Hezi Tenenboim

TRPLSC 1427 No. of Pages 11 Review Omic Relief for the Biotically Stressed: Metabolomics of Plant Biotic Interactions Hezi Tenenboim1 and Yariv Brotman2,* Many aspects of the way plants protect themselves against pathogen attack, or react upon such an attack, are realized by metabolites. The ambitious aim of metabolomics, namely the identification and annotation of the entire cellular metabolome, still poses a considerable challenge due to the high diversity of the metabolites in the cell. Recent advances in analytical methods and data analysis have resulted in improved sensitivity, accuracy, and capacity, allowing the analysis of several hundreds or even thousands of compounds within one sample. Investigators have only recently begun to acknowledge and harness the power of metabolomics to elucidate key questions in the study of plant biotic interactions; we review trends and developments in the field. Metabolomic Approaches to the Study of Plant Biotic Interactions Any article or review discussing plant biotic stress inadvertently includes a stern reminder of the grave challenges humanity is facing in light of a growing population, namely the sustainable production of sufficient food plants and the need to protect these plants against pathogens and herbivores. A debate over the three main strategies to help plants to defend against pathogens– breeding, the use of chemicals, and genetic modification–is increasingly heating up. To make educated decisions on the best strategy it is vital to understand the basis and details of plant biotic-stress responses. Despite an impressive body of work encompassing the topic of plant biotic interactions, many gaps remain in our understanding of the mechanisms underlying plant stress responses as well as of the beneficial interactions that promote plant survival and growth. Metabolites play a crucial role in plant biotic interactions, and may therefore be considered as a phenotypic signature of a living organism, reflecting genetic variance, epigenetic modifications, and transcriptomic and proteomic components. The identification and annotation of the entire cellular metabolome, together with a mechanistic understanding of the roles metabolites play in biotic interactions, hold promise for addressing some of the open questions, but still pose a significant challenge given the high structural and molecular diversity of the metabolites. Three other factors hampering metabolomic discovery are (i) the wide range of physiological metabolite concentrations, (ii) the complex and rapid nature of metabolic regulation, with swift enzymatic reactions that alter the picture within seconds, with such reactions interconnecting with each other, creating elaborate networks, and (iii) the high level of interfering chemical impurities, which are much less relevant in the analysis of nucleic acids or proteins. None of the current metabolite extraction methods or analytical platforms is able to deliver a comprehensive snapshot of the metabolome. One reason is that the huge diversity of metabolites produced by plants govern diverse phenomena in plant biotic interactions: from herbivores preferring specific plants while ignoring others [1] to defense molecules that ward off specific herbivores while attracting others [2], to name but two. The partial picture of the metabolome that we have today leaves many questions unanswered. Trends in Plant Science, Month Year, Vol. xx, No. yy Trends The field of metabolomics is advancing rapidly: machines and platforms are becoming more accurate and sensitive, and public metabolite databases enable improved annotation. Practices previously considered novel or excessive are gradually becoming the norm: these include the use of multiple analytical approaches in parallel in a single study, the metabolomic analysis of single cells, high-resolution time-course metabolomic analyses, metabolomic profiling of samples from nature, and increasingly seamless integration of different types of omics data. Metabolomics has become a tool that is now routinely used to address specific topics relating to plant biotic interactions, including systemic acquired resistance, induced resistance, multiple stresses, allelopathy, and more. Metabolomics is becoming a prominent tool for elucidating the underlying mechanisms. 1 Max-Planck-Institut für Molekulare Pflanzenphysiologie, Potsdam, Germany 2 Department of Life Sciences, Ben Gurion University of the Negev, Beersheva, Israel *Correspondence:-(Y. Brotman). http://dx.doi.org/10.1016/j.tplants- © 2016 Elsevier Ltd. All rights reserved. 1 TRPLSC 1427 No. of Pages 11 Box 1. Metabolomics in the Study of Induced Resistance Glossary Metabolomics has helped to unravel many aspects of induced resistance, or priming. A plant is said to be primed for pathogen or herbivore attack when previous contact with a particular elicitor has made the plant more resistant to future attacks. The elicitor is often a pathogen, in which case the priming is termed systemic acquired resistance (SAR), but it can also be a symbiotic fungus or PGPR–the priming is then dubbed induced systemic resistance (ISR)[4_TD$IF]–or a chemical. Priming often operates at a distance: infection of one leaf or of the roots induces resistance in a remote, unscathed leaf. In addition, it can operate at even longer distances: plant-produced VOCs can prime other, remote plants [12]. For priming, phytohormones and other secondary metabolites carry a great part of the workload. Metabolomic profiling was used to identify a long-elusive SAR-conferring signal molecule, azelaic acid [47]. Likewise, but using beneficial, growth-promoting, and ISR-inducing bacteria, potential volatile ISR metabolites were identified by capturing them from the air around the plant and subjecting them to metabolomic profiling [80]. Of note, some genetic mutants show constitutive priming. Both genetic priming (natural mutants lacking a specific nitrogen transporter that are thus constitutively primed against pathogens) and artificially induced priming [using the non-protein amino acid b-aminobutyric acid (BABA) as elicitor] were investigated in one study using metabolomic profiling [81]. Effector: a compound transferred from a pathogen into a host cell, where it causes modifications that are usually beneficial for the pathogen. Metabolic fingerprinting: metabolic analysis that results in a peak/feature pattern, without annotation of unique metabolites. Metabolomic profiling: qualitative and quantitative metabolic analysis that results in a comprehensive list of metabolites in a sample. Natural variation: the variance in traits seen among different natural strains, cultivars, accessions, or ecotypes of a species. Numerous excellent reviews on plant metabolomics have been published, including reviews with a focus on plant biotic interactions [3,4]. In this review we explore, in addition to the traditional applications of metabolomic research, topics that are unique to plant biotic interactions but have rarely been touched on so far (Figure 1). The traditional applications include technology and machinery [5], analytical challenges and limitations [6], experimental design [3], and annotation and databases [7]. Beyond these aspects we discuss how metabolomics can contribute to the study of induced resistance (Box 1), allelopathy, and the time-course of pathogenic infections. We also explore what special considerations are required when inspecting combinations of different biotic and abiotic stresses (Box 2). Finally, we consider how metabolomic and other omic data can be integrated with each other (Box 3) in the context of plant biotic interactions. How Primary and Secondary Metabolism Change with Plant Biotic Interactions The lion's share of contemporary metabolomic discovery in this field is dedicated to the elucidation of the changes (and often the status quo) that plants undergo when they interact with other organisms. Metabolites are traditionally divided into primary (essential for the viability of the cell) and secondary (required for the viability of the organism in the environment). When considering Box 2. Metabolomics of Stress Combinations Plants in their natural environment often face a combination of stresses such as extreme temperature, drought, and salinity, alongside pathogen infection with one or more pathogens. The response of the plant to combined stress is unique and cannot be directly extrapolated from the response to each stress individually [82]. Metabolomics, so far only scantily exploited for this purpose, has a promising potential in this niche. The number of studies utilizing metabolomics to characterize combinations of abiotic stresses is small, but only a handful have addressed abiotic combined with biotic stress. In an agriculturally[5_TD$IF] relevant study, metabolomic profiling was used on water-stressed and nematode-infected tomato (S. lycopersicum) plants to assay for their effects on nutrients [83]. In another study the effect of drought on the feeding behavior of insect herbivores and on parasites that infect those insects was investigated [84]. Metabolomic profiling was also used to study the combined effect of temperature and aphid density on the plant VOC profile [85]. Instead of applying actual stress, the phytohormones abscisic acid and salicylic acid as representatives of abiotic and biotic stress, respectively, were applied to the plants, followed by metabolomic profiling [86]. It was found that the combined artificial stress induced a different metabolic response from that induced by each stress applied separately. Herbivorous moths were shown to prefer to feed on transgenic plants protected against light stress than on plants lacking photoprotection [87]. The mechanisms for this were partly elucidated by profiling the metabolic changes in the plants. The issue of multiple stresses needs to be taken into account when designing a metabolomics experiment. Even a seemingly innocuous growth chamber can harbor unconsidered stresses: interactions with soil microbes, for instance, even beneficial ones, may completely skew the results of a metabolomics experiment. Similarly, the plastic hood put over the tray and used to protect the plants immediately after spray inoculation with bacteria may cause heat stress. These factors require proper controls. 2 Trends in Plant Science, Month Year, Vol. xx, No. yy TRPLSC 1427 No. of Pages 11 Induced resistance [12,47,80,81] Pathogens Pathogens Allelopathy [1,8,9,66–69] Mulple stresses [82–87] Pathogens O ? OH Beneficial microbes Beneficial interacons [13,36,53,80,97–102] OH ? Metabolite: of plant or pathogen origin? [72,74,75,78,79] Omics integraon [23–30,57,72,88–91, 93,94,97,101] Natural variaon [1,23–41,98,103,104] Figure 1. A Selection of Topics Covered in This Review. From the top, clockwise: in induced resistance (priming) signals pass from a challenged part of the plant to a remote, unchallenged part, inducing defense responses in the latter. Plants often need to cope with multiple stresses, and the plant response is not necessarily additive. Detected metabolites in profiling experiments of infected plants may derive from the plant or the pathogen; several approaches are available to solve the ambiguity. Various types of omics data, in particular genomic, transcriptomic, and proteomic, can be integrated with metabolomic results to obtain a comprehensive picture of defense mechanisms. Natural variation, wherein different accessions of a plant species produce different amounts of metabolites and respond differently to biotic stress, can be used as a valuable tool for elucidating defense mechanisms. Using metabolomics, many mechanisms of beneficial biotic interactions have been elucidated. The same applies to allelopathy–remote, chemical communication between plants and neighboring plants. In brackets: referenced articles. plant biotic interactions, changes in secondary metabolism are at first glance obvious mediators: the plant produces compounds that deter pathogens and herbivores [8], attract the enemies of herbivores [9], form mechanical barriers against pathogen invasion [10], directly kill invading pathogens [11], and signal organs, remote plant parts, and neighboring plants about an infection [12], attract beneficial symbionts [13] (Box 4), and can serve several of these roles concomitantly [14] (note, all of the cited studies [6–12] employed metabolomics). However, the significance of secondary metabolites is not solely limited to their exploitation by plants in the functions described above. On the one hand, pathogens [15] and herbivores [16] themselves occasionally make and use secondary compounds to subvert plant immunity. Coronatine, a bacterial jasmonic acid (JA) analog, provides a well-studied example [15]: it intervenes in the delicate balance between the two main defense pathways in the plant, namely the JA and the salicylic acid (SA) pathways, by ‘pretending’ to be JA, thereby suppressing the SA pathway that is harmful to this bacterium [15]. On the other hand, the plant secondary metabolome often falls Trends in Plant Science, Month Year, Vol. xx, No. yy 3 TRPLSC 1427 No. of Pages 11 Box 3. Integration of Metabolomics with Other Omics Data To fully harness the power of metabolomics and to obtain as comprehensive an overview as possible of cellular processes in a physiological context, it is essential that metabolomics is combined and integrated with other omics approaches (that said, one advantageous feature of metabolomics is the ability to perform it completely standalone, independently of the availability of other omics data for a particular organism; most notably, non-model organisms whose genomes have not yet been sequenced can be utilized for metabolomic studies [34]). Omics integration is not a new concept, but the number of articles employing this in the context of plant biotic interactions is scant. Different methodologies, such as pathway assignment and transcript protein metabolite network analysis, are used to integrate omics data [57,72,88,89]. For example, two wheat lines of highly similar genetic background, but differing in a resistance allele for the pathogenic fungus Fusarium graminearum, were recently used for metabolomic and proteomic profiling, demonstrating that the gene controls resistance mainly by adjusting cell-wall thickness [90]. In a study crucially pertinent to human nutrition, the biosynthetic pathway of steroidal alkaloids, toxic compounds found in potato and tomato, was deciphered by harnessing modern genomic, transcriptomic and networking tools, and supported by metabolomics [91]. As hinted in the main text, the investigation of natural variation is aided by the application of multiple omics techniques: the integration of metabolomic and genomic data from various accessions enables the identification of mQTLs and novel metabolic pathways [23–30]. This methodology can be further expanded by adding other layers of data, most notably transcriptomic (expression), proteomic, and phenotypic, resulting in eQTLs [92], pQTLs [93], and pQTLs [94], respectively. Network analysis utilizes advanced algorithms to illustrate connections and interdependencies. Such networks enable the relation between different types of omics data to be quantitatively displayed. Primary metabolism in two maize (Z. mays) accessions, and in a recombinant inbred population derived from crossing the two, [6_TD$IF]was recently dissected [94]. The authors constructed a correlation network of agronomic traits (plant performance) and metabolites to define the connection between the former and the latter. Online tools, such as KEGG [95] and PlantCyc (plantcyc.org), allow the visualization and layering of experimental transcriptomic and metabolomic data upon metabolic and signaling pathway maps. While not an ‘integration tool’ per se, this provides a convenient overview of affected cellular processes in an experiment. Such tools, aside from being valuable in a plethora of studies and applications, can specifically be used to complement experimentally obtained metabolomic data [88,96]. prey to modification by the pathogen, and this is frequently mediated by effector (see Glossary) proteins that intervene in plant metabolite biosynthetic pathways [17]. Notwithstanding, changes in primary metabolism are not to be trifled with: in addition to reallocation of energy and resources from primary to secondary metabolism such that the plant can cope with the onslaught as best as possible–a reallocation that results in the much-studied trade-off in terms of biomass [18]–a myriad of infection-related primary metabolic activities take place: carbon and nitrogen compounds are removed from the infection site to starve the pathogen [19], and they induce the expression of defense-related genes [20]; the cell wall is thickened with callose to hinder further infection [20], or cells around the infection site are sacrificed and filled with callose to limit the infection; and nitric oxide and reactive oxygen species (ROS) are produced, directly and indirectly harming the pathogen and mediating programmed cell death and the hypersensitive response [21]. In addition, the composition of primary metabolites present in the plant, and the pathogens that normally infect it, have co-evolved, and the pathogens are able, to some extent, to adapt to fast changes in primary metabolite concentrations in the plant, for example during fruit ripening [22]. Natural Variation as a Tool for Metabolomic Discovery Natural variation research investigates the often significant differences between natural strains of one species. There are dramatic differences, for example in the abundance of flavonoids– secondary compounds involved in, among others, biotic interactions–between different Arabidopsis thaliana accessions [23]. Natural variation in metabolite abundance and in other biotic interaction-related traits has been used as a prominent tool for identifying key defense determinants [23–33]. 4 Trends in Plant Science, Month Year, Vol. xx, No. yy TRPLSC 1427 No. of Pages 11 Box 4. Metabolomics of Beneficial Biotic Interactions Several systems of beneficial biotic interactions in plants have been well studied, most notably mycorrhyza, PGPR, and several types of filamentous fungi, such as (some strains of) Trichoderma. Many conceptual aspects–nutrient absorption, transfer of effectors, recognition by receptors–are shared between deleterious infection and beneficial interactions. Several interactions are three-way: a symbiont (1st partner) associating with a plant (2nd partner) primes the plant against pathogens (3rd partner; Box 1). Metabolomics has helped to elucidate the connection and the divide between infection and symbiosis. In an elegant work that also serves as an excellent example of integration of different types of omics data (Box 3), interactions between Arabidopsis, PGPR, and various pathogens were explored [97]. Using metabolomic profiling and microarray, the authors showed that defense and signaling molecules in the plant, namely salicylic acid, glucosinolates, and camalexin, mediate the PGPR-induced priming of the plant against the various pathogens. The association between plant and symbiont can be very specific, as shown using metabolomics: two strains of PGPR isolated from two specific cultivars of rice (O. sativa) were shown to promote rice growth preferentially in the cultivar from which they were isolated, and this was reflected at the metabolic level [98]. The same principle was shown even more comprehensively for arbuscular mycorrhiza (AM): five different plant species were colonized with a single common AM– metabolomic profiling revealed, except for a conserved ‘core metabolome’, highly-divergent and species-specific metabolic changes [99]. In a more straightforward approach, induced metabolites in AM-colonized plants were screened for; the discovered metabolites were then applied to the AM fungus in vitro, and some of them were found to induce colonization-related genes in the fungus [100]. The nature of the proverbial three-way interaction was further characterized in another study showing that the metabolic fingerprint of an infected plant and that of the same plant associated with beneficial Trichoderma fungi do not simply add up when the plant is challenged with both, exacerbating the complex nature of the relationship [101]. In maize (Z. mays), volatiles emitted from root bacteria were profiled and shown to influence maize resistance to pathogens and herbivores, as well as to attract natural enemies of the herbivores in a tritrophic interaction [13]. In an example of a different type of symbiotic interaction involving plants, the metabolomic footprint of epiphytic, commensal microorganisms residing on aboveground plant parts was characterized using an environmental metabolomic approach [102]. The complex mechanisms by which plants ‘choose’ the microbes with which they associate are gradually being uncovered [103,104]; this provides fertile ground for the application of metabolomics. Metabolites, in addition to merely being measured, can be mapped to so-called metabolic quantitative trait loci (mQTL). In other words, novel genes controlling the biosynthesis or regulation of defense metabolites can be localized in the genome and identified by taking advantage of natural variation. The recent surge in the use of advanced genomic tools has contributed immensely to this field [24]. There are several examples of this approach in the study of plant biotic interactions: metabolic profiling of A. thaliana glucosinolates, prominent secondary plant defense compounds synthesized mostly by Brassicaceae, has been used for mQTL identification [25,26], leading to the identification of many novel defense-related genes and providing support for the function of previously studied genes. A decade earlier the same group had performed a similar type of analysis–but with a twist–measuring glucosinolate and myrosinase (the enzyme that cleaves and activates glucosinolates) levels in two accessions and correlating them to resistance to herbivory [27]. The first mQTL studies in the context of plant biotic interactions were performed mostly in model, genome-sequenced species. As more and more genomes become sequenced, and with advances in marker and transcriptomic technology, mQTL studies in crop species are becoming more common. For example, glucosinolate-related genes have been identified in cabbage (Brassica oleracea) [28], phenolamide-related genes in rice (Oryza sativa) [29] and in maize (Zea mais) [30], and phenylpropanoid-related genes in apple (Malus domestica) [31]. All these compounds are known defense metabolites. Indeed, mQTL studies, with the variety of gene types that they unravel (biosynthetic as well as regulatory), be they in model organisms or crops, can help to resolve the elaborate life cycle of defense metabolites. As described above, in addition to metabolite-related genes, any quantitative or qualitative trait can be mapped, including the pathogen resistance [[7_TD$IF]32,33]. This is another way to gain information about the role of metabolites in defense, as was done using the resistance levels Trends in Plant Science, Month Year, Vol. xx, No. yy 5 TRPLSC 1427 No. of Pages 11 of maize (Z. mays) inbred lines to southern and northern leaf blight, respectively [32] [31], leading to the identification of numerous defense-related genes, some of which are metabolite biosynthetic genes. In a report demonstrating metabolite discovery from a different angle, natural variation between 37 accessions of a wild, non-model tobacco (Nicotiana attenuata) species was exploited, in combination with a molecular network methodology, to identify novel defense metabolites and to link them to genes and pathways [32]. Natural variation is not solely a plant matter: variation in pathogens also [8_TD$IF]modulates their capacity for infection. For example, different isolates of the pathogenic fungus Fusarium graminearum produce different metabolic responses in resistant and susceptible barley genotypes [35]. The same principle was demonstrated in plant growth-promoting rhizobacteria (PGPR), emphasizing the relevance of natural variation to beneficial interactions as well [36] (Box 4). Last but not least, natural variation is a major evolutionary phenomenon with implications for related topics including fitness, diversity, and speciation. Metabolomic research in the double context of natural variation and plant biotic interactions has helped to address key evolutionary questions. The issue of why defense metabolites are so numerous and diverse has been shown to be linked, at least in part, to plant co-evolution with pathogens and herbivores, and to[9_TD$IF] the concept of an ‘arms race’ [37,38]. Elaborate, large-scale field experiments, conducted in different environments, with subsequent metabolomic profiling of ellagitannins and glucosinolates, two classes of defense secondary metabolites, have added greatly to our understanding of the evolution of the diversity observed in these metabolites; of how evolutionary pressure by, in these cases, insect herbivores, as well as environmental conditions, can influence plant fitness; and of the relative contributions of genetic makeup versus herbivore challenge and the environment to, among other, fitness and metabolite diversity [39–41]. Developments and Trends Several new trends have developed since the inception of the field of metabolomics. We list these trends in the context of plant biotic interaction research. Studies of the plant metabolome during biotic stress have shifted over the years from frill-free listings of metabolites (such as flavonoid profiling following application of a biotic elicitor in white lupin (Lupinus albus) [42] and metabolomic fingerprinting of periwinkle (Catharanthus roseus) infected with phytoplasma [43]) to studies that put metabolites in a mechanistic context: the elucidation, for instance, of antifungal defense mechanisms in maize (Z. mays) using metabolomic profiling of secondary metabolites [44]; the metabolomic analysis of tobacco (Nicotiana tabacum) after application of ergosterol, a fungal compound that elicits plant defense responses [45]; the application of the defense elicitor chitin followed by metabolomic profiling, which revealed that phytoalexin induction is mediated by a MAPK signaling pathway [46]; and the identification of azelaic acid [47], pipecolic acid [48,49], and glycerol-3-phosphate [50] as prominent plant defense molecules. While the latter metabolite had earlier been implicated in defense [50], the former two had had no known defense roles, and were detected as highly infection-induced compounds in metabolomic profiling experiments [47–49]. Another trend in metabolomics is a switch from single- to multi-omics: metabolomic profiling on its own seems to suffice less and less, and fully fledged systems biology is gradually taking over (Box 3). Metabolomic studies, initially limited to single techniques, now increasingly utilize several platforms in parallel. The two most prominent analytical techniques in metabolomics, namely mass spectrometry (MS) and nuclear magnetic resonance (NMR), complement each other with their 6 Trends in Plant Science, Month Year, Vol. xx, No. yy TRPLSC 1427 No. of Pages 11 respective advantages and caveats. NMR and two different MS variants were used, for example, to document metabolic changes in Pseudomonas-infected A. thaliana [51]. The three methods confirmed each other, but each also facilitated the detection of several unique features. Micrometabolomics, the metabolomic analysis of isolated tissues and single cells, has been gaining popularity. Collecting sufficient material from specific, at times ill-defined, tissues for metabolomic analysis is not trivial, and methods to this end are being constantly updated. Laser microdissection was used to show that terpenoids accumulate in specific tissues of spruce (Picea glauca) trees in response to methyl jasmonate application, a widely used defense response-inducing treatment [52]. More recently, and using a beneficial biotic interaction (Box 4), mycorrhizal root tissue of barrel medic (Medicago truncatula) plants was laser-dissected and subjected to MS analysis, leading to novel insights into how root cells metabolically adapt– mainly via amino acids and sugars–to build and benefit from the mutualistic interaction with the fungus [53]. Both examples used microdissection to excise and collect cells of a specific type; the descriptor ‘single-cell metabolomics’ is somewhat misleading. However, there are techniques that truly focus on single cells [54]. While such studies in plants have blossomed over the past decade, bona fide single-cell metabolomics has not yet been used in the study of plant biotic interactions. The course of infection, disease, and immune response is highly dynamic, and high-resolution time-course experiments therefore deliver indispensable information. Notably, transcriptomic studies generally have much higher timepoint resolution than metabolomic studies (e.g., 24 timepoints in [55]). Metabolomic analyses of infection, once limited to a single snapshot of the system, now often incorporate a time-course methodology [51,56,57]. In addition to temporal metabolomic studies, spatial studies are also becoming popular. MS imaging (MSI) techniques, most prominently matrix-assisted laser desorption–ionization (MALDI), have flourished since the beginning of the millennium and are increasingly utilized, although the number of plant studies remains modest. Less than a handful of studies have used this method in the context of plant disease and immunity: one such study demonstrated the immense value of spatial metabolomics by showing why herbivorous larvae avoid, when feeding, particular areas of the A. thaliana leaf–namely the areas that accumulate glucosinolates [58]. Glucosinolates were systematically catalogued and quantified on the leaf surface, and the specific glucosinolate composition provided valuable insights into how some herbivorous insects, having developed mechanisms to neutralize glucosinolate toxicity, use glucosinolates as attractants and stimulants for oviposition [2]. A different study used MALDI-MSI on the glucosinolate-resistant insects described above to demonstrate how glucosinolates pass rapidly through the larval gut to escape activation and toxicity [59]. Glucosinolates were also analyzed in flowers and siliques, showing that layers of glucosinolate-laden cells protect these delicate organs [60]. In infected tomato (Solanum lycopersicum) leaf areas, the pathogenic fungus was shown in situ to be able to degrade the otherwise fungitoxic alkaloid /-tomatine [61]. In a recent study, infection sites of a citrus (Citrus sinensis grafted onto C. limonia) disease, known as citrus variegated chlorosis, spatially correlated with the accumulation of the defense-related flavanone hesperidin [62]. The chemical interfaces of soybean (Glycine max)–aphid (Aphis glycines) and rice (O. sativa)–bacteria (Xanthomonas oryzae) interactions were studied by means of localized metabolomic profiling of infected plant leaves [63]. MSI analyses are generally performed directly on intact plant material, thereby reflecting the immediate and unbiased[3_TD$IF] physiology and [10_TD$IF]metabolite distribution. Direct analysis can also be implemented without the spatial aspect: in a report from 2010, tobacco (N. tabacum) leaves were inoculated with a fungus, allowed to incubate for different times, then excised and directly profiled [64]. The results gave valuable insights on the time-course of infection. Ecometabolomics, similar to the parallel field known as metagenomics (or ecogenomics), focuses on the analysis of samples from nature/the environment, expanding the scope offered Trends in Plant Science, Month Year, Vol. xx, No. yy 7 TRPLSC 1427 No. of Pages 11 by growth-controlled model organisms in laboratories and aiming to decipher the metabolomics of ecological communities and systems. A fungal disease that afflicts grapevine (Vitis vinifera) was studied using metabolomic profiling [65], as was a compound present in the latex excreted from dandelion (Taraxacum officinale) roots that deters herbivorous beetles [66]. The latter example was the first demonstrated instance of a belowground defense metabolite that not only merely deters herbivores but also, by doing so, positively contributes to the fitness of the plant in its natural setting. Bioremediation is also a prominent niche for ecometabolomic studies. Many soil microorganisms used for toxic compound degradation associate with plant roots, and it was shown, with the help of metabolomic profiling, that plant root exudates–specifically their flavonoid components–induce these bacteria to degrade the toxic compounds [67]. The two latter reports [66,67] touch on an adjacent research topic that is also on the rise, namely allelopathy–long-distance chemical communication between organisms. It has been known for decades that plants release compounds to the environment, especially volatiles into the air and the soil, for the purpose of herbivore deterrence [8], warning neighboring plants about an attack [8], and attracting the herbivore's enemies [9], but metabolomics has only recently been recruited to the study of allelopathy [68]. In one of the cited reports [8] it was shown that a volatile metabolite emitted by tomato (S. lycopersicum) plants infested with the herbivorous common cutworm (Spodoptera litura) reaches neighboring tomato plants, where it converts to a defense-active compound that deters the herbivore. Interestingly, the received signal does not induce a reaction cascade in the receiving plant, but is instead simply and directly metabolized into the active herbivore-deterring compound [8]. In another example of allelopathy, metabolomic profiling was used to show that plants change their defense strategy based on the identity of their plant neighbors [69]. Metabolomics was used to characterize the types of information that plants advertise by emitting volatile organic compounds (VOCs) [1]: these were profiled in 52 intact and wounded species of oak (Quercus) to show that, although unique VOC compositions can be used by herbivores to identify their favorite plant species, plants choose at times to [1_TD$IF]advertize and at other times to hide their identity, for reasons that remain to be clarified [1]. Because allelopathic signal chemicals in plants pass through the air, soil, or water, and affect the ecology of groups or communities of plants, they are best studied on-site rather than in the lab– hence the relation to ecometabolomics. VOCs, in addition to their relevance for allelopathy, have recently been developed as biomarkers. This honor, however, is not exclusively reserved for volatile compounds. Metabolites of all types– at times single metabolites, at times metabolic fingerprints–may serve as biomarkers for several aspects of plant pathogen interactions: for the existence and identification of disease [70], for its progression and severity [71], for predicted resistance levels in different accessions or cultivars (usually in the context of a breeding endeavor), and as a measure of the efficiency of pest management. To summarize, the field of metabolomics has made great technological and conceptual strides over the past decade, also in the narrower context of plant biotic interactions. Technology has allowed more elaborate studies that use multiple platforms, exploring the temporal, spatial, cellular, and ecological dimensions, and integrating data from adjacent omics fields. Metabolite: Of the Plant or the Pathogen? Countless metabolomics articles have generated data from pathogen-infected plants, but little thought is often given to the question: which detected metabolite is derived from the plant and which from the pathogen? Although not a pressing issue in transcriptomic or proteomic studies, where sequence analysis can facilitate source attribution, metabolites bear no such identity card. In a minority of the studies the cultured pathogen is profiled separately, allowing conclusions to be drawn for some metabolites [72]. [12_TD$IF]Regarding the typical origin of a given metabolite can be 8 Trends in Plant Science, Month Year, Vol. xx, No. yy TRPLSC 1427 No. of Pages 11 informative (e.g., ergosterol is exclusively fungal [73]), but can also introduce unwanted bias. The application of a virus (which does not produce any metabolites) or of a purified pathogen-derived elicitor (such as chitin, flagellin, or an effector protein), a plant-derived immunity inducer (such as salicylic acid [74] or methyl jasmonate [75]), or even an artificial elicitor such as Flg22 [76] in place of bona fide infection can eliminate the problem, and also permits focused dissection of the function of the elicitor. In the same vein, exogenous challenge may be avoided completely by using mutants in which the plant immune response is constitutively activated. Both approaches constitute, however, a poor imitation of a full-featured infection, and the metabolome is invariably affected by the difference. Two additional potential approaches for distinguishing plant- from pathogen-derived metabolites have been suggested. Isotope labeling of the entire metabolome has been successfully performed in many organisms, including plants [77]. In the context of pathogen infection, isotope labeling has already been performed for pathogens of mammals [78], and may also prove successful for plant pathogens. This technique needs to be used with caution in the case of pathogens feeding on plant material because pathogen metabolism gradually eliminates the distinction between labeled and non-labeled nutrient sources. This method can therefore be primarily used for the identification of secondary metabolites shortly after infection. The second approach includes spectroscopic methods, such as MALDI MS imaging, in which the spatial movement of metabolites can be observed [61]. For example, the issue of distinguishing the source of the metabolite was addressed directly by co-incubating plant cell culture and pathogen, then separating them post-infection and analyzing (fingerprinting rather than profiling) the metabolome of each [79]. Concluding Remarks and Future Perspectives The several hundred metabolites identified so far merely represent the tip of the iceberg. While some groups of secondary metabolites have been extensively researched, others remain neglected. Although the lack of annotation of plant metabolites is a major bottleneck, great strides have been taken in recent years, including standardization of the way in which metabolomic studies are reported. The technology is constantly advancing but, as many investigators in the field point out, technology is of little use if contributors do not collaborate sufficiently to standardize their results to make data more transferable (see also Outstanding Questions). An improvement in this regard will not only benefit active investigators in the field but will also improve the accessibility of metabolomics and facilitate widespread application. Outstanding Questions How can the annotation of metabolites be improved and standardized? How can the different analytical techniques complement each other better, and how can data generated by different methods be reconciled? How can a better integration be achieved between metabolomics and the gradually increasing body of genomic, transcriptomic, and proteomic data? How can conclusions from metabolomic experiments be applied to other species, especially crops? How can it be used for crop improvement? How are different pathogen lifestyles manifested in the metabolic fingerprint? How can this be applied to exposure to multiple stresses? How can we better utilize natural variation in combination with metabolomics to understand the ways in which different plant strains, accessions, or cultivars respond to pathogens? Acknowledgments This research was supported by the I-CORE Program of the Planning and Budgeting Committee and The Israel Science Foundation (grant 757/12). References 1. Pearse, I.S. et al. (2013) Headspace volatiles from 52 oak species advertise induction, species identity, and evolution, but not defense. J. Chem. Ecol. 39, 90–100 7. Tohge, T. and Fernie, A.R. (2009) Web-based resources for mass-spectrometry-based metabolomics: a user's guide. Phytochemistry 70, 450–456 2. Shroff, R. et al. (2015) Quantification of plant surface metabolites by matrix-assisted laser desorption–ionization mass spectrometry imaging: glucosinolates on Arabidopsis thaliana leaves. Plant J. 81, 961–972 8. Sugimoto, K. et al. (2014) Intake and transformation to a glycoside of (Z)-3-hexenol from infested neighbors reveals a mode of plant odor reception and defense. Proc. Natl. Acad. Sci. 111,- 3. Kushalappa, A.C. and Gunnaiah, R. (2013) Metabolo-proteomics to discover plant biotic stress resistance genes. Trends Plant Sci. 18, 522–531 9. Cusumano, A. et al. (2015) Attraction of egg-killing parasitoids toward induced plant volatiles in a multi-herbivore context. Oecologia 179, 163–174 4. Feussner, I. and Polle, A. (2015) What the transcriptome does not tell–proteomics and metabolomics are closer to the plants’ patho-phenotype. Curr. Opin. Plant Biol. 26, 26–31 10. Bourdenx, B. et al. (2011) Overexpression of Arabidopsis ECERIFERUM1 promotes wax very-long-chain alkane biosynthesis and influences plant response to biotic and abiotic stresses. Plant Physiol. 156, 29–45 5. Tugizimana, F. et al. (2013) Plant metabolomics: a new frontier in phytochemical analysis. S. Afr. J. Sci. 109, 1–11 6. Hegeman, A.D. (2010) Plant metabolomics–meeting the analytical challenges of comprehensive metabolite analysis. Brief. Funct. Genomics 9, 139–148 11. Prince, D.C. et al. (2014) The leucine-rich repeat receptor-like kinase BRASSINOSTEROID INSENSITIVE1-ASSOCIATED KINASE1 and the cytochrome P450 PHYTOALEXIN DEFICIENT3 contribute to innate immunity to aphids in Arabidopsis. Plant Physiol. 164,- Trends in Plant Science, Month Year, Vol. xx, No. yy 9 TRPLSC 1427 No. of Pages 11 12. Erb, M. et al. (2015) Indole is an essential herbivore-induced volatile priming signal in maize. Nat. Commun. 6, 6273 13. D’Alessandro, M. et al. (2014) Volatiles produced by soil-borne endophytic bacteria increase plant pathogen resistance and affect tritrophic interactions. Plant. Cell Environ. 37, 813–826 35. Khan, S.A. et al. (2012) Genetic analysis of metabolites in apple fruits indicates an mQTL hotspot for phenolic compounds on linkage group 16. J. Exp. Bot. 63,- 14. Clay, N.K. et al. (2009) Glucosinolate metabolites required for an Arabidopsis innate immune response. Science 323, 95–101 36. Walker, V. et al. (2011) Host plant secondary metabolite profiling shows a complex, strain-dependent response of maize to plant growth-promoting rhizobacteria of the genus Azospirillum. New Phytol. 189, 494–506 15. Geng, X. et al. (2012) The coronatine toxin of Pseudomonas syringae is a multifunctional suppressor of Arabidopsis defense. Plant Cell 24,- 37. Gilbert, G.S. and Webb, C.O. (2007) Phylogenetic signal in plant pathogen-host range. Proc. Natl. Acad. Sci. U. S. A. 104,- 16. Helms, A.M. et al. (2013) Exposure of Solidago altissima plants to volatile emissions of an insect antagonist (Eurosta solidaginis) deters subsequent herbivory. Proc. Natl. Acad. Sci. 110, 199–204 38. Speed, M.P. et al. (2015) Coevolution can explain defensive secondary metabolite diversity in plants. New Phytol. 208,- 17. Djamei, A. et al. (2011) Metabolic priming by a secreted fungal effector. Nature 478,-. Kempel, A. et al. (2011) Tradeoffs associated with constitutive and induced plant resistance against herbivory. Proc. Natl. Acad. Sci. U. S. A. 108,-. Chen, H.-Y. et al. (2015) The Arabidopsis vacuolar sugar transporter SWEET2 limits carbon sequestration from roots and restricts Pythium infection. Plant J. 83,-. Sun, L. et al. (2014) Sugar homeostasis mediated by cell wall invertase GRAIN INCOMPLETE FILLING 1 (GIF1) plays a role in pre-existing and induced defence in rice. Mol. Plant Pathol. 15,-. Scheler, C. et al. (2013) Nitric oxide and reactive oxygen species in plant biotic interactions. Curr. Opin. Plant Biol. 16, 534–539 39. Agrawal, A.A. et al. (2012) Insect herbivores drive real-time ecological and evolutionary change in plant populations. Science 338,-. Prasad, K.V.S.K. et al. (2012) A gain-of-function polymorphism controlling complex traits and fitness in nature. Science 337,-. Kerwin, R. et al. (2015) Natural genetic variation in Arabidopsis thaliana defense metabolism genes modulates field fitness. Elife 2015, e-. Bednarek, P. et al. (2001) Profiling changes in metabolism of isoflavonoids and their conjugates in Lupinus albus treated with biotic elicitor. Phytochemistry 56, 77–85 43. Choi, Y.H. et al. (2004) Metabolic discrimination of Catharanthus roseus leaves infected by phytoplasma using 1H-NMR spectroscopy and multivariate data analysis. Plant Physiol. 135,-. Balmer, D. et al. (2013) Induced resistance in maize is based on organ-specific defence responses. Plant J. 74, 213–225 22. Bi, F. et al. (2015) Carbon regulation of environmental pH by secreted small molecules that modulate pathogenicity in phytopathogenic fungi. Mol. Plant Pathol http://dx.doi.org/10.1111/ mpp.12355 45. Tugizimana, F. et al. (2014) Multi-platform metabolomic analyses of ergosterol-induced dynamic changes in Nicotiana tabacum cells. PLoS ONE 9, e87846 23. Routaboul, J.M. et al. (2012) Metabolite profiling and quantitative genetics of natural variation for flavonoids in Arabidopsis. J. Exp. Bot. 63,- 46. Kishi-Kaboshi, M. et al. (2010) A rice fungal MAMP-responsive MAPK cascade regulates metabolic flow to antimicrobial metabolite synthesis. Plant J. 63, 599–612 24. Scossa, F. et al. (2015) Genomics-based strategies for the use of natural variation in the improvement of crop metabolism. Plant Sci. 242, 47–64 47. Jung, H.W. et al. (2009) Priming in systemic plant immunity. Science 324, 89–91 25. 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, e- 48. Návarová, H. et al. (2012) Pipecolic acid, an endogenous mediator of defense amplification and priming, is a critical regulator of inducible plant immunity. Plant Cell 24,-. Bernsdorff, F. et al. (2016) Pipecolic acid orchestrates plant systemic acquired resistance and defense priming via salicylic aciddependent and -independent pathways. Plant Cell 28, 102–129 26. Wentzell, A.M. et al. (2007) Linking metabolic QTLs with network and cis-eQTLs controlling biosynthetic pathways. PLoS Genet. 3,- 50. Chanda, B. et al. (2011) Glycerol-3-phosphate is a critical mobile inducer of systemic immunity in plants. Nat. Genet. 43, 421–427 27. Kliebenstein, D. et al. (2002) Comparative analysis of quantitative trait loci controlling glucosinolates, myrosinase and insect resistance in Arabidopsis thaliana. Genetics 161, 325–332 51. Ward, J.L. et al. (2010) The metabolic transition during disease following infection of Arabidopsis thaliana by Pseudomonas syringae pv. tomato. Plant J. 63, 443–457 28. Sotelo, T. et al. (2014) Identification of metabolic QTLs and candidate genes for glucosinolate synthesis in Brassica oleracea leaves, seeds and flower buds. PLoS ONE 9, e91428 52. Abbott, E. et al. (2010) Laser microdissection of conifer stem tissues: isolation and analysis of high quality RNA, terpene synthase enzyme activity and terpenoid metabolites from resin ducts and cambial zone tissue of white spruce (Picea glauca). BMC Plant Biol. 10, 106 29. Dong, X. et al. (2015) Spatiotemporal distribution of phenolamides and the genetics of natural variation of hydroxycinnamoyl spermidine in rice. Mol. Plant 8,-. Wen, W. et al. (2014) Metabolome-based genome-wide association study of maize kernel leads to novel biochemical insights. Nat. Commun. 5, 3438 31. Poland, J.A. et al. (2011) Genome-wide nested association mapping of quantitative resistance to northern leaf blight in maize. Proc. Natl. Acad. Sci. U. S. A. 108,-. Li, D. et al. (2015) Navigating natural variation in herbivoryinduced secondary metabolism in coyote tobacco populations using MS/MS structural analysis. Proc. Natl. Acad. Sci. U. S. A. 112,-. Kumaraswamy, G.K. et al. (2012) Differential metabolic response of barley genotypes, varying in resistance, to trichothecene-producing and -nonproducing (tri5 ) isolates of Fusarium graminearum. Plant Pathol. 61,-. Kump, K.L. et al. (2011) Genome-wide association study of quantitative resistance to southern leaf blight in the maize nested association mapping population. Nat. Genet. 43, 163–168 10 Trends in Plant Science, Month Year, Vol. xx, No. yy 53. Gaude, N. et al. (2015) Symbiosis dependent accumulation of primary metabolites in arbuscule-containing cells. BMC Plant Biol. 15, 234 54. Oikawa, A. and Saito, K. (2012) Metabolite analyses of single cells. Plant J. 70, 30–38 55. Windram, O. et al. (2012) Arabidopsis defense against Botrytis cinerea: chronology and regulation deciphered by high-resolution temporal transcriptomic analysis. Plant Cell 24,-. Parker, D. et al. (2009) Metabolomic analysis reveals a common pattern of metabolic re-programming during invasion of three host plant species by Magnaporthe grisea. Plant J. 59,-. Tzin, V. et al. (2015) Dynamic maize responses to aphid feeding are revealed by a time series of transcriptomic and metabolomic assays. Plant Physiol. 169,-. Shroff, R. et al. (2008) Nonuniform distribution of glucosinolates in Arabidopsis thaliana leaves has important consequences for plant defense. Proc. Natl. Acad. Sci. U. S. A. 105,- TRPLSC 1427 No. of Pages 11 59. Abdalsamee, M.K. et al. (2014) Rapid incorporation of glucosinolates as a strategy used by a herbivore to prevent activation by myrosinases. Insect Biochem. Mol. Biol. 52, 115–123 81. Gamir, J. et al. (2014) Targeting novel chemical and constitutive primed metabolites against Plectosphaerella cucumerina. Plant J. 78, 227–240 60. Sarsby, J. et al. (2012) Mass spectrometry imaging of glucosinolates in Arabidopsis flowers and siliques. Phytochemistry 77, 110–118 82. Suzuki, N. et al. (2014) Abiotic and biotic stress combinations. New Phytol. 203, 32–43 61. Etalo, D. et al. (2015) Spatially-resolved plant metabolomics: some potentials and limitations of laser-ablation electrospray ionization (LAESI) mass spectrometry metabolite imaging. Plant Physiol. 169,- 83. Atkinson, N.J. et al. (2011) Influence of combined biotic and abiotic stress on nutritional quality parameters in tomato (Solanum lycopersicum). J. Agric. Food Chem. 59,- 62. Soares, M.S. et al. (2015) Quantification and localization of hesperidin and rutin in Citrus sinensis grafted on C. limonia after Xylella fastidiosa infection by HPLC-UV and MALDI imaging mass spectrometry. Phytochemistry 115,-. Klein, A.T. et al. (2015) Investigation of the chemical interface in the soybean–aphid and rice–bacteria interactions using maldimass spectrometry imaging. Anal. Chem. 87,-. Ibáñez, A.J. et al. (2010) Rapid metabolic profiling of Nicotiana tabacum defence responses against Phytophthora nicotianae using direct infrared laser desorption ionization mass spectrometry and principal component analysis. Plant Methods 6, 14 65. Lima, M.R.M. et al. (2010) NMR metabolomics of esca disease-affected Vitis vinifera cv. Alvarinho leaves. J. Exp. Bot. 61,-. Huber, M. et al. (2016) A latex metabolite benefits plant fitness under root herbivore attack. PLOS Biol. 14, e-. Toussaint, J.P. et al. (2012) Plant exudates promote PCB degradation by a rhodococcal rhizobacteria. Appl. Microbiol. Biotechnol. 95,-. Scognamiglio, M. et al. (2015) Metabolomics: an unexplored tool for allelopathy studies. J. Allelochem. Interact. 2015, 9–23 69. Broz, A.K. et al. (2010) Plant neighbor identity influences plant biochemistry and physiology related to defense. BMC Plant Biol. 10, 115 70. Aksenov, A.A. et al. (2014) Detection of huanglongbing disease using differential mobility spectrometry. Anal. Chem. 86,-. Laothawornkitkul, J. et al. (2010) Volatile organic compounds as a diagnostic marker of late blight infected potato plants: A pilot study. Crop Prot. 29,-. Rudd, J.J. et al. (2015) Transcriptome and metabolite profiling of the infection cycle of Zymoseptoria tritici on wheat reveals a biphasic interaction with plant immunity involving differential pathogen chromosomal contributions and a variation on the hemibiotrophic lifestyle def. Plant Physiol. 167,-. Mille-Lindblom, C. et al. (2004) Ergosterol as a measure of living fungal biomass: persistence in environmental samples after fungal death. J. Microbiol. Methods 59,-. Mustafa, N.R. et al. (2009) Metabolic changes of salicylic acidelicited Catharanthus roseus cell suspension cultures monitored by NMR-based metabolomics. Biotechnol. Lett. 31,-. Farag, M.A. et al. (2008) Metabolomics reveals novel pathways and differential mechanistic and elicitor-specific responses in phenylpropanoid and isoflavonoid biosynthesis in Medicago truncatula cell cultures. Plant Physiol. 146,-. Sun, Y. et al. (2013) Structural basis for flg22-induced activation of the Arabidopsis FLS2–BAK1 immune complex. Science 342, 624–628 84. Weldegergis, B.T. et al. (2015) Drought stress affects plant metabolites and herbivore preference but not host location by its parasitoids. Oecologia 177,-. Truong, D.H. et al. (2014) Temperature regimes and aphid density interactions differentially influence VOC emissions in Arabidopsis. Arthropod. Plant. Interact. 8,-. Okamoto, M. et al. (2009) Metabolic movement upon abscisic acid and salicylic acid combined treatments. Plant Biotechnol. 26,-. Johansson Jänkänpää, H. et al. (2013) Non-photochemical quenching capacity in Arabidopsis thaliana affects herbivore behaviour. PLoS ONE 8, e-. Sade, D. et al. (2014) Comparative metabolomics and transcriptomics of plant response to tomato yellow leaf curl virus infection in resistant and susceptible tomato cultivars. Metabolomics 11, 81–97 89. Kumar, Y. et al. (2016) Fusarium oxysporum mediates systems metabolic reprogramming of chickpea roots as revealed by a combination of proteomics and metabolomics. Plant Biotechnol. J. http://dx.doi.org/10.1111/pbi-. Gunnaiah, R. et al. (2012) Integrated metabolo-proteomic approach to decipher the mechanisms by which wheat QTL (Fhb1) contributes to resistance against Fusarium graminearum. PLoS ONE 7, e-. Itkin, M. et al. (2013) Biosynthesis of antinutritional alkaloids in solanaceous crops is mediated by clustered genes. Science 341,-. Keurentjes, J.J.B. et al. (2007) Regulatory network construction in Arabidopsis by using genome-wide gene expression quantitative trait loci. Proc. Natl. Acad. Sci. U. S. A. 104,-. Fu, J. et al. (2009) System-wide molecular evidence for phenotypic buffering in Arabidopsis. Nat. Genet. 41,-. Wen, W. et al. (2015) Genetic determinants of the network of primary metabolism and their relationships to plant performance in a maize recombinant inbred line population. Plant Cell 27,-. Kanehisa, M. and Goto, S. (2000) KEGG: Kyoto Encyclopaedia of Genes and Genomes. Nucl. Acids Res. 28, 27–30 96. Aliferis, K.A. and Jabaji, S. (2012) FT-ICR/MS and GC-EI/MS metabolomics networking unravels global potato sprout's responses to Rhizoctonia solani infection. PLoS ONE 7, 1–13 97. van de Mortel, J.E. et al. (2012) Metabolic and transcriptomic changes induced in Arabidopsis by the rhizobacterium Pseudomonas fluorescens SS101. Plant Physiol. 160,-. Chamam, A. et al. (2013) Plant secondary metabolite profiling evidences strain-dependent effect in the Azospirillum–Oryza sativa association. Phytochemistry 87, 65–77 99. Schweiger, R. et al. (2014) High specificity in plant leaf metabolic responses to arbuscular mycorrhiza. Nat. Commun. 5, 3886 77. Giavalisco, P. et al. (2011) Elemental formula annotation of polar and lipophilic metabolites using 13C, 15N and 34S isotope labelling, in combination with high-resolution mass spectrometry. Plant J. 68, 364–376 100. Laparre, J. et al. (2014) Combining metabolomics and gene expression analysis reveals that propionyl- and butyryl-carnitines are involved in late stages of arbuscular mycorrhizal symbiosis. Mol. Plant 7, 554–566 78. Creek, D.J. et al. (2015) Probing the metabolic network in bloodstream-form Trypanosoma brucei using untargeted metabolomics with stable isotope labelled glucose. PLoS Pathog. 11, e- 101. Brotman, Y. et al. (2012) Transcript and metabolite analysis of the Trichoderma-induced systemic resistance response to Pseudomonas syringae in Arabidopsis thaliana. Microbiology 158,- 79. Allwood, J.W. et al. (2010) Dual metabolomics: a novel approach to understanding plant–pathogen interactions. Phytochemistry 71, 590–597 102. Ryffel, F. et al. (2015) Metabolic footprint of epiphytic bacteria on Arabidopsis thaliana leaves. ISME J. 10, 632–643 80. Farag, M.A. et al. (2006) GC-MS SPME profiling of rhizobacterial volatiles reveals prospective inducers of growth promotion and induced systemic resistance in plants. Phytochemistry 67,- 103. Lebeis, S.L. et al. (2015) Salicylic acid modulates colonization of the root microbiome by specific bacterial taxa. Science 349,-. Haney, C.H. et al. (2015) Associations with rhizosphere bacteria can confer an adaptive advantage to plants. Nat. Plants 1, 15051 Trends in Plant Science, Month Year, Vol. xx, No. yy 11