehrlich and Raven [49]. Under this model, plants evolve new resistance traits in response to selective pressure exerted by herbivores; following herbivores then evolve specific counter-adaptations that enable them to overcome the new resistance, and so on. The same ‘zigzag’ model of coevolution of plant resistance with enemy counter-adaptations is also likely to be an important driver of the multiple layers of inducible resistance traits that plants exhibit against pathogens [14,16] and so potentially represents a major driving force in the process of adaptive radiation and ecological specialization. Thus, a common prediction is that coevolution will promote specialization via optimization of performance on a restricted subset of hosts. This outcome implies the existence of trade-offs between the capacity to attack a host and another component of fitness [50]. A necessary prerequisite of any host shift is that the plant enemy initially has the equipment to experience positive net fitness on the new host, before evolving specific adaptations that facilitate its utilization. The concept of ecological fitting was originally formulated by Daniel H. Janzen [51] to scrutinize the fact that observing a functioning plant–enemy interaction in nature does not necessarily indicate any coevolutionary history [52,53]. In fact, each of the thousands of invasive herbivores and plant pathogens represents independent empirical support of

the importance of ecological fitting: all these species have initially experienced positive net fitness on new hosts that were probably never encountered before throughout their evolutionary history. At the ecological level, ecological fitting represents a ‘black box’ that is unlikely to be mechanistically understood, or let alone reliably predicted. However, recent studies have revealed molecular mechanisms that are likely to causally underlie ecological fitting and host shifts (see the section below on effectors and the determination of host specificity and host shifts). Trade-offs Low specificity increases the number of individual hosts available, probably enhances the geographic range that the enemy can occupy, reduces the time required to search for new hosts and lowers the risk of extinction should any one host be unavailable (Table 1). However, all plant enemies show some level of specialization and most are characterized by high phylogenetic conservatism [14,35–39]. The existence of trade-offs is perhaps the most frequently advanced hypothesis regarding the evolution towards specialization in plant–enemy interactions [6]. For example, adaptive radiation requires that a plant enemy that is well adapted to a specific new host will perform suboptimally on the ancestral species, and the concept of enemy-free space assumes different performances on the original versus the new host species [4,49]. In other words, genotypes that perform well on one host should perform relatively poorly on alternate hosts, and specialists should outperform generalists on any given host species (i.e. the ‘jack of all trades – master of none’ principle) [54]. Perhaps the clearest empirical support for these predictions comes from studies of microbial pathogens. For example, a trade-off has been demonstrated between host range and the mean number of infective spores produced by the pathogen Melampsora lini, such that strains infecting a wider range of hosts were generally less fecund [55]. Studies of plant viral pathogens [56,57] further provide evidence that trade-offs can be important for the maintenance of different specialist pathogen lineages, such that experimentally passaged viral populations that evolved increased capacity to exploit novel hosts suffered negative effects on the original hosts. Another study found that less virulent strains of Pseudomonas syringae had a higher probability of survival in non-host conditions than did more virulent strains [58]. For black bean aphids (Aphis fabae), studies comparing different clones reported tradeoffs in lifetime fitness between two different host plants

Table 1. Advantages and disadvantages of specialist versus generalist strategies Advantages

Disadvantages

286

Specialist Higher optimal performance Reduced interspecific competition Better able to respond to changes in host resistance Reliance on fewer hosts increases potential for heterogeneity in terms of resource availability Increased intraspecific competition Reduced capacity to establish in new environments and exploit novel hosts (niche contraction) Increased chance of evolutionary constraint?

Generalist Diet mixing can improve development (herbivores) Decreased resource heterogeneity Increased capacity to establish in new environments and exploit novel hosts (niche expansion) Lower optimal performance (jack of all trades – master of none) Metabolically costly Promiscuous enzymes (e.g. for detoxification) are less efficient

Review [59]. Similarly, fitness of pea aphid (Acyrthosiphon pisum) races was negatively correlated among different hosts and, consequently, it has been argued that antagonistic pleiotropy is likely to be generally important [60]. However, strains of the bacterial pathogen Salmonella typhimurium that were passaged though multiple generations in plant hosts did not alter their virulence for animals cells [24], suggesting that pathogens can evolve to exploit different hosts without measurable reductions in their performance. Specialization originating via trade-offs requires that the fitness of genotypes within populations be negatively correlated among different hosts. However, trade-offs at the interspecific or interpopulation level might be more likely to represent the consequence of adaptation to specific hosts rather than its cause [6,54]. Thus, the generality of tradeoffs as mechanisms that limit the emergence of generalist enemies is the subject of much ongoing debate [10]. Indeed, many studies seeking empirical support for trade-offs within populations report results that conflict with expectations. For example, one study searched for performance trade-offs in a population of the moth Rothschildia lebeau, whose larvae feed on several host species [61], whereas another study compared the performance of various Cephaloeia leaf beetles on native and invasive Zingiberaceae [54]. Both studies found mainly positive rather than negative genetic correlations in cross-host performance, meaning that genotypes that performed particularly well on one host also performed better on the alternative host [54,61]. Surprisingly, this phenomenon applied to both generalist and specialist beetle species [54]. How can these contrasting results be explained? First, the above-cited studies might indicate that trade-offs are more relevant for enemies that are tightly associated with their host plant, such as pathogens and sap suckers. However, the available data are not sufficient to decide whether these contrasting results really indicate a general and biologically relevant difference among guilds of plant enemies. Second, it is possible that trade-offs might only be important under certain conditions and so are difficult to detect. For example, many plant–herbivore studies that find no support for trade-offs have commonly used comparisons among genotypes within populations [54]. Third, many studies have used generalists with an arguably narrow host spectrum. For example, the generalist beetle used in [62] feeds naturally on 15 host species from five families; however, all the species fall within the order Zingiberales. Fourth, because ecologically realistic studies depend on already established plant–enemy interactions, they are focused on enemies that exhibit ecological fitting towards the new host. Finally, detecting trade-offs requires picking the appropriate measure of performance, measured under appropriate ecological conditions; many experiments might fail to meet these criteria. Further mechanisms independent of, or reinforcing, trade-offs Although it might depend on the detailed experimental design whether the capacity to perform on one specific host can be demonstrated to be negatively correlated with the capacity to utilize another one, the empirical evidence is overwhelming: the ‘jack of all trades – master of none’


principle does not apply in all situations and a negative correlation among the performances of a given enemy genotype on different hosts is not sufficient as a sole explanation for the evolution of specificity. Indeed, trade-offs are by no means the only evolutionary mechanism proposed to influence variation in host range. In particular, demographic events and population-level processes, such as bottlenecks and assortative mating, might reinforce, and perhaps even drive, the evolution of specialization [10,63]. For example, populations of plant enemies that utilize different host plants have a reduced probability of encounter and mating even before any genetic isolation mechanisms can act, and this isolation can become total if hosts are geographically separated or exhibit non-overlapping phenologies [64]. Enemies that utilize geographically separated discrete host populations can secondarily specialize via genetic drift and assortative mating, evolve genetic differentiation and, consequently, divergent patterns of specificity towards these host populations [3,65]. The accumulation of deleterious mutations that degrade performance on alternate hosts might further reinforce the effects of drift and assortative mating to the point that specialization may evolve even in the absence of other selective processes [63]. Therefore, major patterns in the specialization of plant enemies can be explained by host–enemy coevolution. However, host shifts are common and can be favored if the new host represents an enemy-free space [4–6]. Insect herbivores might shift to nutritionally suboptimal host species when enemy encounters are less likely to occur on the new hosts. For example, caterpillars of the swallowtail butterfly (Papilio machaon aliaska) were found on new host plants that allowed for lower survival rates than the original hosts, more commonly in habitats with lower antmediated lethality [66]. Such observations indicate the importance of the third trophic level in host choice. Herbivores might also shift onto new hosts on which the encounter rates with competitors, rather than predators, are reduced [6]. These mechanisms also apply to microbial pathogens, whose performance often is impaired by plant endophytic fungi, which outcompete or directly attack the pathogens [23]. Thus, in principle, pathogens might also search for enemy-free space and specialize on new hosts when these contain lower competitor, parasite or predator loads. Biotic heterogeneity and the emergence of specialists As highlighted above, host range can be viewed as the result of a trade-off in the ability to exploit individual hosts optimally and the ability to utilize the maximum number of hosts encountered [10,50]. As soon as plant enemies exhibit at least some intraspecific variability of the underlying traits, selection can act upon the different genotypes and favor an evolution towards more specialized or more generalist species, depending on the detailed selective pressures. In the classic ‘arms-race’ model of adaptive variation, this selection should normally cause the evolution of a more generalist ancestor towards a group of closely related specialists. However, heterogeneities in selective pressure experienced in complex plant communities might make it difficult 287

Review for generalist enemies to counteradapt to the emergence of new host resistance traits. Costs apply when enemies must search for suitable hosts or express specific effectors or digestive enzymes to invade or utilize changing host plants. Therefore, it might be expected that generalist strategies should evolve when suitable hosts are infrequent or ephemeral, whereas specialist strategies should be favored when susceptible hosts are abundant and predictable [67]. It has been argued that interactions between generalist pathogens and rare or ephemeral hosts should favor the host in any evolutionary arms race, because although common generalist enemies might be important to the population biology of rare hosts, the reverse is unlikely to be true [47]. In addition, the demography of enemy populations is likely to be tightly linked to variation in host community structure. In particular, enemies with a large population size have an increased probability of colonizing new hosts and areas, and of generating new mutations [68]. Thus, whereas extreme variability in host availability might favor generalist enemies, the emergence of such strategies may be more probable at intermediate levels of resource heterogeneity [69]. Although decreased population size is likely to be associated with a concomitant decrease in the supply of mutants from which capacity to attack new hosts might emerge, this should be countered by an increased selective advantage to those enemies capable of infecting novel hosts. Thus, the probability that a generalist enemy will evolve and then be maintained may be highest at some intermediate level of host community complexity. Phylogenetic patterns In summary, classic theories predict a general tendency to evolve towards a higher degree of specialization, but certain conditions might also favor a widening of host range. In fact, both scenarios have been reported in phylogenetic studies. Given the general propensity for host range conservatism (see above), many hypotheses advanced to explain the remarkable levels of specificity in interactions between plants and their enemies have been based on the assumption of tightly coupled, pairwise coevolution and subsequent co-speciation [70]. Furthermore, specialization is often predicted to be an evolutionary ‘dead-end’ because, due to the costly accumulation of host-specific adaptations, specialized enemies should have increasingly lower fitness on hosts to which they are not specialized (see [38] for a recent empirical example). In Box 2, we review evidence for such phylogenetic constraints on the evolution of different strategies. By contrast, phylogenetic studies using ancestral state mapping increasingly reveal evidence for host switching [71–77] and examples of generalists that have evolved from a specialist ancestor [78,79]. These observations make it increasingly apparent that life-history evolution in species interactions can be highly dynamic. Molecular mechanisms determine specificity in plant– enemy interactions As we have highlighted above, specialization along one or more axes is inherent to all plant enemies, and trade-offs are one of the key evolutionary mechanisms that are likely to underlie the maintenance of specialized strategies. A specialist that very efficiently utilizes one host is commonly 288


Box 2. Phylogenetic constraints and the evolution of specialization If specialization evolves due to trade-offs in performance, then optimized performance on one host should limit performance on others. Thus, host specialization has classically been thought to be an evolutionary outcome that strongly selects for further specialization [67,78]. If these predictions are correct, then phylogenetic reconstructions should reveal that host shifts are rare in specialists, that transitions from specialist to generalist strategies are uncommon, and that specialist lineages are phylogenetically derived. Despite the intuitive appeal of these predictions, it is becoming increasingly clear that transitions from specialist to generalist strategies are common, host shifts are frequent and generalist lineages are equally likely to be phylogenetically derived [14,70,113]. Several authors report generalists and specialists within the same phylogenetic line [14], and models assuming irreversible evolution of generalists to specialists are usually strongly rejected [78]. Genomic plasticity and rapid evolution of the mechanisms underlying specialization are emerging as key drivers of this evolutionary dynamism. Horizontal gene transfer can radically alter the genomes of microbial pathogens, and many host jumps involve horizontal transfer of large effector complements [14]. Furthermore, conserved effector loci often undergo strong diversifying selection and display unusually high sequence polymorphism, suggesting rapid evolution in genes underlying host specificity [106,114].

less efficient on a second host, generalists are often less successful than specialists on highly defended plants, and the potential to encounter an enemy-free space favors shifts towards hosts that are less optimal as a food source than the original host. However, why is it so difficult in a proximate sense to utilize multiple plant species, or host organs, and which traits promote ecological fitting? Rather than enemies being limited by primary metabolic demands, performance on a host is mainly determined by the interplay of host resistance and enemy counteradaptation. Thus, host resistance traits and the capacity of the enemies to deal with them emerge as central factors in the determination of host ranges. The use of any specific host assemblage (which can be narrow or broad) requires particular adaptations [80]. Due to their different size and mobility, pathogens and herbivores have different strategies to avoid being affected by host resistance traits. Here, we discuss how host resistance traits and the molecular basis of their suppression or avoidance by plant enemies are fundamentally involved in the determination of host spectra and, thus, the evolution of specialist versus generalist strategies. Genetic and physiological trade-offs In specialized interactions, constraints on the use of certain host plants might be evident as pleiotropic trade-offs in the performance on alternate hosts (i.e. genes that promote the ability to utilize one partner, impair the ability to utilize another) [10,81]. For example, in the interaction between the plant Linum usitatissimum and its fungal pathogen M. lini, several interacting host resistance (R) and pathogen effector gene loci, which provide alternate resistance and infectivity specificities, have been identified [82]. Importantly, allelic variants at the AvrP123 effector locus that escape recognition from one R gene, usually confer recognition to a different R gene [82]. Thus, specialization via antagonistic pleiotropy is seemingly built into the system. Analogous trade-offs might also

Review mediate broader patterns of resistance specificity, such as resistance to insect herbivores versus microbial pathogens [83,84], and biotrophic versus necrotrophic pathogens [85]. Interestingly, evidence is mounting that some enemies can directly exploit constraints imposed on hosts by such pleiotropic costs [86–88]. Genetic constraints can also limit the success of generalist strategies, via trade-offs between the capacity to utilize a wide range of hosts and optimal performance on any one host [81,88]. A mechanism that causally underlies the ‘jack of all trades – master of none’ principle would be the fact that specialist herbivores can utilize specialized enzymes for the detoxification of the ingested food whereas generalists require either multiple [20–22] or widely effective digestive enzymes [19,89]. Because promiscuous enzymes are less efficient than those that catalyze only one distinct chemical reaction [18], and the synthesis of multiple enzymes comes at a high metabolic cost, generalists are usually less efficient than specialists in utilizing any given host species. Effectors are used to overcome host resistance Pathogens and herbivores have evolved some common molecular mechanisms to evade or suppress host resistance. Perhaps most universal is the concept of the ‘effector’: a term used to denominate all molecules that are released from plant enemies for host manipulation [13,14]. The concept of effectors and their role in host invasion and activation of resistance is most advanced in the context of plant–pathogen interactions [14,16,82,90,91]. In general, plants have evolved the capacity to perceive two classes of molecule (elicitors) that indicate attack by a pathogen. Conserved microbial molecules, known as ‘pathogen-associated molecular patterns’ (PAMPs) or ‘microbe-associated molecular patterns’ (MAMPs), are perceived by host receptor proteins known as pattern recognition receptors (PRRs). PAMPS are typically common structural components of a class of pathogen, such as chitin and flagellin, and their recognition causes PAMP-triggered immunity (PTI). To overcome this problem, many successful pathogens have evolved the capacity to deliver effectors into host cells to suppress PTI and other defence responses. In turn, many hosts have acquired the capacity to recognize either the changes that are inflicted by the action of these elicitors (‘modified-self recognition’), or to recognize directly and specifically the effectors via their interaction with a class of plant receptor proteins that contain nucleotide-binding (NB) domains and leucine-rich repeat (LRR) receptor kinases. The recognition of effectors by plant NB-LRR proteins results in further layers of more specific (‘gene-for-gene’) resistance responses denominated ‘effector-triggered immunity’ (ETI) [16,82]. Plant NB-LRR proteins confer resistance to both microbial pathogens and insects [82]. The emerging pattern is that R genes in general confer resistance to herbivores in a similar manner to that described above for pathogens [12,13,15], although it is likely that the relative importance of effectors for host utilization is higher for insect herbivores that are more intimately associated with their host plant, such as leaf miners, phloem feeders and single-cell feeders. By contrast, classic, leaf-chewing folivores are likely to depend more on behavioral, detoxification and


sequestration strategies, although specific effectors, such as b-glucosidases and proteases, are likely to play a crucial role in this process. As for pathogens, plants can recognize herbivore-derived effectors either directly, or by monitoring their action. In the first case, the recognition of ‘herbivore-associated patterns’ (HAMPs) induces relatively general resistance responses, which often depend, at least partly, on induction of a jasmonic acid (JA) pathway. Plants also recognize the action of herbivores, similar to the above-mentioned ‘modified-self recognition’ strategy, by perceiving fragmented or delocalized molecules associated with damage caused by their action. The responses that are elicited by plant ‘damagedself recognition’ are very general ones and are commonly based on JA induction [92,93]. By contrast, sucking insects cause little mechanical damage. However, they are intimately associated with the plant cells and need multiple effectors to evade recognition of their HAMPs [13]. As such, herbivores such as the hessian fly (Mayetiola destructor) and aphids were the first for which resistance mediated by R genes was reported [13,15,94,95]. The targets of these R genes are likely to be insect-derived effectors. For example, based on their similarity to pathogen effectors, 48 effector candidates have been identified in the green peach aphid (Myzus persicae) [96]. At least one of these functions as the target of recognition in certain plant hosts [96]. Correspondingly, several NB-LRR proteins have been identified that are required for a successful resistance induction against insect herbivores, including whitefly and aphids ([15,97] and references therein). Effectors and the determination of host specificity and host shifts As essential microbial components, PAMPS are highly conserved within and among pathogen species and plant PRRs are also highly conserved [98]. By contrast, the overall effector repertoire of pathogens can be highly variable, particularly among species and host-specific lineages [14,90] and, when conserved, often displays unusually high levels of sequence polymorphism [3,75]. Similarly, NBLRR resistance protein repertoires are variable among species and can be highly polymorphic [14] or deleted entirely within host species [99]. As we note above, there are well-established patterns of evolutionary conservatism in the range of hosts used by any given pathogen. It has been hypothesized that the interplay among highly conserved PRR-triggered immunity, and highly specific NB-LRR protein-triggered immunity can explain the phylogenetic specialization of plant enemies [14]. In particular, as phylogenetic distances among hosts increase, the effector repertoire carried by a pathogen becomes increasingly ineffective, first at suppressing specific NB-LRRmediated resistance, followed by increasing basal PRRmediated resistance. Together with trade-offs and ongoing antagonistic coevolution, such dynamics could strongly promote evolution towards increasing specialization in plant–enemy interactions. Despite the general appeal of the above scenario, some enemies do have truly wide host ranges, and host shifts can involve quite distantly related hosts [70,76]. What mechanisms facilitate the ecological fitting that underlies these 289

Review seeming anomalies? One strategy that may be important to the maintenance of wide host ranges is to suppress host resistance mechanisms at an early stage of induction. In pathogens, the type III secretion system is used to inject multiple effectors into host cells and helps Salmonella to colonize plant and animal hosts [17,24]. This system represents a common trait of numerous plant and animal pathogens, many of which are characterized by wide host ranges [100–104]. Because it can be also used by specialists to inject specific cocktails of coevolved effectors [105], the type III secretion system does not represent a ‘generalist strategy’ per se, but represents instead an apparatus that might facilitate a true ‘generalists’ strategy. For example, many Pseudomonas strains inject coronatine, a JA mimic. Coronatine manipulates the crosstalk between the JA and salicylic acid (SA) pathways [84], resulting in the suppression of SA-dependent responses. Thus, this process renders hosts generally susceptible to this pathogen [11]. Similarly, many insects and necrotrophic pathogens release hormones that suppress JA-dependent defence responses [11,84]. Such strategies mean that enemies can avoid the consequences of the expression of hundreds of defence-related genes, thereby greatly enhancing their ability to utilize a wide range of hosts. Other generalists may rely on carrying a broad spectrum of effectors, only a subset of which might be effective against any given host (e.g. Botrytis cinerea), and generalist herbivores often use multiple, or highly promiscuous, enzymes to detoxify their food [18–22], although such strategies presumably come at a cost [106]. Thus, host shifts and ecological fitting likely involve mechanisms that suppress resistance strategies that are shared between the old and new host. Moreover, shifts within and among closely related species may be achieved by mutations or deletion of single effector genes [106], whereas more distant jumps often seem to involve horizontal transfer of large complements of effectors [14]. Perspectives: new approaches to studying specificity in plant–enemy interactions In the above section, we reviewed the most common molecular mechanisms that underlie the specificity in host use by herbivores and plant pathogens and discussed how recently developed molecular concepts can help to explain classic ecological and evolutionary hypotheses, such as adaptive radiation, phylogenetic conservatism and ecological fitting. However, molecular tools remain underutilized in the ecological and evolutionary disciplines and more could be done to identify the molecular determinants of specificity of host use by plant enemies. As we highlight in Box 1, the various axes along which plant enemies evolve specificity are important because they provide insight into the underlying ecological and evolutionary mechanisms. However, there is a real deficiency of empirical data on host range under natural conditions. Such data will be required to inform theory and to develop capacity to predict host shifts and potential for invasion of plant enemies. For pathogens, one way to develop a better understanding of host utilization under field conditions might be intensive sampling and unbiased sequencing of microbial DNA resident in plants. This approach has been applied recently to discover asymptomatic endophytes, 290


many of which are pathogens in crops [27–30]. However, to determine precise outcomes, such studies will need to be accompanied by experiments examining the effects of colonization under common environmental conditions. For herbivores, realized host ranges can be assessed by unbiased collection strategies (fogging, etc.) and food web construction, preferably accompanied by feeding trials (e.g. [107]). An alternative way to determine realized host ranges of herbivores will be DNA barcoding or another sequencingbased approach to determine species ranges of ingested food items in the digestive tracts of animals living in the wild. The continued development of next-generation sequencing platforms will revolutionize research into the functional and evolutionary genetics of specialization in plant–enemy interactions. As well as the identification of realized host ranges of herbivores and pathogens, DNA barcoding and other sequencing-based strategies can be used to identify cryptic species and patterns in the association of certain genotypes of plant enemies with specific hosts [42,44]. Recently, these techniques have successfully been applied to understand the specificity and virulence of the over 50 pathovars of the ‘generalist’ pathogen, P. syringae [104,108]. Large-scale phylogenies are increasingly becoming available and can be subjected to ancestral trait mapping to identify host shifts and truly ‘phylogenetically conservative’ plant enemies [71–77]. Enemies from different populations, or species that have recently diverged and specialized onto different hosts [109], can be compared at the genomic, transcriptomic and phenotypic level, to investigate directly the genetic changes that are involved in host specialization. Perhaps the most powerful tool is represented by phylogenetically controlled comparisons among transcriptomes of specialists and generalists or in enemies that have recently been subject to a major shift in their host range. In particular, pathogens that have evolved higher specialization following a host shift [73,108], pathogens that have changed their life style from pathogen to asymptomatic endophyte [110] or vice versa [111], and related herbivores that represent the same feeding guild but differ strongly in host range [22], are promising models to screen for adaptations that allow generalists and specialists to fulfill successfully all the specific tasks that are required for their respective strategy. As is the case for many other disciplines, research into host ranges of plant enemies urgently requires multidisciplinary approaches to gain a causal understanding of why a particular enemy can, or cannot, successfully attack certain hosts and to predict potential host shifts and changes among specialist and generalist strategies. Acknowledgments We thank Carlos Garcıá-Robledo for discussing critical parts of the manuscript and CONACyT de Me´xico (project 129678) and the Australian Research Council (DP1097256) for financial support. The following people generously provided photographs for Figure 1: Eric M. Coombs, W. Billen, S. Bauer, David Cappaert, Edith Freitag, Matthias Goeke, Marc E. Ellis and Anurag A. Agrawal.

References 1 Thompson, J.N. (2005) Coevolution: the geographic mosaic of coevolutionary arms races. Curr. Biol. 15, R992–R994

Review 2 Dyer, L.A. et al. (2007) Host specificity of Lepidoptera in tropical and temperate forests. Nature 448, 696–699 3 Barrett, L.G. et al. (2009) Continua of specificity and virulence in plant host–pathogen interactions: causes and consequences. New Phytol. 183, 513–529 4 Berdegue, M. et al. (1996) Is it enemy-free space? The evidence for terrestrial insects and freshwater arthropods. Ecol. Entomol. 21, 203– 217 5 Jeffries, M.J. and Lawton, J.H. (1984) Enemy free space and the structure of ecological communities. Biol. J. Linnean Soc. 23, 269–286 6 Futuyma, D.J. and Moreno, G. (1988) The evolution of ecological specialization. Annu. Rev. Ecol. Syst. 19, 207–233 7 Connell, J.H. (1978) Diversity in tropical rain forests and coral reefs. Science 199, 1302–1310 8 Janzen, D.H. (1970) Herbivores and the number of tree species in tropical forests. Am. Nat. 104, 501–528 9 Parker, I.M. and Gilbert, G.S. (2004) The evolutionary ecology of novel plant–pathogen interactions. Annu. Rev. Ecol. Evol. Syst. 35, 675–700 10 Poisot, T. et al. (2011) A conceptual framework for the evolution of ecological specialisation. Ecol. Lett. 14, 841–851 11 Pieterse, C.M.J. and Dicke, M. (2007) Plant interactions with microbes and insects: from molecular mechanisms to ecology. Trends Plant Sci. 12, 564–569 12 Erb, M. et al. (2012) Role of phytohormones in insect-specific plant reactions. Trends Plant Sci. 17, 250–259 13 Hogenhout, S. and Bos, J. (2011) Effector proteins that modulate plant–insect interactions. Curr. Opin. Plant Biol. 14, 422–428 14 Schulze-Lefert, P. and Panstruga, R. (2011) A molecular evolutionary concept connecting nonhost resistance, pathogen host range, and pathogen speciation. Trends Plant Sci. 16, 117–125 15 Walling, L.L. (2008) Avoiding effective defenses: strategies employed by phloem-feeding insects. Plant Physiol. 146, 859–866 16 Jones, J.D.G. and Dangl, J.L. (2006) The plant immune system. Nature 444, 323–329 17 Schikora, A. and Hirt, H. (2012) Plants as alternative hosts for Salmonella. Trends Plant Sci. 17, 245–249 18 Li, W. et al. (2003) Diversification of furanocoumarin-metabolizing cytochrome P450 monooxygenases in two papilionids: specificity and substrate encounter rate. Proc. Natl. Acad. Sci. U.S.A. 100, 14593–14598 19 Li, X. et al. (2004) Structural and functional divergence of insect CYP6B proteins: from specialist to generalist cytochrome P450. Proc. Natl. Acad. Sci. U.S.A. 101, 2939–2944 20 Pauchet, Y. et al. (2008) Mapping the larval midgut lumen proteome of Helicoverpa armigera, a generalist herbivorous insect. J. Proteome Res. 7, 1629–1639 21 Pauchet, Y. et al. (2010) Pyrosequencing the Manduca sexta larval midgut transcriptome: messages for digestion, detoxification and defence. Insect Mol. Biol. 19, 61–75 22 Ramsey, J.S. et al. (2010) Comparative analysis of detoxification enzymes in Acyrthosiphon pisum and Myzus persicae. Insect Mol. Biol. 19, 155–164 23 Partida-Martinez, L.P.P. and Heil, M. (2011) The microbe-free plant: fact or artefact? Front. Plant Sci. 2, 100 24 Schikora, A. et al. (2011) Conservation of Salmonella infection mechanisms in plants and animals. PLoS ONE 6, e24112 25 Winfield, M.D. and Groisman, E.A. (2003) Role of nonhost environments in the lifestyles of Salmonella and Escherichia coli. Appl. Environ. Microbiol. 69, 3687–3694 26 Fisher, P.J. et al. (1994) Fungal endophytes from the leaves and twigs of Quercus ilex L. from England, Majorca and Switzerland. New Phytol. 127, 133–137 27 Albrectsen, B.R. et al. (2010) Endophytic fungi in European aspen (Populus tremula) leaves: diversity, detection, and a suggested correlation with herbivory resistance. Fungal Divers. 41, 17–28 28 Ghimire, S.R. et al. (2011) Biodiversity of fungal endophyte communities inhabiting switchgrass (Panicum virgatum L.) growing in the native tallgrass prairie of northern Oklahoma. Fungal Divers. 47, 19–27 29 Frohlich, J. et al. (2000) Endophytic fungi associated with palms. Mycol. Res. 104, 1202–1212 30 Gazis, R. and Chaverri, P. (2010) Diversity of fungal endophytes in leaves and stems of wild rubber trees (Hevea brasiliensis) in Peru. Fungal Ecol. 3, 240–254


31 Janzen, D.H. (1979) New horizons in the biology of plant defences. In Herbivores: Their interactions with Secondary Plant Metabolites (Rosenthal, G.A. and Janzen, D.H., eds), pp. 331–350, Academic Press 32 Garcıá-Robledo, C. and Horvitz, C.C. (2011) Experimental demography and the vital rates of generalist and specialist insect herbivores on native and novel host plants. J. Anim. Ecol. 80, 976–989 33 Altermatt, F. and Pearse, I.S. (2011) Similarity and specialization of the larval versus adult diet of European butterflies and moths. Am. Nat. 178, 372–382 34 Jin, Y. (2010) Century-old mystery of Puccinia striiformis life history solved with the identification of Berberis as an alternate host. Phytopathology 100, 432–435 35 Novotny, V. and Basset, Y. (2005) Host specificity of insect herbivores in tropical forests. Proc. R. Soc. B 272, 1083–1090 36 Poulin, R. et al. (2011) Host specificity in phylogenetic and geographic space. Trends Parasitol. 27, 355–361 37 Gilbert, G.S. and Webb, C.O. (2007) Phylogenetic signal in plant pathogen–host range. Proc. Natl. Acad. Sci. U.S.A. 104, 4979–4983 38 Rasmann, S. and Agrawal, A. (2011) Evolution of specialization: a phylogenetic study of host range in the red milkweed beetle (Tetraopes tetraophthalmus). Am. Nat. 177, 728–737 39 Ness, J.H. et al. (2011) Phylogenetic distance can predict susceptibility to attack by natural enemies. Oikos 120, 1327–1334 40 Zentmyer, G.A. (1983) The world of Phytophthora. In Phytophthora, its Biology, Taxonomy, Ecology and Pathology (Erwin, D.C. et al., eds), pp. 1–8, American Phytopathological Society 41 Fox, L.R. and Morrow, P.A. (1981) Specialization: species property or local phenomenon? Science 211, 887–893 42 Ormond, E.L. et al. (2010) A fungal pathogen in time and space: the population dynamics of Beauveria bassiana in a conifer forest. FEMS Microbiol. Ecol. 74, 146–154 43 Brown, J.K. et al. (1995) The sweetpotato or silverleaf whiteflies: biotypes of Bemisia tabaci or a species complex? Annu. Rev. Entomol. 40, 511–534 44 Hebert, P.D.N. et al. (2004) Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astraptes fulgerator. Proc. Natl. Acad. Sci. U.S.A. 101, 14812–14817 45 Garrido, E. et al. (2012) Local adaptation: simultaneously considering herbivores and their host plants. New Phytol. 193, 445–453 46 Loxdale, H.D. et al. (2011) The evolutionary improbability of ‘generalism’ in nature, with special reference to insects. Biol. J. Linnean Soc. 103, 1–18 47 Kniskern, J.M. et al. (2010) Maladaptation in wild populations of the generalist plant pathogen Pseudomonas syringae. Evolution 65, 818– 830 48 Bolnick, D.I. et al. (2003) The ecology of individuals: incidence and implications of individual specialization. Am. Nat. 161, 1–8 49 Ehrlich, P.R. and Raven, P.H. (1964) Butterflies and plants: a study in coevolution. Evolution 18, 586–608 50 Thrall, P.H. et al. (2007) Coevolution of symbiotic mutualists and parasites in a community context. Trends Ecol. Evol. 22, 120–126 51 Janzen, D.H. (1985) On ecological fitting. Oikos 45, 308–310 52 Agosta, S.J. (2006) On ecological fitting, plant–insect associations, herbivore host shifts, and host plant selection. Oikos 114, 556–565 53 Agosta, S.J. et al. (2010) How specialists can be generalists: resolving the ‘parasite paradox’ and implications for emerging infectious disease. Zoologia 27, 151–162 54 Garcıá-Robledo, C. and Horvitz, C.C. (2012) Jack of all trades masters novel host plants: positive genetic correlations in specialist and generalist insect herbivores expanding their diets to novel hosts. J. Evol. Biol. 25, 38–53 55 Thrall, P.H. and Burdon, J.J. (2003) Evolution of virulence in a plant host–pathogen metapopulation. Science 299, 1735–1737 56 Agudelo-Romero, P. et al. (2008) The pleiotropic cost of hostspecialization in Tobacco etch potyvirus. Infect. Genet. Evol. 8, 806–814 57 Wallis, C.M. et al. (2007) Adaptation of plum pox virus to a herbaceous host (Pisum sativum) following serial passages. J. Gen. Virol. 88, 2839–2845 58 Barrett, L.G. et al. (2011) Cheating, trade-offs and the evolution of aggressiveness in a natural pathogen population. Ecol. Lett. 14, 1149– 1157 59 Mackenzie, A. (1996) Trade-off for host plant utilization in the black bean aphid, Aphis fabae. Evolution 50, 155–162 291

Review 60 Hawthorne, D.J. and Vian, S. (2001) Genetic linkage of ecological specialization and reproductive isolation in pea aphids. Nature 412, 904–907 61 Agosta, S.J. and Klemens, J.A. (2009) Resource specialization in a phytophagous insect: no evidence for genetically based performance trade-offs across hosts in the field or laboratory. J. Evol. Biol. 22, 907–912 62 Garcıá-Robledo, C. et al. (2010) Larval morphology, development, and notes on the natural history of Cephaloleia ‘rolled-leaf’ beetles (Coleoptera: Chrysomelidae: Cassidinae). Zootaxa 2610, 50–68 63 Kawecki, D. (1994) Accumulation of deleterious mutations and the evolutionary cost of being a generalist. Am. Nat. 144, 833–838 64 Matsubayashi, K.W. et al. (2010) Ecological speciation in phytophagous insects. Entomol. Exp. Appl. 134, 1–27 65 Sicard, D. et al. (1997) Genetic diversity and pathogenic variation of Colletotrichum lindemuthianum in the three centers of diversity of its host, Phaseolus vulgaris. Phytopathology 87, 807–813 66 Murphy, S.M. (2004) Enemy-free space maintains swallowtail butterfly host shift. Proc. Natl. Acad. Sci. U.S.A. 101, 18048–18052 67 Jaenike, J. (1990) Host specialization in phytophagous insects. Annu. Rev. Ecol. Syst. 21, 243–273 68 Normark, B. and Johnson, N. (2011) Niche explosion. Genetica 139, 551–564 69 Benmayor, R. et al. (2009) Host mixing and disease emergence. Curr. Biol. 19, 764–767 70 Janz, N. (2011) Ehrlich and Raven revisited: mechanisms underlying codiversification of plants and enemies. Annu. Rev. Ecol. Evol. Syst. 42, 71–89 71 de Vienne, D.M. et al. (2009) Phylogenetic determinants of potential host shifts in fungal pathogens. J. Evol. Biol. 22, 2532–2541 72 Desprez-Loustau, M-L. et al. (2011) Interspecific and intraspecific diversity in oak powdery mildews in Europe: coevolution history and adaptation to their hosts. Mycoscience 52, 165–173 73 Raffaele, S. et al. (2010) Genome evolution following host jumps in the Irish potato famine pathogen lineage. Science 330, 1540–1543 74 Roy, B.A. (2001) Patterns of association between crucifers and their flower-mimic pathogens: host jumps are more common than coevolution or cospeciation. Evolution 55, 41–53 75 Van Der Merwe, M.M. et al. (2008) Coevolution with higher taxonomic host groups within the Puccinia/Uromyces rust lineage obscured by host jumps. Mycol. Res. 112, 1387–1408 76 Voglmayr, H. (2003) Phylogenetic relationships of Peronospora and related genera based on nuclear ribosomal ITS sequences. Mycol. Res. 107, 1132–1142 77 Woolhouse, M.E.J. et al. (2005) Emerging pathogens: the epidemiology and evolution of species jumps. Trends Ecol. Evol. 20, 238–244 78 Nosil, P. and Mooers, A.O. (2005) Testing hypotheses about ecological specialization using phylogenetic trees. Evolution 59, 2256–2263 79 Branca, A. et al. (2011) Intraspecific specialization of the generalist parasitoid Cotesia sesamiae revealed by polyDNAvirus polymorphism and associated with different Wolbachia infection. Mol. Ecol. 20, 959–971 80 Gomez, J.M. et al. (2010) Ecological interactions are evolutionarily conserved across the entire tree of life. Nature 465, 918–921 81 Forister, M.L. et al. (2012) Revisting the evolution of ecological specialization, with emphasis on insect–plant interactions. Ecology DOI: 10.1890/11-0650.1 82 Dodds, P.N. and Rathjen, J.P. (2010) Plant immunity: towards an integrated view of plant–pathogen interactions. Nat. Rev. Genet. 11, 539–548 83 Erb, M. et al. (2011) Synergies and trade-offs between insect and pathogen resistance in maize leaves and roots. Plant Cell Environ. 34, 1088–1103 84 Thaler, J.S. et al. (2012) Evolution of jasmonate and salicylate signal crosstalk. Trends Plant Sci. 17, 260–270 85 Kliebenstein, D.J. and Rowe, H.C. (2008) Ecological costs of biotrophic versus necrotrophic pathogen resistance, the hypersensitive response and signal transduction. Plant Sci. 174, 551–556 86 Lorang, J.M. et al. (2007) Plant disease susceptibility conferred by a ‘resistance’ gene. Proc. Natl. Acad. Sci. U.S.A. 104, 14861–14866 87 Faris, J.D. et al. (2010) A unique wheat disease resistance-like gene governs effector-triggered susceptibility to necrotrophic pathogens. Proc. Natl. Acad. Sci. U.S.A. 107, 13544–13549

292


88 Loiseau, C. et al. (2008) Antagonistic effects of a MHC class I allele on malaria-infected house sparrows. Ecol. Lett. 11, 258–265 89 Marsh, K.J. et al. (2006) The detoxification limitation hypothesis: where did it come from and where is it going? J. Chem. Ecol. 32, 1247– 1266 90 Arnold, D.L. and Jackson, R.W. (2011) Bacterial genomes: evolution of pathogenicity. Curr. Opin. Plant Biol. 14, 385–391 91 Pieterse, C.M.J. et al. (2009) Networking by small-molecule hormones in plant immunity. Nat. Chem. Biol. 5, 308–316 92 Heil, M. et al. (2012) How plants sense wounds: damaged-self recognition is based on plant-derived elicitors and induces octadecanoid signaling. PLoS ONE 7, e30537 93 Heil, M. (2009) Damaged-self recognition in plant herbivore defence. Trends Plant Sci. 14, 356–363 94 Foster, J.E. et al. (1991) Effects of deploying single gene resistance in wheat for controlling damage by the Hessian fly (Diptera: Cecidomyiidae). Environ. Entomol. 20, 964–969 95 Hogenhout, S.A. et al. (2009) Emerging concepts in effector biology of plant-associated organisms. Mol. Plant Microbe Interact. 22, 115– 122 96 Bos, J.I.B. et al. (2010) A functional genomics approach identifies candidate effectors from the aphid species Myzus persicae (green peach aphid). PLoS Genet. 6, e1001216 97 Dogimont, C. et al. (2010) Host plant resistance to aphids in cultivated crops: genetic and molecular bases, and interactions with aphid populations. C. R. Biol. 333, 566–573 98 Zipfel, C. (2009) Early molecular events in PAMP-triggered immunity. Curr. Opin. Microbiol. 12, 414–420 99 Stahl, E.A. et al. (1999) Dynamics of disease resistance polymorphism at the Rpm1 locus of Arabidopsis. Nature 400, 667–671 100 Coombes, B.K. (2009) Type III secretion systems in symbiotic adaptation of pathogenic and non-pathogenic bacteria. Trends Microbiol. 17, 89–94 101 Grant, S.R. et al. (2006) Subterfuge and manipulation: Type III effector proteins of phytopathogenic bacteria. Annu. Rev. Microbiol. 60, 425–449 102 Killiny, N. and Almeida, R.P.P. (2011) Gene regulation mediates host specificity of a bacterial pathogen. Environ. Microbiol. Rep. 3, 791– 797 103 Marlovits, T.C. and Stebbins, C.E. (2010) Type III secretion systems shape up as they ship out. Curr. Opin. Microbiol. 13, 47–52 104 O’Brien, H.E. et al. (2011) Next-generation genomics of Pseudomonas syringae. Curr. Opin. Microbiol. 14, 24–30 105 Dean, P. (2011) Functional domains and motifs of bacterial type III effector proteins and their roles in infection. FEMS Microbiol. Rev. 35, 1100–1125 106 Barrett, L.G. et al. (2009) Diversity and evolution of effector loci in natural populations of the plant pathogen Melampsora lini. Mol. Biol. Evol. 26, 2499–2513 107 Novotny, V. et al. (2010) Guild-specific patterns of species richness and host specialization in plant-herbivore food webs from a tropical forest. J. Anim. Ecol. 79, 1193–1203 108 Baltrus, D.A. et al. (2011) Dynamic evolution of pathogenicity revealed by sequencing and comparative genomics of 19 Pseudomonas syringae isolates. PLoS Pathog. 7, e1002132 109 Stukenbrock, E.H. et al. (2010) Whole-genome and chromosome evolution associated with host adaptation and speciation of the wheat pathogen Mycosphaerella graminicola. PLoS Genet. 6, e1001189 110 Freeman, S. and Rodriguez, R.J. (1993) Genetic conversion of a fungal pathogen to a nonpathogenic, endophytic mutualist. Science 260, 75– 78 111 Romao, A.S. et al. (2011) Enzymatic differences between the endophyte Guignardia mangiferae (Botryosphaeriaceae) and the citrus pathogen G. citricarpa. Genet. Mol. Res. 10, 243–252 112 Rayachhetry, M.B. et al. (2001) Host range of Puccinia psidii, a potential biological control agent of Melaleuca quinquenervia in Florida. Biol. Control 22, 38–45 113 Roy, B.A. and Kirchner, J.W. (2000) Evolutionary dynamics of pathogen resistance and tolerance. Evolution 54, 51–63 114 Chen, M-S. et al. (2010) Unusual conservation among genes encoding small secreted salivary gland proteins from a gall midge. BMC Evol. Biol. 10, 296


Review


Specialist versus generalist insect herbivores and plant defense Jared G. Ali and Anurag A. Agrawal Department of Ecology and Evolutionary Biology, Cornell University, E425 Corson Hall, Ithaca, NY 14853-2701, USA

There has been a long-standing hypothesis that specialist and generalist insects interact with plants in distinct ways. Although many tests exist, they typically compare only one species of each, they sometimes confound specialization and feeding guild, and often do not link chemical or transcriptional measures of the plant to actual resistance. In this review, we synthesize current data on whether specialists and generalists actually differ, with special attention to comparisons of their differential elicitation of plant responses. Although we find few consistencies in plant induction by specialists versus generalists, feeding guilds are predictive of differential plant responses. We outline a novel set of predictions based on current coevolutionary hypotheses and make methodological suggestions for improved comparisons of specialists and generalists. Why might specialist and generalist herbivores have distinct interactions with plants? ‘Jack of all trades is master of none’. Here lies the theoretical basis for why ecologists and plant scientists have long argued that specialist insect herbivores, as compared with generalists, will have distinct and predictable interactions with their host plants (Box 1). With specialization, it was proposed that alongside the loss of ability to use many host plants, herbivores would gain the ability to tolerate plant defenses, manipulate hosts to their benefit and evolve ways to reduce predation and parasitism [1,2]. This powerful and seductive hypothesis has been a mainstay of coevolutionary studies for over 40 years, and yet little resolution has been reached on certain predictions. In particular, we argue below that ecologists and plant scientists have been too quick to position the specialist–generalist dichotomy as a paradigm, and often uncritically. Below we evaluate the current evidence and provide a roadmap for future studies. There have been several specific predictions made about the specialist–generalist paradigm. First, specialists should be less impacted by a given plant defense compared with a generalist [2] (Figure 1). In addition to being less affected by particular defense traits, some specialist herbivores have even evolved the capacity to use these same traits in host finding or protection from predators (sequestration or fecal shields). Second, generalists should have ‘general’ mechanisms to tolerate an array of plant defenses and also possess mechanisms to manipulate plants via highly conserved plant pathways [1,2]. The notion behind Corresponding author: Ali, J.G. ([email protected]).

this prediction is that although generalists do not master any one defense, many aspects of plant defense can be overcome because plants possess a common evolutionary history leading to shared physiological features in core signal transduction chains [e.g. jasmonate (JA) signaling] [3]. Third, upon damage, induced plant responses to specialists will be distinct compared with responses to generalists. This general prediction is complicated by coevolution: are observed plant responses adaptive for the plant or manipulated by herbivores? The perspective from which we view the interaction distinctly shapes our predictions (Figure 2). Although we will touch on the first prediction above, the focus of our review is on the latter two: how and why specialists and generalists might elicit differential plant responses (or manipulate the plants in different ways). Since the origin of the specialist–generalist paradigm, there have been hundreds of studies of insect tolerance and detoxification of plant defense [4]. However, it is only in the past 20 years that plant biologists have realized that induced responses are a crucial component of plant defense, and ideas about how specialists and generalists differ in this regard are continuing to develop. In addition,

Box 1. Who’s who on the diet breadth continuum? Insect herbivores have been conventionally grouped into categories based on their degree of dietary specialization. When limited to only one or a few closely related plant taxa, often a single genus, herbivores are considered monophagous (or highly specialized). Insect herbivores that feed on several plant species, usually within one botanical family, are designated oligophagous. Finally, polyphagous (or highly generalized) species are insects that feed on species in more than one plant family. Although these terms are helpful for generalizing broad groups of herbivores into simpler categories, their basis is drawn on fairly arbitrary observations and may lead to inherit limitations in their use. Nonetheless, some groups of herbivores, such as aphids, leaf hoppers and leaf miners are dominated (>75%) by monophages [71]. Across all herbivorous insects, it is estimated that 290 empirical studies [4]. A Compounds within the normal host range: for specialists that normally encounter a particular defense, toxicity is lower compared with the impact on generalists. B Novel compounds: for specialists that do not typically encounter a particular defense, toxicity to the specialist is greater than or equal to that compared with the generalist. In other words, it is empirically the case that specialists are less impacted by the toxicity of the plant defenses they typically consume compared with generalists; nonetheless, specialists can be highly susceptible to novel plant secondary compounds.

modern studies that span bioassays of insect preference and performance, plant production of hormonal signals and defensive secondary metabolites, and transcriptional responses have the potential to aid us in making rapid progress in understanding how and why specialists and generalist herbivores differ. Impacts of plant defense on specialists and generalists The notion that specialists are immune to the defenses of the host plant is widespread but incorrect. Cases where specialist herbivores are negatively impacted by defense compounds include: parsnip webworms (Depressaria pastinacella) eating furanocoumarins [5], buckeye caterpillars (Junonia coenia) ingesting iridoid glycosides [6], monarch caterpillars (Danaus plexippus) on cardenolide-containing sandhill milkweed (Asclepias humistrata) [7], cabbage white caterpillars (Pieris rapae) being poisoned by isothiocyanates [8] and tobacco hornworms (Manduca sexta) fed artificial diets containing nicotine [9]. In nearly all of these cases, the specialists do have physiological adaptations to cope with the plant defenses, which allow greater tolerance than most generalists. Indeed, on average, specialist herbivores are less negatively impacted by defense compounds than generalists [4] (Figure 1). Our main message is that tolerance of specialist insects to low levels of toxins is to be expected; however, at higher levels of defense, few insects are immune to the deleterious effects of plant toxins. It is unclear whether certain classes of plant defense are more effective against generalists or specialists. A study 35 years ago suggested that although toxins could be overcome by specialists, digestibility reducers are likely to be effective against all attackers [10]. Others have argued that indirect defense (i.e. attracting enemies of herbivores) is likely to be more difficult to overcome compared with direct defense (e.g. digestibility reducers and toxins) [11]. Although most plants produce all of these classes of 2

Do specialists and generalists elicit different defensive responses? A hypothesis that grew out of the specialist–generalist paradigm is that specialist herbivores will cause distinct induction of plant defenses compared with those induced by generalists [18–20]. Nonetheless, there have been few explicit predictions in the literature about how and why specialists will differ from generalists with regard to elicitation of induced defenses. Given that generalists are typically more sensitive to plant toxins than specialists, from the perspective of the insect, one prediction is that generalists should suppress induced plant responses, whereas specialists should only minimize the induction of high levels of defense (Figure 2). From the perspective of the plant, the predicted responses are less consistent: induction of direct defenses could be variable against specialists (Figure 2), but induction of indirect defenses [e.g. extrafloral nectar and parasite-attracting volatile organic compounds (VOCs)] should be strong if the specialist is not sequestering. Nonetheless, it is presumably adaptive for plants to respond, as strongly as possible, to most generalists (Figure 2). Experiments comparing phenotypic or transcriptional responses to both specialist and generalist herbivores often include only one specialist and one generalist species, making rigorous conclusions impossible; in addition, many studies compare specialist and generalist species from different feeding guilds [21–23]. We found 20 studies comparing the phenotypic or transcriptional responses of a plant to both specialist and generalist herbivores using one feeding guild (Table 1). Although we interpret these results in light of the predictions in Figure 2, we recommend caution because nearly every result can be interpreted in an adaptive context, because what is beneficial for the plant and beneficial for the insect herbivore can be different. In addition, we assumed that the authors were careful to match the amount and timing of damage by the two herbivores; we highly recommend that future studies explicitly address this issue (Box 2). A few generalizations emerged from our review. First, there are few studies linking mechanistic plant response to impacts on herbivores; however, these links are crucial for interpreting specific consequences of plant defenses. For example, some studies in the Brassicaceae found that generalist and specialist elicited a similar plant response [20,24], whereas other studies that only measured impacts on herbivores found differential induction of resistance [19]. Second, of the generalist chewers, 14 out of 16 studies used only noctuid agricultural pests in one of a few genera. All four studies of generalist and specialist aphids used the same two species on Brassicaceae hosts (Table 1). Aside from the potential taxonomic bias in herbivores, there was


Review


Impacts of plant toxins on herbivores (a)

Predicted induced plant responses

Sequestering specialist

Insect performance

Insects benefit from intermediate induction Plants benefit from weak or strong induction (indirect defense may only be effective at low toxin levels)

(b) Insect performance

Non-sequestering specialist Insects indifferent to all but high levels of induction Plants benefit from strong induction and indirect defense

Generalist

Insect performance

(c)

Insects benefit from suppressing induction Plants benefit from any induction

Toxin level produced by plant TRENDS in Plant Science

Figure 2. Three herbivore strategies (a–c) and their expected relationships with plant toxins. Sequestering specialists benefit from the toxins at intermediate levels (via protection from predation) and nonsequestering specialists are tolerant of toxins at low levels; however, in both cases toxins eventually impose a cost. From the perspective of the insect, induction should maximize their own growth, and across all herbivore strategies induction should be low (either intermediate, minimal or suppressed (a–c), respectively). From the perspective of the plant, maximizing defense, induction responses can be more variable and alternative strategies (i.e. indirect defense via induction of volatile organic compounds) might be the most effective defense against specialists. We note that there are special cases that might not fit this model; for example, some generalists benefit from feeding on toxic plants, even if they do not sequester the toxins [59].

Table 1. Comparison of plant defensive response to at least one specialist and one generalist insect herbivore from the same feeding guild Plant (Brassicaceae) A. thaliana

Generalist (Aphididae) Myzus persicae

Specialist (Aphididae) Brevicoryne brassicae

Measure of plant response Transcriptional responses, glucosinolates (GS)

(Brassicaceae) Brassica oleraceae

(Aphididae) M. persicae

(Aphididae) B. brassicae

GS

Results a The generalist caused slightly more changes in gene expression than did the specialist (sequesterer). General stress-responsive genes and octadecanoid and indole GS synthesis genes were similarly induced by generalist and specialist [22,32]. The specialist induced a lower GS response than did the generalist [26]. Induction pattern by the two species depended on water status of the plant [58].

3


Review


Table 1 (Continued ) Plant (Brassicaceae) A. thaliana

Generalist (Noctuidae) Spodoptera exigua

Specialist (Piridae) Pieris rapae

Measure of plant response Transcriptional response, GS

(Brassicaceae) A. thaliana (Brassicaceae) A. thaliana

(Noctuidae) S. littoralis (Noctuidae) S. exigua

Transcriptional response

(Brassicaceae) A. thaliana

(Noctuidae) T. ni, S. exigua

(Brassicaceae) Brassica nigra

(Noctuidae) Mamestra brassicae

(Brassicaceae) B. nigra

(Noctuidae) Trichoplusia ni

(Piridae) P. rapae (Piridae) P. rapae (Plutellidae) P. xylostella (Piridae) P. rapae, (Plutellidae) P. xylostella (Piridae) P. rapae, (Plutellidae) Plutella xylostella (Piridae) P. rapae

(Brassicaceae) Boechera divaricarpa

(Noctuidae) T. ni

(Plutellidae) P. xylostella


(Brassicaceae) Raphanus sativus

(Noctuidae) T. ni, S. exigua

(Piridae) P. rapae, (Plutellidae) P. xylostella

Induced resistance, herbivore performance

(Brassicaceae) Sinapis alba

(Noctuidae) S. frugiperda

(Tenthredinidae) Athalia rosae

GS, myrosinase (MYR)

(Lauraceae) Lindera benzoin

(Noctuidae) S. exigua

(Plantaginaceae) Plantago lanceolata Poaceae Zea mays

(Nymphalidae) Junonia ceonia

(Geometridae) Epimecis hortaria (Erebidae) Spilosoma congra (Chrysomelidae) Diabrotica virgifera virgifera (Sphingidae) Manduca sexta

Peroxidase activity (POD), C/N ratio, protein content, insect bioassays Iridoid GS (IrGS), protein, foliar nitrogen

(Solanacae) Nicotiana attenuata (Solanacae) N. attenuata

(Solanacae) N. attenuata (Solanacae) N. tabacum

a

(Chrysomelidae) Diabrotica balteata (Noctuidae) S. exigua

Parasitoid specificity for herbivore induced plant volatiles (HIPVs) Transcriptional responses, GS

Results a Expression of GS genes was similar for generalist and specialist, but GS levels only showed an increase in response to S. exigua. Mean aliphatic GS levels were equal. P. rapae caused a higher increase in indolyl GS content [22]. Transcription profiles were indistinguishable [24]. Parasitoid attracted to damaged plants over controls for both generalists and specialists. Parasitoids only discriminate between induction by insects in different guilds [21]. Transcriptional responses and GS were not consistently influenced by degree of insect specialization [26].

GS

Indole GS was significantly higher after feeding by P. rapae and M. brassicae than after P. xylostella feeding [60].

Foliar trichomes, sinigrin, foliar nitrogen

Differential induction by specialist versus generalist led to increased trichomes, but the effect reversed on different leaf positions [61]. Specialist induced SA- and ethylene-associated genes, whereas generalist induced JA and ET genes [36]. The specialist might be well adapted, but the plant defends against the generalist. Variation in induction was found, but it was not associated with insect specialization. P. xylostella and S. exigua induced resistance to all, whereas P. rapae only induced resistance to P. rapae and S. exigua. T. ni did not induce resistance [19]. Specialist (sequesterer) and mechanical wounding induced GS and MYR threefold, whereas generalist induced only GS (twofold) [37] – generalist might be adaptively suppressing defense. POD activity was more strongly induced by generalist than specialist (no difference in bioassay) [62] – plant might be adaptively defending against generalist. Higher IrGS induced by specialist (sequesterer) compared with generalist [63] – plant might be adaptively defending against generalist. Natural enemies preferred roots attacked by specialist over roots damaged by generalist. The specialist induced significantly more (E)-b-caryophyllene than the generalist. Specialist induced JA/ET burst, generalist induced SA [64] – might be adaptive for generalists to suppress resistance by activating SA. Despite large overlap, the plant response to the generalists was more similar than the response to the specialist. This was correlated to FACs/oral secretions. Both generalists were noctuids and downregulated a large number of similar genes [54]. M. sexta induced a JA and SA response, whereas S. littoralis and T. ni induced stronger SA responses [33].

Parasitoid specificity for herbivore induced plant volatiles Phytohormones

(Noctuidae) Heliothis virescens, S. exigua

(Sphingidae) M. sexta


(Noctuidae) T. ni, S. littoralis (Noctuidae) Helicoverpa armigera

(Sphingidae) M. sexta

Phytohormones

(Noctuidae) Helicoverpa assulta

Lipoxygenase (LOX), proteinase inhibitors (PIs), nicotine, peroxidase (POD), polyphenol oxidase (PPO)

Both herbivores induced a similar defensive response, but response intensity of plants was different: specialist induced a lower PPO response and more intensive nicotine and POD response than generalist (JA, LOX and PIs were not different) [65].

Color-coding reflects consistency with the hypotheses in Figure 2 (green = consistent, but only two species are compared). Yellow indicates no consistent pattern and red indicates that the level of specialization was not predictive of plant responses.

4


Review


Box 2. Testing for differences in induced plant defense among specialist and generalist herbivores If the goal is to test the hypothesis that specialists elicit differential plant resistance compared with generalists, we recommend the following experimental design (Figure I). Ideally, a comparison of more than two species is necessary because any two species will differ in a myriad of ways. We suggest a minimum of comparing four species that are all from one feeding guild (e.g. leaf chewers or phloem-suckers) and in two taxonomic pairs. As an example, consider the plant defense response induced by two Helicoverpa spp. (one specialist and one generalist) and two sawfly species (Tenthredinidae) (one specialist and one generalist). Note that the two pairs in this case are grouped at differing taxonomic levels (within genus versus within family). Nonetheless, the comparisons are both valid because within each group, a specialist and generalist are compared. A common shortcoming of studies is that both specialists (or both generalists) are from one group (e.g. noctuids), confounding the comparison between specialists and generalists and taxonomic grouping. A benefit of having the four species in two taxonomic groups is that a two-way analysis of variance approach can be used to partition the relative impact of herbivore specialization and taxonomic grouping in the plant response.

To test for the differences in the induction of the defense response, it is crucial to conduct all treatments at the same time and intermixed within the experimental arena, for example, a growth chamber (in our scenario of four species there would be six treatments: a control, mechanical damage and damage by each of the four herbivores). The reason for this approach is that differences between the induced defense responses are often subtle and, thus, it is important to have treatments intermixed. The timing, location, extent of herbivory (and mechanical damage), developmental stage and diet on which the insects are raised must also be highly controlled because differences can arise because of differences in feeding style unrelated to specialization. Finally, we strongly recommend some measure of the plant responses (e.g. chemical and transcriptional) be coupled with some biological effect (i.e. a bioassay). An important benefit of this approach is the connection between complex (often multivariate) response measures being linked to the hypothesized effect on organisms. We note that although the proposed experimental design appears onerous, there should be possibilities, particularly for crop plants and trees with well-known insect faunas.

Relative significance of host range comparisons

Herbivore specialists

Optimal comparison

A

Herbivore generalist

Sub-optimal comparison

Herbivore specialists

B

Herbivore generalist

Time TRENDS in Plant Science

Figure I. A phylogenetic representation for suggested comparisons in studies comparing herbivores with different levels of specialization. In one scenario, A represents a chewing herbivore lineage and B represents a piercing-sucking lineage; here, the optimal comparison between specialists and generalists is within guild (and also within lineage). In a second scenario, all represented herbivores are chewers, but A are Lepidoptera and B are sawflies; again, the within lineage comparison is superior to the across lineage comparison because it controls for many other differences between the species.

also a very limited range in the plant species, as all species were herbaceous, and most were representatives of the Brassicaceae or Solanaceae. Third, few studies compared induction of indirect defenses [21,25]. This area requires further studies because the adaptive value of indirect defenses, particularly VOCs, can be associated with the ability of a specialist to sequester toxins (Figure 2). The additional trophic level further complicates generalizations of plant–herbivore interactions because the involvement

of a natural enemy incorporates dynamics of foraging behavior and signal reliability (see review by Jonathan Gershenzon and colleagues in this special issue). Finally, despite efforts to align appropriate comparisons (i.e. within taxon and guild), we found no consistent pattern of differential elicitation based on the degree of host plant specialization (Table 1). None of the studies that compared more than two herbivores showed consistency with regard to responses associated with insect specialization. 5


Review The strongest studies compared at least two herbivores each from both the generalist and specialist categories and within the same feeding guild; only two studies met this criteria, and neither found consistent differential induction by specialists and generalists [19,26]. Although one of the studies [19] found specificity of induction among four caterpillar species [the beet armyworm (Spodoptera exigua) and cabbage looper (Trichoplusia ni), which are generalists, and the diamondback moth (Plutella xylostella) and the cabbage white (Pieris rapae), which are specialists] damaging wild radish (Raphanus sativus), this was not associated with diet specialization. The other study [26] approached the paradigm with a rigorous analysis of plant response to three specialists [diamondback moth, small cabbage white and the cabbage aphid (Brevicoryne brassicae)] and three generalists [cabbage looper, beet armyworm and the green peach aphid (Myzus persicae)], taking into consideration the role that feeding-guild might play. This excellent study on Arabidopsis (Arabidopsis thaliana) was able to partition the relative effects of specialists and generalists and simultaneously compare the induction by two guilds [26]. Nonetheless, an examination of genome-wide transcriptional responses, major defenserelated pathways and phenotypic responses in terms of glucosinolate levels revealed that plant responses were not consistently influenced by the degree of specialization. In summary, given the methodological issues with testing the generalist–specialist hypothesis, it is premature to draw any strong conclusions about differential induction based on host plant specialization (Box 2). To address whether alternative categorizations (irrespective of specialization) of herbivores, namely feedingguild (e.g. chewers versus phloem-feeders), can have consistent predictive value for differential induction, we reviewed the recent literature (Table 2). Indeed, it has been widely suggested that depending on the feeding mode of a herbivore, different plant responses will be induced, resulting in the activation of different plant defense mechanisms [26,27]. Many studies have suggested the involvement of salicylic acid (SA) in defense against phloemsucking insects [27,28], whereas chewing larvae (mainly Lepidopterans) are often shown to cause extensive tissue damage and JA and ethylene (ET) induction [29,30]. Of the 13 studies that directly compared chewers and suckers, there was a strong trend for phloem-feeding insects to induce fewer genes associated with the JA pathway, whereas the chewers induced fewer genes associated with the SA pathway. This is consistent with the prediction that phloem-feeding herbivores, such as aphids, leafhoppers and whiteflies, cause only minor tissue damage and induce defense signaling pathways resembling those activated against pathogens (SA regulated) [27,31,32]. A second emerging trend is that phloem-feeders cause a less drastic, more subtle response in the plant. Often they suppress more genes than the chewing herbivores (e.g. [23,29,33]), suggesting that they minimize the activation of plant defenses. Again (Table 1), we found few studies linking observations of plant responses to herbivore performance [34,35]. An exception is a study that compared adult potato aphids (Macrosiphum euphorbiae) and beet armyworm caterpillars attacking tomato (Solanum lycopersicum) 6


[35], where alongside experiments that compared transcriptional and chemical responses of the plant, bioassays were conducted on the caterpillars. Aphid feeding changed the level of expression of 2.8 times more plant genes than caterpillar feeding, downregulating significantly more genes, and yet increasing the expression of fewer herbivore defense-related genes (and secondary metabolites). Accordingly, caterpillars were heavier and had a lower mortality on aphid-damaged plants compared with controls, but weighed less and had increased mortality on plants previously damaged by caterpillars compared with controls [35]. By linking plant responses to herbivore performance, the authors provided evidence of aphids minimizing the magnitude of induction by reducing the ability of the plant to respond to caterpillar feeding. Who is in charge: insects or plants? Interpreting which party is ‘in charge’ is of crucial importance when attempting to understand the induction of the plant defense response by specialists and generalists. For example, observing a minimal induced response might be adaptive for the plant because a sequestering herbivore benefits from plant toxins (Figure 2). However, this response might be adaptive for a generalist insect that suppresses the potentially harmful defenses of the plant. It is important to consider the respective qualities of the herbivore (e.g. sequestering or stealthy) and the consequences of a given plant response in the context of each herbivore in order to distinguish the roles of a plant response. Moreover, the fitness impact of each herbivore is likely to dictate the extent to which a coevolutionary process is likely between any two herbivores. We suggest that one way to specifically address this problem of the response being interpreted as beneficial to different parties is to include an extra control treatment in induction studies. In particular, treatments that provide a baseline for induction in the absence of herbivore-specific cues allow for greater interpretation of differential induction by different herbivores. Such controls can involve: (i) mechanical damage, typically realistic maceration of leaves, exactly matching the amount of maceration to that in real herbivory treatments, and with treatments that span the timing of real herbivory; (ii) a JA or other phytohormone treatment; or (iii) insect manipulations that reduce the salivary activity of the herbivore (e.g. ablation of the spinnerets). For example, in a study comparing the transcript profiles after insect herbivory, wounding and response to JA, SA and ET in Boechera divaricarpa (Brassicaceae), analyses revealed that responses to the specialist diamondback moth (P. xylostella) were determined by effects associated with the ET and SA pathways, whereas responses to the generalist cabbage looper (T. ni) were determined by the ET and JA pathways [36]. Mechanical damage induced all three pathways, yet was dominated by a JA effect. Thus, each herbivore appears to elicit a distinct response from mechanical damage. Another study investigated specificity in induction patterns of chemical defenses from plants damaged by a sequestering specialist herbivore (turnip sawfly, Athalia rosae), a generalist herbivore (fall armyworm, Spodoptera frugiperda) or mechanical wounding (cork borer) in


Review


Table 2. Comparisons of plant defense elicitation by chewing versus phloem-feeding insectsa Herbivores Plutella xylostella, Pieris rapae, Spodoptera exigua, Brevicoryne brassicae, Myzus persicae M. persicae, B. brassicae, S. exigua, P. rapae M. persicae, P. rapae

Measure of plant response Transcriptional responses

Results b Chewers upregulated defense-related pathways involving JA signaling, sulfate metabolism and aliphatic glucosinolate biosynthesis. Phloem-feeders downregulated the above [26].

Transcriptional responses, glucosinolates (GS)

Phloem-feeders increased aliphatic GS. Chewers increased indolyl and aliphatic GS (P. rapae did not induce aliphatics) [22].

Phytohormones, transcriptional responses Parasitoid specificity for herbivore-induced plant volatiles (HIPVs)

(Brassicaceae) Brassica nigra

P. xylostella, P. rapae, S. exigua, M. persicae P. rapae, B. brassicae

Phloem-feeders downregulated genes significantly and did not induce detectable changes in SA, JA and ET, whereas chewers induced JA-dependent responses [29]. Parasitoids preferred chewer damaged over phloem-feeder damaged plants [21].

(Fabaceae) Glycine max

Cerotoma trifurcata, Spissistilus festinus

Oxidative enzymes

(Malvaceae) Gossypium hirsutum (Plantaginaceae) Plantago lanceolata

Bemisia tabaci, S. exigua

HIPVs

Dysaphis cf. Plantaginea, Grammia incorrupta, Heliothis virescens Spodoptera littoralis, Rhopalosiphum maidis Macrosiphum euphorbiae, Helicoverpa zea

Secondary metabolites

Plant (Brassicaceae) Arabidopsis thaliana

(Brassicaceae) A. thaliana

(Brassicaceae) A. thaliana (Brassicaceae) A. thaliana

(Poaceae) Zea mays (Solanacae) Lycopersicon esculentum

Transcriptional responses

HIPVs Oxidative enzymes, herbivore performance

(Solanacae) S. lycopersicum

Macrosiphum euphobiae, S. exigua

Transcriptional responses, biochemistry, herbivore performance

(Solanacae) Solanum tuberosum

M. persicae, Leptinotarsa decemlineata

HIPVs, oxylipin synthesis

(Solanacae) N. attenuata

Manduca sexta, S. littoralis, Trichoplusia ni, Myzus nicotianae

Transcriptional responses

Caterpillars induced more genes (JA-dependent), repressed fewer genes (SA dependent), whereas phloem-feeder repressed ET-dependent genes [28]. Phloem-feeders caused increases in the activities of LOX, POD, ascorbate oxidase and PPO, the chewers induced LOX only [66]. Phloem-feeders did not induce volatile emissions or affect the density of pigment glands, whereas chewers strongly induced volatiles [67]. Chewers had stronger effects and upregulated many compounds. Aphids mainly downregulated compounds [23]. Chewers induced many volatiles, whereas aphids induced no measurable emissions (even after heavy infestation) [68]. Aphid feeding induced POD and LOX, but had no effect on PI and reduced PPO activities; the chewers induced PPO, PI and LOX, but did not induce POD. Prior aphid feeding had decreased resistance to S. exigua. Prior chewer feeding increased resistance to S. exigua [69]. Aphids changed the expression of more genes than caterpillars, yet caterpillar defense induction was higher (PIs). Prior aphid feeding decreased resistance. Prior chewers increased resistance via JA-regulated genes. Aphid feeding had weak JA pathway responses [35]. Chewers induced fewer genes (no JA-dependent responses), whereas the phloem-feeders induced JA-dependent responses. Volatile signatures and biochemical precursors associated with stress signaling were distinct [70]. Chewers induced JA-dependent genes, whereas the phloemfeeders reduced some JA-dependent genes and increased SA-dependent genes [33].

a

Each comparison is from a single study.

b

Color-coding reflects consistency with the hypothesis that the phloem-feeders induced a weaker defensive response than the chewers (green). Yellow indicates no consistent pattern and red indicates rejection of the hypothesis.

white mustard (Sinapis alba) [37]. Specialist feeding and mechanical damage induced threefold increases of the glucosinolate–myrosinase system, whereas generalist feeding induced up to twofold increases in glucosinolate only. Although these studies did not have replication at the level of specialists and generalists, because of the additional controls, we can speculate that specialists might have different mechanisms based on their strategy to evade (diamondback moth) or sequester (turnip sawfly) the plant’s defenses. Although the herbivore treatments alone in both experiments would have demonstrated differences between the two species, having relative bases of comparison allows for a stronger interpretation. Ultimately a link

between these differential induced responses and the impacts on the herbivores would be needed to assess which parties benefit. One of our major predictions is that generalist herbivores use mechanisms to suppress plant defenses more so than specialists, allowing them to feed on a broad range of species (Figure 2). This hypothesis was advocated some time ago with regards to behavioral trenching, a method by which some generalist herbivores attack plants that exude latex [16]. Generalists that trench were able to feed on a diversity of host plants with latex, whereas generalists that did not trench had poor performance on these same plants. Recent developments confirm other, less visually 7


Review apparent, mechanisms by which generalists can also suppress plant defense [38]. Mandibular glands of the noctuid caterpillar Helicoverpa zea were found to secrete salivary glucose oxidase (GOX) [38,39], which has been implicated as an effector responsible for suppression of defense by eliciting an SA burst (which, in turn, attenuates JA and ET levels). When the ability of caterpillars to introduce GOX to their host plants is removed (via ablating the saliva-producing spinnerets), tobacco (Nicotiana tabacum) plants mount a response that reduces herbivore performance, thus demonstrating a benefit for generalists to reduce the ability of the plant to respond to herbivore attack [40]. A recent survey of GOX levels in 85 species (across 23 families of Lepidoptera) found that highly polyphagous species have relatively higher levels of GOX compared with more specialized species [39]. Thus, the production of GOX as a suppressor of induced plant defenses appears to follow our prediction of generalists being more suppressive of plant defense than specialists. An additional example of generalists suppressing plant defense was found in Arabidopsis plants infested by the phloem-feeding silverleaf whitefly (Bemisia tabaci). Whitefly feeding increased SA-responsive gene transcripts, whereas JA- and ET-dependent pathways were repressed or not modulated [41]. Mutant plants with higher activity of JA defenses or impaired in SA defenses slowed nymphal development, whereas those that activate SA and impair JA increased nymphal development [41]. Thus, generalist whitefly feeding strategies appear to benefit the whiteflies at the expense of plant defense. Given the similarity of this result with that of the generalist potato aphid on tomato discussed above [35], we advocate a critical comparison of plant responses to generalist versus specialist aphids. All of the examples of comparisons thus far have been between the generalist green peach aphid (M. persicae) and the specialist cabbage aphid (B. brassicae) on Brassicaceae, and none have linked plant responses with aphid performance (Table 1). But wait, are specialists not specialists? Given that specialist herbivores share an intimate evolutionary history with their host plants, are specialists more manipulative as herbivores than generalists? The answer to this question is complicated by three issues: (i) specialists can be somewhat tolerant of defenses (and, thus, might not need to be manipulative); (ii) specialists can maximize their fitness in nonobvious ways (e.g. phenology, location of feeding); and (iii) from a coevolutionary standpoint, plants might recognize specialists (particularly those with strong fitness impacts on the plant) and defend appropriately. Of course there are examples of specialists that manipulate their hosts [42,43]. Insect gallers perhaps epitomize highly manipulative specialist herbivores. The conventional view is that gallers reprogram both primary and secondary plant metabolism to their benefit [43]. Indeed, most gallers are highly specialized, more so even then their endophagous (but nongalling) relatives [44]. Thus, specialists can either be highly manipulative or not so manipulative. As discussed above, directly comparing specialist induction to some mechanical damage and to 8


elicitor-free control should aid in addressing this issue. More generally, we are in need of direct contrasts of specialists and generalists, testing whether generalists are more sensitive to particular defenses and, hence, can manipulate them effectively. Plants in charge? Fatty acid amino acid conjugates and beyond Plants are able to perceive a wide range of herbivoreassociated elicitors resulting in the activation of specific plant responses, although the adaptive value of such specificity is unclear [39]. Most elicitors and their respective responses differ from responses to mechanical damage and appear to be restricted to particular plant–insect associations [45]. There have been four documented elicitors produced by insects: b-glucosidase [46], fatty acid amino acid conjugates (FACs) [47], inceptins [48] and caeliferins [49]. The most broadly investigated and described elicitors to date have been FACs from lepidopteran larvae (generalists and specialists) and these constituents (typically obtained from oral secretions or regurgitate) are thought to betray the insects presence (and perhaps identity) to the plant [45,47,50]. The first well-characterized FAC was volicitin [N-(17-hydroxylinolenoyl)-L-glutamine], which was identified from the beet armyworm [47] and induces direct and indirect plant defenses in several plants [51]. FACs (particularly volicitin) have a strong impact on plant hormone levels as well as on the induction of plant volatiles in a variety of plant species, unlike caeliferin and inceptin, two newly identified elicitors that appear to be more restricted in the plants for which they are active [48,52]. We presume that insect elicitors, although potentially harmful to the insect in the plant–herbivore interaction, are produced (and not lost because of natural selection) because they are an essential part of the primary metabolism of the insect. For example, FACs in the noctuid moth Spodoptera litura play an active role in nitrogen assimilation by regulating the amount of glutamine in the larval midgut [53]. A recent FAC screen of 29 Lepidoptera species found that some species do not produce these elicitors [51]. Additional categories of elicitors are combinations of plant and insect constituents, which might be a highly stable mechanism for plant recognition of attack. For example, inceptins are derived from fragments of digested plant tissues. Peptides released from proteolytic fragments of chloroplastic ATP synthase were found in the oral secretions of the fall armyworm (S. frugiperda) [48], thus giving the plant a direct role in the perception of a specific attacker. It is unclear if generalist and specialist herbivores differ in their elicitors. A study has shown that the transcriptional responses of Nicotiana attenuata to attack from two generalist herbivores [the tobacco budworm (Heliothis virescens) and the beet armyworm (S. exigua)] was more similar than that of the tobacco hornworm (M. sexta), which is a specialist herbivore, and that this difference was linked to their FACs (although in this case, the two generalists were closely related and thus shared many traits) [54]. Regurgitates of the generalists were virtually identical [55], whereas that of the specialist differed, lacking volicitin and dominated by fatty acid–glutamic acid


Review conjugates that were not present in the regurgitates of the generalists [56,57]. FACs from the specialist M. sexta are involved in suppressing the nicotine response in tobacco, but do not suppress indirect defensive responses (VOCs), and this has been interpreted as adaptive on the part of the plant [58,11]. It would be interesting to evaluate the degree of specificity of insect recognition in plants and to assess whether plants tend to have more fine-tuned degrees of recognition (e.g. via mechanisms specifically associated with FACs or saliva produced by labial and mandibular glands) for specialists and more broad feedback mechanisms (plant-derived byproducts of herbivore digestion, e.g. inceptins or regurgitants) for generalists. Concluding remarks For plants to ‘be in charge’ we assume that after integrating signals from a given attack they will activate pathways that provide the most defensive response. The predictions of the specialist–generalist paradigm suggest that there can be consistency in herbivore elicitation and plant recognition among different types of attackers. Yet, to date, evidence for distinct groupings of generalists and specialists is not so clear, in part because of methodological limitations. A ubiquitous problem with interpreting the specialist–generalist paradigm is that there are two sides to every story (that of the herbivore and of the plant), there are also potentially different predictions based on the type of specialist (sequestering or not?) and the fact that coevolutionary interactions can modify the dynamics in space and time. Nonetheless, we are optimistic. As detailed in this review, we advocate the use of real species level replication, strong controls and links between measures of plant responses with insect performance. It is premature to kill the specialist–generalist paradigm, but perhaps also too early to celebrate its generality. Acknowledgments We thank Martin Heil, Sergio Rasmann, Andre Kessler, Jennifer Thaler and the Plant-Interactions Group at Cornell for helpful comments and the United States National Science Foundation (DEB-1118783) for financial support.

References 1 Krieger, R.I. et al. (1971) Detoxication enzymes in the guts of caterpillars: an evolutionary answer to plant defenses? Science 172, 579–580 2 Whittaker, R.H. and Feeny, P.P. (1971) Allelochemics: chemical interactions between species. Science 171, 757–770 3 Katsir, L. et al. (2008) Jasmonate signaling: a conserved mechanism of hormone sensing. Curr. Opin. Plant Biol. 11, 428–435 4 Cornell, H. and Hawkins, B. (2003) Herbivore responses to plant secondary compounds: a test of phytochemical coevolution theory. Am. Nat. 161, 507–522 5 Berenbaum, M. et al. (1989) Chemical barriers to adaptation by a specialist herbivore. Oecologia 80, 501–506 6 Adler, L.S. et al. (1995) Genetic variation in defensive chemistry in Plantago lanceolata (Plantaginaceae) and its effect on the specialist herbivore Junonia coenia (Nymphalidae). Oecologia 101, 75–85 7 Zalucki, M.P. et al. (2001) Detrimental effects of latex and cardiac glycosides on survival and growth of first-instar monarch butterfly larvae Danaus plexippus feeding on the sandhill milkweed Asclepias humistrata. Ecol. Entomol. 26, 212–224 8 Agrawal, A.A. and Kurashige, N.S. (2003) A role for isothiocyanates in plant resistance against the specialist herbivore Pieris rapae. J. Chem. Ecol. 29, 1403–1415


9 Harvey, J.A. et al. (2007) Effects of dietary nicotine on the development of an insect herbivore, its parasitoid and secondary hyperparasitoid over four trophic levels. Ecol. Entomol. 32, 15–23 10 Feeny, P. (1976) Plant apparency and chemical defense. Recent Adv. Phytochem. 10, 1–40 11 Kahl, J. et al. (2000) Herbivore-induced ethylene suppresses a direct defense but not a putative indirect defense against an adapted herbivore. Planta 210, 336–342 12 Bernays, E.A. (1986) Diet-induced head allometry among foliagechewing insects and its importance for graminivores. Science 231, 495–497 13 Broadway, R.M. (1995) Are insects resistant to plant proteinase inhibitors? J. Insect Physiol. 41, 107–116 14 Gruden, K. et al. (1998) The cysteine protease activity of Colorado potato beetle (Leptinotarsa decemlineata Say) guts, which is insensitive to potato protease inhibitors, is inhibited by thyroglobulin type-1 domain inhibitors. Insect Biochem. Mol. 28, 549–560 15 Hartmann, T. et al. (2005) Acquisition, transformation and maintenance of plant pyrrolizidine alkaloids by the polyphagous arctiid Grammia geneura. Insect Biochem. Mol. 35, 1083–1099 16 Dussourd, D.E. and Denno, R.F. (1994) Host range of generalist caterpillars: trenching permits feeding on plants with secretory canals. Ecology (Tempe) 75, 69–78 17 Pelser, P.B. et al. (2005) Frequent gain and loss of pyrrolizidine alkaloids in the evolution of Senecio section Jacobaea (Asteraceae). Phytochemistry 66, 1285–1295 18 Bowers, M. and Stamp, N. (1993) Effects of plant-age, genotype, and herbivory on Plantago performance and chemistry. Ecology 74, 1778– 1791 19 Agrawal, A.A. (2000) Specificity of induced resistance in wild radish: causes and consequences for two specialist and two generalist caterpillars. Oikos 89, 493–500 20 Poelman, E.H. et al. (2008) Performance of specialist and generalist herbivores feeding on cabbage cultivars is not explained by glucosinolate profiles. Entomol. Exp. Appl. 127, 218–228 21 van Poecke, R.M.P. et al. (2003) Attraction of the specialist parasitoid Cotesia rubecula to Arabidopsis thaliana infested by host or non-host herbivore species. Entomol. Exp. Appl. 107, 229–236 22 Mewis, I. et al. (2006) Gene expression and glucosinolate accumulation in Arabidopsis thaliana in response to generalist and specialist herbivores of different feeding guilds and the role of defense signaling pathways. Phytochemistry 67, 2450–2462 23 Sutter, R. and Muller, C. (2011) Mining for treatment-specific and general changes in target compounds and metabolic fingerprints in response to herbivory and phytohormones in Plantago lanceolata. New Phytol. 191, 1069–1082 24 Reymond, P. et al. (2004) A conserved transcript pattern in response to a specialist and a generalist herbivore. Plant Cell 16, 3132–3147 25 Rasmann, S. and Turlings, T.C.J. (2008) First insights into specificity of belowground tritrophic interactions. Oikos 117, 362–369 26 Bidart-Bouzat, M.G. and Kliebenstein, D. (2011) An ecological genomic approach challenging the paradigm of differential plant responses to specialist versus generalist insect herbivores. Oecologia 167, 677–689 27 Walling, L.L. (2000) The myriad plant responses to herbivores. J. Plant Growth Regul. 19, 195–216 28 Broekgaarden, C. et al. (2011) Transcriptional responses of Brassica nigra to feeding by specialist insects of different feeding guilds. Insect Sci. 18, 259–272 29 De Vos, M. et al. (2005) Signal signature and transcriptome changes of Arabidopsis during pathogen and insect attack. Mol. Plant Microbe Interact. 18, 923–937 30 Kessler, A. and Baldwin, I.T. (2002) Plant responses to insect herbivory: the emerging molecular analysis. Annu. Rev. Plant Biol. 53, 299–328 31 Moran, P.J. and Thompson, G.A. (2001) Molecular responses to aphid feeding in Arabidopsis in relation to plant defense pathways. Plant Physiol. 125, 1074–1085 32 Kusnierczyk, A. et al. (2007) Transcriptional responses of Arabidopsis thaliana ecotypes with different glucosinolate profiles after attack by polyphagous Myzus persicae and oligophagous Brevicoryne brassicae. J. Exp. Bot. 58, 2537–2552 9


Review 33 Heidel, A. and Baldwin, I. (2004) Microarray analysis of salicylic acidand jasmonic acid-signalling in responses of Nicotiana attenuata to attack by insects from multiple feeding guilds. Plant Cell Environ. 27, 1362–1373 34 Stout, M.J. et al. (1998) Specificity of induced resistance in the tomato, Lycopersicon esculentum. Oecologia (Berlin) 113, 74–81 35 Rodriguez-Saona, C. et al. (2010) Molecular, biochemical, and organismal analyses of tomato plants simultaneously attacked by herbivores from two feeding guilds. J. Chem. Ecol. 36, 1043–1057 36 Vogel, H. et al. (2007) Different transcript patterns in response to specialist and generalist herbivores in the wild Arabidopsis relative Boechera divaricarpa. PLoS ONE 2, e1081 37 Travers-Martin, N. and Mueller, C. (2008) Matching plant defence syndromes with performance and preference of a specialist herbivore. Funct. Ecol. 22, 1033–1043 38 Eichenseer, H. et al. (1999) Salivary glucose oxidase: multifunctional roles for Helicoverpa zea? Arch. Insect Biochem. Physiol. 42, 99–109 39 Erb, M. et al. (2012) Role of phytohormones in insect-specific plant reactions. Trends Plant Sci. 17, This Special issue 40 Musser, R.O. et al. (2002) Caterpillar saliva beats plant defences: a new weapon emerges in the evolutionary arms race between plants and herbivores. Nature 416, 599–600 41 Zarate, S.I. et al. (2007) Silverleaf whitefly induces salicylic acid defenses and suppresses effectual jasmonic acid defenses. Plant Physiol. 143, 866–875 42 Dussourd, D.E. and Eisner, T. (1987) Vein-cutting behavior: insect counterploy to the latex defense of plants. Science 237, 898–900 43 Karban, R. and Agrawal, A.A. (2002) Herbivore offense. Annu. Rev. Ecol. Syst. 33, 641–664 44 Miller, W. (2004) Host breadth and voltinism in gall-inducing Lepidoptera. J. Lepidopterists Soc. 58, 44–47 45 Bonaventure, G. et al. (2011) Herbivore-associated elicitors: FAC signaling and metabolism. Trends Plant Sci. 16, 294–299 46 Mattiacci, L. et al. (1995) Beta-glucosidase: an elicitor of herbivoreinduced plant odor that attracts host-searching parasitic wasps. Proc. Natl. Acad. Sci. U.S.A. 92, 2036–2040 47 Alborn, H.T. et al. (1997) An elicitor of plant volatiles from beet armyworm oral secretion. Science 276, 945–949 48 Schmelz, E.A. (2006) Fragments of ATP synthase mediate plant perception of insect attack. Proc. Natl. Acad. Sci. U.S.A. 103, 8894– 8899 49 Alborn, H.T. et al. (2007) Disulfooxy fatty acids from the American bird grasshopper Schistocerca americana, elicitors of plant volatiles. Proc. Natl. Acad. Sci. U.S.A. 104, 12976–12981 50 Mori, N. and Yoshinaga, N. (2011) Function and evolutionary diversity of fatty acid amino acid conjugates in insects. J. Plant Interact. 6, 103– 107 51 Yoshinaga, N. et al. (2010) Fatty acid-amino acid conjugates diversification in lepidopteran caterpillars. J. Chem. Ecol. 36, 319–325 52 Schmelz, E.A. et al. (2009) Phytohormone-based activity mapping of insect herbivore-produced elicitors. Proc. Natl. Acad. Sci. U.S.A. 106, 653–657 53 Yoshinaga, N. et al. (2008) Active role of fatty acid amino acid conjugates in nitrogen metabolism in Spodoptera litura larvae. Proc. Natl. Acad. Sci. U.S.A. 105, 18058–18063 54 Voelckel, C. and Baldwin, I.T. (2004) Generalist and specialist lepidopteran larvae elicit different transcriptional responses in Nicotiana attenuata, which correlate with larval FAC profiles. Ecol. Lett. 7, 770–775

10


55 Pohnert, G. et al. (1999) New fatty acid amides from regurgitant of Lepidopteran (Noctuidae, Geometridae) caterpillars. Tetrahedron 55, 11275–11280 56 Halitschke, R. et al. (2001) Molecular interactions between the specialist herbivore Manduca sexta (Lepidoptera, Sphingidae) and its natural host Nicotiana attenuata. III. Fatty acid-amino acid conjugates in herbivore oral secretions are necessary and sufficient for herbivore-specific plant responses. Plant Physiol. 125, 711–717 57 Alborn, H.T. et al. (2003) Differential activity and degradation of plant volatile elicitors in regurgitant of tobacco hornworm (Manduca sexta) larvae. J. Chem. Ecol. 29, 1357–1372 58 Winz, R.A. and Baldwin, I.T. (2001) Molecular interactions between the specialist herbivore Manduca sexta (Lepidoptera, Sphingidae) and its natural host Nicotiana attenuata. IV. Insect-induced ethylene reduces jasmonate-induced nicotine accumulation by regulating putrescine N-methyltransferase transcripts. Plant Physiol. 125, 2189–2202 59 Cornelius, M.L. and Bernays, E.A. (1995) The effect of plant chemistry on the acceptability of caterpillar prey to the argentine ant iridomyrmex humilis (hymenoptera: formicidae). J. Insect Behav. 8, 579–593 60 Poelman, E.H. et al. (2008) Early season herbivore differentially affects plant defence responses to subsequently colonizing herbivores and their abundance in the field. Mol. Ecol. 17, 3352–3365 61 Traw, M.B. and Dawson, T.E. (2002) Differential induction of trichomes by three herbivores of black mustard. Oecologia 131, 526– 532 62 Mooney, E.H. et al. (2009) Differential induced response to generalist and specialist herbivores by Lindera benzoin (Lauraceae) in sun and shade. Oikos 118, 1181–1189 63 Stamp, N.E. and Bowers, M.D. (1994) Effects of cages, plant age and the mechanical clipping on plantain chemistry. Oecologia (Berlin) 99, 66–71 64 Diezel, C. et al. (2009) Different Lepidopteran elicitors account for cross-talk in herbivory-induced phytohormone signaling. Plant Physiol. 150, 1576–1586 65 Zong, N. and Wang, C-Z. (2007) Larval feeding induced defensive responses in tobacco: comparison of two sibling species of Helicoverpa with different diet breadths. Planta 226, 215–224 66 Felton, G.W. et al. (1994) Oxidative responses in soybean foliage to herbivory by bean leaf beetle and three-cornered alfalfa hopper. J. Chem. Ecol. 20, 639–650 67 Rodriguez-Saona, C. et al. (2003) Volatile emissions triggered by multiple herbivore damage: beet armyworm and whitefly feeding on cotton plants. J. Chem. Ecol. 29, 2539–2550 68 Turlings, T.C.J. et al. (1998) The induction of volatile emissions in maize by three herbivore species with different feeding habits: possible consequences for their natural enemies. Biol. Control 11, 122–129 69 Stout, M.J. et al. (1998) Effect of nitrogen availability on expression of constitutive and inducible chemical defenses in tomato, Lycopersicon esculentum. J. Chem. Ecol. 24, 945–963 70 Gosset, V. et al. (2009) Attacks by a piercing-sucking insect (Myzus persicae Sultzer) or a chewing insect (Leptinotarsa decemlineata Say) on potato plants (Solanum tuberosum L.) induce differential changes in volatile compound release and oxylipin synthesis. J. Exp. Bot. 60, 1231–1240 71 Schoonhoven, L.M. et al. (2005) Insect-Plant Biology, Oxford University Press 72 Bernays, E.A. and Graham, M. (1988) On the evolution of host specificity in phytophagous arthropods. Ecology 69, 886–892

Review


The specificity of herbivore-induced plant volatiles in attracting herbivore enemies Andrea Clavijo McCormick, Sybille B. Unsicker and Jonathan Gershenzon Department of Biochemistry, Max Planck Institute for Chemical Ecology, Hans-Kno¨ll Strasse 8, D-07745 Jena, Germany

Plants respond to herbivore attack by emitting complex mixtures of volatile compounds that attract herbivore enemies, both predators and parasitoids. Here, we explore whether these mixtures provide significant value as information cues in herbivore enemy attraction. Our survey indicates that blends of volatiles released from damaged plants are frequently specific depending on the type of herbivore and its age, abundance and feeding guild. The sensory perception of plant volatiles by herbivore enemies is also specific, according to the latest evidence from studies of insect olfaction. Thus, enemies do exploit the detailed information provided by plant volatile mixtures in searching for their prey or hosts, but this varies with the diet breadth of the enemy. Plant volatiles from an herbivore enemy perspective Plants emit complex blends of volatile compounds from many of their organs, and there is considerable variation in volatile composition among species and other taxonomic levels [1]. For these reasons, volatiles are commonly used by herbivores as cues in choosing host plants. However, as well as herbivores, organisms at the next trophic level, herbivore enemies, also use plant volatiles as cues. Two types of herbivore enemy, predatory arthropods and parasitoids, are known to exploit plant odors to find plants harboring their prey and hosts [2–4]. These enemies take advantage of the fact that, after herbivory, many plants emit blends of volatiles differing in both quantity and composition from those emitted before herbivory. Over 50 different species of plants are reported to produce distinct blends of herbivore-induced volatiles, and these are known to attract a range of herbivore enemies, including predators and parasitoids drawn from five insect orders, plus predatory mites, nematodes and birds [3,4]. Much of the research on volatile-mediated attraction of herbivore enemies to damaged plants has centered on the ecological and evolutionary significance of volatile emissions to plants. There has been an ongoing debate in the literature about whether volatiles increase plant reproductive fitness by attracting herbivore enemies, and so act as defenses [2,5,6] (Box 1), or instead play other roles in the lives of the plants that produce them [7,8]. By contrast, herbivoreassociated plant volatiles have always been assumed to have Corresponding author: Gershenzon, J. ([email protected]).

real value for herbivore enemies, although this has rarely been subject to detailed analysis. Over 20 years ago, L.E.M. Vet and M. Dicke [9] proposed that plant volatiles are useful cues for herbivore enemies at medium to long range because they can be detected more readily than volatile cues emitted from herbivores. The authors reasoned that plants usually have a greater biomass than their herbivores and typically emit volatiles systematically from a greater area than that actually damaged. Shortly thereafter, Dicke and colleagues also considered the potential of volatiles to provide specific information to enemies about the prey and hosts present [10]. Yet, knowledge of the information content of volatile blends has advanced little since then. It is sometimes asserted that the composition of the volatile blend might furnish specific details for enemies about the kinds of herbivore(s) present and their age, developmental state and abundance, but this possibility is not easy to demonstrate. In this review, we consider the specificity of herbivoreinduced plant volatiles as cues for herbivore enemies in light of the latest results on enemy behavior and insect olfaction. We cover specificity at three stages of enemy attraction: in the blend of volatiles emitted from the plant, in the response of enemies to volatiles, and in how enemies perceive volatiles. Our goal is to understand whether the complex volatile mixtures emitted from plants under herbivore attack have significant information value to herbivore enemies. Specificity in plant volatile emission Given the enormous variety of substances reported in volatile mixtures [1,11], it is easy to imagine that each plant species could emit a distinct blend of volatile compounds, and thus be recognizable to herbivores and their enemies. However, perusal of the major herbivore-induced volatiles shows that the same constituents (Box 2) are released by most plant species, irrespective of their taxonomic affinities. For example, the monoterpenes (E)-b-ocimene and linalool, the sesquiterpenes (E, E)-a-farnesene and (E)-b-caryophyllene, the C11 homoterpene (E)-4,8-dimethyl-1,3,7-nonatriene (DMNT), and the fatty acid derivatives known as green leaf volatiles (GLVs), including (Z)-3-hexen-1-ol and (Z)-3-hexenyl acetate, are frequent components of volatile blends released after herbivore damage from a wide range of plant species [12–17]. Nevertheless, the relative amounts of these substances vary greatly among species and there are

1360-1385/$ – see front matter ß 2012 Elsevier Ltd. All rights reserved. doi:http://dx.doi.org/10.1016/j.tplants.2012.03.012 Trends in Plant Science, May 2012, Vol. 17, No. 5

303

Review Box 1. Herbivore enemy attraction to volatiles from the plant perspective: ecological and evolutionary ambiguities Despite all the reports on volatile-mediated attraction of herbivore enemies to damaged plants, its ecological and evolutionary significance for plants is still uncertain. Unlike for herbivore enemies, where exploitation of volatile cues attracting them to their prey or hosts is assumed to bring benefits, there has been no clear proof of whether volatiles increase plant reproductive fitness and so can actually be considered a defense [2,6,5]. In part, this is because most studies have been carried out under laboratory conditions or with agricultural systems where true evolutionary inference is not possible. Field studies have demonstrated that herbivore-induced volatiles increase the attraction of herbivore enemies [93,94], and this can result in decreased herbivore survival and performance [95], but there is still no evidence that volatile emission positively affects plant fitness. Another difficulty in showing the fitness benefits of herbivore enemy recruitment to plant volatiles is that emission has also been reported to have additional roles for plants in direct defense, internal signaling or communication with neighboring plants [8,11,96,97]. These may complement or conflict with a role in herbivore enemy attraction, and so it is hard to assign precise costs or benefits to volatile emission. Furthermore, there is some doubt as to the effectiveness of one type of enemy frequently attracted by herbivore-mediated volatiles: koinobiont parasitoids. These do not kill their herbivore hosts immediately, but wait until they are about to pupate or become an adult. Koinobiont parasitoids might even decrease the fitness of individual plants by stimulating herbivores to feed more [98], but could benefit plants by reducing future herbivore populations. Finally, recent research at the community level reveals an unexpected number of plant volatile-mediated interactions among species at various trophic levels causing further complications in assessing the fitness benefits of volatiles. [99,100].

typically many differences in less abundant compounds that could contribute to specificity. If these differences are perceived by herbivore enemies, they could facilitate species recognition. Within a single species, plant volatile emission can vary with the herbivore present, as noted many years ago by Dicke and colleagues [10]. This variation provides herbivore enemies with valuable information on the identity of prey or hosts available on a plant and their feeding guilds [18–21]. For example, Brassica rapa (turnip rape) plants damaged by the root herbivore Delia radicum (cabbage rootfly) emit a distinct blend of volatiles from their aboveground tissue that differs significantly from the blend that is released when the plants are attacked aboveground by caterpillars of Pieris brassicae (large cabbage white butterfly) [22]. 4-Methyltridecane and salicylaldehyde are dominant compounds in the blend of D. radicum-damaged plants, whereas methyl salicylate is characteristic for P. brassicae-damaged cabbage. When roots and shoots are attacked simultaneously, the GLV hexyl acetate is released in high relative amounts [22]. The distinctive profiles induced by various herbivores could be caused by specific elicitors in oral secretions applied during the process of feeding. In addition to the well-known fatty acid–amino acid conjugates and b-glucosidases in the oral secretions of lepidopteran larvae, several new elicitors associated with herbivore oviposition and feeding have recently been reported [23–26]. However, the elicitors known so far seem insufficient to explain most of the differences in plant volatile emission patterns. Specificity in 304


elicitor recognition could arise from differential induction of defense-signaling pathways and associated phytohormones [27]. In general, leaf-chewing insects activate jasmonic acid-dependent signaling, whereas phloem feeding induces salicylic acid-dependent signaling, which sometimes antagonizes jasmonic acid signaling [28,29]. More research on hormones, signaling pathways, new elicitors and other factors affecting the biosynthetic pathways of volatile emission is required to judge what drives specificity in plant volatile emission. Further evidence of specificity in volatile release comes from plants attacked by different developmental stages of the same insect herbivore species [30,31]. For instance, larvae of Plagiodera versicolora (the willow leaf beetle) trigger young Salix eriocarpa (wollypod willow) trees to emit six out of 17 detected volatile compounds in significantly higher amounts than after adult beetle herbivory. Larval feeding also results in higher overall emission rates than does adult beetle damage [31]. Egg deposition has also been reported to affect volatile blends [24,32,33], with egg-induced volatile blends distinct from those induced by larval feeding [34]. The volatile blend released from plants can also give valuable information on the number of herbivores that are currently feeding on a plant because the rate of emission of particular compounds is often positively correlated to the amount of inflicted damage [35,36]. In addition, volatile release can even indicate whether herbivores have already been attacked by parasitoids and the identity of the attacking species [37], which could represent valuable information for other parasitoids. Despite these impressive examples of specificity in the herbivore-induced emission of plant volatiles, there are several reports in which different herbivore species, feeding guilds, developmental stages and number of attackers were not found to alter volatile emission significantly [6,38,39]. For example, the spectrum of Nicotiana attenuata (coyote tobacco) volatiles induced by herbivory from a lepidopteran, Manduca quinquemaculata (five-spotted hawkmoth), a beetle, Epitrix hirtipennis (tobacco flea beetle) and an hemipteran, Tupiocoris notatus (suckfly) is similar with compounds being released in only slightly different proportions [39]. To judge whether herbivore-induced emission is ultimately specific enough to be useful to herbivore enemies, experimental methods have to be improved. A more thorough chemical analysis of plant volatile mixtures is necessary that includes even minor compounds, because the abundance of individual compounds may not be correlated with their information value. In addition, explicit statistical procedures are needed to establish whether blends differ significantly. These must take into account the fact that abundance and composition change at the same time, that compounds derived from the same biosynthetic pathway may be autocorrelated, and that data are often non-normally distributed and heteroscedastic [40]. A major improvement in analyzing induced volatile emission would come if collections were made in the field to assess how blend composition is affected by the typical biotic and abiotic factors prevailing there. Most herbivore-induced volatile collections have been carried out under the controlled conditions of the laboratory

Review Box 2. Plant volatiles associated with herbivore enemy attraction Plants produce a large range of metabolites that are volatile because of their high vapor pressure under standard conditions. Aside from simple gases, such as O2, CO2, water vapor and ethylene, over 1700 volatiles are reported from plants [1], but only a fraction of these are emitted by individual plants after herbivore damage. These can be grouped into four categories. Terpenes Comprising the largest class of plant volatiles, terpenes or terpenoids are classified by the number of branched C5 units in their structures. Major terpene volatiles emitted from vegetative tissue include the C5 compound isoprene (one C5 unit), C10 monoterpenes, such as (E)-b-ocimene and linalool (two C5 units), and C15 sesquiterpenes, such as (E)-b-caryophyllene and (E,E)-a farnesene (three C5 units). Two terpenes that occur frequently after herbivore damage have irregular structures, the C11 (E)-4,8dimethyl-1,3,7-nonatriene (DMNT) and the C16 (E,E)-4,8,12-trimethyl- 1,3,7,11-tridecatetraene (TMTT). Fatty acid derivatives The oxidation of fatty acids leads to the formation of a large family of volatile derivatives. After herbivore damage, sequential lipoxygenase and hydroperoxide lyase action results in the production of C6 compounds, such as (E)-2-hexenal, (Z)-3-hexenal, (Z)-3-hexen-1ol and (Z)-3-hexenyl acetate, called GLVs because they impart the typical odor of green leaves. Aromatic compounds The metabolism of phenylalanine leads to a group of compounds with simple aromatic rings and C1–C3 side chains, whereas an offshoot of tryptophan biosynthesis leads to indole derivatives. The most important representatives of this group after herbivore damage are methyl salicylate and indole. Amino acid derivatives After herbivore damage, various amines, oximes, nitriles, isothiocyanates and sulfides are released that are produced from amino acids. These compounds are often not as well recovered in standard headspace collections as terpenes, GLVs and aromatics, and may be more abundant than is currently realized. For chemical structures and more information on the chemistry and biochemistry of herbivore-induced volatiles, several good references are available [101,102].

or greenhouse, where demonstration of specificity is only inferential. Ultimately, the specificity of volatile blends for herbivore enemy attraction must be determined by the enemy and not the plant. In the following two sections, we focus on the behavior of herbivore enemies and their perception of plant volatiles. Specificity in exploitation of plant volatile cues by herbivore enemies We have discussed above that herbivore-damaged plants may release distinct volatile blends that vary with the plant species, the herbivore species, and the age, stage and number of herbivores present. Do herbivore enemies take advantage of these cues to find the prey or hosts they are seeking? A review of the literature reveals many examples in which herbivore enemies use herbivore-induced plant volatiles to discriminate among different plant species [41,42], different cultivars or varieties of the same species [43,44] and even different phenological states of the plant [45,46]. The behavior of enemies is also influenced by qualitative changes in volatile profiles caused by different types [47,48] and growth stages of herbivores [30], and by quantitative


changes in volatiles associated with varying host or prey density [35,49]. It can be inferred that this behavior helps enemies to find prey and hosts. Although herbivore enemies are ultimately seeking herbivores rather than plants, their attraction to plants has been rationalized on the basis of the detectability and reliability of volatile cues [9]. Volatile signals from herbivore-induced plants, which are often systemically emitted, are stronger and easier to detect than are signals from the arthropod prey or the hosts themselves, even if they are not as reliable. Enemies able to perceive the cues released by the plant upon attack can optimize their foraging efficiency by concentrating their searching behavior [50]. Whether differential attraction towards plant volatiles is directly connected to increased herbivore enemy fitness is so far unknown, but it has been reported that the species, genotype or nutritional status of plants upon which herbivores develop can have a great impact on enemy performance [51–55]. For example, parasitization of herbivores that feed on plants with a high level of defense compounds, such as alkaloids, iridoid glycosides or furanocoumarins, may have adverse effects on the growth or survival of the parasitoid, and so parasitoids may be under selection to avoid chemically well-defended plants in host searching and to be differentially attracted by volatiles to plants with lower levels of defense compounds [55]. By contrast, some parasitoids perform better on hosts feeding on chemically well-defended plants, perhaps because the ability of the host to encapsulate the parasitoid is compromised [56,57]. Although there is little evidence for the ability of herbivore enemies to discriminate among plants based on their nutritional value to herbivores, enemies are sometimes observed to prefer healthy, infested plants compared with those under abiotic stress [58,59], and have reduced attraction to plants that are simultaneously infested by other herbivores or pathogens [60–62]. For example, predatory mite (Phytoseiulus persimilis) attraction to lima bean (Phaseolus lunatus) infested with spider mites (Tetranychus urticae) was reduced when the plants were also infested with tobacco whitefly (Bemisia tabaci), which reduced monoterpene emission [21]. The attraction of the hymenopteran parasitoid Diadegma semiclausum to the P. rapae-induced volatiles of Arabidopsis thaliana is diminished by the presence of methyl salicylate, which might serve as an indicator of salicylate-signaling arising from coinfestation of a pathogen or phloem-feeder [63]. Herbivore enemies living below ground show similar behavior. Attraction of the entomopathogenic nematode, Heterorhabditis megidis, to its prey, the larvae of Diabrotica virgifera virgifera (the western corn rootworm) feeding on maize roots is reduced by shoot herbivory, which decreases emission of the principal volatile, the sesquiterpene (E)-b-caryophyllene [60]. The presence of additional, non-target herbivores on a plant may decrease its attractiveness for enemies owing to induction of defenses or nutrient reallocation, which can negatively impact the development of target herbivores. It seems likely that the use of plant volatile cues by herbivore enemies to choose plants for prey or host searching has fitness value. However, studies on the attraction of herbivore predators and parasitoids to volatile cues rarely 305

Review incorporate measurements of growth, development or other aspects of performance. Thus, more research is needed on the relationship between enemy response to volatile cues and performance (e.g. [64]) before one can appreciate the ecological and evolutionary significance of this behavior. The way in which herbivore enemies exploit plant volatile cues to locate their prey or hosts depends, at least for parasitoids, on their degree of host specificity [65,66]. For example, the parasitoid Cardiochiles nigriceps, which is specialized on the noctuid Heliothis virescens (tobacco budworm), can discriminate between plant volatiles emitted after host feeding versus feeding of a non-host, the closely related Helicoverpa zea (corn earworm) [67]. By contrast, the generalist parasitoid Campoletis chlorideae was equally attracted to the volatiles emitted by the feeding of two other noctuid moth larvae, Helicoverpa armiguera (cotton bollworm) and Pseudaletia separata (oriental armyworm), even though the composition of the two blends was not the same [68]. Among aphid enemies, the specialist parasitoid Aphidius ervi recognizes its host Acyrthosiphon pisum (pea aphid) over the non-host Aphis fabae (bean aphid) on Vicia faba (broad bean), whereas a generalist aphid parasitoid Diaeretiella rapae did not discriminate between two aphid species based on plant volatiles alone [69]. It has already been mentioned that plants attacked by different developmental stages of the same herbivore can release different blends of volatiles. These differences appear to be exploited more by specialists than by generalists. For example, the specialist parasitoid Cotesia kariayi was more attracted to maize plants fed upon by earlyinstar larvae of P. separata, which are suitable for parasitization, than to plants fed upon by late-instar larvae, which are less suitable for parasitization [30]. A similar situation was observed for the specialist predator Aiolocaria hexaspilota [31]. By contrast, the generalist parasitoids Microplitis rufiventris and Cotesia glomerata, which also attack early stages of their hosts, were unable to discriminate between different instars based on plantemitted volatiles [70,71]. Within a plant, parasitoids may also use volatile information to locate suitable hosts more precisely and, once again, specialists exploit this cue more than generalists do. Females of the specialist parasitoid, Cotesia rubecula, were observed to land more frequently on leaves infested with unparasitized caterpillars of their host Pieris rapae (lesser cabbage white) than on leaves infested with parasitized caterpillars, whereas the generalist parasitoid C. glomerata could not make this discrimination [72]. Taken together, these examples demonstrate that the ability of parasitoids to exploit plant-derived volatiles as cues to locate their hosts is higher when host ranges are narrow. Nevertheless, other factors such as habitat complexity and competition might be important selective pressures on parasitoid behavior, and more experiments in this area are warranted. Specificity in herbivore enemy perception of volatiles Herbivore enemies can only respond to variations in plant volatile blends if their sensory apparatus can perceive the 306


differences among them. Thus, it is important to understand how enemies detect volatiles and discriminate among different types and quantities of compound. Olfactory perception in arthropods occurs when volatile compounds bind to olfactory receptors in the antennae, maxillary palps or legs. One or few olfactory receptors are expressed in each olfactory neuron, and neurons bearing similar receptors converge in distinct structures called glomeruli in the olfactory bulb or antennal lobe, which are functional units for odorant coding and processing [73,74]. The number of such glomeruli is variable among insect taxa, with approximately 40 in the fruit fly Drosophila melanogaster [75], approximately 160 in the honeybee Apis mellifera [76], between 14 and 21 in the predatory mite P. persimilis [74] and approximately 190 in the parasitoid wasps C. glomerata and C. rubecula [77]. Thus, at least some herbivore enemies have complex machinery for olfactory perception. By extrapolating from studies with herbivores, herbivore enemies are likely to employ olfactory cues to locate the food plants of their prey and hosts via two different modes of perception (Figure 1). In one mode, called ‘speciesspecific odor recognition’, single compounds that are characteristic of a certain plant species or a group of related species are used for plant recognition, whereas in a second mode, called ‘ratio-specific odor recognition’, enemies detect a set of plant volatiles ubiquitous to many families and are tuned to discriminate differences in ratios among these compounds to find their hosts [78]. Evidence for species-specific odor recognition in herbivores has been obtained from studies with herbivores specialized on Brassicaceae [79]. These insects are attracted

[(Figure_1)TD$IG]

? Attractive blend

Non-attractive blend

Species-specific odor recognition

Presence versus absence 4

: 6

Ratio-specific odor recognition 3

Whole-blend odor recognition

: 4

: 1

versus : 4

: 4

: 1

versus


Figure 1. Possible modes of odor discrimination by herbivore enemies. Mechanisms of odor discrimination are depicted using colored symbols to represent individual volatile compounds in an odor blend. In species-specific odor recognition, discrimination is based on compounds restricted to a single species or group of related species. In ratio-specific odor recognition, discrimination is based on the ratios of blend compounds. In whole-blend odor recognition, discrimination is based on the entire blend or many of its components perceived as a whole.

Review to isothiocyanates, hydrolysis products of the characteristic glucosinolate metabolites of the Brassicaceae that are sometimes volatile, and have specific olfactory receptor neurons tuned to isothiocyanate perception. Among herbivore enemies associated with the Brassicaceae, species-specific odor recognition might also occur, because the specialist parasitoid Diaeretiella rapae, a braconid wasp that attacks aphids living on Brassicaceae, displays electroantennogram (EAG) responses to various isothiocyanates and is behaviorally attracted to them [69,80]. It was suggested that D. rapae has specific receptors for isothiocyanates and uses these compounds as cues to locate its hosts. Interestingly, isothiocyanate receptors are also present in herbivores that do not use Brassicaceae as hosts [81], suggesting the role of single compound recognition not only in attracting herbivores to hosts, but also in avoidance of non-host plants, which may make host searching more efficient. For herbivore enemies, the role of repellents has so far received less attention, but might also play a crucial role in host selection. For example, an investigation of the parasitoid Diadegma semiclausum, an ichneumonid wasp, revealed that the volatile compound isoprene is perceived by olfactory receptors and acts in a repellent manner [82]. Where herbivore enemies do not use plant-specific compounds as attractants or repellents, the second mode of host perception, ratio-specific odor recognition, may be dominant (Figure 1). EAG and gas chromatography-electroantennographic detection (GC-EAD) have shown that most enemies tested respond to a comparable list of widely occurring plant volatiles (Box 2), including linalool, DMNT, (E, E)-a-farnesene, (E)-2-hexenal, (Z)-3-hexen-1-ol and, (Z)-3-hexenyl acetate, but to different extents [83–87]. Thus, their ability to discriminate among odor blends is not likely to be due to the presence or absence of individual substances, but instead to the differences in the relative proportions of individual constituents. In addition, enemies such as hymenopteran parasitoids have different sensitivities to volatiles, and do not necessarily respond most strongly to the most abundant compounds. Differences in relative perception have been linked to the range of host choice, with generalist parasitoids having stronger GC-EAD responses to compounds emitted by a large variety of species immediately after damage, such as the green leaf volatiles, and specialist parasitoids responding more strongly to volatiles emitted several hours after damage [84,87,88]. Ratio-specific odor recognition of a plant by an herbivore enemy is well illustrated by the case of Closterocerus ruforum, an egg parasitoid of the pine sawfly Diprion pini [89]. The parasitoid is more attracted to pine foliage on which D. pini eggs have been laid than to foliage on which no eggs have been laid. The volatile blend of the egg-laden foliage differs from that of the no-egg foliage only in an increased amount of (E)-b-farnesene, but this sesquiterpene alone is not attractive. However, (E)-b-farnesene is attractive when present as a mixture with five other terpenes [b-phellandrene, (E)- and (Z)-b-ocimene, (E)-bcaryophyllene, and a-humulene], whose amounts are not increased on D. pini oviposition. The mixture of these five compounds plus (E)-b-farnesene was significantly attractive to the parasitoid, but only when the ratios mimicked those emitted from egg-laden pine foliage. The parasitoid


Cotesia vestalis is also attracted to a complex mixture of volatiles. In this case, a response requires four herbivoreinduced volatiles [n-heptanal, a-pinene, sabinene, and (Z)3-hexenyl acetate] present in ratios similar to those emitted by a cabbage plant infested with diamondback moth (Plutella xylostella) and presented against a background of non-infested cabbage [90]. None of the compounds alone is attractive. These results suggest that even when there are one or a few key attractive compounds, these may only be active as host-finding cues when present in natural ratios against a background of other volatiles emitted by the plant. A study involving an herbivore predator, the mite P. persimilis, shows that mixtures are also important to this class of herbivore enemies. When tests were performed involving several major herbivore-induced lima bean volatiles, (E)- and (Z)-b-ocimene, (Z)-3-hexenyl acetate, DMNT, TMTT and methyl salicylate, only the last of these compounds was attractive when presented alone. However, a mixture of all five was more attractive than the individual compounds or partial mixtures, as long as it was tested with background odor from lima bean without herbivory. Interestingly, the GLV (Z)-3-hexenyl acetate that was repellent when tested alone, was a key component of the attractive blend [91,92]. The results suggest that predatory mites perceive plant volatiles not as a mixture of individual attractive and repellent compounds, but as a synthetic whole blend (Figure 1). Thus, in future behavioral studies with herbivore enemies, individual plant volatile compounds should be tested not only alone, but also as part of mixtures and in the context of background odors. In addition, the perception of Box 3. The importance of naı¨ve versus learned responses The value of plant volatiles as cues for herbivore enemies may depend on whether enemies respond innately or only after learning. All herbivore enemies are likely to have fixed naı¨ve responses towards a subset of odors that are most relevant for host or habitat location [9]. Naı¨ve responses have been hypothesized to be associated with highly reliable cues, such as host or host-related odors (i.e. pheromones, honeydew or frass) [9,103], but there is accumulating evidence that some parasitoid or predator species also use naı¨ve responses in association with plant odors [41,104–107]. Associative learning of odors is also widespread among both classes of herbivore enemy, parasitoids and predators [108–110]. Rewarding experiences, such as successful oviposition (parasitoids) or prey capture (predators), in association with plant odors, have a positive impact on foraging behavior by enhancing innate or naı¨ve attraction to the compounds involved or generating attraction to blends that are not attractive to inexperienced individuals. Unrewarding experiences do not seem to affect naı¨ve responses or produce aversive learning [47,110], but encountering sequential unrewarding and then rewarding experiences enhances the learning value of positive associations [111,112]. Initial predictions assumed that learning of plant odors would have a high adaptive value for generalist predators and parasitoids. By contrast, specialized species were expected to exhibit a naı¨ve response to odors and have little learning abilities [9]. However, it now appears that all herbivore enemies make use of both naı¨ve and learned responses, and their relative importance cannot be explained solely on the basis of dietary specialization [50]. Furthermore the great variation in learning rate and memory formation between closely related parasitoid species [65,113,114] hints that the trade-off between learned and naı¨ve responses reflects adaptations to specific ecological constraints [115]. 307

Review plant volatiles may not be static for a single insect, given that enemies may be able to learn to associate odors with prey or hosts (Box 3). Concluding remarks and outlook Ever since the first reports of the attraction of herbivore enemies to volatile blends released from plants after herbivore damage, researchers have marveled at the complexity of these blends and wondered whether blend composition could be an informative cue for enemies in host and prey searching. Here, we highlighted the potential information content of volatile blends for herbivore enemies as well as the ability of herbivore enemies to detect them. Enemies not only detect individual major and minor compounds of herbivore-induced blends, but also respond to the blend as a whole depending on the ratios of the components present. The sensory capabilities of herbivore enemies should facilitate their use of complex plant volatile information in searching for prey and hosts, as documented in many reports in the literature. The greater use of plant volatiles by specialist as opposed to generalist enemies is consistent with the greater need for detailed information in locating specific target prey or hosts. To further understand the value of herbivore-induced volatiles in herbivore enemy attraction, it is essential to discover whether herbivore enemy behavior is linked with performance: Can the response to plant volatiles be directly correlated with increased success in prey and host searching and, ultimately, reproductive fitness? It is also important to learn whether the plants releasing the volatiles obtain any fitness benefits for themselves by recruiting herbivore enemies (Box 1). If so, volatile blends may be under selection to maximize signal intensity, detectability, reliability and information content to herbivore enemies. In most reviews of plant–herbivore–herbivore enemy interactions, researchers are encouraged to conduct more experiments under natural conditions to better understand the ecological significance of the results obtained. This should also hold true for investigations of herbivore enemy attraction to plant volatiles, because emission patterns often differ between laboratory and field conditions [3,4,6]. The use of naturally co-occurring species in such studies is also desirable as it would allow inferences to be made about the evolution of herbivore enemy–plant volatile interactions. Additional work with herbivore predators is also called for given that plant volatiles have usually been studied in relation to herbivore parasitoids rather than predators. Because predators kill herbivores immediately, they may be of greater value to plants than koinobiont parasitoids and plants may have been under particular selection pressure to attract them with volatiles. Acknowledgments The authors gratefully acknowledge the International Max Planck Research School in Jena, Germany and other Max Planck Society funds for support during the preparation of this review. We also thank Andreas Reinicke for helpful comments on the manuscript.

References 1 Knudsen, J.T. et al. (2006) Diversity and distribution of floral scent. Bot. Rev. 72, 1–120 308


2 Kessler, A. and Heil, M. (2011) The multiple faces of indirect defences and their agents of natural selection. Funct. Ecol. 25, 348–357 3 Mumm, R. and Dicke, M. (2010) Variation in natural plant products and the attraction of bodyguards involved in indirect plant defense. Can. J. Zool. 88, 628–667 4 Turlings, T.C.J. and Wackers, F.L. (2004) Recruitment of predators and parasitoids by herbivore-injured plants. In Advances in Insect Chemical Ecology (Carde, R.T. and Millar, G.J., eds), pp. 21–75, Cambridge University Press 5 Dicke, M. and Baldwin, I.T. (2010) The evolutionary context for herbivore-induced plant volatiles: beyond the ‘cry for help’. Trends Plant Sci. 15, 167–175 6 Hare, J.D. (2011) Ecological role of volatiles produced by plants in response to damage by herbivorous insects. Ann. Rev. Entomol. 56, 161–180 7 Holopainen, J.K. and Gershenzon, J. (2010) Multiple stress factors and the emission of plant VOCs. Trends Plant Sci. 15, 176–184 8 Vickers, C.E. et al. (2009) A unified mechanism of action for volatile isoprenoids in plant abiotic stress. Nat. Chem. Biol. 5, 283–291 9 Vet, L.E.M. and Dicke, M. (1992) Ecology of infochemical use by natural enemies in a tritrophic context. Annu. Rev. Entomol. 37, 141–172 10 Dicke, M. et al. (1993) Herbivore-induced plant volatiles mediate plant–carnivore, plant–herbivore and plant–plant interactions: talking plants revisited. In Plant Signals in Interactions with Other Organisms (Schultz, J. and Raskin, I., eds), pp. 182–196, American Society of Plant Physiologists 11 Gershenzon, J. and Dudareva, N. (2007) The function of terpene natural products in the natural world. Nat. Chem. Biol. 3, 408–414 12 Arimura, G. et al. (2004) 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. Plant J. 37, 603–616 13 Danner, H. et al. (2011) Four terpene synthases produce major compounds of the gypsy moth feeding-induced volatile blend of Populus trichocarpa. Phytochemistry 72, 897–908 14 De Moraes, C.M. et al. (2001) Caterpillar-induced nocturnal plant volatiles repel conspecific females. Nature 410, 577–579 15 Kigathi, R.N. et al. (2009) Emission of volatile organic compounds after herbivory from Trifolium pratense (L.) Under laboratory and field conditions. J. Chem. Ecol. 35, 1335–1348 16 Schaub, A. et al. (2010) Real-time monitoring of herbivore induced volatile emissions in the field. Physiol. Plant. 138, 123–133 17 Turlings, T.C.J. et al. (1998) The induction of volatile emissions in maize by three herbivore species with different feeding habits: possible consequences for their natural enemies. Biol. Control 11, 122–129 18 Delphia, C.M. et al. (2007) Induction of plant volatiles by herbivores with different feeding habits and the effects of induced defenses on host-plant selection by thrips. J. Chem. Ecol. 33, 997–1012 19 Piesik, D. et al. (2011) Cereal crop volatile organic compound induction after mechanical injury, beetle herbivory (Oulema spp.), or fungal infection (Fusarium spp.). J. Plant Physiol. 168, 878–886 20 Soler, R. et al. (2007) Root herbivores influence the behaviour of an aboveground parasitoid through changes in plant-volatile signals. Oikos 116, 367–376 21 Zhang, P.J. et al. (2009) Whiteflies interfere with indirect plant defense against spider mites in Lima bean. Proc. Natl. Acad. Sci. U.S.A. 106, 21202–21207 22 Pierre, P.S. et al. (2011) Differences in volatile profiles of turnip plants subjected to single and dual herbivory above- and belowground. J. Chem. Ecol. 37, 368–377 23 Alborn, H.T. et al. (2007) Disulfooxy fatty acids from the American bird grasshopper Schistocerca americana, elicitors of plant volatiles. Proc. Natl. Acad. Sci. U.S.A. 104, 12976–12981 24 Hilker, M. et al. (2005) Insect egg deposition induces defence responses in Pinus sylvestris: characterisation of the elicitor. J. Exp. Biol. 208, 1849–1854 25 Schmelz, E.A. et al. (2006) Fragments of ATP synthase mediate plant perception of insect attack. Proc. Natl. Acad. Sci. U.S.A. 103, 8894–8899 26 Schmelz, E.A. et al. (2007) Cowpea chloroplastic ATP synthase is the source of multiple plant defense elicitors during insect herbivory. Plant Physiol. 144, 793–805

Review 27 Schmelz, E.A. et al. (2009) Phytohormone-based activity mapping of insect herbivore-produced elicitors. Proc. Natl. Acad. Sci. U.S.A. 106, 653–657 28 Dicke, M. et al. (2009) Chemical complexity of volatiles from plants induced by multiple attack. Nat. Chem. Biol. 5, 317–324 29 Zarate, S.I. et al. (2007) Silverleaf whitefly induces salicylic acid defenses and suppresses effectual jasmonic acid defenses. Plant Physiol. 143, 866–875 30 Takabayashi, J. et al. (1995) Developmental stage of herbivore Pseudaletia separata affects production of herbivore-induced synomone by corn plants. J. Chem. Ecol. 21, 273–287 31 Yoneya, K. et al. (2009) Can herbivore-induced plant volatiles inform predatory insect about the most suitable stage of its prey? Physiol. Entomol. 34, 379–386 32 Mumm, R. et al. (2003) Chemical analysis of volatiles emitted by Pinus sylvestris after induction by insect oviposition. J. Chem. Ecol. 29, 1235–1252 33 Kopke, D. et al. (2008) Does egg deposition by herbivorous pine sawflies affect transcription of sesquiterpene synthases in pine? Planta 228, 427–438 34 Hilker, M. and Meiners, T. (2011) Plants and insect eggs: how do they affect each other? Phytochemistry 72, 1612–1623 35 Girling, R.D. et al. (2011) Parasitoids select plants more heavily infested with their caterpillar hosts: a new approach to aid interpretation of plant headspace volatiles. Proc. R. Soc. B: Biol. Sci. 278, 2646–2653 36 Horiuchi, J. et al. (2003) A comparison of the responses of Tetranychus urticae (Acari: Tetranychidae) and Phytoseiulus persimilis (Acari: Phytoseiidae) to volatiles emitted from lima bean leaves with different levels of damage made by T. urticae or Spodoptera exigua (Lepidoptera: Noctuidae). Appl. Entomol. Zool. 38, 109–116 37 Poelman, E.H. et al. (2011) Parasitoid-specific induction of plant responses to parasitized herbivores affects colonization by subsequent herbivores. Proc. Natl. Acad. Sci. U.S.A. 108, 19647–19652 38 Hare, J.D. and Sun, J.J. (2011) Production of induced volatiles by Datura wrightii in response to damage by insects: effect of herbivore species and time. J. Chem. Ecol. 37, 751–764 39 Kessler, A. and Baldwin, I.T. (2001) Defensive function of herbivoreinduced plant volatile emissions in nature. Science 291, 2141–2144 40 van Dam, N.M. and Poppy, G.M. (2008) Why plant volatile analysis needs bioinformatics-detecting signal from noise in increasingly complex profiles. Plant Biol. 10, 29–37 41 Geervliet, J.B.F. et al. (1996) Innate responses of the parasitoids Cotesia glomerata and C. rubecula (Hymenoptera: Braconidae) to volatiles from different plant–herbivore complexes. J. Insect Behav. 9, 525–538 42 Gohole, L.S. et al. (2003) Role of volatiles emitted by host and non-host plants in the foraging behaviour of Dentichasmias busseolae, a pupal parasitoid of the spotted stemborer Chilo partellus. Entomol. Exp. Appl. 107, 1–9 43 Kappers, I.F. et al. (2011) Variation in herbivory-induced volatiles among cucumber (Cucumis sativus L.) varieties has consequences for the attraction of carnivorous natural enemies. J. Chem. Ecol. 37, 150–160 44 Krips, O.E. et al. (2001) Comparison of cultivars of ornamental crop Gerbera jamesonii on production of spider mite-induced volatiles, and their attractiveness to the predator Phytoseiulus persimilis. J. Chem. Ecol. 27, 1355–1372 45 Vanbergen, A.J. et al. (2007) Consequences for a host–parasitoid interaction of host-plant aggregation, isolation, and phenology. Ecol. Entomol. 32, 419–427 46 Jonsson, M. et al. (2005) Behavioural responses in three ichneumonid pollen beetle parasitoids to volatiles emitted from different phenological stages of oilseed rape. Entomol. Exp. Appl. 115, 363–369 47 De Boer, J.G. et al. (2005) Predatory mites learn to discriminate between plant volatiles induced by prey and nonprey herbivores. Anim. Behav. 69, 869–879 48 De Boer, J.G. et al. (2004) Identification of volatiles that are used in discrimination between plants infested with prey or nonprey herbivores by a predatory mite. J. Chem. Ecol. 30, 2215–2230 49 Geervliet, J.B.F. et al. (1998) Long-distance assessment of patch profitability through volatile infochemicals by the parasitoids


50

51

52

53

54

55

56

57

58 59

60

61

62

63

64

65

66

67 68

69

70

71

Cotesia glomerata and C. rubecula (Hymenoptera: Braconidae). Biol. Control 11, 113–121 Steidle, J.L.M. and van Loon, J.J.A. (2003) Dietary specialization and infochemical use in carnivorous arthropods: testing a concept. Entomol. Exp. Appl. 108, 133–148 Gols, R. et al. (2009) Are population differences in plant quality reflected in the preference and performance of two endoparasitoid wasps? Oikos 118, 733–743 Moreau, J. et al. (2009) Host plant cultivar of the grapevine moth Lobesia botrana affects the life history traits of an egg parasitoid. Biol. Control 50, 117–122 Sarfraz, M. et al. (2009) Host plant nutritional quality affects the performance of the parasitoid Diadegma insulare. Biol. Control 51, 34–41 Schadler, M. et al. (2010) Host plant genotype determines bottom-up effects in an aphid–parasitoid–predator system. Entomol. Exp. Appl. 135, 162–169 Ode, P.J. (2006) Plant chemistry and natural enemy fitness: effects on herbivore and natural enemy interactions. Annu. Rev. Entomol. 51, 163–185 Bukovinszky, T. et al. (2009) Consequences of constitutive and induced variation in plant nutritional quality for immune defence of a herbivore against parasitism. Oecologia 160, 299–308 Smilanich, A.M. et al. (2009) Immunological cost of chemical defence and the evolution of herbivore diet breadth. Ecol. Lett. 12, 612–621 Olson, D.M. et al. (2009) Nitrogen and water affect direct and indirect plant systemic induced defense in cotton. Biol. Control 49, 239–244 Winter, T.R. and Rostas, M. (2010) Nitrogen deficiency affects bottomup cascade without disrupting indirect plant defense. J. Chem. Ecol. 36, 642–651 Rasmann, S. and Turlings, T.C.J. (2007) Simultaneous feeding by aboveground and belowground herbivores attenuates plantmediated attraction of their respective natural enemies. Ecol. Lett. 10, 926–936 Soler, R. et al. (2007) Foraging efficiency of a parasitoid of a leaf herbivore is influenced by root herbivory on neighbouring plants. Funct. Ecol. 21, 969–974 Pierre, P.S. et al. (2011) Aboveground herbivory affects indirect defences of brassicaceous plants against the root feeder Delia radicum Linnaeus: laboratory and field evidence. Ecol. Entomol. 36, 326–334 Snoeren, T.A.L. et al. (2010) The herbivore-induced plant volatile methyl salicylate negatively affects attraction of the parasitoid Diadegma semiclausum. J. Chem. Ecol. 36, 479–489 Puente, M. et al. (2008) Impact of herbivore-induced plant volatiles on parasitoid foraging success: a spatial simulation of the Cotesia rubecula, Pieris rapae, and Brassica oleracea system. J. Chem. Ecol. 34, 959–970 Tamo, C. et al. (2006) A comparison of naive and conditioned responses of three generalist endoparasitoids of lepidopteran larvae to host-induced plant odours. Anim. Biol. 56, 205–220 Cortesero, A.M. et al. (1997) Comparisons and contrasts in hostforaging strategies of two larval parasitoids with different degrees of host specificity. J. Chem. Ecol. 23, 1589–1606 De Moraes, C.M. et al. (1998) Herbivore-infested plants selectively attract parasitoids. Nature 393, 570–573 Yan, Z.G. and Wang, C.Z. (2006) Similar attractiveness of maize volatiles induced by Helicoverpa armigera and Pseudaletia separata to the generalist parasitoid Campoletis chlorideae. Entomol. Exp. Appl. 118, 87–96 Blande, J.D. et al. (2007) A comparison of semiochemically mediated interactions involving specialist and generalist Brassica-feeding aphids and the braconid parasitoid Diaeretiella rapae. J. Chem. Ecol. 33, 767–779 Gouinguene, S. et al. (2003) Induction of volatile emissions in maize by different larval instars of Spodoptera littoralis. J. Chem. Ecol. 29, 145–162 Mattiacci, L. and Dicke, M. (1995) The parasitoid Cotesia glomerata (Hymenoptera, Braconidae) discriminates between first and 5th larval instars of its host Pieris brassicae, on the basis of contact cues from frass, silk, and herbivore-damaged leaf tissue. J. Insect Behav. 8, 485–498 309

Review 72 Fatouros, N.E. et al. (2005) Herbivore-induced plant volatiles mediate in-flight host discrimination by parasitoids. J. Chem. Ecol. 31, 2033–2047 73 Klowden, M.J. (2007) Physiological Systems in Insects, Elsevier 74 van Wijk, M. et al. (2006) Morphology of the olfactory system in the predatory mite Phytoseiulus persimilis. Exp. Appl. Acarol. 40, 217–229 75 Laissue, P.P. et al. (1999) Three-dimensional reconstruction of the antennal lobe in Drosophila melanogaster. J. Comp. Neurol. 405, 543–552 76 Galizia, C.G. et al. (1999) A digital three-dimensional atlas of the honeybee antennal lobe based on optical sections acquired by confocal microscopy. Cell Tissue Res. 295, 383–394 77 Smid, H.M. et al. (2003) Three-dimensional organization of the glomeruli in the antennal lobe of the parasitoid wasps Cotesia glomerata and C. rubecula. Cell Tissue Res. 312, 237–248 78 Bruce, T.J.A. et al. (2005) Insect host location: a volatile situation. Trends Plant Sci. 10, 269–274 79 Han, B.Y. et al. (2001) Electrophysiology and behavior feedback of diamondback moth, Plutella xylostella, to volatile secondary metabolites emitted by Chinese cabbage. Chin. Sci. Bull. 46, 2086–2088 80 Pope, T.W. et al. (2008) Comparative innate responses of the aphid parasitoid Diaeretiella rapae to alkenyl glucosinolate derived isothiocyanates, nitriles, and epithionitriles. J. Chem. Ecol. 34, 1302–1310 81 Nottingham, S.F. et al. (1991) Behavioral and electrophysiological responses of aphids to host and non-host plant volatiles. J. Chem. Ecol. 17, 1231–1242 82 Loivamaki, M. et al. (2008) Isoprene interferes with the attraction of bodyguards by herbaceous plants. Proc. Natl. Acad. Sci. U.S.A. 105, 17430–17435 83 Gouinguene´, S. et al. (2005) Antennal electrophysiological responses of three parasitic wasps to caterpillar-induced volatiles from maize (Zea mays mays), cotton (Gossypium herbaceum), and cowpea (Vigna unguiculata). J. Chem. Ecol. 31, 1023–1038 84 Ngumbi, E. et al. (2009) Comparative GC-EAD responses of a specialist (Microplitis croceipes) and a generalist (Cotesia marginiventris) parasitoid to cotton volatiles induced by two caterpillar species. J. Chem. Ecol. 35, 1009–1020 85 Verheggen, F.J. et al. (2008) Aphid and plant volatiles induce oviposition in an aphidophagous hoverfly. J. Chem. Ecol. 34, 301–307 86 Hegde, M. et al. (2011) Identification of semiochemicals released by cotton, Gossypium hirsutum, upon infestation by the cotton aphid, Aphis gossypii. J. Chem. Ecol. 37, 741–750 87 Smid, H.M. et al. (2002) GC-EAG-analysis of volatiles from Brussels sprouts plants damaged by two species of Pieris caterpillars: olfactory receptive range of a specialist and a generalist parasitoid wasp species. Chemoecology 12, 169–176 88 Ngumbi, E. et al. (2010) Electroantennogram (EAG) responses of Microplitis croceipes and Cotesia marginiventris and their lepidopteran hosts to a wide array of odor stimuli: correlation between EAG response and degree of host specificity? J. Insect Physiol. 56, 1260–1268 89 Beyaert, I. et al. (2010) Relevance of resource-indicating key volatiles and habitat odour for insect orientation. Anim. Behav. 79, 1077–1086 90 Shiojiri, K. et al. (2010) Herbivore-specific, density-dependent induction of plant volatiles: honest or ‘cry wolf’ signals? PLoS ONE 5, e12161 91 van Wijk, M. et al. (2008) Predatory mite attraction to herbivoreinduced plant odors is not a consequence of attraction to individual herbivore-induced plant volatiles. J. Chem. Ecol. 34, 791–803 92 van Wijk, M. et al. (2011) Complex odor from plants under attack: herbivore’s enemies react to the whole, not its parts. PLoS ONE 6, e21742

310


93 Degenhardt, J. et al. (2009) Restoring a maize root signal that attracts insect-killing nematodes to control a major pest. Proc. Natl. Acad. Sci. U.S.A. 106, 13213–13218 94 Halitschke, R. et al. (2008) Shared signals – ‘alarm calls’ from plants increase apparency to herbivores and their enemies in nature. Ecol. Lett. 11, 24–34 95 Kessler, A. and Baldwin, I.T. (2004) Herbivore-induced plant vaccination. Part I. The orchestration of plant defenses in nature and their fitness consequences in the wild tobacco Nicotiana attenuata. Plant J. 38, 639–649 96 Holopainen, J.K. (2004) Multiple functions of inducible plant volatiles. Trends Plant Sci. 9, 529–533 97 Heil, M. (2009) Damaged-self recognition in plant herbivore defence. Trends Plant Sci. 14, 356–363 98 Coleman, R.A. et al. (1999) Parasitism of the herbivore Pieris brassicae L. (Lep. Pieridae) by Cotesia glomerata L. (Hym. Braconidae) does not benefit the host plant by reduction of herbivory. J. Appl. Entomol. Z: Angew. Entomol. 123, 171–177 99 Kessler, A. and Halitschke, R. (2007) Specificity and complexity: the impact of herbivore-induced plant responses on arthropod community structure. Curr. Opin. Plant Biol. 10, 409–414 100 Bruinsma, M. et al. (2008) Differential effects of jasmonic acid treatment of Brassica nigra on the attraction of pollinators, parasitoids, and butterflies. Entomol. Exp. Appl. 128, 109–116 101 Dudareva, N. (2004) Biochemistry of plant volatiles. Plant Physiol. 135, 1893–1902 102 Maffei, M.E. et al. (2011) Plant volatiles: production, function and pharmacology. Nat. Prod. Rep. 28, 1359–1380 103 Turlings, T.C.J. et al. (1992) Learning of host-finding cues by hymenopterous parasitoids. In Insect Learning: Ecological and Evolutionary Perspectives (Papaj, D.R. and Lewis, A.C., eds), pp. 51–78, Chapman & Hall 104 McAuslane, H.J. et al. (1991) Influence of adult experience on host microhabitat location by the generalist parasitoid, Campoletis sonorensis (Hymenoptera, Ichneumonidae). J. Insect Behav. 4, 101–113 105 Potting, R.P.J. et al. (1997) Absence of odour learning in the stemborer parasitoid Cotesia flavipes. Anim. Behav. 53, 1211–1223 106 Pe´rez-Maluf, R. et al. (2008) Differentiation of innate but not learnt responses to host-habitat odours contributes to rapid host finding in a parasitoid genotype. Physiol. Entomol. 33, 226–232 107 Sznajder, B. et al. (2011) Innate responses of the predatory mite Phytoseiulus persimilis to a herbivore-induced plant volatile. Exp. Appl. Acarol. 54, 125–138 108 Fukushima, J. et al. (2001) Learning of host-infested plant volatiles in the larval parasitoid Cotesia kariyai. Entomol. Exp. Appl. 99, 341–346 109 Meiners, T. et al. (2003) Associative learning of complex odours in parasitoid host location. Chem. Senses 28, 231–236 110 Costa, A. et al. (2010) Effects of rewarding and unrewarding experiences on the response to host-induced plant odors of the generalist parasitoid Cotesia marginiventris (Hymenoptera: Braconidae). J. Insect Behav. 23, 303–318 111 Tentelier, C. and Fauvergue, X. (2007) Herbivore-induced plant volatiles as cues for habitat assessment by a foraging parasitoid. J. Anim. Ecol. 76, 1–8 112 de Boer, J.G. et al. (2008) Prey and non-prey arthropods sharing a host plant: effects on induced volatile emission and predator attraction. J. Chem. Ecol. 34, 281–290 113 Bleeker, M.A.K. et al. (2006) Differences in memory dynamics between two closely related parasitoid wasp species. Anim. Behav. 71, 1343–1350 114 Smid, H.M. et al. (2007) Species-specific acquisition and consolidation of long-term memory in parasitic wasps. Proc. R. Soc. B: Biol. Sci. 274, 1539–1546 115 Hoedjes, K.M. et al. (2011) Natural variation in learning rate and memory dynamics in parasitoid wasps: opportunities for converging ecology and neuroscience. Proc. R. Soc. B: Biol. Sci. 278, 889–897

Review


Association mapping of plant resistance to insects Karen J. Kloth1,2,3*, Manus P.M. Thoen1,2,3*, Harro J. Bouwmeester2, Maarten A. Jongsma3 and Marcel Dicke1 1

Laboratory of Entomology, Wageningen University, P.O. Box 8031, 6700 EH Wageningen, The Netherlands Laboratory of Plant Physiology, Wageningen University, Droevendaalsesteeg 1, 6708 PB Wageningen, The Netherlands 3 Business Unit Bioscience, Plant Research International, Wageningen University and Research Center, P.O. Box 619, 6700 AP Wageningen, The Netherlands 2

Association mapping is rapidly becoming an important method to explore the genetic architecture of complex traits in plants and offers unique opportunities for studying resistance to insect herbivores. Recent studies indicate that there is a trade-off between resistance against generalist and specialist insects. Most studies, however, use a targeted approach that will easily miss important components of insect resistance. Genome-wide association mapping provides a comprehensive approach to explore the whole array of plant defense mechanisms in the context of the generalist–specialist paradigm. As association mapping involves the screening of large numbers of plant lines, specific and accurate highthroughput phenotyping (HTP) methods are needed. Here, we discuss the prospects of association mapping for insect resistance and HTP requirements. Enhancing host–plant resistance against generalist and specialist insects Host–plant resistance is one of the cornerstones of environmentally benign pest management systems [1,2]. Devastating pests and diseases only rarely occur in nature, which is due to the tremendous degree of natural variation in plant defense mechanisms [3,4]. Only a relatively small degree of such variation is contained in cultivated crop populations [5], but wild populations provide ample opportunities for discovering novel mechanisms responsible for resistance to insects. A wide range of resistance mechanisms against herbivorous insects has been described [1,2], and the impact of mechanisms depends on the characteristics of the herbivore, such as insect diet breadth [6,7]. Although specialist insects, feeding on one or a few plant species within one family, are considered to be resistant to toxic compounds of their host [8], generalist insects are thought to thrive on a wider range of hosts with relatively low levels of allelochemicals [9,10]. Toxins, however, affect the performance of specialists as well [11], and generalists can cope with variable levels of secondary metabolites [10], implying a more complex relationship between insect host range and plant defense. More insight into plant defenses against specialist and generalist insects is needed to *

Corresponding author: Dicke, M. ([email protected]) These authors contributed equally to this review.

understand how plants deal with herbivorous insects that differ in the degree of specialization and to improve host– plant resistance of economically important crops against insect pests. Most studies have addressed this topic with a targeted approach, focusing on only one or a few types of secondary metabolites and a restricted amount of natural variation therein. To unravel the paradigm about resistance against specialists and generalists and to identify new plant defense mechanisms, comprehensive technologies are needed that can explore the apparent natural variation in multiple resistance mechanisms at the level of the genotype and phenotype. Association mapping (see Glossary) allows screening of many different wild and cultivated populations for genes involved in complex plant traits. Although association mapping has hardly been used in plant–insect studies thus far, it has the potential to allow new developments in eco-genomic studies of plant–insect interactions. One of the major prospects is the possibility to do genome-wide association (GWA) mapping to retrieve functional genetic loci involved in plant defenses against herbivorous insects in an untargeted way. GWA mapping involves the screening of large numbers of plant lines, which is currently a bottleneck because of the costs involved in this time- and

Glossary Association mapping: a population-based method of mapping quantitative trait loci (QTLs) that takes advantage of historic linkage disequilibrium to link phenotypes to genotypes (also known as ‘linkage disequilibrium mapping’). Candidate gene: a chromosome region suspected of being involved in the expression of a trait of interest. Genome-wide association (GWA) mapping: comprehensive approach to systematically search the genome for causal genetic variation, using a large number of markers, by association between genotypes at each locus and a given phenotype. High-throughput phenotyping (HTP): experimental setup in which large amounts of specimens can be phenotypically screened, preferably automatic, fast, accurate and with low costs. Linkage disequilibrium: two loci that are in linkage disequilibrium (LD) are inherited together more often or less often than would be expected by chance. QTL: quantitative trait locus, a region in the genome that is responsible for variation in the quantitative trait of interest. QTL mapping: a family-based mapping method using well-known pedigrees to generate crosses in which the genetic architecture of traits can be explored (also known as traditional linkage mapping). Quantitative genetics: the study of the heritability of quantitative traits, which are the products of two or more genes.


311

Review


labor-intensive methodology. The large number of plant lines to be screened in insect resistance studies will require high-throughput phenotyping (HTP) techniques that succeed in accurately identifying different resistance traits. Particularly in view of the high diversity in insectresistance mechanisms and their degree of specificity towards their enemies, this will pose some challenges. In this review, we discuss the perspectives of GWA mapping and HTP techniques in the context of insect resistance, with special reference to strategies against specialist and generalist insects. Association studies and linkage mapping Understanding the genetic basis of phenotypic variation is one of the key goals in evolutionary biology. Family-based quantitative trait locus (QTL) mapping (which uses wellcharacterized pedigrees [12–14]) and association mapping (which uses linkage disequilibrium among numerous individuals of different populations [15,16]) are the most commonly used tools for dissecting the genetic basis of phenotypic trait variation. In QTL mapping only a limited number of recombination events that have occurred within families and pedigrees can be studied, whereas with association mapping the recombination events that have accumulated over thousands of generations can be exploited [17]. Since the 1980s, QTL mapping has been used most frequently, but association mapping is a promising alternative method for dissecting complex traits [18,19]. Increased mapping resolution, reduced research time and larger allele numbers have been put forward as main advantages over traditional QTL mapping [17,20]. Association studies can be divided into two broad categories: (i) candidate gene association mapping, in which variation in a gene of interest is tested for correlation with the

phenotypic trait of interest and (ii) GWA mapping, where genetic variation is explored within the whole genome, aiming to find signals of association with the complex trait [17] (see Table 1 for an overview). Because GWA mapping is less dependent on prior information about candidate genes than QTL mapping and candidate gene association mapping, this is a promising method to identify novel loci involved in complex phenotypic traits. However, GWA mapping should not be regarded as a replacement of traditional QTL mapping. In fact, GWA mapping and QTL mapping have complementary advantages and disadvantages, which can lead to a better understanding of causal genetic polymorphism when these approaches are combined [18,21]. Association mapping in plant sciences In the past decade, GWA mapping has emerged as a tool for studying the genetics of natural variation and economically important traits in plants [16]. Flowering time, chemical composition, disease resistance, taste and many other economically and evolutionarily important traits have been studied in crop species (see [17] for an overview). Apart from agriculturally relevant crops, the model plant Arabidopsis (Arabidopsis thaliana) is of great value for understanding complex traits using GWA mapping (Box 1). The presence of recombination events that have accumulated in plants over thousands of generations is both an advantage as well as a potential pitfall of GWA mapping, because functional QTLs that are correlated with population structure can result in many false positives [21]. Several statistical methods have been developed that use neutral genotypic information to account for confounding effects of population structure in GWA studies [22–24]. However, inadequate use of these models can lead to

Table 1. Comparison of family-based (QTL) and population-based (association mapping) methods that aim to unravel the genetic basis of complex traits in plantsa Main advantages

Main disadvantages

General requirements

Recent case study in Arabidopsis

QTL mapping b No population structure effects Identification of rare alleles Few genetic markers required Limited genetic diversity Not always possible to create crosses Cannot distinguish between pleiotropic and physically close genes Small ‘original population size’, low number of genetic markers, many replicates needed Generated mapping material, e.g. F2 population, (AI-)RILs, MAGIC lines, NILs, HIFs, etc.

QTL mapping with AI-RILs on flowering time [14] two AI-RIL populations (approximately 280 individuals each) 181 and 224 markers 12 to 70 replicates

Candidate gene association mapping Allows fine mapping Relatively low costs Detailed functional knowledge of trait is required No novel traits will be found Large population size, small number of genetic markers, the bigger the population size, the less replicates needed Prior genetic and biochemical knowledge on trait of interest Prior knowledge on LD, nucleotide polymorphism, breeding system and population structure Candidate gene approach on flowering time [86] 251 accessions 51 SNPs 10 replicates per accession

Genome-wide association mapping Allows untargeted fine mapping (blind approach) Detection of common alleles Confounding effects due to population structure Will miss rare and weak effect alleles Large population size, many genetic markers, the bigger the population size, the fewer replicates needed Prior knowledge on LD, nucleotide polymorphism, breeding system and population structure

Whole-genome approach on multiple phenotypic traits [16] 199 accessions in total 216 150 SNPs Four replicates in general

a

Combinations of these three approaches can allow the identification of false positives and negatives, but is much more laborious: a recent dual QTL mapping–GWA study [26] involved phenotyping nearly 20 000 individual plants, including 184 worldwide natural accessions genotyped for 216 509 SNPs and 4366 RILs derived from 13 independent crosses. See [25] for an overview of different linkage mapping populations mentioned in this table.

b

Abbreviations: AI-RIL, advanced intercross-recombinant inbred line; HIFs, heterogeneous inbred family; LD, linkage disequilibrium; MAGIC, multiparent advanced generation intercross; NIL, near-isogenic line; QTL, quantitative trait locus; RIL, recombinant inbred line; SNPs, single nucleotide polymorphisms.

312

Review Box 1. Arabidopsis–insect interactions as a model for GWA studies The model species, Arabidopsis (Arabidopsis thaliana), is often used in plant–insect studies for obvious reasons, such as the availability of extensive information about genetic variation and physiology, and numerous mutants. Even though Arabidopsis is not a crop, there are numerous devastating crop pest insects (such as the generalist insect herbivores Frankliniella occidentalis and Myzus persicae and the specialist insect herbivores Pieris rapae, Plutella xylostella and Brevicoryne brassicae) that readily feed on Arabidopsis [41,62,87–89]. However, one disadvantage in the light of insect–plant biology is that many accessions of Arabidopsis are winter annuals, so the life cycle of Arabidopsis does not temporally overlap with the life cycle of many herbivorous insects. It is known that herbivore performance (quantified in terms of mortality and developmental time) is commonly better on plants with such a ‘pausing’ strategy, indicating that such plants may invest less in defense traits [90]. This has probably influenced the evolution of signaling pathways in Arabidopsis, because the main biotic stresses probably comprise pathogens such as oomycetes, bacteria and fungi. Still, Arabidopsis is of great interest for studying insect resistance, because many insect defense mechanisms have been evolved within the Brassicaceae family, such as glucosinolates [6,7,29], and many defense mechanisms against pathogens are also effective against herbivorous insects. Leaf toughness is, for example, effective against both microbial pathogens and insects [2], and salicylic acid-, jasmonic acid- and ethylene-regulated defenses are involved in defenses against both pathogen and insect infestations [6,87,91,92].

overcorrection, resulting in false negatives which are equally problematic [21]. Studies that have combined GWA and QTL mapping strategies (dual linkage association mapping) revealed a false positive rate of 40% and a false negative rate of 24% in assays that solely involved GWA mapping [25]. A major drawback of such a dual linkage association mapping, however, is that it requires phenotyping of several thousands of individual plants and the genesis of numerous linkage mapping populations [26]. GWA mapping in regional mapping populations (instead of GWA mapping at the species scale) is an alternative approach to reduce confounding due to population structure [25]. Another major impediment in GWA studies is the phenomenon of missing heritability. Often, the associated QTLs can explain very little of the phenotypic variation, even after accounting for the effects of population structure. This phenomenon is attributed to several factors, including a scattered signal across numerous QTLs, each contributing to only a marginal proportion of the phenotype. Complex traits, such as insect resistance, are likely to encounter this problem [27,28]. Integrating association mapping with transcriptional network analysis can decrease high false positive rates and increase the resolution in scattered associations [19]. The scattering of genotype– phenotype associations can also be reduced by phenotyping multiple component traits instead of one multifactorial trait, as will be further discussed in the section ‘Requirements for phenotyping’. Association mapping of plant–insect interactions The complexity in the orchestration of insect resistance and its evolution in plants make it a difficult trait to study in a genomic context [4]. So far, only few GWA studies have


been reported that deal explicitly with plant defense mechanisms against herbivorous insects (see [16] for an example on aphids). One such study on glucosinolates (GSL) – secondary defense metabolites within the Brassicaceae family involved in resistance against herbivorous insects [6,7,29] – was conducted using 96 Arabidopsis accessions exhibiting 43 distinct GSL phenotypes and 230 000 single nucleotide polymorphisms (SNPs) [18]. In this study, GWA analysis successfully identified two major polymorphic loci controlling GSL variation in natural populations, but variation in resistance to specialist and generalist insects remains to be investigated for these accessions. This would require an experimental setup in which GWA mapping and HTP of insect resistance are integrated (Figure 1). GWA mapping of insect resistance will probably encounter similar obstacles as recognized in other GWA studies. Because insect resistance is generally under strong positive selection pressure, GWA mapping of insect resistance might, however, unlike GWA studies of human diseases [15,27], be less affected by rare alleles that are not included in the haplotype map. Nevertheless, a good representation of all (sub)populations is indispensable for

[(Figure_1)TD$IG]

(a)

(b)

(c)

Preference

#Eggs

Performance

Feeding damage

(d)

Mobility

ATCCAGCT ATCCTGCT

(e) Overexpression

Gene silencing


Figure 1. Screening plants for insect resistance through GWA mapping. This simplified overview shows how the genetic architecture underlying insect resistance can be determined in five steps, using GWA mapping. (a) Genotype SNPs for numerous accessions of the plant of interest; (b) develop HTP choice and no-choice experiments to screen for insect preference and performance (using leaf discs in this example); (c) screen for relevant insect-resistance parameters; (d) find the genetic basis of phenotypic differences, using GWA mapping; (e) validate candidate genes with reverse genetic tools, such as overexpression and gene silencing.

313

Review detecting variation in host–plant resistance and preventing them from having a too low allele frequency in the experimental setup. Particularly, the confounding effects of population structure can have a large effect on the success of GWA studies of host–plant resistance, because resistance against specific insects could have evolved independently and be based on different mechanisms in different populations and habitats [30]. Moreover, confounding effects due to strong population differences can be severe, when an intense evolutionary arms race between plant and herbivore has occurred as may be the case for specialist herbivorous insects and their host plants [30–33]. This will require statistical correction of population structure, which can enhance the chance of false negatives due to overcorrection. This problem is expected to be less evident with generalists, because they lack a reciprocal evolutionary interaction with specific plants [9]. Resistance against specialist versus generalist insect herbivores Specialist and generalist insect herbivores have different ways to deal with the defensive mechanisms of their host plants, and this is expected to result in different associations. In addition to morphological and structural aspects, chemical defenses involving secondary metabolites play a major role in plant defense against insects [2]. Secondary metabolites can be divided into two broad functional categories, based on their modes of actions: qualitative compounds, which can be interpreted as toxins, and quantitative defensive compounds, with a dose-dependent effect, such as digestibility reducers [9]. Recent studies show that qualitative compounds (e.g. GSL and alkaloids) often fail to affect specialist insects, because specialist insects evolved ways to detoxify or tolerate these compounds [9]. In other words, if secondary metabolites play a role in defense against specialist insects, predominantly quantitative defensive compounds that reduce the digestibility are expected to be functional, whereas defense against polyphagous insects is mainly achieved by qualitative compounds. Toxins are even used by specialist insects to locate their host plants or sequester these toxins for their own defense [1,2]. Thus, plants have to ‘choose’ between investing in substantial concentrations of qualitative compounds to deter polyphagous insects or marginal concentrations of the same compounds to decrease preference by specialist insects [34,35]. The evolution of defensive traits against generalists could, therefore, lead to an increased host–plant preference by specialists and vice versa. This trade-off between resistance to specialists and generalists is expected to be reflected in genotype– phenotype associations of the host plant. There are, however, many examples that do not support the qualitative–quantitative dichotomy. The generalist aphid Myzus persicae feeds on herbaceous plants in over 40 plant families, including families such as the Solanaceae that are well-known producers of toxic alkaloids [10,36]. Moreover, specific toxins do affect specialist herbivores. For instance, silencing nicotine production in tobacco (Nicotiana tabacum) results in improved performance of the specialist herbivore Manduca sexta [37] and overexpression of the lectin agglutinin in tobacco negatively affected the larval 314


performance of M. sexta [11]. Isothiocyanates, breakdown products of GLS, negatively affect the performance of the specialist herbivore Pieris rapae [38]. The performance of P. rapae on the coi1 mutants of Arabidopsis, which is compromised in the JA signal transduction pathway, is significantly improved in comparison to wild-type plants, showing that even a specialist is affected by inducible plant defenses [39]. Interestingly, the effects of quantitative and qualitative defenses may interact: nicotine prevents a compensatory response of the generalist herbivore Spodoptera exigua to proteinase inhibitors and thus counters an insect adaptation to a qualitative defense [40]. The main deficiency in addressing the defense mechanisms of plants against specialist and generalist insect herbivores is that most studies have used a targeted approach, focusing on only one or a few types of secondary metabolites in a limited number of plant lines. Because resistance and tolerance are likely to be phenotypic traits that are composed of multiple factors, a targeted approach will easily miss important components. This is true for resistance to both generalists and specialists, but comparing the components and their relative strength of resistance to specialists and generalists may reveal how these traits are balanced. A more comprehensive approach is, for example, taken in transcript profiling studies, where gene expression signatures of infested plants and/or herbivorous insects are analyzed in different treatments [39,41–43]. Although several studies did not find a different plant response to specialist and generalist insects [39,41], one study [42] found a differential response in the insects that foraged on wild-type and mutant Nicotiana attenuata. The specialist M. sexta showed diet-specific alterations in gene expression, whereas the generalist Heliothis virescens regulated similar transcripts over different diets, indicating that the specialist is better adapted to both qualitative (nicotine) and quantitative (trypsin protease inhibitor) compounds of the host [42]. Another explorative approach is taken in a recent study, where metabolite fingerprints of Plantago lanceolata leaves differed after they were attacked by specialist or generalist herbivores, and by insects belonging to different taxa [43]. These examples show that untargeted approaches, such as transcript profiling, metabolic fingerprinting and GWA mapping, allow exploring a large array of plant defense mechanisms in many plant lines. Requirements for phenotyping Phenotyping is a prime factor in GWA mapping of host plant resistance. Among vast numbers of genome-wide markers, the aim is to achieve significant statistical power for only those molecular markers that are located close to the genes that influence the phenotypic trait of interest. In reality, functional associations between phenotype and molecular markers are often confounded, both in association and QTL mapping studies [18,44,45]. In the discussion about missing heritability of associations, where the identified genetic loci explain only little of the phenotypic variation, little attention has been paid to the role of phenotypes and phenotyping techniques. Some association studies of crop yield, for example, resulted in

Review


the characterization of numerous minor functional genes [46]. This confirms the infinitesimal model of Fischer [47], which assumes a very large number of loci to be involved in quantitative genetics, each with a marginal effect on the phenotype. It is to be expected, however, that the number of functional (low effect) QTLs involved is trait-specific. A complex trait is generally the result of numerous processes, which will result in a scattered association across multiple genetic loci: numerous QTLs are involved, which have a reduced statistical significance and each contribute to only a marginal proportion of the effect size of the phenotypic variation (Figure 2). Although a multifactorial character is inherent to complex traits, the efficiency of association mapping can be optimized by dissecting the phenotype into quantitative components with a minimum expected number of responsible mechanisms [48]. A genome-wide screening within the scope of only a few mechanisms attributing to the trait of interest will increase the success of finding novel functional genes. A drawback is that it narrows the scope of a genome-wide survey. Complex traits are generally based on gene networks; therefore, the assessment of individual components will probably overlook interactions between components, and the network as a whole and its environment [49,50]. In insect resistance studies, typically multiple traits are phenotyped and reduced to one resistance variable, R. Most often, the total of life history parameters of the insect are summarized in the variable rm, the intrinsic rate of population increase [51,52]. This summary statistic is an accurate parameter of the effect of resistance mechanisms on the herbivorous insects. However, insect performance is typically dependent on multiple plant traits (e.g. nutritional components of the host plant and multiple resistance mechanisms of the plant [52]). Hence, rm may lack resolution in association studies and using this parameter may

[(Figure_2)TD$IG]

(a)

result in a high proportion of missing heritability due to scattered signals (Figure 2). We expect that dissecting the complex parameter in multiple specific phenotypic components, e.g. host preference, time interval before the insect starts feeding, reproduction, larval development time and mortality, will contribute to solving the problem of missing heritability and will help to identify multiple underlying mechanisms (Figure 2). The combination of these individual mechanisms will ultimately allow plant breeding to achieve sustainable host–plant resistance in crops. Indeed, multiparameter approaches, using a combination of phenotypic traits, for example both concentration of secondary metabolites and insect performance, have been postulated to deliver more significant relations to functional genetic data [50,53]. Apart from the parameterization of the phenotype(s), increasing the number of plant lines is of major importance for the statistical support of relevant associations [15,27]. So far, most studies have used sample sizes of approximately 100 to 500 plant lines, but more genetic lines will increase the number and frequency of functional alleles and thereby improve the statistical power to detect them [18,24,44,54,55]. Secondly, a larger number of replicates within plant lines will increase the accuracy of the phenotype and the statistical support of genotype–phenotype associations. Particularly phenotyping insect resistance, involving the interaction among two or more organisms and species, is sensitive to stochastic errors and could result in relatively high levels of missing heritability. Although there is an example of successful GWA mapping by assessing aphid offspring in four replicates on 96 Arabidopsis lines [16], more replicates will reduce confounding effects. Moreover, the quality of phenotypic data can be improved by eliminating noise induced by the environment [50,56]. Many studies have shown that insect resistance is an adaptive response to several biotic and

(b)

Preference

Time before feeding

Leaf toughness

VOCs

Leaf toughness VOCs Plant surface

Feeding deterrents

Plant surface Nutrients

Feeding deterrents

Insect life history parameter rm

Nutrients

Larval development Key: Association signal

Egg

Larva

Pupa

Adult

Genome Markers associated with functional gene TRENDS in Plant Science

Figure 2. Dissecting insect resistance into component traits. Association mapping of a complex trait such as insect resistance can result in numerous associations with low statistical power. This is illustrated in (a) where the life history parameter rm of the insect is associated with many genetic loci. One approach to improve resolution in genotype–phenotype associations is to dissect the complex phenotype into component traits (b), e.g. insect preference (detection of repellent VOCs), time before the insect starts feeding (screening for the influence of leaf toughness and deterrent structures on the plant surface) and larval development (detection of, e.g. feeding deterrents, toxins and nutrient content). Whereas the genetic architecture can overlap to some degree due to similar underlying processes, mapping these component traits will result in fewer genotype–phenotype associations with larger statistical power and a higher proportion of functional associations. Genotype–phenotype associations can be further elucidated with, for example, metabolite fingerprints of VOCs, plant tissues or epicuticular waxes.

315

Review Box 2. Plant resistance to herbivorous insects Host–plant resistance against herbivorous insects is generally defined as ‘the relative amount of heritable qualities possessed by the plant which influence the ultimate degree of damage done by the insect in the field’ [1]. Herbivorous insects use host plants for oviposition, feeding and shelter. Plants can achieve protection against herbivorous insects by both indirect defense, i.e. the attraction and facilitation of natural enemies of the insect herbivore, and direct defense against the pest insect [93–95]. Three main categories of resistance against insect herbivores are (i) antixenosis, (ii) antibiosis and (iii) tolerance [1]. Antixenosis mechanisms deter the insect or, after the insect has arrived on the plant, prevent it from settling. Generally, the insect ‘decides’ not to colonize the plant due to the absence or low availability of an attractant, or the presence or quantity of a deterrent. A wide range of components can act as attractants or deterrents: volatile organic compounds (VOCs), color, topology of the plant, chemicals and morphology of the plant surface (e.g. trichomes, epicuticular waxes, substrate texture), and physical and chemical characteristics of internal plant tissues (e.g. secondary metabolites, nutrient content, toughness of the cell wall) [2]. Herbivorous insects use olfactory and visual cues in the prealighting stage, and assess olfactory, visual, tactile and gustatory traits after arriving on the host plant. Plants that exhibit antixenosis have a reduced number of initial colonizers and a relatively small population of herbivorous insects. After the insect has ‘decided’ to utilize the host, antibiosis mechanisms of the host can affect insect performance (e.g. growth, development, reproduction and survival) by toxins released after tissue damaging, feeding deterrents (e.g. protease inhibitors), nutritional imbalance or tissue toughness. Antibiosis causes a decrease in the insect population size [1]. Plants can display antixenosis and antibiosis mechanisms constitutively or after induction by, e.g. herbivory or egg deposition [93,96]. Finally, tolerance represents the plant’s ability to compensate insect damage by increased growth, reproduction or repair of the damage. In contrast to antixenosis and antibiosis, tolerance does not severely affect the insect herbivore, but rather minimizes the impact of herbivory on the performance of the plant itself [1,2].

abiotic factors (Box 2) [19,35,57,58]. For example, it has been shown that the developmental stage of the plant altered the outcome of the GWA analysis, resulting in the identification of different functional genetic loci in different developmental stages [19]. This underlines the need for an experimental setup with uniform conditions among the genetic lines (Figure 1). Some noise will be inevitable for plant species harboring a high diversity of ecotypes that differ in optimal growth conditions and development time. By contrast, ‘uniform’ laboratory assays can deliver functional associations different from field conditions [16] due to genotype-by-environment interactions [49]. Including several (a)biotic treatments or an additional field assay could yield more field-predictive outcomes. High-throughput phenotyping For quantitative traits such as insect resistance, reliable phenotyping requires a substantial amount of space, time and manpower, and this will be increasingly so in the context of association studies that require large sample sizes. There is, thus, a need for HTP methods that are accurate and yet predictive of field performance. Particularly, in view of the differential impact of mechanisms to specialist and generalist insects as discussed earlier, insect and plant performance are not necessarily correlated with 316


each other, as high levels of deterrent compounds do not always negatively affect the performance of herbivorous insects [34,35], and good insect performance does not always result in reduced plant performance (Box 2). Therefore, both plant and insect traits are relevant for assessing the underlying mechanisms of insect resistance. Because association mapping requires at least hundreds of plant lines to be screened, it poses some challenges to the phenotyping efforts. Below some potential HTP techniques for assessing insect resistance are discussed. High-throughput phenotyping of plant defense In the past decades, plant phenotyping techniques have gone through major developments [59,60]. Several of these methods can be applied to detect antixenosis, antibiosis or tolerance against insects and the benefits and costs involved for the plant (Box 2). Metabolite profiling techniques, such as mass spectrometry and nuclear magnetic resonance, are the most obvious methods for screening primary and secondary proteins and metabolites in large-scale experiments [60]. However, image processing techniques are also highly suitable for HTP platforms [59– 61]. These techniques translate changes in the spectral signature of a plant to quantify characteristics concerning plant growth, yield and (a)biotic stress. In the visible spectrum it is possible to detect damage caused by leafchewing insects or for example silver damage due to thrips feeding [62]. Multicolor fluorescence imaging has been used to assess feeding damage of mites and stylet penetrations of whiteflies [63]. In the near-infrared spectrum, stress-related changes in plants and changes in organic compounds can be detected [64–67]. High-throughput phenotyping of insect performance and preference Assessing insect performance rather than that of the host plant, delivers the opportunity to study the direct and indirect impacts of plant nutritional quality and defense mechanisms on the dynamics of the herbivore population [52]. Although a wide variety of insect phenotyping techniques is available, only a marginal proportion of these techniques is translated into high-throughput devices. This field in particular faces some challenges in developing methodologies that have low demands in terms of space, time and labor but yet are accurate and predictive of field performance. Most insect studies have focused on insect performance, e.g. population density, insect growth, development rate, fecundity, survival and the intrinsic rate of population increase (rm) [52]. These parameters are correlated to both antixenosis and antibiosis. Assessing insect performance can be time consuming, depending on the generation time and life cycle of the insect, and is usually done in a nonautomated way [51,68]. Image analysis of photographs or videos represents potential for automated indexing of insect performance parameters (e.g. the number of eggs, larvae and surviving adults). A behavioral assay can, in contrast to just monitoring insect performance or plant traits, result in a detailed chronological dataset of the process of host selection and food uptake. An additional advantage is that a behavioral

Review assay can potentially be much shorter than an end-point measurement of reproduction and survival [69,70]. Food uptake is an important aspect of insect behavior, related to insect performance and host–plant resistance [52]. Electronic monitoring of probing behavior in piercing–sucking insects has proven to be successful in finding feeding deterrents [71,72], but is hardly feasible in large-scale experiments necessary for association mapping. Alternatively, automated tracking of insect behavior allows measuring multiple factors involved in host selection, e.g. host preference, mobility of the insect and the timing and duration of food uptake. An additional advantage is that it allows screening the behavior of multiple individual insects and multiple arenas simultaneously [73–78]. The major challenge of high-throughput video-monitoring of insect behavior is to realize two- or three-dimensional arenas predictive of field performance [79]. In large-field trials, a mark–release–recapture technique can be a cost-effective method to assess host preference and population growth of insects [80]. Ultimately, the choice of a phenotyping technique will largely depend on the study system and research focus. Future perspectives The development of accurate and field-predictive HTP will allow GWA mapping to increase insight into the genetic architecture of plant resistance to generalist versus specialist insects that will contribute to the development of host–plant resistance in crops. ‘Blind’ screening, unbiased by parental phenotypes and candidate genes, is the basis of this method and opens the opportunity to analyze the full scope of existing natural variation in resistance mechanisms. Although current studies mainly focus on one or a few candidate mechanisms, the untargeted nature of GWA mapping will include multiple factors that contribute to resistance against generalist and specialist herbivores. We expect that the current assumptions about differential resistance mechanisms against specialists and generalists can be addressed more comprehensively using such an unbiased approach. A further step forward will be the integration of association mapping with transcriptomics, proteomics and metabolomics, to assess insect resistance at the levels of the genotype, gene expression, and metabolite and protein networks [15,19,27,81–84]. However, a major determinant of finding phenotype–genotype associations is imposed by the plant species itself. At present, next generation sequencing technologies result in an increasing amount of sequenced plant species and lines within a species, so that the scope of plant–insect association studies will be expanded to additional biological systems with a wider array of plant–insect interactions and resistance mechanisms. In the near future, the genomes and genetic variation of an increasing number of insect herbivores will also become available [85]. Comparing functional mechanisms in insect and plant populations at the genomic level will allow the development of ecological insights into the evolution of plant–herbivore interactions and will take host–plant resistance studies to the next level. Acknowledgments We thank Fred van Eeuwijk and three anonymous reviewers for stimulating discussions and comments on a previous version of the


manuscript. The Netherlands Organization for Scientific Research (NWO) is acknowledged for funding through the Technology Foundation (grant STW10989, Perspectief Programme ‘Learning from Nature’).

References 1 Panda, N. and Khush, G.S. (1995) Host Plant Resistance to Insects, CABI International 2 Schoonhoven, L.M. et al. (2005) Insect-Plant Biology, Oxford University Press 3 Alonso-Blanco, C. and Koornneef, M. (2000) Naturally occurring variation in Arabidopsis: an underexploited resource for plant genetics. Trends Plant Sci. 5, 22–29 4 Anderson, J.T. and Mitchell-Olds, T. (2011) Ecological genetics and genomics of plant defences: evidence and approaches. Funct. Ecol. 25, 312–324 5 Gols, R. et al. (2008) Performance of generalist and specialist herbivores and their endoparasitoids differs on cultivated and wild Brassica populations. J. Chem. Ecol. 34, 132–143 6 Mewis, I. et al. (2006) Gene expression and glucosinolate accumulation in Arabidopsis thaliana in response to generalist and specialist herbivores of different feeding guilds and the role of defense signaling pathways. Phytochemistry 67, 2450–2462 7 Rohr, F. et al. (2011) Impact of hydroxylated and non-hydroxylated aliphatic glucosinolates in Arabidopsis thaliana crosses on plant resistance against a generalist and a specialist herbivore. Chemoecology 21, 171–180 8 Karban, R. and Agrawal, A.A. (2002) Herbivore offense. Annu. Rev. Ecol. Syst. 33, 641–664 9 Price, P.W. et al. (2011) Insect Ecology: Behavior, Populations and Communities, Cambridge University Press 10 Loxdale, H.D. et al. (2011) The evolutionary improbability of ‘generalism’ in nature, with special reference to insects. Biol. J. Linn. Soc. 103, 1–18 11 Vandenborre, G. et al. (2010) Nicotiana tabacum agglutinin is active against Lepidopteran pest insects. J. Exp. Bot. 61, 1003–1014 12 Brotman, Y. et al. (2011) Identification of enzymatic and regulatory genes of plant metabolism through QTL analysis in Arabidopsis. J. Plant Physiol. 168, 1387–1394 13 Doboń, A. et al. (2011) Quantitative genetic analysis of salicylic acid perception in Arabidopsis. Planta 234, 671–684 14 Balasubramanian, S. et al. (2009) QTL mapping in new Arabidopsis thaliana advanced intercross-recombinant inbred lines. PLoS ONE 4, e4318 15 Ingvarsson, P.K. and Street, N.R. (2011) Association genetics of complex traits in plants. New Phytol. 189, 909–922 16 Atwell, S. et al. (2010) Genome-wide association study of 107 phenotypes in Arabidopsis thaliana inbred lines. Nature 465, 627–631 17 Zhu, C. et al. (2008) Status and prospects of association mapping in plants. Plant Gen. 1, 5–20 18 Chan, E.K.F. et al. (2010) Understanding the evolution of defense metabolites in Arabidopsis thaliana using genome-wide association mapping. Genetics 185, 991–1007 19 Chan, E.K.F. et al. (2011) Combining genome-wide association mapping and transcriptional networks to identify novel genes controlling glucosinolates in Arabidopsis thaliana. PLoS Biol. 9, e1001125 20 Yu, J.M. and Buckler, E.S. (2006) Genetic association mapping and genome organization of maize. Curr. Opin. Biotechnol. 17, 155–160 21 Mitchell-Olds, T. (2010) Complex-trait analysis in plants. Genome Biol. 11, 1–3 22 Price, A.L. et al. (2006) Principal components analysis corrects for stratification in genome-wide association studies. Nat. Genet. 38, 904–909 23 Yu, J.M. et al. (2006) A unified mixed-model method for association mapping that accounts for multiple levels of relatedness. Nat. Genet. 38, 203–208 24 Zhao, K. et al. (2007) An Arabidopsis example of association mapping in structured samples. PLoS Genet. 3, e4 25 Bergelson, J. and Roux, F. (2010) Towards identifying genes underlying ecologically relevant traits in Arabidopsis thaliana. Nat. Rev. Genet. 11, 867–879 26 Brachi, B. et al. (2010) Linkage and association mapping of Arabidopsis thaliana flowering time in nature. PLoS Genet. 6, e1000940 317

Review 27 Myles, S. et al. (2009) Association mapping: critical considerations shift from genotyping to experimental design. Plant Cell 21, 2194–2202 28 Visscher, P.M. (2008) Sizing up human height variation. Nat. Genet. 40, 489–490 29 Mewis, I. et al. (2005) Major signaling pathways modulate Arabidopsis glucosinolate accumulation and response to both phloem-feeding and chewing insects. Plant Physiol. 138, 1149–1162 30 Poelman, E.H. et al. (2008) Consequences of variation in plant defense for biodiversity at higher trophic levels. Trends Plant Sci. 13, 534–541 31 Becerra, J.X. (2007) The impact of herbivore–plant coevolution on plant community structure. Proc. Natl. Acad. Sci. U.S.A. 104, 7483–7488 32 Thompson, J.N. (2005) The Geographic Mosaic of Coevolution, University of Chicago Press 33 Vermeer, K.M.C.A. et al. (2011) The potential of a population genomics approach to analyse geographic mosaics of plant–insect coevolution. Evol. Ecol. 25, 977–992 34 Van der Meijden, E. (1996) Plant defence, an evolutionary dilemma: contrasting effects of (specialist and generalist) herbivores and natural enemies. Entomologia Experimentalis et Applicata 80, 307–310 35 Poelman, E.H. et al. (2010) Herbivore-induced plant responses in Brassica oleracea prevail over effects of constitutive resistance and result in enhanced herbivore attack. Ecol. Entomol. 35, 240–247 36 Blackman, R.L. and Eastop, V.F. (2006) Aphids on the World’s Herbaceous Plants and Schrubs, John Wiley & Sons, Ltd and Natural History Museum 37 Steppuhn, A. et al. (2004) Nicotine’s defensive function in nature. PLoS Biol. 2, e217 38 Agrawal, A.A. and Kurashige, N.S. (2003) A role for isothiocyanates in plant resistance against the specialist herbivore Pieris rapae. J. Chem. Ecol. 29, 1403–1415 39 Reymond, P. et al. (2004) A conserved transcript pattern in response to a specialist and a generalist herbivore. Plant Cell 16, 3132–3147 40 Steppuhn, A. and Baldwin, I.T. (2007) Resistance management in a native plant: nicotine prevents herbivores from compensating for plant protease inhibitors. Ecol. Lett. 10, 499–511 41 Bidart-Bouzat, M. and Kliebenstein, D. (2011) An ecological genomic approach challenging the paradigm of differential plant responses to specialist versus generalist insect herbivores. Oecologia 167, 677–689 42 Govind, G. et al. (2010) Unbiased transcriptional comparisons of generalist and specialist herbivores feeding on progressively defenseless Nicotiana attenuata plants. PLoS ONE 5, e8735 43 Sutter, R. and Muller, C. (2011) Mining for treatment-specific and general changes in target compounds and metabolic fingerprints in response to herbivory and phytohormones in Plantago lanceolata. New Phytol. 191, 1069–1082 44 Aranzana, M.J. et al. (2005) Genome-wide association mapping in Arabidopsis identifies previously known flowering time and pathogen resistance genes. PLoS Genet. 1, e60 45 Nemri, A. et al. (2010) Genome-wide survey of Arabidopsis natural variation in downy mildew resistance using combined association and linkage mapping. Proc. Natl. Acad. Sci. U.S.A. 107, 10302–10307 46 Schon, C.C. et al. (2004) Quantitative trait locus mapping based on resampling in a vast maize testcross experiment and its relevance to quantitative genetics for complex traits. Genetics 167, 485–498 47 Fischer, R.A. (1918) The correlation between relatives on the supposition of Mendelian inheritance. Trans. R. Soc. Edinb. 52, 399–433 48 Li, X. et al. (2011) Mapping QTLs for improving grain yield using the USDA rice mini-core collection. Planta 234, 347–361 49 Hammer, G. et al. (2006) Models for navigating biological complexity in breeding improved crop plants. Trends Plant Sci. 11, 587–593 50 Benfey, P.N. and Mitchell-Olds, T. (2008) Perspective – from genotype to phenotype: systems biology meets natural variation. Science 320, 495–497 51 Krips, O.E. et al. (1998) Intrinsic rate of population increase of the spider mite Tetranychus urticae on the ornamental crop gerbera: intraspecific variation in host plant and herbivore. Entomologia Experimentalis et Applicata 89, 159–168 52 Awmack, C.S. and Leather, S.R. (2002) Host plant quality and fecundity in herbivorous insects. Annu. Rev. Entomol. 47, 817–844 53 Eberius, M. and Lima-Guerra, J. (2009) High-throughput plant phenotyping – data acquisition, transformation, and analysis. In 318


54

55 56

57

58 59 60 61

62 63

64 65

66 67 68

69

70

71

72

73

74

75

76

77

Bioinformatics: Tools and Applications (Edwards, D. et al., eds), pp. 259–278, Springer Science and Business Media Li, Y. et al. (2010) Association mapping of local climate-sensitive quantitative trait loci in Arabidopsis thaliana. Proc. Natl. Acad. Sci. U.S.A. 107, 21199–21204 Kang, H.M. et al. (2008) Efficient control of population structure in model organism association mapping. Genetics 178, 1709–1723 Hall, D. et al. (2010) Using association mapping to dissect the genetic basis of complex traits in plants. Brief. Funct. Genomics 9, 157–165 Ballare´, C.L. (2009) Illuminated behavior. Phytochrome as a key regulator of light foraging and plant herbivore defense. Plant Cell Environ. 32, 713–725 Holopainen, J.K. and Gershenzon, J. (2010) Multiple stress factors and the emission of plant VOCs. Trends Plant Sci. 15, 176–184 Montes, J.M. et al. (2007) Novel throughput phenotyping platforms in plant genetic studies. Trends Plant Sci. 12, 433–436 Fernie, A.R. and Schauer, N. (2009) Metabolomics-assisted breeding: a viable option for crop improvement? Trends Genet. 25, 39–48 Kokorian, J. et al. (2010) An ImageJ based measurement setup for automated phenotyping of plants. In Proceedings of the ImageJ User and Developer Conference 2010 (Jahnen, A. and Moll, C., eds), pp. 178– 182, Centre de Recherche Public Henri Tudor Abe, H. et al. (2008) Function of jasmonate in response and tolerance of Arabidopsis to thrip feeding. Plant Cell Physiol. 49, 68–80 Buschmann, C. and Lichtenthaler, H.K. (1998) Principles and characteristics of multi-colour fluorescence imaging of plants. J. Plant Physiol. 152, 297–314 Cozzolino, D. (2009) Near infrared spectroscopy in natural products analysis. Planta Med. 75, 746–756 Rutherford, R.S. and van Staden, J. (1996) Towards a rapid nearinfrared technique for prediction of resistance to sugarcane borer Eldana saccharina Walker (Lepidoptera: Pyralidae) using stalk surface wax. J. Chem. Ecol. 22, 681–694 Kramer, K.J. et al. (2000) Transgenic avidin maize is resistant to storage insect pests. Nat. Biotechnol. 18, 670–674 Chaerle, L. and Van der Straeten, D. (2000) Imaging techniques and the early detection of plant stress. Trends Plant Sci. 5, 495–501 Poelman, E.H. et al. (2008) Performance of specialist and generalist herbivores feeding on cabbage cultivars is not explained by glucosinolate profiles. Entomologia Experimentalis et Applicata 127, 218–228 Hardie, J. et al. (1992) The combination of electronic monitoring and video-assisted observations of plant penetration by aphids and behavioural effects of polygodial. Entomologica Experimentalis et Applicata 62, 233–239 Foster, S.P. et al. (2005) Reduced response of insecticide-resistant aphids and attraction of parasitoids to aphid alarm pheromone; a potential fitness trade-off. Bull. Entomol. Res. 95, 37–46 Tjallingii, W.F. and Hogen Esch, T. (1993) Fine structure of aphid stylet routes in plant tissues in correlation with EPG signals. Physiol. Entomol. 18, 317–328 Backus, E.A. and Bennett, W.H. (2009) The AC–DC correlation monitor: new EPG design with flexible input resistors to detect both R and emf components for any piercing–sucking hemipteran. J. Insect Physiol. 55, 869–884 Allemand, R. et al. (1994) Behavioral circadian-rhythms measured in real-time by automatic image-analysis – applications in parasitoid insects. Physiol. Entomol. 19, 1–8 Noldus, L.P.J.J. et al. (2002) Computerised video tracking, movement analysis and behaviour recognition in insects. Comput. Electron. Agric. 35, 201–227 Reynolds, D.R and Riley, J.R. (2002) Remote-sensing, telemetric and computer-based technologies for investigating insect movement: a survey of existing and potential techniques. Comput. Electron. Agric. 35, 271–307 Beeuwkes, J. et al. (2008) 3-D flight behaviour of the malaria mosquito Anopheles gambiae s.s. inside an odour plume. Proc. Neth. Entomol. Soc. Meet. 19, 137–146 Lacey, E.S. and Carde, R.T. (2011) Activation, orientation and landing of female Culex quinquefasciatus in response to carbon dioxide and odour from human feet: 3-D flight analysis in a wind tunnel. Med. Vet. Entomol. 25, 94–103

Review


78 Pistori, H. et al. (2010) Mice and larvae tracking using a particle filter with an auto-adjustable observation model. Pattern Recognit. Lett. 31, 337–346 79 Prasifka, J.R. et al. (2010) Video-tracking and on-plant tests show Cry1Ab resistance influences behavior and survival of neonate Ostrinia nubilalis following exposure to Bt maize. J. Insect Behav. 23, 1–11 80 Hagler, J.R. and Jackson, C.G. (2001) Methods for marking insects: current techniques and future prospects. Annu. Rev. Entomol. 46, 511–543 81 Keurentjes, J.J.B. et al. (2006) The genetics of plant metabolism. Nat. Genet. 38, 842–849 82 DellaPenna, D. and Last, R.L. (2008) Perspective – genome-enabled approaches shed new light on plant metabolism. Science 320, 479–481 83 Macel, M. et al. (2010) Metabolomics: the chemistry between ecology and genetics. Mol. Ecol. Resour. 10, 583–593 84 Keurentjes, J.J.B. et al. (2011) Redefining plant systems biology: from cell to ecosystem. Trends Plant Sci. 16, 183–190 85 Whiteman, N.K. and Jander, G. (2010) Genome-enabled research on the ecology of plant–insect interactions. Plant Physiol. 154, 475–478 86 Ehrenreich, I.M. et al. (2009) Candidate gene association mapping of Arabidopsis flowering time. Genetics 183, 325–335 87 De Vos, M. et al. (2005) Signal signature and transcriptome changes of Arabidopsis during pathogen and insect attack. Mol. Plant Microbe Interact. 18, 923–937

88 De Vos, M. and Jander, G. (2009) Myzus persicae (green peach aphid) salivary components induce defence responses in Arabidopsis thaliana. Plant Cell Environ. 32, 1548–1560 89 Bruessow, F. et al. (2010) Insect eggs suppress plant defence against chewing herbivores. Plant J. 62, 876–885 90 Van Poecke, R.M.P. (2007) Arabidopsis–insect interactions. In The Arabidopsis Book (Somerville, C.R. and Meyerowitz, E.M., eds), doi:10.1199/tab.0107, www.aspb.org/publications/arabidopsis/, pp. 1–34, American Society of Plant Biologists 91 Pieterse, C.M.J. et al. (2009) Networking by small-molecule hormones in plant immunity. Nat. Chem. Biol. 5, 308–316 92 Verhage, A. et al. (2011) Rewiring of the jasmonate signaling pathway in Arabidopsis during insect herbivory. Front. Plant Sci. 2, 1–12 93 Dicke, M. (1999) Evolution of induced indirect defence of plants. In The Ecology and Evolution of Inducible Defenses (Tollrian, R. and Harvell, C.D., eds), pp. 62–88, Princeton University Press 94 Heil, M. (2008) Indirect defence via tritrophic interactions. New Phytol. 178, 41–61 95 Dicke, M. and Baldwin, I.T. (2010) The evolutionary context for herbivore-induced plant volatiles: beyond the ‘cry for help’. Trends Plant Sci. 15, 167–175 96 Hilker, M. and Meiners, T. (2006) Early herbivore alert: insect eggs induce plant defense. J. Chem. Ecol. 32, 1379–1397

Plant Science Conferences in 2012 10th International Congress on Plant Molecular Biology 21 – 26 October, 2012 Jeju, South Korea http://www.ipmb2012.org/ ABIC 2012 2 – 6 September, 2012 Rotorua, New Zealand http://www.abic2012.com

Plant Science Conferences in 2013 24th International Conference on Arabidopsis research (ICAR) 24 – 28 June, 2013 Sydney, Australia http://www.icar2013.org/ Plant Biology 2013 20 – 24 July, 2013 Providence, USA http://www.aspb.org/meetings ----------------------------------------------------------------------------------------------------------------------------------Suggest a conference Please use the form at http://www.cell.com/conferences/SuggestConference to suggest a conference for Cell Press the Conference Calendar.

319