8. Biotic Stress

8.1.1 Types of Pathogens: Viruses, Bacteria, Fungi, Oomycetes and Nematodes

The specificity of plant pathogens can vary widely. Often the host range is restricted to only one species. This applies, in particular, to biotrophic pathogens. However, there are also examples of pathogens infecting multiple hosts. Several necrotrophic fungi (e.g. Botrytis species) cause disease in hundreds of plant species.

Viruses basically consist of a protein coat enclosing nucleic acids (DNA or RNA) that carry the information for the synthesis of only a handful of proteins. For their replication they have to force a host cell to produce these proteins, which then assemble and give rise to new virus particles. Unlike viruses infecting bacterial or animal cells, plant viruses do not have to destroy their host cell in order to release newly formed virus particles, which then attack neighbouring cells. Instead, they move through plasmodesmata from cell to cell and spread systemically via the phloem. Initial infection of a host plant mostly occurs through inoculation of the phloem via virus-carrying phloem-feeding insects such as aphids. Typical symptoms of virus-infected plants are yellowing or a mosaic-like colouration of leaves, stunted growth and sometimes the death of a plant.

Microbial pathogens typically infect only specific tissues or organs. Phytopathogenic bacteria, for example, can cause root rot, leaf chlorosis or stem dieback, depending on the pathogen and its infection strategy. Bacterial pathogens usually enter a plant through wounds or the stomata and multiply in the apoplast. Only a minute fraction of bacterial species have evolved the ability to infect plants. No more than a few hundred species have been described as pathogens of crop plants. Most disease-causing bacteria are gram positive and belong to a limited number of clades (e.g. Xanthomonas, Pseudomonas, Erwinia and Dickeya).

Economically the most important groups of plant pathogens are the oomycetes and fungi. Oomycetes were for a long time regarded as fungi, but phylogenetically they belong to the Chromalveolata and are thus very distant from fungi. Arguably the most famous plant pathogen is an oomycete. Phytophthora infestans is the causal agent of potato blight and was responsible for the Irish famine in the mid-nineteenth century, which left millions of people dead and triggered a mass emigration from Ireland to North America. Downy mildews are oomycetes too.

Among the possibly millions of fungal species living in association with plants, only a small fraction cause disease (Crous et al. 2015). Nonetheless, no other group of pathogens accounts for more yield loss than fungi. Important genera with phytopathogenic species include Botrytis, Alternaria, Colletotrichum, Fusarium and Puccinia.

A few taxa of nematodes are plant pathogens. Presumably because they almost exclusively attack the root system, they are less well studied. Phytopathogenic nematodes are obligate biotrophs—that is, they are absolutely dependent on a plant host for existence. Endo- and ectoparasites are differentiated according to whether they penetrate root tissue or not, respectively. Most of the considerable nematode damage to crop plants is due to infection by cyst and root-knot nematodes.

8.1.2 Pathogenicity Mechanisms

Pathogens have to gain access to the plant’s resources and be able to grow and reproduce rapidly within plant tissues. Both processes require effective suppression of the plant’s immune system. Invasion of plants is achieved through widely different strategies depending on the type of pathogen. As mentioned, viruses are transmitted by insect vectors. The same applies to some phytopathogenic bacteria such as phytoplasma. Most bacteria do not have the ability to actively penetrate plant surfaces. However, after entry through wounds or stomata into the apoplast they can degrade plant cell walls via the secretion of a suite of enzymes such as cellulases and pectinases. Cell wall degradation causes disease symptoms such as softening of tissues. For some Erwinia species it is known that the bacteria “wait” to secrete these enzymes only when a sufficiently large number of bacterial cells are present at the infection site to overwhelm the plant. Earlier release would only alert the plant immune system (as is discussed in Sect. 8.2.2). The “waiting” is mediated by quorum-sensing mechanisms, which bacteria use to monitor the density of a population.

Two other elements of bacterial virulence are common to both plant and animal pathogens—namely, toxins and effectors. The latter are molecules that bacteria transfer into host cells via secretion systems (e.g. type III secretion system; Fig. 8.2). Effectors influence host cells through specific interaction with target structures. They interfere with the plant immune system and render the plant environment more favourable for the bacteria. An example is the protein avrBs3 secreted by the tomato pathogen Xanthomonas campestris. Even though it is encoded by the bacterial genome, it has the features of a eukaryotic transcription factor. After release into the plant cell it enters the nucleus and interacts with specific promoter elements to activate a group of genes that collectively cause hypertrophy of cells and thereby disease (Kay and Bonas 2009) (Fig. 8.2). A particularly famous example of an effector molecule is the transfer DNA (T-DNA) of Agrobacterium tumefaciens that triggers tumour growth and the synthesis of organic nutrients for the bacteria (Box 8.1).

Like many effectors, toxins appear to often function as suppressors of plant defences. This is illustrated by coronatine, a molecule synthesised by Pseudomonas syringae. Coronatine mimics the phytohormone jasmonate and interacts with the jasmonate receptor COI1. Activation of the jasmonate pathway reverses the closure of stomata, which is an integral part of the plant’s defence response (Melotto et al. 2006).

Box 8.1: Agrobacterium tumefaciens

The gram-negative soil bacterium A. tumefaciens is the causal agent of crown gall disease, which affects predominantly dicotyledonous species. Infection by A. tumefaciens leads to the formation of tumours—that is, the rapid proliferation of cells and the growth of undifferentiated tissue. Tumours are induced by transfer DNA (T-DNA), which the bacterium uses to transform the genome of a plant cell. The T-DNA carries genes encoding enzymes involved in the synthesis of the phytohormones auxin and cytokinin. Also, the T-DNA leads to the synthesis of opines, amino compounds that A. tumefaciens can use as sole carbon and nitrogen source. The emergence of such a favourable habitat for the bacteria requires the stable integration of the T-DNA into the host cell genome. The required machinery is encoded on the tumour-inducing plasmid (Ti plasmid) of A. tumefaciens. The vir (for virulence) genes on the Ti plasmid become activated when Agrobacterium senses the presence of plant metabolites such as acetosyringone. These and other phenolic compounds are released by damaged plant cells at a wound site. Products of the vir genes synthesise a copy of the T-DNA, transport it into the plant host cell and integrate it randomly into the host nuclear genome (Fig. 8.3).

When the molecular mechanisms underlying the A. tumefaciens–mediated gene transfer were elucidated in the 1980s, it was immediately realised that it could be utilised to genetically transform plants with other foreign DNA of interest (Caplan et al. 1983). The T-DNA localised between the so-called right border and left border sequences on the tumour-inducing plasmid (Ti plasmid) can be replaced by any DNA, using standard recombinant DNA techniques. Transformed plant cells do not give rise to tumours anymore but instead express genes encoded by the foreign DNA. A typical plant transformation employing A. tumefaciens encompasses the co-cultivation of A. tumefaciens cells carrying a modified Ti plasmid with wounded plant tissue (e.g. leaf discs), selection of transformed plant cells—for example, based on co-transformed antibiotic resistance—and regeneration of a whole plant, using tissue culture techniques.

Although fungi and oomycetes are unrelated, their infection strategies and growth forms as plant pathogens share several similarities, so they can be described together. An infection starts from a spore adhering to the surface of a plant and the germination of this spore. Biotrophic pathogens form an appressorium (Fig. 8.4). Within this appressorium, pressure builds up, exerting a mechanical force that, together with the activity of secreted enzymes (cutinases, cellulases and pectinases), disrupts the plant cuticle and cell wall. A feeding structure called a haustorium then develops behind an invagination of the plant cell plasma membrane (Fig. 8.4). The plant cell is kept alive but now supplies organic and inorganic nutrients to the invading fungus.

Necrotrophic fungi do not build haustoria. Instead they overpower a plant with non-specific or host-specific toxins and with cell wall–degrading enzymes. The plant cells die and the fungi utilise the dead organic material. An intensively studied potent toxin is fusicoccin, a diterpene synthesised by Fusarium species. It causes rapid wilting of exposed plants by constitutively activating the plasma membrane H⁺-ATPase via the modulation of 14-3-3 proteins. In guard cells this activation prevents the closure of stomata because the cells remain hyperpolarised (Chap. 6, Sect. 6.​3).

Like bacteria, fungi and oomycetes produce effectors to suppress the plant defence response and to disturb plant development or metabolism. The phytohormone gibberellin was discovered as a molecule the fungus Gibberella fujikuroi uses to trick rice seedlings into abnormal growth. This increases their risk of lodging and thereby falling victim to the fungus. Ustilago maydis, a biotrophic pathogen of maize (causing maize smut, characterised by large tumours), secretes more than 100 effector proteins during its attack. One of them is the enzyme chorismate mutase. Following uptake into plant cells it modulates the phenylpropanoid pathway in a way that synthesis of the defence hormone salicylic acid (Sect. 8.2.3) is disturbed (Djamei et al. 2011).

Nematodes penetrate plant cell walls with feeding stylets. Movement through root tissue is aided by the release of cell wall–degrading enzymes (Davis and Mitchum 2005). Inside plant cells, feeding tubes are formed that extract nutrients from the host. Molecules secreted by the nematode trigger dramatic changes in the regulation and metabolism of root cells, which eventually give rise to giant cells or symplastically connected cells that form syncytial feeding structures for the nematode.

8.2.1 Preformed Defences Against Bacteria, Fungi and Oomycetes

Features of a plant’s morphology and anatomy serve as a first line of defence against invasion by pathogens. The cuticle and cell walls form mechanical barriers against pathogen ingress. Many potential pathogens never gain access to plant cells and their resources, because they lack the means to overcome these barriers.

A characteristic of plant biology is the synthesis of thousands of secondary metabolites—that is, compounds that, in contrast to primary metabolites, are not found in every plant cell and species. Instead, different plants produce different spectra of secondary metabolites, which nonetheless serve important functions in the interaction of a plant with its environment (Bednarek and Osbourn 2009). Because “secondary” might be misunderstood as meaning “of lesser importance”, “specialised metabolites” was proposed as an alternative and perhaps more meaningful term (Pichersky et al. 2006). Nonetheless, throughout this chapter the more common term “secondary metabolites” is used.

Another sensible distinction between primary and secondary metabolites is this: primary metabolites are involved in nutrition and essential metabolic processes; secondary metabolites are involved in the interaction of a plant with its environment (Buchanan et al. 2015). A prominent role is preformed chemical defence against pathogens. Plants produce a rich cocktail of compounds with antimicrobial activity, which are mostly stored in vacuoles or in specialised cells such as trichomes. Sometimes they are referred to as phytoanticipins (Fig. 8.6) and are distinguished from phytoalexins, which are synthesised de novo upon pathogen attack (Sect. 8.2.2).

Biotic interactions are regarded as an important driver of the evolution of an extremely rich diversity of secondary metabolites (Dixon 2001), a resource that provides important services to humans as a major source of pharmaceuticals. However, the spectrum of secondary metabolites produced by a given plant species is far from completely known even in the model systems. Also, because most compounds act as ingredients in a complex cocktail, few examples of a directly demonstrated antimicrobial effect exist. Saponins are widely distributed amphipathic triterpenoid or steroid glucosides with soap-like behaviour in aqueous solutions. One example is avenacin, found in oats (Fig. 8.6). Saponins provide broad-range protection against microbial pathogens. Mutant oat genotypes lacking saponin synthesis in root epidermal cells become susceptible to Gaeumannomyces graminis var. tritici, a pathogen that usually cannot infect oat but is able to colonise wheat and barley—two species that do not synthesise saponins. Conversely, successful pathogens have evolved the ability to metabolise saponins. The main tomato saponin, α-tomatine, is metabolised by the tomato pathogen Septoria lycopersici via secretion of a tomatinase. This enzyme is crucial for the pathogenicity of S. lycopersici (Bouarab et al. 2002). Another well-studied example of preformed chemical defence is the synthesis of glucosinolates (Fig. 8.6), which will be described in more detail as part of a plant’s herbivory defence (Sect. 8.3.1).

A rare alternative to chemical defence is so-called elemental defence. Metal hyperaccumulating plants apparently fend off pathogens and herbivores by accumulating metals such as Zn or Cd to toxic levels in their leaves (Chap. 7, Sect. 7.​5.​2).

8.2.2 Inducible Local Defences

Notwithstanding the presence of preformed defences, the outcome of an interaction between a plant and a potential pathogen (scenarios 2–4 in Fig. 8.5) is largely determined by the presence or absence and the activities of specific molecules on the host and pathogen sides. The major components are (1) molecules characteristic of potential pathogens (traditionally termed elicitors) and (2) plant receptors that detect such molecules, as well as (3) pathogen effectors (see above), and (4) plant proteins, encoded by resistance genes, which recognise the presence or activity of these effectors.

A characteristic feature of the plant immune system is cell autonomy. Practically every plant cell is capable of mounting defence responses. Thus, many interactions remain highly localised; this is apparent, for instance, as tiny necrotic lesions on leaves, stemming from cell death.

Depending on the severity of a pathogen attack, two principal amplitudes of defence based on two different recognition strategies can be distinguished. This is summarised in the zig-zag scheme of plant immunity (Jones and Dangl 2006) (Fig. 8.7), which provides a “unifying theory” integrating the multitude of plant defence responses and plant–pathogen interactions. The first layer fends off all potential foes that have not evolved specialised weapons to attack a particular plant. The respective pathogens are sometimes referred to as non-adapted pathogens, and the phenomenon is called non-host resistance (Sect. 8.2 and scenario 2 in Fig. 8.5).

The second, more intensive layer of defence becomes activated when a plant cell recognises an enemy equipped with effectors that could potentially overcome the plant’s defence. The responses are stronger and faster, and can eventually lead to the hypersensitive reaction—that is, localised cell death (scenario 4 in Fig. 8.5). Inability to respond to the effectors results in susceptibility (scenario 3 in Fig. 8.5).

In cases where preformed defences are not sufficient to completely prevent the activity of a potential pathogen, the first layer of inducible defence, the innate immunity, becomes active. Plant cells respond to the presence of a potential foe with the induction of various defence mechanisms. Antimicrobial compounds (phytoalexins) are synthesised, cell walls are locally reinforced and defence proteins such as enzymes (e.g. chitinases) to attack microbial cell walls are produced.

Plants do not only constitutively store a wide range of secondary metabolites that are disadvantageous for potential pathogens coming in contact with these molecules (see preformed defence and phytoanticipins, Sect. 8.2.1 and Fig. 8.6). In addition, the recognition of a potential threat triggers the de novo synthesis of compounds with antimicrobial activity against a broad spectrum of fungi, oomycetes and bacteria. Collectively they are termed phytoalexins (Greek alekein = “to fend off”) (Ahuja et al. 2012). The molecular mechanisms underlying their toxicity are largely unknown. Phytoalexins comprise a heterogeneous group of metabolites from different classes of secondary plant products. The predominant ones are isoprenoids, flavonoids and stilbenes. Examples include the stilbene resveratrol in grapevine and several isoflavonoids (e.g. medicarpin, pisatin) in legumes. Phytoalexin synthesis and exocytotic release are often highly localised at sites of pathogen attack. This is exemplified by vesicles filled with reddish defence compounds accumulating around a fungal penetration site on Sorghum leaves (Snyder and Nicholson 1990) (Fig. 8.8). Such localised response requires polarisation of the host cell, movement of particular organelles by the cytoskeleton and targeted secretion (Kwon et al. 2008).

Likewise, cell walls are strengthened locally—for instance, right underneath an attempted penetration site of a biotrophic fungus (Fig. 8.4b). Small papillae are formed by deposition of callose (a β-1,3-glucan) and lignin. Cross-linking of extracellular proteins with the cell wall matrix further reinforces the mechanical barrier to fungal ingress.

A third element of the defence response is the synthesis of many pathogenesis-related proteins (PR proteins) which, contrary to what the name suggests, are in fact proteins thought to have antimicrobial activity. As for a range of other stress-related proteins in plants (see dehydrins in Chap. 6), their actual biochemical function is poorly understood. Exceptions are chitinases and glucanases, which attack fungal cells walls.

Defence responses are dependent on changes in gene activity. Genes encoding PR proteins or enzymes involved in phytoalexin biosynthesis and in the deposition of callose are induced. The responses are mediated by pattern recognition receptors (PRRs) residing in the plasma membrane of plant cells (Fig. 8.9). PRRs bind pathogen-associated molecular patterns (PAMPs) or, more generally, microbe-associated molecular patterns (MAMPs). PAMPs/MAMPs correspond to elicitors in the older literature. Defence responses elicited by PAMPs/MAMPs result in PAMP-triggered immunity (PTI) (Figs. 8.7 and 8.9). PAMPs/MAMPs represent molecules that are highly conserved in microbial species but absent from plants. A classic example is flagellin, the major protein of bacterial flagella. Receptor-mediated perception of flg22, a 22–amino acid peptide that is part of flagellin, reliably indicates the presence of flagellated bacteria. Similarly, chitin fragments give away fungal cells in the vicinity because chitin is a major polymer of the fungal cell wall and PRRs recognising chitin fragments are known. A third group of molecules perceived by PRRs are so-called danger-associated molecular patterns (DAMPs). These are characteristic structures not of microbes but, rather, of microbial activity. Examples are pectin fragments released by cell wall–degrading polygalacturonases of necrotrophic fungi (Boller and Felix 2009).

The number of PRRs per plant species is estimated at around 100–200. Together they constitute an effective surveillance system that enables plant cells to sense the extracellular presence of many different potential microbial pathogens. The innate immunity based on this surveillance system is mechanistically very similar to animal innate immunity. However, plant cells express far more PRRs than animal cells. This is regarded as one way to compensate for the absence of an adaptive immune system in plants (Dodds and Rathjen 2010).

Activation of PRRs triggers signal transduction events that (1) result in the production of reactive oxygen species (ROS) by plasma membrane-localised reduced nicotinamide adenine dinucleotide phosphate oxidases (NADPH oxidases) and extracellular peroxidases, causing an oxidative burst; and (2) result in altered gene regulation (Fig. 8.10). The oxidative burst has at least three different functions. ROS production in the apoplast supports local strengthening of the cell wall. H₂O₂ can be used as a substrate to cross-link molecules of the cell wall matrix. ROS can also directly damage pathogens. Finally, ROS may serve as signalling molecules and modulate the host cell response.

Transduction of the signal from activated PRRs to ROS-producing enzymes and to transcription factors proceeds via multiple phosphorylation steps. PRRs themselves are receptor kinases. Upon ligand binding they interact with co-receptors and become activated (Fig. 8.10). Ca²⁺ spikes and the activation of mitogen-activated protein kinase cascades (MAP kinase cascades) eventually modulate a complex network of transcriptional regulators that ensures an adequate response—that is, a reprogramming of the cellular metabolism, which is adjusted to the severity of the pathogen threat (Tsuda and Somssich 2015). Like all plant stress responses, PTI represents a balancing act to optimise resource allocation between defence and growth.

The second layer of defence is activated when pathogen effectors are recognised. At this level of immunity most of the co-evolutionary dynamics take place because unlike MAMPs/PAMPs/DAMPs, effectors are highly variable and also dispensable—that is, not essential for normal cell functioning (Dodds and Rathjen 2010). Therefore, effector recognition requires large sets of resistance genes (R genes), which are plant species specific and even genotype specific, while many PRRs are conserved across plant families because they perceive PAMPs, which are widely conserved as well. R gene products mediate the effector perception and the resulting effector-triggered immunity (ETI; see zig-zag scheme in Fig. 8.7).

As described in Sect. 8.1.2, during infection, plant pathogens synthesise and release up to several hundred effector molecules, which target a variety of sites in a host cell. Successful manipulation of the host defence system by well-documented examples such as avrBs3 (Fig. 8.2) and chorismate mutase results in effector-triggered susceptibility (ETS; see zig-zag scheme Fig. 8.7). A comparison across many different plant pathogen systems, which is now possible because of genome-wide studies, reveals that in spite of the huge diversity of effector molecules produced by bacteria, fungi, oomycetes or nematodes, the number of host processes that are predominantly targeted is rather small (Mukhtar et al. 2011; Dou and Zhou 2012). The list provides deep biological insight into key processes of defence and pathogenicity. It includes PRR-dependent immune responses, hormonal regulation, secretory processes and cell death programmes. Interference with PTI, for instance, is achieved either by effector proteins that inactivate PRRs directly or by proteins that reduce the concentration of DAMPs. The Pseudomonas syringae effector avrPto, for example, directly interacts with the flagellin receptor and other PPRs (Xiang et al. 2008). The Cladosporium fulvum effector Ecp6 binds chitin fragments in the apoplast and thereby inhibits activation of the respective PRRs sensing these DAMPs (de Jonge et al. 2010) (Fig. 8.11).

Upon recognition of an effector, a plant cell mounts a strong and rapid response that confers ETI. The most extreme manifestation of this response is the hypersensitive reaction, meaning programmed cell death, which effectively blocks further spread of an invading pathogen.

ETI is dependent on R genes. R genes can confer resistance to fungi, oomycetes, bacteria, nematodes and even viruses. They have been used extensively in crop breeding to achieve resistance against commercially important pathogens. Their molecular nature, however, was not known until 1995. Most R genes encode proteins that share certain structural domains—namely, leucine-rich repeats (LRRs), often involved in intermolecular interactions—and nucleotide-binding (NB) domains. There are, however, intriguing exceptions. The R gene Bs3 in pepper “outsmarts” pathogens carrying the effector avrBs3 (Fig. 8.2). When activated, Bs3 triggers a hypersensitive reaction (Römer et al. 2007). Bs3 is activated by avrBs3 binding because the Bs3 promoter carries the same cis element that is targeted by avrBs3 to cause disease. In other words, by attacking a plant cell with Bs3 in its genome, Xanthomonas campestris terminates its own infection attempt.

R gene products have two principal functions: recognition of pathogen effectors and triggering of a cell death programme upon activation (Figs. 8.7, 8.9 and 8.10). The gene-for-gene interaction between pathogen effectors and host R genes raises the question: How do plants keep up with pathogens in the evolutionary race? Microbial pathogens can evolve new effectors much faster than plants can evolve R genes. Their mutation rate is higher and their generation time much shorter. In fact, in an agricultural context, resistance is often broken by the emergence of new pathogen races (Box 8.2). The adaptive immune system of animals generates a seemingly unlimited number of immune receptors and antibodies to cope with the genetic diversity and rapid evolution of pathogens. Part of the answer for plants lies in the way R gene products recognise effectors. Not all of them directly interact with effectors. As postulated by the “guard hypothesis” (Jones and Dangl 2006), many R gene products guard critical virulence targets and recognise not the effector itself but the abnormal modification of a host protein caused by an effector (Fig. 8.12). In this way a particular R gene can sense several different effectors, in some cases even from pathogens as diverse as fungi and nematodes. Structurally unrelated effectors of the fungal pathogen C. fulvum and the plant-parasitic nematode Globodera rostochiensis both perturb the function of an extracellular protease (Rcr3) that is important for resistance of tomato cells. The R gene Cf2 encodes an immune receptor that guards Rcr3—that is, it senses changes in Rcr3 and confers resistance against genotypes of the fungus and the nematode that express the respective effectors (Lozano-Torres et al. 2012).

Another reason why plants are not overwhelmed in the “arms race” with microbial pathogens is the diversity of R genes, which is driven by the continuous emergence of pathogen effectors. Plant genomes carry large numbers of R genes, many of them organised in clusters facilitating recombination events. R genes are consistently the most polymorphic class of genes in plant genomes. Thus, within a given plant population, hundreds or thousands of R gene variants exist that can sense effector-dependent pathogen attacks. Genetic diversity within a population reduces the vulnerability to pathogens.

The second function of R gene products, besides direct or indirect recognition of pathogen effectors (i.e. signalling to activate cell death programmes (Figs. 8.9 and 8.10)), is molecularly poorly understood.

Box 8.2: Ug99 Stem Rust on Wheat

Among the important traits of elite bread wheat varieties that made the dramatic yield increases of the “Green Revolution” possible was resistance against wheat stem rust, caused by the basidiomycete Puccinia graminis. Wheat stem rust can cause severe losses. Resistance was introduced into wheat cultivars by plant breeders around Norman Borlaug, who was awarded the Nobel Peace Prize in 1970 for his contributions to the Green Revolution. Because pathogens constantly evolve, new races can arise that escape the effects of prevailing resistance genes. In 1999 a highly virulent new race of P. graminis was discovered (named Ug99 because it was first detected in Uganda), which potentially threatens most of the currently cultivated wheat varieties in the world. Initially inspired and coordinated by the late Norman Borlaug, global efforts were launched to contain Ug99’s spread and to protect the world’s wheat cultivation against this threat. In 2013 the cloning of resistance genes against Ug99 was reported. These genes were found in the wild wheat relative Aegilops tauschii (Periyannan et al. 2013) and in Triticum monococcum (Saintenac et al. 2013). These genes—Sr33 and Sr35, respectively—can now be transferred into elite cultivars either via more time-consuming conventional breeding or via genetic engineering to confer Ug99 resistance (Fig. 8.13).

8.2.3 Inducible Systemic Resistance

Besides the cell-autonomous plant immune system there is an additional layer of defence. Cells under attack do not only mount a local response; they also send out signals that activate defence in remote tissues and organs. This phenomenon is called systemic acquired resistance (SAR) (Fig. 8.14). SAR can last for weeks or months and protect a plant from secondary infection—that is, render an otherwise compatible interaction incompatible. SAR is non-specific in the sense that (1) any potential pathogen (avirulent or virulent) triggers SAR, and (2) protection is achieved against a wide range of pathogens. Furthermore, SAR prepares a plant to respond more rapidly and strongly to subsequent pathogen attack. The plant becomes primed for defence (defence priming).

The main changes in systemic tissue involve the synthesis of PR proteins (Fig. 8.15). Unlike ETI, SAR does not involve cell death. Most of the PR proteins are either secreted into the apoplast (i.e. into the space colonised by potential pathogens) or stored in vacuoles. As mentioned above, the actual molecular mechanisms explaining PR protein–mediated protection are barely known. PR protein function is nearly impossible to dissect genetically, because many proteins act in concert and PR genes are often clustered in plant genomes, so multiple knockout mutants are practically impossible to obtain (Fu and Dong 2013).

The trigger for SAR is local ETI—that is, the response of cells under direct attack by a pathogen. ETI leads to an increase in the concentration of the defence hormone salicylic acid. This in turn elicits the generation of phloem-mobile signals, whose molecular nature has not been unequivocally shown (Fig. 8.16). Several compounds are candidates, among them azelaic acid (Jung et al. 2009) and pipecolic acid (Návarová et al. 2012).

Arrival of the signals in the systemic tissue again results in an increase in the salicylic acid concentration, albeit smaller than that in the local tissue. Salicylic acid is perceived by different receptors locally and systemically. In cells under direct pathogen attack, salicylic acid positively regulates cell death, while in systemic tissue it suppresses cell death and sets signalling cascades in motion that result in PR gene activation and SAR.

8.2.4 Defence Against Viruses via Gene Silencing

Defence against the systemic spread of viruses is conferred not only by the R gene–mediated hypersensitive response or by inducible systemic resistance; in fact, the main immune system against viruses in plants is based on gene silencing. For decades it has been observed in the field and under laboratory conditions that virus-infected plants develop resistance against a second infection by the same virus or a closely related virus. Weeks after an infection the newly developed leaves remain symptom-free even when inoculated with virus particles (Soosaar et al. 2005). This phenomenon, called “recovery”, resembles the immunisation of humans or other mammals through treatment with an inactivated virus or a harmless related virus. We now know that a gene-silencing mechanism is behind the recovery from virus infection (Baulcombe 2004). This silencing targets virus-derived double-stranded RNA (dsRNA), which also explains why resistance covers not only the originally infecting virus but also viruses with a genome that is similar in sequence.

Most plant viruses are RNA viruses—that is, they carry genetic information as either single-stranded RNA (ssRNA) or double-stranded RNA instead of DNA. Different processes during the infection process give rise to the existence of dsRNA even when the genome is DNA or ssRNA. Replication catalysed by RNA-dependent RNA polymerases involves the formation of double-stranded intermediates. ssRNA can fold back and base pair into dsRNA. Bidirectional transcription of viral DNA can produce overlapping transcripts (Fig. 8.17). The presence of dsRNA triggers cleavage into small interfering RNAs (siRNAs) with a length of 21–24 nucleotides. The cleavage is catalysed by endonucleases called Dicer-like. Discovery of siRNAs in virus-infected plants was a major step in unravelling the existence of a complex world of regulatory small RNAs (Hamilton and Baulcombe 1999). One of the strands of the siRNA is then loaded onto another endonuclease, Argonaute (AGO), which is part of the so-called RNA-induced silencing complex (RISC). RISC targets RNA similar in sequence to the siRNAs and causes its degradation, thereby inhibiting replication or transcription of viral RNA. An alternative silencing mechanism is the sequence-specific methylation of DNA directed by siRNA. This mechanism is important for resistance against DNA viruses.

The protection of newly developed leaves requires the spreading of the RNA silencing throughout the plant. The action of RNA-dependent RNA polymerases on single-stranded viral RNA results in a strong amplification of the siRNAs. They can move through plasmodesmata and the phloem. In receiving cells they then trigger RNA-dependent RNA polymerases to produce siRNAs (Fig. 8.18). Viruses in turn have independently evolved a variety of proteins that suppress dsRNA-dependent virus resistance by interfering with components of the RNA silencing pathway. For example, suppressor protein p19, encoded in the genome of the tomato bushy stunt virus, binds to siRNA and prevents its integration into RISC. Thus, the corresponding viral RNAs can no longer be targeted for degradation (Fig. 8.17).

8.3.1 Constitutive Defences

Plants rely heavily on chemical defences against herbivores. As discussed in Sect. 8.2.1, plants produce a plethora of secondary metabolites. More than 200,000 different chemical structures have been identified and presumably many more exist in nature. Per plant species, several thousand of these metabolites can be expected. A major biological role is protection against herbivore attack.

The major compound classes implicated in herbivore defence are alkaloids, phenolic compounds, terpenoids (including steroids), cyanogenic glucosides and glucosinolates (Mithöfer and Boland 2012) (Fig. 8.21). Their protective function can be due to toxicity, anti-digestive effects or the fact that they make a plant tissue unpalatable for herbivores. The modes of action or target sites in the herbivores, however, are barely understood at the molecular level. The few existing examples illustrate principal mechanisms such as interference with neurotransmission, inhibition of metabolic processes or disruption of signal transduction pathways. A negative effect of toxic compounds on the producing plant—that is, self-intoxication—is usually prevented by storage in vacuoles. Furthermore, many defence compounds are stored in an inactive form. It is only after mechanical damage has caused contact with processing enzymes localised elsewhere in the cell or in different cells that the active compounds are released (see cyanogenic glucosides and glucosinolates below, Fig. 8.25).

About 20% of all plant species produce alkaloids—N-containing metabolites synthesised from various amino acids. Many alkaloids are known as potent toxins, effective against both arthropods and vertebrates. Modes of toxicity vary. Some alkaloids affect the nervous system. For example, caffeine (an alkaloid of Coffea arabica and many other plant species) can have a paralysing effect on insects by inhibiting phosphodiesterase activity at synapses. Another famous example of a toxic alkaloid with a defence function is nicotine (Fig. 8.21). Nicotine is produced in the roots of Nicotiana tabacum and transported via the xylem to leaf cells, where it is stored in vacuoles. In insects it inhibits abundant postsynaptic receptors, the nicotinic acetylcholine receptors, thereby affecting control over muscle movement. Nicotine can, in addition, act as a deterrent. Using transgenic Nicotiana attenuata plants with blocked nicotine biosynthesis, it was shown that nicotine in floral nectar reduces nectar robbery and at the same time modulates the behaviour of pollinators in a way that increases reproductive success. Hummingbirds feeding on nicotine-containing nectar are discouraged from staying too long at any one flower by the unpleasant taste of nicotine. This results in more cross-pollination of tobacco plants and thereby promotes genetic diversity (Kessler et al. 2008).

Phenolic compounds (or phenolics) comprise a huge and chemically diverse group of metabolites sharing an aromatic ring with at least one hydroxyl group. They are found in all plant groups and have a wide range of functions—for example, as major constituents of plant cell walls (lignin, suberin), as flower colours (anthocyanins) or as ultraviolet protectants (flavones, flavonols). Many phenolics act as toxins or antifeedants. Tannins are synthesised by the majority of plant species. In the leaves of many woody dicots they accumulate to concentrations of up to 10% of dry weight (Barbehenn and Constabel 2011). Two major types are distinguished: hydrolysable tannins are polymers of simple phenolics and sugars and are localised mainly in cell walls; condensed tannins are formed by the polymerisation of flavonoids and are stored in vacuoles. Tannins can protect plants because they act as strong feeding deterrents for invertebrates and vertebrates. Beavers, for instance, prefer low-tannin poplars to high-tannin poplars. The toxic effects of tannins are not understood in molecular detail and are thought to be due either to non-specific interaction with proteins or to pro-oxidant activities that result in oxidative stress in the gut of herbivorous insects. Overall, tannins can serve as an example of the lack of direct evidence as to the exact activity and effects of compounds widely implicated in chemical defence. In the absence of clearly defined genotypes that differ only in the molecules in question, it is difficult to infer the actual consequences of these molecules for the interaction of a plant with its herbivores. Such ecological research greatly benefits from genetic modification, as demonstrated by the nicotine example described above (Kessler et al. 2008).

Nonetheless, studies on tannins in poplar were among the first to illustrate concepts of community or ecosystem genetics (Whitham et al. 2006). For poplars differing in a quantitative trait locus (QTL) controlling total condensed tannin concentrations, it was found in common garden and field experiments that a variation in one trait (tannin concentration) has far-reaching consequences for the communities living around such a foundation species. For instance, the composition of arthropod communities and soil microbial communities is associated with the variation in condensed tannins controlled by this one QTL. Thus, one can refer to “community phenotypes” of individual tree genotypes. These phenotypes show considerable heritability (Whitham et al. 2008) (Fig. 8.22).

Terpenoids represent another major class—with over 30,000 known structures in fact the largest class—of plant secondary metabolites. Terpenoids are derived from five carbon isoprene units. Besides the terpenoids that are essential players in plant biology as phytohormones (abscisic acid, gibberellins, brassinosteroids, cytokinins) or pigments (carotenoids), thousands of terpenoids are involved in plant–animal interactions, albeit with less defined functions for individual molecules. Terpenoids contribute to both direct and indirect defences. Many smaller terpenoids (monoterpenes, sesquiterpenes) such as linalool and α-farnesene are main components of plant volatiles. Plant volatiles are often released upon herbivore attack, differ in their composition from the blend emitted before herbivory and can therefore attract parasitoids and predators of plant pests. Since the first reports on the attraction of parasitic wasps to maize plants emitting terpenoids in 1990 (Turlings et al. 1990) (Fig. 8.23), many more corresponding examples have been documented. Even the transfer of the ability to synthesise volatile terpenoids from one plant to another by transformation with a terpene synthase gene is sufficient to guide the choice of herbivore parasites (Fig. 8.24).

Direct defence is mediated, for instance, by the feeding deterrence of terpenoids in essential oils or resins, as well as the terpenoid volatiles emitted by some plants (Unsicker et al. 2009).

Steroids are triterpenes and thus are synthesised from isoprene units too. Among them are not only saponins (Fig. 8.6) but also cardenolides, the so-called cardiac glucosides (Sect. 8.2.1). They inhibit Na⁺-/K⁺-ATPase and thereby interfere with maintenance of the membrane potential in neurons. Related molecules with a likewise known mode of toxicity are phytoecdysons. They mimic insect hormones such as ecdysone and thereby interfere with developmental processes in insects.

Hydrogen cyanide (HCN) is a classic example of a compound that can be toxic to the producing plants themselves and therefore has to be stored in a conjugated, inactive form. HCN affects mitochondrial respiration by inhibiting the binding of oxygen to cytochrome c oxidase and is thus a non-specific toxin. Many plant species (cassava, almonds, apricots and more than 2500 others) store cyanogenic glucosides in their vacuoles. When tissue is damaged by herbivore feeding, the glucosides come into contact with cytosolic glucosidases, the sugar moiety is cleaved off and cyanohydrin is released, which spontaneously or enzymatically is converted to HCN (Fig. 8.25).

A second important class of conjugated defence compounds is the glucosinolates—sulphur-containing specialised metabolites typical of the order Capparales. Because Brassicaceae belong to this order and the model species A. thaliana belongs to the Brassicaceae, the mechanisms, evolution and diversity of glucosinolate biosynthesis are molecularly better understood than those of most other defence compounds (Box 8.3). Glucosinolates are sometimes referred to as the “mustard oil bomb”. They are synthesised from various amino acids (e.g. methionine, alanine, phenylalanine, tyrosine and tryptophane) and stored in a glycosylated form in vacuoles. Upon tissue damage, glucosinolates come in contact with myrosinases—thioglucosidases that remove the sugar. Myrosinases are present only in a few scattered cells (Fig. 8.25). The aglycones arising from the sugar cleavage are unstable and spontaneously rearrange into a variety of bioactive compounds, including isothiocyanates and nitriles. Isothiocyanates, in particular, are highly reactive and toxic to insects, mammals and many other herbivores. Volatile isothiocyanates are, in addition, known to attract organisms of the third trophic level (i.e. parasitoids of herbivores) and thus contribute to indirect defence (Hopkins et al. 2009). On the other hand, some herbivores specialising in Brassicaceae have evolved the ability to use isothiocyanates as cues for finding their host plant. These multiple roles of glucosinolates are major drivers of genetic diversity in the biosynthesis machinery (Box 8.3).

Besides the amino acids used for protein biosynthesis, plants produce numerous non-proteinogenic amino acids as important primary metabolites (e.g. S-adenosylmethionine) or as secondary metabolites. Most of the latter are specific to particular plant species. Fabaceae tend to produce more non-common amino acids than other plant families. Among the functions of these metabolites is herbivore defence. The best-documented example is canavanine, an important nitrogen storage compound in the seeds of jack beans and other legume species. Canavanine is an arginine analogue, can be incorporated into proteins in place of arginine and is therefore highly toxic to organisms ranging from bacteria to man (Huang et al. 2011).

Constitutively synthesised defence compounds are found in all plant organs. Concentrations can vary strongly depending on the developmental stage, the environmental conditions (see inducible defences in Sect. 8.3.2) and also the individual genotype (e.g. the tannin example discussed above). A substance particularly rich in defence compounds is latex, an emulsion exuded by about 10% of all plant species when mechanical damage occurs (Agrawal and Konno 2009). Probably the most famous example of a defence compound in latex is morphine in Papaver species. In fact, many of the compound classes discussed above can be present at higher concentrations in latex than in leaves. This is why plant latex is sometimes compared with animal venoms, a complex mixture of low molecular weight and proteinaceous defence molecules, which exert a multitude of effects on a range of species. To date, few of the effects have been resolved molecularly. Mulberry leaves are toxic to insects other than the silk-producing Bombyx mori. Part of the toxicity can be attributed to alkaloids in the latex that resemble sugar molecules and thereby inhibit glucosidases. They are present in high concentrations (>1.5%; about 100-fold higher than in leaves) and are toxic to caterpillars (Konno et al. 2006).

An alternative to chemical defence has evolved in a limited number of plant species: hyperaccumulation of potentially toxic elements in leaves (= elemental defence; Sect. 8.2.1). Selenium hyperaccumulation is the best-documented example; hyperaccumulation of Cd, Zn and Ni may serve similar functions (Cappa and Pilon-Smits 2013).

8.3.2 Inducible Defences Against Herbivores

The previous subchapter provided a glimpse of the huge variety of molecules synthesised by plants that influence the interaction with herbivores, as toxins, deterrents or attractants. While plants produce a complex cocktail of secondary metabolites constitutively, there is, in addition, the induced chemical defence, which comprises not only metabolites but also defence proteins and changes in the storage sites of carbohydrates to minimise the consequences of tissue loss. The advantage of induced defences is that the expenditure of resources is restricted to periods when defence is actually needed, thereby reducing the costs of defence, in favour of growth and reproduction. On the other hand, they have to be activated quickly in order to be effective.

Plant tissues respond locally to herbivore feeding—that is, the cells surrounding the damaged site reprogramme their metabolism and up-regulate the synthesis of defence molecules. In addition, such up-regulation of herbivore defence occurs in tissues distant from the site of attack. As in the case of pathogen attack, long-distance signalling passes the information on herbivore presence to sites not exposed to feeding. The incoming signal (Fig. 8.26) is then converted into changes in gene activities that turn the systemic tissue into a more hostile environment for herbivores. The systemic response to wounding was originally discovered in tomato (Green and Ryan 1972) and later found in a wide range of species.

Induced herbivore defences—and, in particular, the systemic responses—add another layer of complexity to interactions within the communities of arthropods, etc., around a plant. Inducible defences connect organisms through space and time. Activities of root-feeding herbivores, for example, influence the chemical composition of above-ground tissues and, with that, the behaviour of organisms responsive to plant volatiles, etc. Similarly, herbivory early in the growing season has long-lasting systemic effects on a plant’s biochemistry and, again, significantly impacts the interacting organisms later in the year.

Plant volatiles emitted by damaged plants can have yet another community consequence. They can result in the so-called priming of neighbouring plants. Perception of airborne signals indicating a herbivore attack on a plant leads to an up-regulation of defence in as yet uninfested plants or to more rapid and robust defence activation when those plants are attacked by herbivores (Fig. 8.27). This has been shown for lima beans (Phaseolus lunatus) exposed to spider mites (Tetranychus urticae). The bean plants released terpenoids, which triggered the activation of defence genes in plants exposed to the volatiles but not to the herbivores (Arimura et al. 2000). Similarly, the treatment of maize plants with volatiles typical of damaged leaves resulted in a much stronger defence response when these plants came in contact with caterpillars (Engelberth et al. 2004).

Most of the defence molecules described above are synthesised throughout the life cycle of a plant. Upon herbivore attack, production of some can be strongly up-regulated, so they are part of both the constitutive and the inducible defences. As a consequence, the attacked plant becomes more toxic and unpalatable. A further step towards this state is the synthesis of proteins that can affect the fitness and growth of herbivores. What is well documented is the production of protease inhibitors that interfere with an insect’s digestive system. A range of different proteins complements the plant defence arsenal. Oxidative enzymes such as lipoxygenases and polyphenol oxidases have the potential to cause damage to insect tissues. Lectins and other carbohydrate-binding proteins can inhibit sugar-containing molecules in the insect gut. Cysteine proteases damage the peritrophic membrane that protects the gut epithelium. Arginine and threonine deaminases degrade essential amino acids. A feature that many of the plant defence proteins have in common is that they are comparatively resistant to gut proteases. This enables their biochemical identification via selective enrichment during gut passage (Howe and Jander 2008).

Indirect defences can be induced as well (Fig. 8.27). Changes in the blend of volatiles released by a plant upon herbivore feeding are either due to the processing of stored conjugated specialised metabolites (e.g. glucosinolates; Fig. 8.25) or attributable to newly synthesised volatile compounds. A different type of indirect defence relies on rewarding the enemies of herbivores. Secretion of extrafloral nectar attracts non-herbivorous ants and other predators of herbivores (Heil 2008) (Fig. 8.27). The exact chemical composition of such nectars is not known. In an analogy to floral nectaries the presence of sugars such as glucose and fructose, as well as secretion mediated by sugar transporters, can be assumed.

8.3.2.1 Recognition of Herbivore Attack

Rapid induction of direct and indirect herbivore defences is of course dependent on reliable cues perceived by the plant. Insect feeding causes mechanical damage, and plants indeed respond to this (wounding response). However, detailed comparative studies—for instance, of transcriptome changes—have clearly shown that a plant differentiates between the tissue loss due to mechanical wounding and the tissue loss caused by a feeding insect. The pattern of gene induction and repression is clearly distinct for the two conditions (Howe and Jander 2008). Thus, there are apparently signals associated with the insect’s activity that provide information for the plant. Sources of such signals can be insect saliva, regurgitant (another form of oral secretions produced by lepidopteran larvae), oviposition fluids or faeces. In analogy to the term MAMPs used in the context of pathogen defence, these signals are now referred to as HAMPs (herbivory associated molecular patterns) or insect elicitors. In addition, the more general term DAMPs is used as well, summarising signals arising from the plant because of enemy activities (e.g. mechanical damage). In fact, the concept of DAMP recognition unifies induced plant immunity to pathogens and herbivores.

The first HAMP identified was volicitin, a fatty acid–amino acid conjugate isolated from oral secretions of beet army worm caterpillars (Spodoptera exigua) (Alborn et al. 1997). The conjugate is derived from a plant fatty acid, while the glutamine and the conjugation reaction stem from the insect. Treatment of artificially wounded maize plants with volicitin activates the production of volatiles typical of the herbivore response, while the wounding alone does not lead to production of the same volatile blend. Since the discovery of volicitin, several other HAMPs have been molecularly identified: a range of other fatty acid–amino acid conjugates, breakdown products of plant adenosine triphosphate (ATP) synthases called inceptins and sulphated fatty acids called caeliferins (Acevedo et al. 2015).

HAMP receptors are not known yet. Generally, recognition of HAMPs by plant cells is molecularly less well understood than MAMP recognition. The same is true of the early signal transduction events triggered by HAMP sensing, even though the existing knowledge shows extensive principal similarity to pathogen responses. The same basic modules are implicated: transient cytosolic Ca²⁺ spikes, membrane potential changes, ROS as signalling molecules, and mitogen-activated protein kinase cascades (Wu and Baldwin 2010).

As in the plant–pathogen interaction, there is evidence that herbivores produce effectors to overcome plant defences, and that plants have evolved the ability to sense such effectors or their activities. Again, the molecular understanding is far less advanced, at least partly because plant–herbivore interactions are less amenable to the type of genetic analysis that led to the gene-for-gene concept. An example of herbivore effectors are ATPases in insect saliva. Mechanical damage of plant cells causes a release of ATP into the extracellular space. This extracellular ATP can serve as a DAMP for the plant, and insect ATPases degrade this potential signal. Thus, a plant receptor for extracellular ATP presumably functions in DAMP recognition (Choi et al. 2014).

In contrast to the early events of herbivore sensing, the main regulators of plant herbivore responses have been known for a long time. Jasmonic acid (JA) and its derivatives are recognised as plant hormones that play a central role in balancing a plant’s investments in growth versus defence in rapidly changing environments. JA generally inhibits growth, as well as assimilate export from vegetative tissues, and activates defensive and reproductive processes. Most prominent among the defensive roles of JA is protection from attack by insect herbivores and necrotrophic fungi. JA promotes the synthesis of practically all classes of chemical defence compounds, underpinning its universal role in the plant kingdom. For several plant species it has been shown that a loss of JA synthesis or JA perception severely compromises the ability to limit damage caused by all kinds of insect herbivores and even some vertebrates (Fig. 8.28). Field studies on transgenic N. attenuata plants with a defect in JA synthesis showed far-reaching consequences of JA signalling in the natural environment (Kessler et al. 2004). Not only were specialist herbivores more successful when the capacity of the plants to induce defences was compromised by the silencing of a lipoxygenase gene essential for JA biosynthesis, but also plants were chosen for oviposition and severely damaged by opportunistic herbivores (e.g. Empoasca sp. leafhoppers) that rarely attack N. attenuata plants in the wild. Also, JA-deficient plants were preferred to wild-type plants in choice experiments. Thus, plants fall victim to herbivores that they normally are able to fend off. Later field experiments revealed that leafhoppers probe plants during initial feeding, and those they select for extended feeding are the ones with the lowest level of JA-dependent activation of defence responses (Kallenbach et al. 2012).

The central position of JA in defence can be described in distinct steps. HAMP or DAMP recognition triggers the early signal transduction chains that activate the synthesis of JA and its derivatives, among them the actual receptor-active JA–isoleucine (JA-Ile) conjugate (Fig. 8.29). JA synthesis proceeds from the ubiquitous fatty acid linolenic acid via enzymatic reactions localised in plastids and peroxisomes (the core JA module). Perception of JA then activates the herbivore defence genes encoding enzymes and regulatory proteins involved in the synthesis of defence metabolites, or the antifeeding defence proteins such as protease inhibitors.

Upon pathogen or herbivore detection, the active signal JA-Ile, which is barely present in unstressed cells, can be synthesised rapidly (i.e. within minutes) in the cytosol from JA and isoleucine, and then sensed by a receptor in the nucleus (Fig. 8.30). Like several other phytohormone receptors, the JA-Ile receptor COI1 (for “coronatine insensitive”; the name refers to the initial identification of a receptor mutant insensitive to the jasmonate analogue coronatine) is part of an E3 ligase complex that tags specific proteins for degradation in the 26S proteasome by adding ubiquitin chains. These proteins are the JAZ repressors. They prevent the activation of JA-Ile responsive genes by interaction with transcription factors (e.g. MYC2) that bind the promoters of early JA-Ile response genes. Degradation of the JAZ proteins relieves the repression and the JA response is activated (Fig. 8.30).

Herbivore defence responses are analogously switched on in tissues distant from the site of herbivore-inflicted damage. JA or JA conjugates are phloem-mobile signals, which may travel to as yet unaffected tissues. In addition, volatile methyljasmonate serves as an airborne long-distance signal (Fig. 8.26). The JA-dependent activation appears to be universal in the plant kingdom. Besides this central mechanism, there are other known signals and signalling modes. The systemic wound response was discovered and intensively investigated in tomato plants. An early breakthrough was the identification of systemin, a peptide signal derived from a pre-protein through proteolytic processing (Pearce et al. 1991). Systemin is produced in wounded tomato leaves and activates the JA pathway in neighbouring vascular tissues. It can thus be considered as a DAMP released by wounding. Orthologous peptides with comparable activity have never been found, suggesting that systemin and systemin-like peptides are restricted to a small group of plant species. Nonetheless, defence-activating peptides of a different molecular nature have in the meantime been identified, and they appear to be widespread in the plant kingdom (Bartels and Boller 2015).

A systemic wound response dependent on JA as the mobile signal takes a few hours to materialise. An alternative mode of long-distance signalling could be electrical. Indeed, the up-regulation of JA synthesis in distal tissues within minutes of wounding appears to depend on electrical signalling (Mousavi et al. 2013). Both insect feeding and wounding elicit surface potential changes in A. thaliana leaves, which travel through the rosette to distal leaves at an estimated speed of around 6 cm/min. The amplitude of the signal correlates with the strength of the JA response activation in these leaves. Conversely, the injection of current into leaves systemically triggers JA synthesis in the absence of wounding. Ion channels involved in the generation and propagation of the electrical signal are glutamate-receptor-like proteins.

8.3.3 How Plant–Herbivore Interactions Drive Genetic Diversity

The interaction of plants with the plethora of herbivorous animals surrounding them is generally regarded as a major driver of biodiversity. This has in a few cases been possible to demonstrate experimentally. In field experiments with a native plant, Oenothera biennis, it was shown that the suppression of insect herbivores in one half of the plants by insecticide treatment not only suppressed herbivore-inflicted damage but also resulted in rapid evolutionary change, apparent as genetic differentiation between the populations. Over five growing seasons, herbivore defence declined in the protected populations, for instance, because of lower elagitannin levels in fruits (Agrawal et al. 2012). Conversely, competitive ability increased. Such experiments document evolutionary consequences of a trade-off between defence and growth in real time.

Furthermore, plant–herbivore interactions are seen as a major reason for the rich diversity of plant secondary metabolism. As discussed, plant families or plant species synthesise a characteristic cocktail of secondary metabolites. Importantly, also, the spectrum of secondary metabolites produced by an individual plant is strongly influenced by its genotype—that is, the combination of allelic variants of genes encoding regulatory proteins, enzymes or other proteins controlling the spectrum and synthesis rate of secondary metabolites. The molecular dissection of chemical herbivore defence mechanisms provided the first detailed insights into how synthesis and diversity of secondary metabolites arose and which evolutionary forces maintain such diversity. A comparatively well-understood system is glucosinolate biosynthesis in Brassicaceae. Accessions of A. thaliana differ in the mixture of constitutively produced glucosinolates and of glucosinolate breakdown products released when glucosinolates come into contact with myrosinases (Fig. 8.25). The diversity arose through gene duplications and the neo-functionalisation of gene copies. The blend of breakdown products has various consequences for plant–herbivore interactions. The more volatile breakdown products are more likely to attract specialist herbivores or the parasitoids and predators of herbivores. Nitriles are less toxic than isothiocyanates but, at the same time, do not carry the risk of alerting specialists to the presence of a suitable host plant, because specialists use isothiocyanates as clues. Such contrasting consequences of defence molecules maintain the diversity of genotypes, since different habitats select for different genotypes. For example, variation in the polymorphic GS-ELONG locus controlling the side chain length of methionine-derived glucosinolates in A. thaliana (Box 8.3) is explained by the distribution of two specialised aphid species feeding on A. thaliana (Züst et al. 2012). Multigeneration selection experiments with artificial populations demonstrated that the herbivore pressure exerted by the two aphid species selects distinct chemotypes—that is, it causes rapid changes in the relative abundance of the alleles controlling these chemotypes (Fig. 8.31).

Box 8.3: The GS-ELONG Locus in Arabidopsis thaliana

The diversity of glucosinolates and glucosinolate breakdown products in A. thaliana is in part explained by allelic variation in four genetic loci. One of them is the GS-ELONG (or MAM) locus. The dominant glucosinolates in A. thaliana are methionine derived. They differ in the length of their side chain. This influences the volatility of the breakdown products and thus the interaction with insects of the second and third trophic levels. Most A. thaliana individuals carry either a MAM gene (MAM2) that results in the synthesis of a three-carbon side chain (C3) or a MAM gene (MAM1) that gives rise to a four-carbon side chain (C4). The gene enabling C4 synthesis arose by gene duplication and underwent neo-functionalisation. Sequence analysis revealed an over-representation of non-synonymous changes affecting the part of the MAM protein that controls substrate specificity. This suggested that changes in substrate specificity have been positively selected (Benderoth et al. 2009), meaning that a combination of gene duplication, neo-functionalisation and positive selection explains the generation of metabolic diversity (Fig. 8.32).

The logical next questions are: Which processes maintain the diversity? Why did the positively selected MAM1 gene not completely replace the ancestral MAM2 gene? Most likely, it was because under particular conditions it is advantageous for a plant to carry the MAM2 gene instead of the MAM1 gene. For example, MAM2 provides an advantage against certain generalist herbivores because MAM2 activity is associated with larger quantities of glucosinolates. Thus, balancing selection maintains both variants. The composition of local herbivore communities varies considerably over time. The relative frequency of generalists, specialists, parasitoids, etc., temporally changes within a given habitat. With these changes, different glucosinolate cocktails are beneficial and therefore are selected by the herbivore pressure.