Black Mustard and the Butterfly Effect Metabolomics of Plant-Insect Interactions Under Multiple Stress Conditions

Home , Brassica nigra, Elina (butterfly), Samea multiplicalis

Black Mustard And The Butterfly Effect Metabolomics of Plant-Insect Interactions under Multiple Stress Conditions

Stefano Papazian

Black Mustard and The Butterfly Effect Metabolomics of Plant-Insect Interactions under Multiple Stress Conditions

Stefano Papazian

Umeå Plant Science Centre Fysiologisk Botanik Umeå 2017

This work is protected by the Swedish Copyright Legislation (Act 1960:729) ISBN: 978-91-7601-728-9 Front Cover by: Stefano Papazian Electronic version available at http://umu.diva-portal.org/ Printed by: KBC Service Centre, Umeå University Umeå, Sweden 2017

To my parents

“When we speak of Nature, it is easy to forget that we are ourselves a part of Nature. We have to view ourselves with the same curiosity and openness with which we study a tree, the sky or a thought, because we too are linked to the entire Universe.”

Henri Matisse

Henri Matisse, Dance (1909), Oil on canvas, 260 × 391 cm (The Hermitage Museum, St. Petersburg).

Preface

When we think about nature, we can ask ourselves: “What makes organisms adapted to their environment and able to interact with each other?”

Inside an ecosystem, organisms coexist and interact in communities, cycling nutrients and exchanging energy and matter across the food-web. Non-living components (water, soil, air, light, temperature, etc.) are also important parts of ecosystems, which constantly re-shape the surrounding environment. Throughout evolution, environmental pressure increases the success (fitness) of organisms by selecting random favourable mutations inside their DNA. Thus, extending the “dogma” of molecular biology, we can follow the flow of information bottom-up, until the level of ecosystems, with feed-back on DNA.

Thanks to novel techniques that allow for the high-throughput detection of thousands of molecules and the systemic screening of organisms, we can now think in comprehensive terms and study species and ecosystems while looking at several layers of biological organization (Keurentjes et al., 2011). “Metabolomics” studies all the chemical processes producing metabolites, which are the molecules considered to be the ultimate response of biological systems towards environmental perturbations (Fiehn 2002). In this thesis, I combine the disciplines of ecology and metabolomics in order to study the way plants and insects interact with each other, trying to answer:

1. Which metabolites make a plant able to interact with insects and defend against herbivore attacks?

2. How environmental factors such as air pollutants, multiple herbivores, or plant signalling hormones, can influence the plant response?

3. And finally, can we use the rising technology of metabolomics to predict how plants respond to multiple stress conditions?

List of Publications

I. Eliezer Khaling, Stefano Papazian, Erik H. Poelman, Jarmo K. Holopainen, Benedicte R. Albrectsen, James D. Blande (2015). Ozone affects growth and development of Pieris brassicae on the wild host plant Brassica nigra. Environmental Pollution 199: 119-129.

II. Stefano Papazian, Eliezer Khaling, Christelle Bonnet, Steve Lassueur, Philippe Reymond, Thomas Moritz, James D. Blande, Benedicte R. Albrectsen (2016). Central Metabolic Responses to Ozone and Herbivory Affect Photosynthesis and Stomatal Closure. Plant Physiology 172: 2057–2078.

III. Camille Ponzio*, Stefano Papazian*, Benedicte R. Albrectsen, Marcel Dicke, Rieta Gols (2017). Dual herbivore attack and herbivore density affect metabolic profiles of Brassica nigra leaves. Plant, Cell and Environment. [Epub ahead of print] * These authors contributed equally to the work

IV. Stefano Papazian, Tristan Girdwood, Mátyás Ripszam, Erik H. Poelman, Marcel Dicke, Thomas Moritz, Benedicte R. Albrectsen (2017). Herbivore-Induced Metabolic Responses in Brassica nigra are shaped by Leaf Ontogeny. (Manuscript)

Abstract

One main goal of ecological research is to understand nature´s complexity, in order to predict the potential impact of environmental perturbations. In this thesis, I investigate the ecological interactions between some of the most ancient organisms living on our planet: plants and insects.

Focus of my research is the interaction between the wild brassicaceous plant black mustard (Brassica nigra L.) and its specialist insect herbivore, the large white cabbage butterfly (Pieris brassicae L). Both organisms are well characterized model species used in chemical ecology research.

Using different analytical techniques, such as liquid and gas chromatography coupled to mass-spectrometry (LC- and GC-MS) and headspace collection of volatile organic compounds (VOCs), I apply the approach of metabolomics and systems biology to the field of ecology to explore the metabolic changes occurring inside the plants exposed to biotic and abiotic stresses.

Particularly, I study the plant metabolic responses against P. brassicae chewing caterpillars during sequential treatment exposure to: abiotic stress by the oxidative air pollutant ozone (O3); dual herbivory with specialist Brevicoryne brassicae piercing-sucking aphids; and chemical induction of plant defences with the oxylipin phytohormone methyl-jasmonate (MeJA).

Results show how during herbivore-induced responses, changes in defence- and growth-metabolic processes are tightly connected to stress protection mechanisms, indicating that plants actively reprogram their inner metabolic networks in order to adapt to consecutive changes in the environment.

This thesis illustrates how evaluating the plant metabolome in its entirety rather than single metabolites, can help us understanding plant responses towards abiotic and biotic stresses, and improve our ability to predict how constant shifts in the environment affect plant physiology and ecology.

Sammanfattning

Ett huvudsyfte för ekologisk forskning är att förstå naturens komplexitet för att kunna förutse effekter av störningar i miljön. I min avhandling har jag fokuserat på ekologiska interaktioner mellan växter och insekter, två av de äldsta terrestra organismgrupperna på jorden.

I mina studier har jag undersökt interaktioner mellan den korsblommiga växten svartsenap (Brassica nigra L.) och den specifika herbivoren kålfjäril (Pieris brassicae L.). Båda är väl karaktäriserade modellarter i kemisk- ekologisk forskning.

De metaboliska förändringar som sker när växten utsätts för biotisk och abiotisk stress har analyserats hjälp av metabolomik, det vill säga analyser av metabolomet i sin helhet med hjälp av tekniker som vätske- och gaskromatografi kopplad till masspektrometri (LC- och GC-MS), och så kallad headspace-uppsamling av flyktiga organiska föreningar (VOCs).

Jag har särskilt undersökt de metaboliska förändringar som sker när växten betas av kålfjärilslarver vid samtidig exponering för: abiotisk stress i form av ozon (O3), en oxidativ luftförorening; ytterligare betning i form av stickande och sugande bladlus (Brevicoryne brassicae); tillsats av oxylipinfytohormon metyl-jasmonat (MeJA), ett ämne som inducerar växtens försvar.

Resultaten visar att de metaboliska förändringar som sker i växten vid herbivori med konsekvenser för dess försvar och tillväxt är nära kopplade till de metaboliska förändringar som sker vid stress, vilket visar att växten kan fortlöpande och aktivt omprogrammera sina metaboliska nätverk för att anpassa sig till förändringar i miljön.

Avhandlingen visar att genom att utvärdera växtmetabolomet i sin helhet, snarare än att studera enskilda metaboliter, vi kan få bättre förståelse för hur växter reagerar på olika former av stress och därmed också bidra till att vi kan göra förutsägelser för hur förändringar i miljön kan påverka växters fysiologi och ekologi.

iii

Abbreviations

O2 Oxygen O3 Ozone H2O Water CO2 Carbon dioxide ROS Reactive oxygen species UV Ultra-violet ppb Parts per billion myr Million years

JA Jasmonic acid SA Salicylic acid ET Ethylene MeJA Methyl jasmonate VOCs Volatile organic compounds GLVs Green leaf volatiles ITCs Isothiocyanates GSLs Glucosinolates TCAs Tricarboxylic acids G3P Glycerol-3-phospate GABA Gamma (γ)-amino butyric acid NAD(P)+ Nicotinamide adenine dinucleotide (phosphate) NAD(P)H NAD (phosphate), reduced form ATP Adenosine triphosphate PSI-II Photosystem I and II ETC Electron transport chain

MVA Multivariate analysis HCA Hierarchical cluster analysis PCA Principal component analysis PLS Partial least square OPLS Orthogonal projection of latent structures DA Discriminant analysis VIP Variable importance in the projection LV Latent variable GO Gene ontology GLM General linear model ANOVA Analysis of variance MANOVA Multivariate analysis of variance

LC Liquid chromatography GC Gas chromatography TOF Time of flight MS Mass spectrometry TD Thermal desorption

Contents

Preface

List of Publications i Abstract ii Sammanfattning iii Abbreviations iv Contents v

1. Introduction 1 Plant Metabolism 2 Primary metabolism 2 Secondary metabolism 3 Eco-metabolomics 5 Plant-insect interactions 5 Plant defence responses 7 Study system 8 Black mustard 8 Large white cabbage butterfly 12 Multiple stresses 15 O3: a tropospheric pollutant 17 Aphids: sap-sucking herbivores 18 MeJA: herbivore mimicking 18 2. Objectives 21 3. Results and Discussion 22 O3 effects on plant-insect interactions 22 Effects of O3 on herbivory and plant defences 22 Omics responses to sequential O3 and herbivory 27 Physiological responses to O3 and herbivory 34 Dual herbivore attack 38 Plant responses versus aphids and caterpillars 38 Effects of dual herbivory and aphid density 41 Induced plant resistance to herbivory 45 Effects of MeJA on herbivore-induced responses 46 Ontogenic patterns of induced responses 52 4. Conclusions 55 The “Butterfly effect” 55

5. Acknowledgements 58 6. Literature 60

1. Introduction

The sum of all chemical reactions in an organism is defined as metabolism, from the Greek word μεταβολή [metabolē], meaning "change”. All living organisms harbour an active metabolism, which is the condition that defines life itself as a continuous state of transformation. Inside the cell, chemical changes from a molecular compound into another are regulated by enzymes and connected through long biochemical pathways, very much like subway maps, where each “station” represents a different metabolite (Fig.1).

As conserved features of life evolutionary history, metabolic pathways that are essential for cellular growth and development are shared among most organisms (Weng, 2014). The ensemble of these central pathways is usually referred to as “primary” metabolism, in contrast to “secondary” pathways which later evolved from it (Firn & Jones, 2009).

Figure 1. Metabolism. Cellular network of biochemical pathways (KEGG).

1.1 Plant Metabolism

Plants are magnificent organisms. They literally connect planet Earth to the sky, as “primary producers” of organic matter inside terrestrial ecosystems. Through a series of highly coordinated metabolic processes, plants capture sunlight and transform it into energy and biomass. All the energy that we liberate burning most common fuels (e.g. coal, oil, wood, or biofuels) was indeed originally captured from the Sun. Concealed inside the plant, and accessible to higher levels of the food-web, this energy is what sustains life and ecosystems on our planet.

Plants present an astonishing chemical biodiversity. In the plant kingdom there are estimated more than 250.000 structurally different secondary metabolites (Viant et al., 2017), which are a precious source of bioactive “natural products”. Since the ancient times of traditional medicine, humans have exploited the healing properties of plant chemistry, and plants still constitutes a great potential for pharmaceutical drug discovery and for the development of agrochemicals. Inside the plant, these metabolites serve different biological functions, including defence against pathogens, insects, and protection towards other environmental stresses (Morrisey 2009).

For the scope of this thesis, I will refer to the small molecules involved in plant central energy and growth processes as “primary metabolites”, and I will refer to specialized molecules involved in plant defence and interaction with the environment as “secondary metabolites”. However, it is difficult to draw clear biosynthetic and functional boundaries within the plant metabolic network, and I am aware that the traditional division between primary and secondary metabolism is neither completely satisfactory nor meaningful.

1.1.1 Primary metabolism

Plant primary metabolism starts with the process of photosynthesis, the conversion of CO2 into carbohydrates driven by sunlight. Leaves absorb CO2 through stomata, small pores on the epidermis surface which control plant water transpiration and gas exchange. Solar energy is meanwhile trapped by light-harvesting complexes and chlorophyll inside the chloroplasts. Here, via photosystem centres (PSI and PSII) and the electron transport chain (ETC), plants use excited electrons to create chemical bonds (e.g. the reducing agent NADPH). This “chemical energy” is transferred to the Calvin-Benson cycle, where ribulose-1,5-bisphosphate carboxylase (RuBisCO) traps the CO2 into simple sugars (Fig.2), later incorporated into mono- or polysaccharides for energy storage (starch) or for building the plant cell wall (cellulose).

Sugars are then oxidized with NAD+ along catabolic pathways of glycolysis in the cytosol, and in the mitochondria via the tricarboxylic acid (TCA) cycle, coupled to ETC and cellular respiration. Throughout this process, electrons are passed to oxygen (O2) reducing it to water (H2O), while energy is stored into ATP and transferred to other metabolic activities. Mitochondria are the centre of energy metabolism in all higher organisms (eukaryotes), but they play a double role in plants, where they are tightly connected to chloroplast, sustaining the assimilation CO2 and the redox balance (Blanco et. al, 2014). Glycolysis and TCA central metabolic pathways also provide the precursors necessary for the biosynthesis of amino acids, fatty acids, and other organic acids, which are used as “building blocks” in anabolic pathways for the final assembly of cellular macromolecules (e.g. DNA, proteins and membranes) (Stephanopoulos et al., 1998; Noor et al., 2010) and for the biosynthesis of secondary metabolites (Morrisey 2009) (Fig.2).

1.1.2 Secondary metabolism

Secondary metabolism evolved as groups of new pathways branching from the central processes of photosynthesis, cellular respiration, and carbon- nitrogen metabolism (Weng, 2014). Approximately 15-20% of plant genes are predicted to be involved in the biosynthesis of secondary metabolites (Pichersky, 2011), which always derive from sugar and/or amino acid precursors (Fig.2). For instance, one of the major enzymatic steps connecting primary and secondary metabolism involves the highly conserved enzyme phenylalanine ammonia lyase (PAL), which converts the aromatic amino acid phenylalanine (Phe) into trans-cinnamic acid, at the beginning of the phenylpropanoid pathway (Fraser & Chapple, 2011). Depending on the plant family and species, secondary metabolites range from different kinds of phenolics, terpenoids, or nitrogen containing compounds, such as alkaloids (e.g. Solenacea, Rubiacea), cyanogenic glucosides (e.g. Rosacea, Fabacea), and glucosinolates (GSLs) (Brassicaeae) (Halkier & Gershenzon, 2006). Most of these compounds are normally stored as cell-bound metabolites (Wink, 1997), but they can also be released as volatile organic compounds (VOCs) (Laothawornkitkul et al., 2009; Dudareva et al., 2013). Initially, these phytochemicals were considered to be only “by-products” of primary metabolism and not necessary for the plant survival (Bourgaud et al., 2001; Firn & Jones, 2009). However, secondary metabolites mediate almost all interactions within an ecosystem, and directly contribute to the plant fitness allowing processes such as plant-plant communication (Farmer & Ryan, 1990; Karban et al., 2014), competition (allelopathy) (Morrisey 2009), insect pollination (entomophily) (Burger et al., 2010), and plant defence against insect herbivores (Howe & Jander, 2008).

Figure 2. Plant metabolism. A simplified overview of the plant primary (yellow) and secondary (pink) metabolism, and plant volatile organic compounds (VOCs) (blue). Primary metabolism is involved in central processes of growth and development, starting with routes of photosynthesis and carbon metabolism (Calvin- Benson cycle). Carbohydrate and amino acid synthesis are linked via tricarboxylic acid (TCA) cycle. These central pathways produce chemical energy (ATP) and simultaneously provide all the metabolic precursors (“building-blocks”) required for cellular components (e.g. nucleotides, proteins, cell-membranes and cell-wall) and for the biosynthesis of plant secondary metabolites, e.g. phenolics or Brassicaea specialized glucosinolates (GSLs). Secondary metabolites are usually involved in plant protection and defence against environmental stresses. VOCs derived from primary or secondary metabolism can mediate ecological interactions with other biota (e.g. GLSs derived thiocyanates, isothiocyanates, and nitriles). In addition, phytohormones (in red) regulate both growth and defence processes, and can also be found as VOCs.

1.2 Eco-metabolomics

The complete set of metabolites in an organism is called the “metabolome” (Tweeddale et al., 1998), and in analogy with other systems biology “omics” approaches such as genomics, transcriptomics, and proteomics, the unbiased analysis of the entire metabolome is named metabolomics. Metabolomics is regarded as the link between genotypes and phenotypes (Fiehn, 2002) and it represents a powerful tool to generate new biological hypotheses. Because all plant metabolites can virtually function as info-chemicals in ecosystems, the comprehensive approach of metabolomics is particularly suited for the study of plants´ ecological interactions with the environment, in this case referred to as “eco-metabolomics” (Sardans et al., 2011).

Plants are important members of the ecological community, connecting organisms into different relationships and providing shelter and resources. During the interaction with the surrounding environment, eco-physiological responses will reconfigure the plant inner metabolic networks (Bundy et al., 2008; Maag et al., 2015) and will shape the communication of organisms at higher levels of the trophic network (Bourgaud et al., 2001). Interestingly, the concept of “networks” is shared between the field of ecology and systems biology (Butts, 2009; Ings et al., 2009; Keurentjes et al., 2011). Adopting a systems perspective, the flow of energy and matter through ecosystems can be visualized as the perpetuation of metabolic pathways through organisms (Capra, 1996). Thus, likewise cellular metabolism, plant interactions with other organisms can be seen as a network of interconnected relationships subjected to constant change in a dynamic environment.

1.2.1 Plant-insect interactions

Among the most complex chemical communication networks that we can observe in nature, are those that define plant-insect interactions (Fig.3). Insects and plants colonized land roughly at the same time, during the Early Ordovician, and thus coexisted and interacted in the terrestrial environment for about 470 myr (Misof et al., 2014). Of the entire animal kingdom, insects are by far the most diverse group, with 900.000 species already described and more than five million estimated (Stork et al., 2015). As the theoretical ecologist Robert May once famously noted: “by a good approximation, all species are insects” (in Dawkins, 1976).

The relationship between plants and insects is complicated. About 80% of higher plants develop flowers (angiosperms) (Pimm & Joppa, 2015) and most of them rely on insects as their main source of pollinators, such as bees, bumblebees, flies, or butterflies (Munguia-Rosas et al., 2011; Bruinsma et al., 2014). Combining colorful visual cues and aromatic chemical blends, plants hack into the insects’ communication system and lure them into their petals (Burger et al., 2010) (Fig.3). This process usually results in a win-win (mutualistic) interaction, where insects are rewarded with sugar-rich floral nectar, but some rare dishonest plants (4-6% of angiosperms; Renner 2006), can cheat and attract insects exploiting fake-food rewards (Aristolochiaceae; Oelschlägel et al., 2015) or even sexual deception (Orchidaceae; Ayasse et al., 2011). However, floral odors and other cues from the plants can attract insect herbivores (Karban & Agrawal, 2002; Howe & Jander, 2008; Sauve et al., 2016) which can be extremely harmful and destroy the plant (Crawley, 1989; Adhikari & Russell, 2014; Stephens & Westoby, 2015; Lemoine et al., 2017).

Figure 3. Chemical networks define plant-insect interactions. Multiple levels of connection between plant metabolic networks and interaction with insects.

1.2.2 Plant defence responses

Without possibility to move, plants developed a variety of defence strategies to cope with insect herbivores. Aside from physical barriers (e.g. modified leaf and branch structures, such as thorns and spines) plants highly rely on chemistry to fight their enemies. Most specialized secondary metabolites are produced for defence purposes as part of their constitutive (steady-state) metabolism, and thus compete with primary metabolic demands for growth (Stamp, 2003; Neilson et al., 2013)(Fig.2). Secondary metabolites constitute an effective direct chemical defence which can impair the performance of generalist (polyphagous) herbivores (Howe & Jander, 2008). However, specialist (monophagous) herbivores that co-evolved with the plant are often attracted by these secondary metabolites which they are capable to detoxify and digest (Ali & Agrawal, 2012; Edger et al., 2015).

Upon herbivore attack, plant defences can be further induced, providing an energy-saving solution to maximize the plant fitness (Heil & Baldwin, 2002; Neilson et al., 2013; Moore et al., 2014). The induction process is regulated by signal-transduction pathways which involve three major plant phytohormones: jasmonic acid (JA), salicylic acid (SA), and ethylene (ET). JA and SA pathways are reciprocally antagonistic (Thaler et al., 2012) and in a complex cross-talk with ET (and other phytohormones, including auxins, cytokinins, gibberellins, ABA, etc.) they fine-tune the herbivore-induced responses (Ali & Agrawal, 2012; Erb et al., 2012), similarly to the process induced during plant immunity against pathogens (Pieterse et al., 2009). This hormonal network will eventually determine the metabolic strategy of the plant, in order to defend against generalist and specialist or different herbivore guilds. For instance, chewing herbivores, such as caterpillars, typically induce the biosynthesis of the oxylipin JA, which is derived from fatty acid metabolism along the octadecanoid pathway (e.g. α-linolenic acid) (Fig.2) (Schaller & Stintzi, 2009; Wasternack, 2015; Lortzing & Steppuhn, 2016). On the other hand, sucking herbivores that feed on the plant phloem, such as aphids, induce synthesis of SA (Glazebrook, 2005; Pieterse & Dicke, 2007), a phenolic metabolite derived from phenylalanine (Chen et al., 2009).

Different defensive metabolites have different biological activity and can either reduce the nutritional quality inside the leaves or can be emitted as a bouquet of VOCs, such as phenylpropanoids, terpenoids (C5-C20), and fatty acid-derived green leaf volatiles (GLVs) (Laothawornkitkul et al., 2009; Dudareva et al., 2013). These VOCs constitute an indirect defence for the plant, often functioning as a “cry-for-help” which attracts enemies of the herbivore, such as predators and parasitoids (Dicke & Baldwin, 2010; Ponzio et al., 2014; Bruce, 2015) (Fig.3).

1.3 Study system

For the work presented in this thesis, together with my colleagues I studied the plant-insect interaction between the Black mustard and its specialist herbivore, the Large white cabbage butterfly. This plant-herbivore system was reproduced in collaboration with different universities which combined complementary expertise in plant biology, entomology, and ecology. My role in the projects was to analyze the plant metabolic changes relative to primary and secondary metabolism, which were induced by environmental stresses.

1.3.1 Black mustard

The kingdom of heaven is like a grain of mustard seed, which a man took, and sowed in his field. Which indeed is the least of all seeds, but when it is grown, it is the greatest among herbs, and becomes a tree, so that the birds of the air come and lodge in the branches thereof.

Matthew 13:31-32

The Black Mustard, i.e. Brassica nigra L. (Brassicacea), is an annual herbaceous plant originally from Middle East and Mediterranean regions, which has been extensively cultivated by humans for millennia (Prakash et al., 1980). Although a minor crop nowadays, gradually replaced by the Ethiopian mustard B. juncea, it is often found as a weed and as an invasive species widely spread across the globe (Westman & Kresovich, 1999), thriving in both cold and hot geographical regions (60ºN - 40ºS) and surviving extreme temperatures of 45-50º C (Kask et al., 2016). In the field, B. nigra can stand two meters in height with long reaching branches. Lower leaves are usually dentate with wide lobes, while the upper leaves are narrow and lanceolate. Flowers are hermaphrodite, bright and yellow, with four petals oriented in a cross shape common to all cruciferous plants (from Latin, “crux” = cross). As an obligate out-crossing species in nature, B. nigra is pollinated by several insects, including bees and butterflies (Conner & Neumeier, 1995).

(a) (b)

(c)

Figure 4. The Black mustard (Brassica nigra). (a) yellow flower with petals disposed in the typical cross-shape (cruciferous) of Brassicacea, (b) whole inflorescence, and (c) B. nigra plants, approximately four-week old, as grown for the experiments of this thesis conducted at Umeå University (Wallenberg greenhouse).

The genus Brassica comprehends 38 species, six of which are crops of high economic importance and are cultivated for the production of vegetables, condiments, or oilseeds (Table 1) (Das et al 2007). These include turnip (Brassica rapa), cabbage, broccoli, cauliflowers (Brassica oleracea), as well as rapeseed (Brassica napus) used for biodiesel production (Golovitchev & Yang, 2009). B. nigra is mostly cultivated for its seeds, which are ground and mixed with water and other ingredients (e.g., vinegar and lemon) to produce the mustard paste, commonly used in many countries use as a cooking spice.

Table 1. Brassica species and their economic importance (Das et. al 2007). Species and variety Common name Uses B. nigra Black mustard Condiment B. oleracea var. acephala Kale Vegetable var. alboglabra Chinese kale Vegetable var. botrytis Cauliflower Vegetable var. capitata Cabbage Vegetable var. italica Broccoli Vegetable B. rapa var. chinensis Pak choi Vegetable var. olifeira Turnip rape Oilseeed var. rapifera Turnip Vegetable var. yellow sarson Yellow sarson Oilseeed var. toria Toria Oilseeed var. brown sarson Brown sarson Oilseeed var. nipposinica Mizunami Salad B. carinata Ethiopian mustard Condiment B. juncea var. oleifera Indian mustard Oilseeed B. napus var. oleifera Rapeseed Oilseeed var. rapifera Sweed Fodder

Brassicaeous plants contain high amounts of secondary metabolites called glucosinolates (GSLs). GSLs are sulfur and nitrogen containing glycosides derived from amino acids. According to the class of amino acid from which they derive, GSLs are divided in:

- aliphatic (from methionine, alanine, leucine, isoleucine, or valine). - aromatic (from phenylalanine) - indolic (indeed aromatic, but from tryptophan)

The central carbon of the GSL structure is bound via a nitrogen atom to a sulfate group and via a sulfur atom to a glucose moiety (Fig.5). In addition, a side group (R) which is derived from the amino acid precursor differs according to the kind of GLS and is responsible for the variation in the biological activities (Halkier & Gershenzon, 2006).

In B. nigra, the most abundant secondary metabolite is the aliphatic GSL derived from methionine, sinigrin (2-propenyl-GSL), which represents 90- 99% of the entire plant’s GLS profile (Lankau & Strauss, 2007).

Breaking B. nigra seeds and mixing with water confers the flavor to the mustard paste, similar to that of horseradish. This flavor is due to VOCs released from the degradation of sinigrin during a hydrolytic reaction called the “mustard oil bomb” (Fig.5).

Figure 5. The “Mustard oil bomb”. The hydrolysis of GSLs (e.g. sinigrin in B. nigra) is catalyzed by myrosinase, producing different volatile compounds (according to the different environmental conditions, such as pH) some of which have toxic properties, for instance ITCs (e.g. allyl isothiocyanate). The general GSL structure is shown. Side R-group for sinigrin: CH2=CH=CH2-. See GSLs list in Clarke (2010).

The hydrolysis of GSLs is catalyzed by the enzyme myrosinase (Rask et al., 2000) which is produced by all Brassicaeae plants and was first reported in B. nigra (Bussy, 1840). Myrosinase is a glycosidase (EC 3.2.1.147), normally kept separate inside plastids or in the apoplast (Morant et al., 2008). Upon any tissue disruption - by herbivores or by mechanical damage – myrosinase gets in contact with GSLs, cleaving their glucose moiety and converting them into VOCs such as thiocyanates, isothiocyanates (ITCs), and nitriles (Winde & Wittstock, 2011).

Particularly ITCs have toxic properties and constitute an efficient direct and indirect defence against insect herbivores (Burow et al., 2009; Hopkins et al., 2009; Kos et al., 2012). Arabidopsis thaliana L. knocked-out mutants impaired in the formation of ITCs (Burow et al., 2006) or in the biosynthesis of GSLs (Beekwilder et al., 2008) are more sensitive to feeding by caterpillars, especially those of generalists butterflies such as Mamestra brassicae or Spodoptera littoralis. On the other hand, specialist herbivores of Brassicaceae, such as Pieris butterflies, are immune to these metabolites (Blatt et al., 2008; but see arguments in the study on B. nigra and P. rapae by Traw & Dawson, 2002).

1.3.2 Large white cabbage butterfly

The Large White cabbage butterfly, i.e. Pieris brassicae L. (Pieridae), is a Lepidoptera commonly found in Eurasia, and it is a serious pest of all cultivated Brassica vegetables, including B. nigra.

Butterflies of P. brassicae can feed on B. nigra floral nectar (Fig.6a) and then oviposit on the lower side of leaves, in batches of 50-100 eggs (Fig.6b). After hatching, neonate (first instar) caterpillars move gregariously and feed on leaves (Fig.6b,c). However, late second and early third instars prefer buds and flowers, which contain high concentrations of GSLs (Smallegange et al., 2007). After the fifth (last) instar, caterpillars leave the plant for a safe location to mutate into pupae (chrysalis) (Fig.6d), from which they reemerge about 14 days later as butterflies. In our greenhouse, the complete life cycle usually lasted 5-6 weeks (Fig.6e), but the duration in the field largely depends on the geographical location and environmental conditions. As a specialist of Brassicaceae, female butterflies are attracted by ITCs and nitriles, which they use to locate their host (Hopkins et al., 2009).

Egg deposition creates necrotic zones on B. nigra leaves and can suppress JA-induced defences via SA accumulation (Bruessow et al., 2010). However, studies in Arabidopsis showed that egg deposition can induce JA defences and trigger mechanisms similar to those induced by pathogen-associated molecular patterns (PAMPs) (Gouhier-Darimont et al., 2013). P. brassicae oviposition on B. nigra can also alter the plant induced responses before and after caterpillar feeding (Pashalidou et al., 2015), whereas male pheromones left on mated female butterflies can cause chemical modifications of the leaf surface that eventually attract the egg parasitoid Trichogramma brassicae L. (Fig.7a-c)(Fatouros et al., 2005; Fatouros et al., 2008).

P. brassicae caterpillars can digest GSLs which they degrade into non-toxic nitriles (Smallegange et al., 2007; Winde & Wittstock, 2011) and can even interfere with the induction of plant responses (Reymond et al., 2000; Karban & Agrawal, 2002). For instance, oral secretions of P. brassicae suppress wound-induced responses in Arabidopsis (Consales et al., 2012). On the other hand, elicitors in these same secretions are also recognized by the plant as molecular signals of the insect-attack, leading to the induction of plant direct and indirect defences (Mattiacci et al., 1995; Reymond et al., 2004). Induced emission of ITCs and nitriles will attract the parasitoid wasp Cotesia glomerata L. (Fig.7 d-g) which lay its eggs inside the caterpillars (Agrawal & Kurashige, 2003; Pashalidou et al., 2015).

Figure 6. The Large white (P. brassicae L.). Development of P. brassicae in stages: (a) Fully developed butterfly, (b) hatching eggs with neonate first instar caterpillars, (c) fully developed caterpillars feeding on leaves of Brassica plants, (d) pupae stage and (e) full life-cycle. Photo: T. Bukovinszky and H. Smid.

All these fascinating processes are the result of 470 myr interaction between plants and insects (Thaler et al., 2012; Misof et al., 2014; Bruce, 2015) and specifically of the 70 myr arms-race between Brassicacea and Pieridae (Edger et al., 2015). However, in the climate change scenario, large environmental shifts are shaping plant-insect interactions on a much shorter time-scale.

Figure 7. Indirect defences. Parasitoid wasps are attracted by B. nigra as an indirect defence mechanism. (a,b) Trichogramma brassicae is attracted by chemical modifications on the leaf surface and lays its eggs inside P. brassicae´s eggs. (c) T. brassicae (arrow) hitchhiking on P. brassicae butterfly. Photo: N. Fatouros. (d) Cotesia glomerata is attracted by VOCs released upon herbivore damage (ITCs and nitriles) and attacks early instar of P. brassicae (e-g) caterpillars to lay eggs. Photo: T. Bukovinszky and H. Smid.

Climate change models predict an increase in the occurrence and in the intensity of drought and heat waves (IPCC), with further concerns for natural ecosystems and agricultural productivity (Fuhrer, 2003; Atkinson & Urwin, 2012; Suzuki et al., 2014). Plants and insects are migrating towards northern latitudes (Parolo & Rossi, 2008; Chen, IC et al., 2011; Bebber et al., 2013), thus altering the evolutionary processes driving the adaptation of both species and the stability of native info-chemical networks (Morriën et al., 2010; Desurmont et al., 2014).

1.4 Multiple stresses

A realistic description of plant-insect interactions cannot exclude other elements which are integral components of the environment. In nature, chewing damage by caterpillars is only one type of stress that plants encounter, and plant responses to herbivory always concur with response to other stresses. In addition to biotic stress, which is stress caused by any living agent, plants are also exposed to a variety of abiotic stresses caused by non-living environmental factors - e.g. lack or excess of water in the soil (drought or flood), extreme temperatures, UV radiation, or air pollution (Fig.8). The performance of plants in the field thus depends on the ability to perceive different environmental cues and to quickly adapt new responses. Both biotic and abiotic stresses can severely impair growth, and plants developed different mechanisms to reduce the damage (Rejeb et al., 2014; Nguyen et al., 2016).

Among different abiotic stresses, the most studied that affect both natural ecosystems and human agriculture include drought, heat, cold (chilling, freezing), nutrients, high salinity, high light intensity, and ozone (Suzuki et al., 2014). Drought and heat combined cause the most damage to global agricultural production (Hatfield & Walthall, 2015) and are thus among the most investigated abiotic stresses, along with salt stress and heavy metals (Suzuki et al., 2014; Zandalinas et al., 2017). Interaction between biotic and abiotic stresses can trigger convergent phytohormone cross-talk (JA, SA, ET) and signalling pathways (e.g. ROS and Ca++)(Nguyen et al., 2016, Pandey et al., 2016) which often reinforce or inhibit each other and lead to a variety of physiological responses (Atkinson & Urwin, 2012; Rejeb et al., 2014; Suzuki et al., 2014).

Thus, predictions based on individual stress responses only partially help us in understanding the complexity of the plant response to multiple stresses, as the presence of an initial stress can alter the response to a second stress and eventually result in a positive, negative, or neutral interaction (Fig. 8) (Mittler, 2006; Atkinson & Urwin, 2012; Suzuki et al., 2014). For instance, simultaneous exposure to heat and drought stress in tobacco (Nicotiana tabacum L.) has a negative interaction on the plant response, with greater suppression of photosynthesis compared to the individual stresses (Rizhsky et al., 2002). Similar responses are also observed in Arabidopsis, resulting in differential expression of both defence and central metabolic transcripts, enhanced respiration, and an increased accumulation of sucrose and other sugars (Rizhsky et al., 2004).

Relatively few studies have investigated abiotic stress responses in plants combined with biotic stresses (Suzuki et al., 2014). In most of these studies, the biotic stress components was often represented by pathogens (i.e. virus, bacteria or fungi) - possibly as easier to handle than insects (Fig.8), and analyses usually focused on the effects at the gene expression at level or on the role of single transcription factors (Rizhsky et al., 2004; Atkinson & Urwin, 2012; Sewelam et al., 2014; Olivas et al., 2017).

Figure 8. The stress matrix. Combination of different environmental stresses can affect plants in the field, resulting in positive and/or negative interactions on the plant performance. Color-codes indicate the effects of stress combinations on plant growth and yield. Multiple abiotic stress conditions have been thoroughly studied (especially combinations with drought), but only few studies have been performed with biotic components, which were usually limited to pathogenic microorganisms (Suzuki et al., 2014).

When metabolomics was applied to study multiple environmental stress responses in plants it often highlighted a large functional overlaps between pathways, and the importance of both primary and secondary metabolism for plant adaptation to the stresses (Sana et al., 2010; Caldana et al., 2011; Sun et al. 2015; and see reviews by Bundy et al., 2008; Schwachtje & Baldwin, 2008; Nakabayashi & Saito, 2015). In this thesis, changes in plant metabolism were investigated for combined abiotic and biotic stresses which may alter the plant response towards consecutive herbivory, such as elevated ozone exposure, aphid herbivory, and phytohormone treatments (Fig.9-10).

1.4.1 O3: a tropospheric pollutant

Abiotic stress in this thesis has focused on the negative effects of ozone (O3). O3 is the three-atom conformation of oxygen, but whereas O2 is essential for cellular respiration, O3 is toxic to living organisms. However, in comparison to O2 which is extremely abundant in the Earth´s atmosphere (20.9%), O3 is only a trace gas, with ground-level concentrations of 25-30 ppb (Chevalier et al., 2007). Most of O3 (90%) is found at the altitude between 20-50 km (stratosphere), where man-made halocarbon emissions have infamously depleted the layer that protects us and all other living organisms from UV solar radiation. On the other hand, in the atmospheric region from ground- level to ca. 20 km (troposphere), O3 is a strong oxidizing pollutant.

As a result of rapid economic growth and human emissions, tropospheric O3 concentrations have been globally increasing in the last decades (Lin et al., 2017). Burning of fossil fuels releases nitrogen oxides and hydrocarbons which, in the presence of sunlight, react with O2 and VOCs to form O3 (Sitch et al., 2007). The steady emission of these precursors has led to an increase in the background concentrations of O3, which during hot summers in the Northern Hemisphere can peak to 200-400 ppb (Tiwari et al., 2016). Although most of O3 is produced in polluted urban areas and is directly related to traffic (Wennberg & Dabdub, 2008), long-range transport of O3 to rural areas can severely affect crop yields and forest growth (Ashmore, 2005). Depending on the plant species, accumulated exposure to O3 above 40 ppb can result in negative effects on photosynthesis, leaf transpiration, and emission of VOCs (Bagard et al., 2008; Booker et al., 2009; Blande et al., 2010; Tiwari et al., 2016), and can have severe impact on ecological interactions and agricultural systems (Pinto et al., 2010).

1.4.2 Aphids: sap-sucking herbivores

Aphids (Aphidoidea) are small sap-sucking insects that feed on the plant phloem. Among all insect pests, aphids are some of the most destructive and difficult to control, as they rapidly increase in density doubling every few days throughout cycles of asexual reproduction (parthenogenesis) performed by genetically identical female clones (Davis, 2012). When feeding, aphids avoid the massive cell damage typical of chewing herbivores like caterpillars. Instead, they penetrate the leaf epidermis adopting a “stealthy” piercing-sucking strategy which triggers SA-induced responses, similar to those induced by biotrophic pathogens (Walling, 2008; Thaler et al., 2012; Foyer et al., 2016; Nguyen et al., 2016). However, JA-induced responses are also known to be involved in the plant defence against aphids (Ellis et al., 2002; De Vos et al., 2005; De Vos & Jander, 2009).

In this thesis, we studied the cabbage aphid Brevicoryne brassicae L., which is a serious pest of all Brassica crops worldwide (Ellis & Farrell, 1995; Ellis et al., 2000; Broekgaarden et al., 2008). B. brassicae is a specialist herbivore of Brassicacea and is able to sequester the plant GSLs (e.g. sinigrin) and – using its own myrosinase (Jones et al., 2001) – to exploit them for its own defence against predators, such as the ladybirds Adalia bipunctata L. Hence the famous nick-name, “walking mustard oil bomb” (Kazana et al., 2007).

1.4.3 MeJA: herbivore mimicking

Upon herbivore-attack and tissues damage especially by caterpillar, intracellular JA accumulates within the plant to spread the defence response both locally and via the phloem as long distance signal (Glauser et al., 2008; Chauvin et al., 2013; Mousavi et al., 2013). JA signal transduction depends on the interaction of its conjugated form JA-isoleucine (Ja-Ile) with ubiquitin ligase SCFCOI and the jasmonate-zim-domain (JAZ) repressor (Chini et al., 2007; Katsir et al., 2008); JAZ normally blocks the response by binding to the transcription factor MYC2, a master regulator of plant herbivore-induced defences. Accumulation of JA leads to ubiquitination and degradation of JAZ, and to the activation of JA-responses (Lortzing & Steppuhn, 2016).

Initial studies conducted in tomato (Solanum lycopersicum L.) discovered that the JA ester derivative, methyl-JA (MeJA), was able to further spread as airborne signal, allowing plant-plant communication and induced-resistance in non-attacked plants (Farmer & Ryan, 1990; Wasternack, 2015).

Further studies confirmed that exogenous application of MeJA sprayed on plants can stimulate responses very similar to those induced by physical herbivore damage, including induction of genes involved in defence (Kouzai et al., 2016; Yi et al., 2016), synthesis of specialized secondary metabolites (Fritz et al., 2010; Zang et al., 2015) and emission of VOCs (Geervliet & Brodeur, 1992; Gols et al., 2003; Lortzing & Steppuhn, 2016).

In Brassicacea, increased GSL levels are typically observed after application of JA or MeJA (Cipollini & Sipe, 2001; Petersen et al., 2002; van Dam & Oomen, 2008). For these reasons, JA and MeJA are often used as molecular tools to study herbivore-induced resistance and production of secondary metabolites, with applications ranging from bioprospecting (Yukimune et al., 2000; Leonard et al., 2009) to agricultural and forestry pest management (Rodriguez-Saona et al., 2001; Fritz et al., 2010; Lundborg et al., 2016).

Figure 9. Multiple stress conditions. Abiotic and biotic stress combinations studied in this thesis on the model brassicaceous plant B. nigra. The focus of these studies was to understand how metabolic changes induced in the plant in response toward P. brassicae caterpillar herbivory, are affected by abiotic and biotic stresses, such as ozone (O3), B. brassicae aphids, and phytohormone treatment with MeJA.

Figure 10. Experimental set-ups. Stress responses were tested in B. nigra combining herbivory by P. brassicae (green) with other abiotic and biotic stresses. (Paper I): 70-120 ppb O3 stress (blue) and three days herbivory with 20 caterpillars; (Paper II): 70 ppb O3 stress (blue) and one day of herbivory by 30 caterpillars; (Paper III): dual herbivory of three days by 50-100 B. brassicae aphids (yellow), and one day by 30 caterpillars; (Paper IV): three days of phytohormone induction by 1mM methyl-jasmonate (MeJA)(red) and five days of herbivory by 15 caterpillars. The analyses included metabolomics, transcriptomics, and herbivore performance.

2. Objectives

In this thesis I present a collection of studies that aimed at clarifying the way plants regulate their metabolism during the response to caterpillar herbivory and other abiotic and biotic stresses (Fig.9-10).

In the first study (Paper I), together with colleagues at the University of Eastern Finland (Kuopio, Finland), we investigated how exposure to elevated O3 concentrations (70-120 ppb) can affect plant-insect interactions between B. nigra and P. brassicae. Combining herbivore performance tests with analyses of plant secondary metabolism, we were able to discriminate direct and indirect effects of O3 on the plant-herbivore system.

In a follow-up study (Paper II), we investigated the effects of O3 (70 ppb) and herbivory on the plant metabolic responses relative to primary and secondary metabolism. For this project we collaborated with the University of Eastern Finland and the University of Lausanne (Lausanne, Switzerland). Metabolomics was integrated with transcriptomics combining network and pathway analyses, allowing us to suggest a comprehensive model for the plant stress adaptation, which was later verified on the plant physiology.

In the third study (Paper III), we collaborated with the Entomology Department of Wageningen University (the Netherlands) to investigate plant responses towards multiple herbivores. We examined the changes in plant primary and secondary metabolism after single and dual herbivory by both B. brassicae aphids and P. brassicae caterpillars, also considering potential effects of aphid-density.

In the last study (Paper IV), we investigated the effects of MeJA treatment on the plant metabolic response against herbivory. Particularly, we examined how MeJA induction can prepare the plant metabolome towards consecutive caterpillar attack, resulting in an increased plant resistance. Moreover, we evaluated how herbivore-induced responses, relatively for both primary and secondary metabolism, were shaped along the plant leaf ontogeny.

3. Results and Discussion

3.1 O3 effects on plant-insect interactions

In Paper I we investigated the effects of O3 stress (70 - 120 ppb) on feeding preference and performance of caterpillars, combined with the analyses of plant secondary metabolites (GSLs and phenolics).

In a follow-up study (Paper II) we focused on the responses to O3 (70 ppb) and sequential herbivory, linking metabolomics and transcriptomics in a systems biology framework, and examining the plant physiological responses relative to photosynthesis, stomatal conductance, and leaf transpiration.

3.1.1 Effects of O3 on herbivory and plant defences

We initially investigated the effects of elevated O3 on caterpillar herbivory and caterpillar development.

Dual-choice tests were performed to determine feeding preference towards host-plants treated with different O3 concentrations. Inside a Petri dish, caterpillars were offered the youngest fully developed leaves (usually L3-L4) (Fig.11a) and let choose for 12 hours between two differently treated plants: control vs. 70 ppb; or vs. 120 ppb O3; and 70 ppb vs. 120 ppb O3 (Fig.11b). Overall, O3 exposure resulted in increased feeding, with a larger leaf area consumed on plants treated with 70 ppb and 120 ppb O3 (ca. 3.8 cm2) compared to controls (ca. 2.2 cm2), and with the most damage observed on plants exposed to 120 ppb O3, which were also preferred to control plants (Student’s t-test, P = 0.02) (Fig.11b).

At the same time, O3 negatively affected the development of caterpillars. During the treatment, caterpillars directly exposed to O3 while feeding and growing on the host-plant showed a reduced weight after three and six days at 70 ppb (Student’s t-test, P < 0.01) and at 120 ppb O3 (P < 0.05) (Fig.11c). Growth of caterpillars was indirectly affected by O3 stress even when only the plants were exposed to elevated O3 (70-120 ppb) (Fig.11d), with 120 ppb also affecting their pupation time and final pupal weight (Fig.11e,f).

Figure 11. O3 exposure negatively affects herbivory. (a) Dual-choice set-up for feeding preferences and (b) leaf area consumed (Student t-test, P = 0.02). (c) Direct effects on caterpillar development, with whole-system exposure to O3, and (d) indirect effects with only plants exposed to O3. Further indirect effects on caterpillars growing on B. nigra exposed to O3: (e) caterpillars average days to pupation, and (f) final pupal weight. Significance: one-way ANOVA, Tuckey comparisons. Color-codes for O3 treatments: ambient (white), 70 ppb (blue), 120 ppb (dark-blue). After Fig. 1-4 in Paper I.

To understand the change in the herbivore development caused by the treatments, we performed metabolomics of plants exposed to O3 stress. Analysis via LC-TOF-MS and linear ion trap for tandem MSMS (Orbitrap) allowed the detection of over 400 features and the identification of several metabolites as GSLs (Clarke, 2010) and phenolics (hydroxycinnamates and flavonol glucosides)(Lin et al., 2011). Main metabolites are listed in Table 2.

The metabolic responses to O3 stress after five days (70 ppb and 120 ppb) were modelled with supervised multivariate analyses (MVA) (Fig.12a,b). Plants exposed to O3 were affected in GSL and phenolic metabolism in different directions. Effects of O3 included lower levels of glucoiberin (IBE), 4-hydroxyglucobrassicin (4OH), and neoglucobrassicin (NEO). On the other hand, we observed higher levels of glucobrassicin (GBC), gluconasturtiin (NAS), glucojiaputin (JIAP), and particularly sinigrin (SIN) which increased on from 3.9 to 4.7 µmol/g FW (+17%) (ANOVA, P < 0.05) (Fig.12b,c). Phenolic metabolites were also affected, with a decrease in 1-caffeoyl-β-D- glucose (CGLU) and quercetin-3-sinapoylsophoroside (QSPS), whereas isorhamentin-3-sinapoyl-p-coumaroyldiglucoside (ISCDG), kaempferol (KG, KGG), and quercetin (QS, QSG) glucosides increased. The strongest effects were however observed on another metabolite - i.e. feature “421”, which was always induced by elevated O3, but that we were not eventually able to identify (Fig.12b, and Table 2).

The few studies that investigated the effects of O3 stress on Brassicaceae with focus on the plant secondary metabolism and defence, usually showed shifts in concentrations of GSLs (Gielen et al., 2006), or altered profile ratios between aromatics and indolics (Himanen et al., 2008). In a previous study, B. rapa sensitivity towards O3 stress was shown to interact in a complex way with the insect performace: after 21 days exposure to 75 ppb O3, pupation time was negatively affected for P. brassicae caterpillars that fed on O3- sensitive lines, whereas caterpillar performance and final pupal weight were affected on the O3-resistant lines (Jøndrup et al., 2002). In our study, caterpillars preferred to feed on O3 treated plants, although these same plants were suboptimal for their own development. Cumulative changes in the plant metabolome may have affected the leaf palatability for these specialist caterpillars, which usually exploit defence metabolites as feeding stimulants or as chemical deterrents against predators (van Loon & Schoonhoven, 1999; Ferreres et al., 2009; Winde & Wittstock, 2011).

(a) OPLS-DA

O3 0%) LV2 (2 LV2

LV1 (10%)

(b) (c) QS KG 381 QSG KGG SIN JIAP

QSPS GBC CGLU

ISCDG NEO NAS 4OH IBE 421

Figure 12. O3 stress affects plant secondary metabolism. Effects of O3 stresson the plants metabolic profiles of GSLs, phenolics, and two unknown metabolites (OPLS-DA, 1+1+0; R2Xcum=30%, R2Ycum=85%, Q2cum=44%). The effect of O3 is described along the LV1 (10%). (a) Score plot with plant biological replicates. Color-codes for O3 treatments: ambient (white), 70 ppb (blue), 120 ppb (dark-blue). (b) Loading plot with individual metabolites distributed according to their relative concentrations. Metabolite labels (see Table 2) show VIP scores ≥ 1.00. Color-code: GSLs (green), hydroxicinnamic acid derivates (pink), flavonol glucosides (purple), unknowns (lilac). (c) Sinigrin concentration in plants (µmol/g FW) with significance tested by one-way ANOVA and Tuckey comparisons. After Fig.S1 in Paper I.

Defensive metabolites other than GSLs and phenolics can affect herbivory and caterpillar development (Poelman et al., 2008), while reduction in leaf nutritive quality, such as amino acid and lipid content, is known to result in increased consumption via compensatory feeding (Slansky & Feeny, 1977; Jones & Coleman, 1988; Wheeler & Halpern, 1999). During basic elemental analysis, we found that O3 stress lowered the plant total nitrogen content at both concentrations of 70 ppb and 120 ppb, but did not affect total carbon and water contents (Tab.1 in Paper I). Alternatively, O3 stress may have also affected caterpillars by altering the host ability to initiate herbivore- induced responses during sequential feeding. Altogether, these results show that elevated O3 exposure can have strong effects on the plant metabolism with potential effect on associated organisms.

Table 2. B. nigra secondary metabolism relative to glucosinolates and phenolics. [M-H]- Class Metabolite Code Formula Rt MSn min m/z Sinigrin* SIN C10H16NO9S2 0.9 358.03 259 Gluconapin* NAP C11H18NO9S2 1.6 372.04 Glucojiaputin JIAP C12H23NO9S2 2.2 388.05 259 Glucoiberin* IBE C11H21NO10S3 0.8 422.02 Glucosinolates Gluconasturtiin* NAS C15H21NO9S2 3.1 422.06 Glucobrassicin* GBC C16H20N2O9S2 2.6 447.05 259 4-hydroxyglucobrassicin 4OH C16H20N2O10S2 1.7 463.05 Neoglucobrassicin NEO C17H22N2O10S2 3.4 477.06 1-caffeoyl- -D-glucose CGLU C15H18O9 1.9 341.09 203, 179 1-O-sinapoyl-glucose 1SNG C17H22O10 3.2 385.11 267, 223 β Hydroxy- Sinapoylferuloylgentiobiose SPFG C33H40O18 5.3 723.21 529 cinnamate derivatives Disinapoylgentiobiose DSPG C34H42O19 5.2 753.23 529, 487 Disinapoylferuloylgentiobiose DSPF C44H50O22 5.3 929.27 735, 705 Trisinapoylgentiobiose TSPG C45H52O23 5.6 959.28 735 Km-7-glucoside KG C21H20O11 4.9 447.09 Km-3,7-diglucoside KGG C27H30O16 3.2 609.15 Km-3-sinapoylsophorosid.-7-gl. KSG C44H50O25 3.1 977.26 815, 609 Flavonol Qn-7-sophoroside QS C27H30O17 2.9 625.14 glucosides Qn-3-sophoroside-7-glucoside QSG C33H40Q22 2.6 787.19 625 Qn-3-sinapoylsophoroside QSPS C38H40O21 4.8 831.19 Is-3-sinapoyl-p-coumaroyl-digl. ISCDG C48H48O23 5.8 991.25 639 Feature 381 - unidentified 381 (?) 5.1 349.15 Unknowns Feature 421 - unidentified 421 C25H28O5NS2(?) 5.9 485.13 441, 241

Main secondary metabolites detected in B.nigra. Metabolite class, common name, short codes, molecular formula, retention time (Rt min), mass charge ratio (m/z) detected by UHPLC-TOFMS using negative ionization mode [M-H]-, and MSMS product ions (MSn) detected by Orbitrap. Flavonols: Km = Kaempferol, Qn = Quercetin, Is = Isorhamnetin. * = GSL standard was available.

3.1.2 Omics responses to sequential O3 and herbivory

Plants exposed to five days of O3 stress (70-120 ppb) and consecutive three days of herbivory (see in Fig.11d) were analyzed and compared to the single stress treatments. Using supervised MVA, we were able to distinguish the effects of O3 stress on plants that were sequentially treated with herbivory (Fig.13a), and we could further separate between all the treatments for the effects of herbivory (LV1) and O3 (LV2), when accounting for both stresses as explanatory response variables (Fig.13b).

(a) OPLS-DA

%) O3 LV2 (19 LV2

Controls

LV1 (6%)

(b) OPLS-DA

Controls

Caterpillars %) LV2 (6 LV2

O3 + caterpillars O3

LV1 (10%)

Figure 13. Plant secondary metabolism is induced by O3 and herbivory. (a) Effects of O3 (OPLS-DA, 1+2+0; R2Xcum=36%, R2Ycum=86%, Q2cum=42%) compared to (b) O3 and herbivory (OPLS-DA, 2+3+0; R2Xcum=52%, R2Ycum=81%, Q2cum=40%) on secondary metabolism (GSLs and phenolics). Codes: O3 ambient (white), 70 ppb (blue), 120 ppb (dark-blue); herbivory with no O3 (pink); C = no herbivory (circles), DH = damaged by herbivory (squares); plants treated with O3 and sequential herbivory (green cluster). After Fig.5 and 6 in Paper I.

The strongest effects were observed on unidentified metabolites “421” and “381”, respectively induced by O3 and by herbivory (Fig.5-7 in Paper I). Interestingly, the induction of “381” by herbivory was lowered by initial exposure to 120 ppb O3, corresponding to same level which negatively affected growth and pupation of caterpillars. Herbivory seemed to counteract the effects of O3, decreasing GSL levels including sinigrin (Fig.5 and Tab.S1, in Paper I), possibly as a result of tissue damage and degradation into ITCs. Responses to sequential O3 stress and herbivory also affected the levels of hydroxycinnamates (e.g. CGLU, SNG, TSPG) and flavonol glucosides (QS, QSP, KSTG) (VIP scores ≥ 1.00) (Fig.5-6, and Tab.1 in Paper I). Performing a general MANOVA for GSLs, hydroxycinnamates, and flavonols, we could show significant effects of O3 stress (Pillai's Trace P = 0.016) and herbivory (P = 0.042) although no interaction between the treatments. However, using general linear models (GLMs), we found interactions on two indolic GSLs (4OH, NEO) and a quercetin glucoside (QSG) between the two O3 stress concentrations and herbivory (P ≥ 0.01; Tab.S1, in Paper I).

Combined with the previous analyses, these results show that O3 stress can influence plant responses towards herbivory, affecting the host metabolism and shaping plant-insect interactions. Both individual stresses caused very distinct effects in the plant secondary metabolism, but the response towards herbivory always dominated during the multiple stress condition.

In order to further understand how plant responses are affected during the shift from O3 stress to herbivory, we investigated O3 stress (70 ppb) and a shorter period of herbivory (30 caterpillars for 24 hours). VOCs were also collected from the plant head-spaces during the treatments, and leaf samples were shared between metabolomics (screening of primary and secondary metabolism) and transcriptomics (based on Arabidopsis microarrays). Combination of omics data allowed us to propose a model for the plant response to multiple stresses, which was eventually validated by directly measuring the plant physiological state (Paper II).

Figure 14. Herbivore-induced metabolic responses overtake O3 stress. (p.29) - Multivariate effects on primary and secondary metabolism, including VOCs (PLS-DA; R2Xcum=65%, R2Ycum=95%, Q2cum=44%). (a) Score plot (LV1 vs. LV2). Color-codes: controls (C) (white), 70 ppb O3 (O) (blue), P. brassicae herbivory (P) (pink), and sequential O3 and herbivory (OP) (green). (b) Loading plot showing important metabolites for the model (VIP > 1.00). Metabolites color-codes: organic and fatty acids (yellow), amino acids (red), sugars (orange), TCA cycle (lime), amines (dark-blue), hydroxycinnamates (lilla), flavonols (dark-purple), GSLs (green), VOCs (light-blue), unknowns (dark-grey). Abbreviations: α-ketoglutarate (α-KG), 3,8- dimethyl-1,4,7-nonatriene (E-DMNT). After Fig.3 in Paper II.

- Metabolome profiles modelled with supervised MVA (Fig.14) showed patterns very similar to the previous experiment: exposure to elevated O3 stress resulted in a distinct metabolic shift from control plants (LV2), but the strongest response was always induced by caterpillar herbivory, which also dominated the response during the sequential treatment (LV1) (Fig.14a). Analyses relative to secondary metabolism (LC-TOF-MS) showed a general increase in phenolics but a decrease in most GSLs, especially after herbivory (beside glucoraphanin). Likewise in the previous study, the two unidentified metabolites “381” and “421” were again induced respectively after herbivory and O3 stress (Fig.14b and Fig.3c in Paper II).

(a) PLS-DA

Caterpillars %) LV2 (14 LV2

LV1 (16%) (b)

- ITC

Herbivory induced the emission of several VOCs (analysed by TD-GC-MS), such as GSL degradation products (ITCs and nitriles), terpenoids (e.g. E- DMNT), and GLVs (Fig.14b and Fig.3c in Paper II). Responses between O3 and herbivory showed opposite effects on central primary metabolism - e.g. carbohydrates, amino acids, organic acids (GC-TOF-MS), which generally increased after O3 and decreased after herbivory.

Glycerol levels accumulated only in response to O3 (+55%)(ANOVA, P<0.05) but were later restored after herbivory. A common response to all treatments was the increase in γ-aminobutyric acid (GABA) (+300%) (ANOVA, P<0.05) (Fig.14b), a free-amino acid involved in abiotic and biotic stress responses such as photo-inhibition (Bouche & Fromm, 2004; Araujo et al., 2012), leaf senescence (Ansari et al., 2005), and plant-insect interaction (Michaeli & Fromm, 2015).

- Transcriptome profiles were initially examined separately, combining hierarchical clustering (HCA) and gene ontology (GO) enrichment analysis. Similarly to the metabolic profiles, responses to single O3 stress were clearly distinct from responses to herbivory (Fig.15a-c). All treatments strongly affected the plant central energy metabolism, including photosystems and mitochondrial electron transport chain (ETC) (Fig.15a,b). Pathway analysis (Mapman) confirmed a consistent response across several biological levels, with a dominance of up- and down-regulated genes after O3 stress and with highest number of genes shared between the single and sequential herbivore treatments (Fig.15d,e).

Genes involved in light harvesting and carbon assimilation were down- regulated in all scenarios but particularly after initial O3 stress (Fig.15d,f), which also up-regulated the expression of the non-photochemical quenching gene NPQ1 (Fig.15f). Several primary metabolic processes - e.g. amino acid and carbohydrate metabolism, showed opposite patterns of regulation between O3 and herbivory (Fig.15g). In general, response to herbivory dominated the sequential treatment inducing the expression of genes involved in plant defences - e.g. the transcription factor MYC2 (Lortzing & Steppuhn, 2016), lipoxygenases LOX2 and LOX3 (Felton et al., 1994; Halitschke & Baldwin, 2003), the trypsin inhibitor WSCP (Zavala & Baldwin, 2004; Boex-Fontvieille et al., 2015), and mitogen-activated protein kinase MPK3 (Pitzschke & Hirt, 2009) (Fig.15h). Genes involved in phytohormone signalling and biotic stress responses were also differentially regulated by O3 and herbivory (e.g. WRKYs, MYC2, and ERF2), but only O3 was associated to abiotic stress responses, such as drought (MYB44) (Jaradat et al., 2013; Li et al., 2014), senescence (EIN3) (Li et al., 2013), and phosphate starvation (RNS1, PT2/PHT1.4) (Fig.15h,i).

Figure 15. O3 and herbivory responses suppress energy metabolism. (a) HCA and heat-map for transcript changes compared to control levels (970 genes, P <0.05), and gene clusters colored and ranked (I–III) according to (b) GO processes relative abundance. (c) Heat map of Pearson´s correlation coefficient (ρ) calculated for paired samples comparisons from the HCA (darker blue = stronger correlation). (d) MapMan pathway analysis for changes in expression of GO biological functions and (e) Venn diagrams comparing up- and down-regulated genes, for log2 fold- change thresholds > 0.585 (P < 0.05). Color-codes for gene expression: up- (blue) and down-regulation (red). Treatments: O3 stress (O), herbivory by P. brassicae (P), sequential treatment (OP). (f-i) Central metabolism and stress response. Color- codes: O3 (blue), herbivory (pink), and sequential treatment (green). Student´s t- tests: * = P < 0.05; ** = P < 0.01; ***, P < 0.001; **** = P < 0.0001. Error bars=S.E.

- Network analysis and guilt-by-association principle allow for data-driven generation of new hypotheses and for the prediction of unknown gene or metabolite functions (Wei et al., 2006; Yonekura-Sakakibara et al., 2008; Mutwil et al., 2011). In this study, 156 metabolites and 970 gene expressions were correlated using Cytoscape producing a “scale-free” (non-random) network (Fig.16a) dominated by “hubs” (high-degree nodes) (Table 3).

Primary metabolites clustered in the central region of the network, while secondary metabolites clustered in modules at the periphery, connecting to genes involved in herbivore-induced responses, VOC emissions, and GSLs biosynthesis (e.g. CYP71,WRKY4, WRKY46) (Bennett et al., 1993; Schön et al., 2013; Irmisch et al., 2014; Mirabella et al., 2015) (Fig. 16b). At the centre, glycerol connected to the herbivore response variable (P) in a module with two flavin monooxygenases NOGC1 and FMO, and positively correlated (ρ = 0.68, P < 0.01) with the transcription factor MYB44, involved in abiotic responses such as drought and osmotic stress (Jaradat et al., 2013) (Fig. 16c). Notably, NOGC1 is a nitric oxide-dependent guanylate cyclase involved in regulation of stomatal closure and osmotic stress response via Ca++ signalling (Mulaudzi et al., 2011; Joudoi et al., 2013). In our network, MYB44 and NOGC1 connected to several genes involved in induced stress responses, carbohydrate metabolism, lipid transport, and energy metabolism (including chloroplast and mitochondria) (Table 3, and Tab.S3, Paper II).

Table 3. Major network hubs for B. nigra omics responses to O3 and herbivory. AGIa Gene name GO Processb,c At4g27740 - Yippee family putative zinc-binding Nucleus (unknown) At4g16590 CSLA1 Cellulose synthase-like A01 Glycosyltransferase At5g04440 - Unknown function (DUF1997) Chloroplast (unknown) At1g12390 - Cornichon protein (Guard cells) Signal transduction At2g29100 GLR2.9 Glutamate receptor 2.9 Ca++ homeostasis At3g50830 COR413 Cold acclimation WCOR413-like Membrane (unknown) At1g29910 LHCB1.2 Light harvest chlorophyll binding 1.2 Chloroplast/Photosynth. At2g22900 MUCI10 Mucilage-related 10 Galactosyltransferase At2g32990 GH9B8 Glycosyl hydrolase 9B8 Cellulose biosynthesis At3g09360 - TBP-binding protein RNA polymerase II At1g62580 NOGC1 NO-dependent guanylate cyclase 1 Stomatal closure At3g50770 CML41 Calmodulin-like 41 Chloroplast /Ca++ signalling At5g58270 M3 ABC transporter mitochondrion 3 Mitochondrion At3g24100 - SERF (uncharacterized) Nucleus (unknown) At1g07040 - Unknown protein Chloroplast (unknown) At2g02390 GST18 Glutathione S-transferase 18 Amino acid biosynthesis

a Arabidopsis Genome Initiative Number (as reported in TAIR). bGO biological processes and/or molecular functions (as reported in TAIR). c Cellular localization confirmed with the visualization tool ePlant (BAR; University of Toronto).

Figure 16. Centrality of plant responses to abiotic and biotic stresses. Metabolomics and transcriptomics profiles of B. nigra multiple responses to O3 and herbivory integrated in a co-expression network (Cytoscape). Edges connect nodes linked by Pearson´s correlation (ρ values of 0.85 or greater); positive and negative correlations represented by blue and red lines, respectively. Nodes represent: genes (white diamonds/squares, ◊/□) and metabolites (colored circles), and herbivory (P). (a) The network of gene-to-gene, gene-to-metabolite, and metabolite-to-metabolite correlations. (b) Modules of secondary metabolites (GSLs, hydroxycinnamates, flavonol glucosides and VOCs) connecting to central metabolism via stress response genes and transcription factors (e.g. WRK40, WRK46, and CYP707A). (c) Expansion of the gene-to-metabolite module connecting glycerol to the herbivore response (P) via genes coding two flavin monoxygenases (NOGC1 and FMO), a receptor kinase (LRR-RK), and a membrane transport protein (SecY protein). This module was connected via NOGC1 to the abiotic stress response transcription factor (MYB44).

Using GO enrichment pathway and network analyses (KEGG/KaPPA-View4, AraNet, GeneMANIA) combined with MVA of gene expression profiles we found that stress treatments induced a differential regulation of energy and glycerol metabolism during shift in responses from O3 stress to herbivory. Particularly, GO network analysis highlighted a functional link between the succinate dehydrogenase subunit (SDH1-1) of the mitochondrial ETC, and the mitochondrial glycerol-3-phosphate (G3P) shuttle (GPDHc1 and SDP6). These genes were all differentially expressed during the shift from O3 stress to herbivory. Moreover, GO network analyses suggested functional links of these central metabolic processes with other stress adaptation mechanisms, such as photosystem stabilization (DGD1 and SQD2), phosphate transport (PHT2;1), water-glycerol transport (aquaglyceroporins), and regulation of stomatal closure (NOGC1) (see Fig.5-7 and Tab.2 in Paper II).

3.1.3 Physiological responses to O3 and herbivory

On the basis of these network-omics results, we decided to test the effects of O3 and herbivory on processes of photosynthesis, stomatal opening, and leaf transpiration, in order to assess the stress impact on the plant physiology.

We measured photosystem capacity and leaf transpiration (using a LI-COR analyzer) during the same stress treatments and an additional long-term exposure to elevated O3 (70 ppb, 16 days) (Fig.17). Only few plants showed visible symptoms of senescence and chlorosis after five days of O3, although chlorophyll content in leaves decreased 9.5% compared to controls (P<0.05). Even before symptoms of chlorosis were visible in leaves, the senescence process was initiated with the down-regulation of photosynthetic genes (light-harvesting LHCs) and the induction of EIN3 (Fig.15), a transcription factor involved in regulation of leaf senescence (Li et al., 2013 Potuschak et al., 2003). Consistently, after five days of exposure, O3 stress resulted in reduced photosynthetic activity and in lower intracellular CO2 levels, while it negatively affected stomatal conductance and leaf transpiration (Table 4).

The phytotoxic effects of O3 were evident in plants exposed to the long-term treatment, which showed severe symptoms of leaf senescence and chlorosis (Fig.17a) with chlorophyll levels 47.8% lower than control plants of the same age (P < 0.001, Fig.17b). Leaf stomatal conductance was also further reduced by prolonged exposure to O3 stress, although there was an increase in the intracellular levels of CO2, possibly due to the very low photosynthetic activity (Table 4).

O3 can negatively affect photosynthesis (Salvatori et al., 2015; Vainonen & Kangasjarvi, 2015) causing the degradation of chlorophyll and thylakoid membranes in chloroplasts (Goumenaki et al., 2010; Tiwari et al., 2016). The decreased stomatal conductance and low transpiration observed in our study were likely the result of a protection mechanism which (via ROS signalling) prevents O3 from entering the leaf (Castagna & Ranieri, 2009; Vahisalu et al., 2010), but that also limits the leaf uptake of CO2 (Table 4).

Figure 17. O3 and herbivory affect photosynthesis and stomatal closure. (a) Leaf phenotypes of B. nigra exposed to O3 at 70 ppb for five days (O), herbivory by caterpillars for 24 hours (P), sequential O3 stress and herbivory (OP), 70 ppb O3 for 16 days (OL), and controls (C). (b) Chlorophyll content in leaves, Student’s t-test values: C and O (P < 0.05; N = 20), P and OP (P < 0.05; N = 10), and C and OL (P < 0.001; N = 10). (c-e) Stomatal conductance (mmol H2O m-2 s-1) measured via steady- state porometry (three experiments on separate days; N = 10 per treatment per day). (f) Stomatal conductance differences between treatments (one-tailed Student’s t-test for mean differences larger than zero: C versus O (P = 0.07), C versus P (P<0.05), and O versus OP (P < 0.01). Error bars indicate S.E.

Herbivory did not affect photosynthesis, although it slightly decreased both stomatal conductance and leaf transpiration. However, when plants exposed to O3 were later treated with herbivory, photosynthetic activity, stomatal conductance, and leaf transpiration were all reactivated, showing a reversed effect of the sequential stress treatment compared to responses induced by single stresses (Table 4). This effect was further confirmed measuring leaf conductivity by steady-state porometry, thus accounting for the stomata circadian rhythms (Fig.17c-f). Overall, physiological responses mirrored the previous patterns of metabolomics and transcriptomics (Fig. 14-15) where the initial response to O3 was counteracted by herbivory, and the phenotype during the sequential stress resembled the one induced by herbivory alone.

Based on both omics and physiological data, we proposed a model for stress adaptation mechanisms in B. nigra (Fig.18) where both glycerol and energy metabolism (chloroplasts and mitochondria) play central roles in protecting against oxidative stress and facilitating stomatal osmoregulation. Glycerol osmolytic properties can enhance the plant resistance against osmotic stress (Eastmond, 2004), and glycerolipid transfer is important for plant tolerance to different abiotic stresses (Li et al., 2015; Mueller et al., 2015), including O3 stress (Sakaki et al., 1990). Glycerol accumulation in our study possibly derives from the peroxidation of lipids in chloroplast membranes (Bienert et al., 2006), as supported by the activity of chloroplastic lipid transfer proteins (LTPc1 and LTPc2) (Xu et al., 2008) which were correlated with O3 stress and with induction of MYB44 (Fig.S3-S4 in Paper II). MYB44 is known to regulate abiotic and biotic stress responses (Baldoni et al., 2015), including stomatal opening and leaf senescence (Jaradat et al., 2013; Li et al., 2014).

Table 4. Opposite effects of O3 and sequential herbivory on photosynthesis. Treatments: C O (vs. C) P (vs. C) OP (vs. O) OL (vs. C)

Photosynthesis 11.72 ± 0.8 8.63 ± 0.63 11.5 ± 0.58 11.24 ± 0.85 3.29 ± 0.26 (µmol CO2 m-2 s-1) ** n.s. * *** Stomatal conductance 0.31 ± 0.02 0.15 ± 0.04 0.25 ± 0.10 0.24 ± 0.03 0.12 ± 0.01 (mol H2O m-2 s-1) ** n.s. * ** Leaf transpiration 3.45 ± 0.22 1.84 ± 0.22 2.97 ± 0.90 2.66 ± 0.26 1.55 ± 0.19 (mol H2O m-2 s-1) *** n.s. * ***

Intracellular CO2 concentration 300 ± 3.87 264 ± 4.93 291 ± 9.35 288 ± 9.81 333 ± 10.19 (ppm) * n.s. n.s. *** Leaf photosynthesis and transpiration measured using a LI-COR analyzer (LI-6400). Means ± SE (n = 6-10) and variation calculated to relative group comparisons (in brackets). Student´s t-test, P-values > 0.05 (*), > 0.01 (**), > 0.001 (***). Treatments: controls (C), exposure to 70 ppb O3 for five days (O), P. brassicae herbivory for 24 hours (P), sequential 70 ppb O3 and herbivory (OP), and additional long term exposure to 70 ppb O3 for 16 days (OL).

Feed-back regulation between MYB44 and NOGC1 may help plant responses to control stomatal conductance and balance CO2 uptake during O3 stress (Joudoi et al., 2013). Altogether, I suggest that combined regulation of glycerol metabolism (including the G3P shuttle, GPDHc1/SDP6) and energy metabolism via suppression of photosynthesis (Shen et al., 2006; Quettier et al., 2008) and increased mitochondrial activity (SDH1-1, MSD1) (Huang et al., 2013; Martin et al., 2013) may help the plant to dissipate excess energy and prevent formation of oxidative radicals (Dizengremel et al., 2009; Blanco et. al, 2014; Tiwari et al., 2016). The redirection of these pathways during sequential herbivory suggests an active plant response via WRKYs, MYC2, and ERF2 (Huot et al., 2014), but it may also represents a herbivore manipulation of the host’s metabolism which interferes with induced defence and water stress responses (Reymond et al., 2000; Consales et al., 2012).

Figure 18. Plant central metabolic responses towards O3 and herbivory. Omics data linked to physiological responses propose stress adaptation mechanisms (1–5). Blue, O3 stress (five days, 70 ppb); red, sequential herbivory by P. brassicae (24 hours). Up/down arrows = up/down regulation of genes, or increase/decrease in metabolites. (1-2) O3 induces the abiotic stress responses of senescence (EIN3 and ERF2) and stomatal closure (MYB44), with feedback on NOGC1, and possibly FMO. ABA and JA/ET cross-talk integrates responses between O3 and sequential herbivory (MYC2 and ERF2), with opposite effects on stomatal closure. (3-4) Photosystem suppression (LHCs) and non-photochemical quenching (NPQ1) in response to O3 are linked to the regulation of glycerol metabolism (LPTs, LIP1, and SQD2). Glycerol derived from degraded chloroplast membranes enters the G3P shuttle (GPDHc1/SDP6) to sustain NAD+ recycling and mitochondrial activity (SDH1-1, MSD1, and GABA) as antioxidative stress mechanisms. A role of glycerol and sugars as osmolytes also is suggested. Herbivory interacts with glycerol pathways for alternative source-sink priorities and induced JA responses (MYC2, ERF2, LOXs). (5) GABA roles in plant stress adaptation: Ca++ homeostasis, carbon-nitrogen metabolism, senescence, ROS scavenging, and plant-insect interactions.

3.2 Dual herbivore attack

In Paper III we investigated the metabolic responses of B. nigra versus two different kinds of herbivory: the chewing caterpillars of Pieris brassicae and the piercing-sucking aphids of Brevicoryne brassicae. Herbivore-induced responses in plants were evaluated for effects of single and dual herbivory on both local and systemic leaves, while accounting for aphid-density effects. Aphids (nymphs, first and second instar) were let feed on the youngest fully expanded leaf of B. nigra for three days, while caterpillars (30, first instar) fed on the same leaf, alone or with aphids, for 24 hours. In total, plants were exposed to five treatments: control undamaged plants (C), herbivory by 50 or 100 aphids alone (50A, 100A), herbivory by caterpillars alone (P), or dual herbivory by 50 or 100 aphids plus caterpillars (50A+P, 100A+P).

Metabolomics was targeted for primary and secondary metabolism using GC-TOF-MS and LC-TOF-MS, respectively. In total, 221 metabolic features were detected, of which 110 were fully identified between GSLs, phenolics (hydroxycinnamates and flavonols), amino acids, sugars, and organic acids. Here I will only focus on the herbivore responses induced in locally damaged leaves, which always showed stronger effects compared to systemic leaves.

3.2.1 Plant responses versus aphids and caterpillars

Changes in the plant metabolic responses differed according to the kind of herbivory, as shown by the general MVA model for the effects of caterpillar and aphid herbivory in all treatments (C, P, 50A, 100A, 50A+P, 100A+P) (Fig. 19a). The strongest plant responses were observed for dual herbivory by caterpillars and aphids during high-aphid density (100A+P) (Fig. 19a). A further MVA classification based on either presence or absence of aphids or caterpillars (i.e. not including the aphid-density variable) showed a clear difference between the two types of herbivore responses (Fig. 19b). In this model, plants infested with only aphids were separated from plants with caterpillars alone or during dual herbivory. Together, these results indicated a dominant effect of caterpillars over the effect of aphids during the plant response to dual herbivory (Fig.19a,b).

Figure 19. Plant responses versus caterpillar and aphid herbivory. (p.39) - MVA models describing the effects of caterpillars and aphids on B. nigra primary and secondary metabolism. (a) Score plot (OPLS-DA 3+2+0; R2Xcum=57%, R2Ycum=52%, Q2cum=21%). Treatments abbreviations: controls (C), plants infested with B. brassicae 50 aphids (50A) or 100 aphids (100A), single herbivory by P. brassicae caterpillars (P), dual herbivory with 50 aphids and caterpillars (50A+P), or with 100 aphids and caterpillars (100A+P).

(a)

OPLS-DA Low-aphids

%) Caterpillars LV2 (5 LV2

Controls High-aphids

LV1 (12%)

(b)

OPLS-DA Aphids + caterpillars

%) AphidsAphids CaterpillarsCaterpillars LV2 (10 LV2

LV1 (6%)

(c)

(continued) (b) Score plot, all plants grouped for herbivore treatments by only aphids, only caterpillars, or dual herbivory (OPLS-DA 1+1+0; R2Xcum=12%, R2Ycum=49%, Q2cum=20%). (c) Loading plot for metabolite contribution (only identified metabolites). VIP > 1.00: (1) fructose, (2) fructose/sorbose, (3) maltose, (4) ribitol, (5) glucose, (6) trehalose, (7) sorbose, (8) p-coumaric-acid, (9) GSH, (10) α-tocopherol, (11) glucoiberin, (12) gluconapin, (13) glucobrassicin, (14) neoglucobrassicin, (15) sinigrin, (16) gluconasturtin, and (17) Quercetin-3- coumaroylsophoroside-7-glucoside. After Fig.1 and 2 in Paper III.

Caterpillars primarily characterized an increase in GSLs (ANOVA, P <0.001) (Fig.19c, Fig.20a), with an increase in sinigrin from 3.8 to 5.9 µmol/g FW (+55%) (ANOVA, P < 0.01). Overall, aphid-density always affected the plant response during caterpillar and dual treatments, but while low-aphid density (50A, 50A+P) increased GSL levels similarly to herbivory by caterpillars (P), high-aphid density did not further induced GSLs levels in the presence of caterpillars (100A+P) (Fig.20a).

Aphid herbivory induced higher levels of carbohydrates (i.e. trehalose, fructose, maltose, ribitol, glucose and sorbose) as well as other metabolites, e.g. glutathione (GSH), α-tocopherol, p-coumaric-acid, and quercetin-3- coumaroylsophoroside-7-glucoside (Fig.19c, VIPs > 1.00). Trehalose was a particularly important metabolite. It was never induced in response to caterpillars alone, while it increased during single (50A, 100A) and dual (50A+P, 100A+P) aphid feeding (ANOVA, P < 0.001) (Fig. 20b). Trehalose also correlated to aphid-density (ρ = 0.64, P < 0.001), with the highest levels observed after high-aphid infestations (100A, 100A+P) (Fig. 20b).

Figure 20. Dual herbivory and aphid-density affect GSLs and trehalose. (a) Total GSLs levels and (b) trehalose levels, detected in B. nigra leaves damaged by combination of single and dual herbivory treatments. Relative abundance based on integrated chromatogram peak areas of GSLs (LC-TOF-MS), sinigrin, gluconapin, glucoerucin, glucobrassicin, neoglucobrassicin, 4-methoxyglucobrassicin, glucoiberin, glucojiaputin, glucobrassicanapin, gluconasturtiin, 4-hydroxy-glucobrassicin) and of trehalose (GC-TOF-MS). Mean ± S.E., n = 6. ANOVA with Tuckey test comparisons. Treatments abbreviations: controls (C), plants infested with 50 aphids of B. brassicae (50A) or 100 aphids (100A), single herbivory by P. brassicae caterpillars (P), and dual herbivory with 50 aphids+caterpillars (50A+P), or 100 aphids+caterpillars (100A+P). After Fig.5 in Paper III.

3.2.2 Effects of dual herbivory and aphid density

To distinguish between the effects of dual herbivory and aphid density, we compared the treatment using pair-wise MVA models (Table 5 and Fig.S1 in Paper III). The strongest effects were observed for the increase from low- to high-aphid density (i.e. from 50 to 100 aphids) (model 1), and for the shift from single herbivory by either the caterpillars (P) or the aphids (50A, 100A), toward dual herbivory (respectively model 4, 5, and 6) (Table 5). In addition, significant effects of aphid-density were also found in systemic leaves which were not directly damaged by herbivores (model 8 and model 9, Table 5).

In keeping with the results from our previous multiple stress studies, it was rarely an individual metabolite that contributed to the pair-wise differences across treatments, and changes in both primary and secondary metabolites characterized the plant response. For instance, when comparing herbivory by aphids at different density levels, the addition of caterpillars resulted in changes of sugars and phenolics in plants treated at low-aphid density - i.e. 50A versus 50A+P (model 5), whereas it mostly affected GSLs and several other unknown metabolites in plants treated at high-aphid density – i.e. 100A versus 100A+P (model 6) (see Fig.S1 in Paper III for score and loading plots of MVA models listed in Table 5).

Table 5. Effects of dual herbivory and aphid-density on the plant metabolism. Model parameters R2X R2Y Q2 Treatment effects Comparisons LVs (cum) (cum) (cum) 1 Aphid-density + Aphid (50) 50A 100A 1+3+0 62% 99% 51% 2 Aphid-density + Aphid (50) 50A+P 100A+P 1+0+0 31% 58% 19% 3 Dual herbivory + Aphid (50) P 50A+P n.s. n.s. n.s. n.s. 4 Dual herbivory + Aphid (100) P 100A+P 1+3+0 68% 99% 14% 5 Dual herbivory + Caterpillar 50A 50A+P 1+3+0 52% 98% 36% 6 Dual herbivory + Caterpillar 100A 100A+P 1+2+0 52% 98% 39% 7 Systemic aphid + Aphid (50) 50A 100A n.s. n.s. n.s. n.s. 8 Systemic aphid + Aphid (50) 50A+P 100A+P 1+3+0 65% 99% 38% 9 Systemic aphid + Aphid (50) P 50A+P 1+1+0 69% 99% 56%

Pair-wise comparisons targeting for effects of dual herbivory and aphid-density on B. nigra metabolome, MVA models (OPLSDA). See relative score and loading plots for each model in Fig. S1, Paper III. Orthogonal variation was standardized at maximum three latent variables (LVs). R2X (cum) = cumulative metabolic variation (x) explained by the MVA model; R2Y (cum) = cumulative treatment effect explained by the MVA model (y); Q2 (cum) = cumulative predictive capacity of the MVA model. Models (1-6) refer to effects on locally damaged leaves, with strongest models highlighted in blue. Models (7-9) refer to effects on systemic leaves, and discussed in Paper III.

We compared responses to dual herbivory at high-aphid density (100A+P) with the two respective single herbivore treatments (100A and P) (Fig.21). This model showed a clear separation between single and dual herbivory (Fig.21a) with a main increase of GSLs in response to caterpillar herbivory (P, 100A+P) and only an increase in 4-methoxy-glucobrassicin associated to dual herbivory (Fig.21b). Several metabolites characterized the response to high-aphid density (100A) which was consistent during dual herbivory (100A+P), with increase in glycolic acid, phosphoric acid, glutathione (GSH), and the sugars fructose and trehalose (Fig.21b). These results showed how both aphids and caterpillars have strong effects on the plant metabolome, but also how responses during dual herbivory are affected at high-aphid density (Fig.21; see also pair-wise model 4 and 6 in Table 5).

(a) Aphids+ caterpillars OPLS-DA

%) High-aphids Caterpillars LV2 (8 LV2

LV1 (6%)

(b)

Figure 21. Responses to caterpillar are shaped by high-aphid density. Combined effects on B. nigra primary and secondary metabolism (OPLS-DA 2+3+0; R2Xcum = 64%, R2Ycum = 95%, Q2cum = 40%). (a) Score plot, with plants infested with P. brassicae caterpillars (P), B. brassicae 100 aphids (100A), or dual herbivory (100A+P). (b) Loading plot for metabolite contribution (only identified metabolites). VIPs > 1.00: (1) glucoiberin, (2) neoglucobrassicin, (3) glucobrassicin, (4) sinigrin, (5) gluconapin, (6) gluconasturtin, (7) 4-hydroxy-glucobrassicin, (8) 4-methoxy- glucobrassicin, (9) glycolic acid, (10) phosphoric acid, (11) glutathione GSH, (12) fructose, and (13) trehalose. After Fig.3 in Paper III.

To reveal other potential aphid-density effects on single and dual herbivory, we created an MVA model with no single caterpillar herbivory (P) (Fig.22). This model showed the effect of aphids on dual herbivory by separating the treatments first for aphid-density (LV1) and then for presence or absence of caterpillars (LV2) (Fig.22a). These results confirmed an effect of aphid- density on the plant response to caterpillar herbivory. Once again, the major effect of caterpillar herbivory was found on GSLs, while especially single aphid herbivory led to increased levels of sugars (Fig.22b). Specific aphid- density effects were observed during high-aphid treatments, which led to increased levels of sugars, i.e. trehalose, ribose, ribitol, xylose, and maltose, whereas low-aphid treatments increased the levels of fructose (Fig.22b).

(a)

OPLS-DA High-aphids

%) Low-aphids LV2 (19 LV2

Caterpillars

LV1 (12%)

(b)

Figure 22. High- and low-aphid density affect response to dual herbivory. Effects of high- and low-aphid density during single and dual herbivory on B. nigra primary and secondary metabolism (OPLS-DA 1+4+0; R2Xcum = 52%, R2Ycum = 75%, Q2cum = 15%). (a) Score plot, plants infested with only 50 (50A) or 100 (100A) of B. brassicae, or in combination with P. brassicae caterpillars (50A+P or 100A+P). (b) Loading plot (only identified metabolites). VIP > 1.00: (1) glucoiberin, (2) sinigrin, (3) gluconapin, (4) glucobrassicin, (5) neoglucobrassicin, (6) 4-hydroxyglucobrassicin, (7) gluconasturtin, (8) ribose, (9) ribitol, (10) trehalose, (11) maltose/cellobiose (12) xylose, (13) fructose, (14) cis,cis-linoleic acid, (15) α-linolenic acid, and (16) glycolic acid. After Fig.4 in Paper III.

Chewing caterpillars induced the strongest metabolic responses in plants, especially with increased levels of GSLs; on the other hand, sucking aphids resulted in density-dependent effects which differentially induced the plant metabolic profile during both single and dual herbivory. Dual caterpillar and aphid herbivory can affect GSLs levels through different processes, such as:

1) de novo GSL biosynthesis as herbivore-induced response during defence against chewing caterpillars (Hopkins et al., 2009);

2) degradation of GSLs into VOCs after leaf damage by caterpillars, which can affect herbivory as an indirect defence system (Ponzio et al., 2014);

3) manipulation of plant JA-induced responses by piercing-sucking aphids (Broekgaarden et al., 2008; Walling, 2008; Thaler et al., 2012)

4) increase or decrease in GSLs after aphid herbivory (Mewis et al., 2006; Kim & Jander, 2007; Khan et al., 2011);

5) sequestration of GSLs by the specialist aphid of Brassicacea, B. brassicae (Jones et al., 2001; Kazana et al., 2007; Broekgaarden et al., 2008).

Interaction between P. brassicae and B. brassicae on cultivated B. oleracea was previously shown to have no effects on GSLs, but SA-mediated defences induced by aphids interfered with JA-mediated defences, facilitating the development of caterpillars (Soler et al., 2012). Parasitoids also performed better on both hosts during dual herbivory compared to single herbivory, suggesting a higher level of complexity in the system that involves VOCs and indirect defences, as confirmed by other studies in B. nigra (Ponzio et al., 2014; Ponzio et al., 2016).

In our study, higher aphid-density increasingly induced levels of trehalose and other sugars (19-22), although only in locally damaged leaves. Sugar levels may have been concentrated by the aphid sucking activity, or alternatively been derived from honeydew excretions on leaves (Lamb, 1959; Yao & Akimoto, 2001). On the other hand, changes in primary metabolites can also represent a strategic re-allocation of plant resources or be directly involved in defence (Schwachtje & Baldwin, 2008; Steinbrenner et al., 2011). Similarly to our study, trehalose is induced in Arabidopsis during herbivory by the generalist aphid Myzus persicae (Singh et al., 2011; Hodge et al., 2013). Trehalose may thus be directly induced by plants to respond to aphid infestation, as suggested by studies showing its role in protection against biotic and abiotic stresses (Fernandez et al., 2010; Singh & Shah, 2012), where it regulates sugar signalling, carbon partitioning, and accumulation of starch (Singh et al., 2011; Nuccio et al., 2015; Foyer et al., 2016).

3.3 Induced plant resistance to herbivory

In Paper IV we investigated the herbivore-induced responses of plants after phytohormone treatment with MeJA and sequential chewing by caterpillars. We tested whether MeJA may enhance plant resistance against herbivory, while we evaluated other potential effects on the plant metabolome. Finally, we studied how herbivore-induced responses were shaped by leaf ontogeny.

In all experiments, plants were sprayed only once with MeJA (1mM, 5mL) and let respond for 72 hours. Herbivore treatments consisted in caterpillars feeding on leaves for five consecutive days. The four treatments included: single MeJA or herbivory, sequential treatment, and controls (Fig. 23a).

Figure 23. Plant induced treatments with MeJA and caterpillar herbivory. (a) Experimental set-up. Treatment abbreviations: controls undamaged plants (C), plants sprayed with MeJA (1mM, 5 mL) and induction for 72 hours (M), herbivory by 15 first instar caterpillars of P. brassicae feeding for five days (P), and sequential MeJA and herbivory (MP). (b) B. nigra phenotype across leaf ontogeny and effects of MeJA treatment. P. brassicae caterpillars were placed on the youngest fully developed leaves, usually corresponding to L4-L5. (c) Effects of MeJA on stem pigmentation, between 24-72 hours. (d) Effect of MeJA induced responses on caterpillars´ weight after five days of feeding (Student´s t-test, p-value < 0.05).

3.3.1 Effects of MeJA on herbivore-induced responses

Symptoms of early leaf senescence appeared in plants treated with MeJA already after three days, most commonly in lower leaves (L6-L10) (Fig.23b). These plants also displayed a dark, purple pigmentation along their stems (Fig.23c). After five days of herbivory treatment, MeJA-treated plants showed enhanced resistance against P. brassicae caterpillars, which gained less weight than those feeding on control plants (-27%) (P<0.05) (Fig.23d).

From plants treated with herbivory (P, MP), damaged leaves were sampled and compared to same leaves from undamaged plants (C, M). A supervised MVA model of the plant primary and secondary metabolism (analysed using GC-MS and LC-MS) showed distinct metabolic shifts from steady-state levels of control plants (C) after each one of the treatments (M, P, MP) (Fig.24). Induced responses were distributed along the first latent variable (LV1) and described similar magnitudes between single MeJA (M) and P. brassicae (P), with an increased response during the sequential treatment (MP). However, the effects of MeJA (M, MP) differed from real caterpillar herbivory (P, MP) and were described by a shift along the second latent variable (LV2) (Fig.24; see also Fig.2b, in Paper IV for loading plot of metabolite contribution).

LV2 (9 LV2 Pieris caterpillars Controls

MeJA

LV1 (14%)

Figure 24. Plant herbivore-induced responses are shifted by MeJA. MVA model for the effects of MeJA and herbivory treatments on primary and secondary metabolism (PLS-DA model, four components; R2Xcum = 42%, R2Ycum = 75%, Q2cum = 56%). (a) Score plot. Treatment abbreviations: controls (C), MeJA (M), P. brassicae herbivory (P), and sequential MeJA and herbivory (MP). (b) Loading plot with metabolites contribution (VIP > 1.00). After Fig.2 in Paper IV.

GSLs were induced to the same extend by MeJA and herbivory, although the effect of MeJA was more uniform with less variation (Tab.1 in Paper VI). In the sequential treatment, initial MeJA application increased the following induction of GSLs by caterpillar herbivory (Fig.25, Table 6). Particularly, sinigrin increased from ca. 3 to 5 µmol/g FW after either single MeJA or herbivory, to 9 µmol/g FW after sequential treatment (P > 0.001) (Fig.26a).

Indolic and aromatic GSLs such as neoglucobrassicin and glucotropaeolin were largely induced after the MeJA treatments, but not particularly after herbivory. Phenylpropanoids were also differentially affected by the MeJA treatment, with a decrease of hydroxycinnamic acid derivatives (e.g. 1- caffeoyl-β-glucose, sinapic acid, and 1-O-sinapoyl-glucose) (Milkowski & Strack, 2010) and of flavonol glucosides (Table 6). On the contrary, several phenylpropanoids and flavonols were found to increase in response to real herbivory (e.g. sinapic acid, disinapoyl-gentiobiose, and quercetin-3- sinapoylsophoroside-7-glucoside) (Fig.26b and Table 6). Sequential MeJA and herbivory also affected the levels of the two unidentified metabolites previously described in Paper I-II (features “381” and “421”), which were important for defining biotic and abiotic stress interactions in B. nigra (Khaling et al., 2015; Papazian et al., 2016).

At the level of primary metabolism, MeJA and herbivory had different effects on sugars, amino acids, and organic acids (Tab.1 in Paper VI). MeJA resulted in a stronger decrease of fructose, an increase in both myo-inositol and maltotriose, but did not affect the levels of gentiobiose, instead induced by caterpillar chewing (Fig. 25b, 26c-e and Table 6). Herbivory also negatively affected the level of almost all amino acids, while MeJA only affected tryptophan (indolic GSLs precursor; Halkier & Gershenzon, 2006) and had no effect on phenylalanine (aromatic GSLs precursor) (Fig.26f,g).

Both MeJA treatment and herbivory had similar effects on the levels of several TCA cycle intermediates, resulting in an overall increase of citric acid and α-ketoglutaric acid (i.e. C5-C6 metabolites of the first half-cycle) and a decrease in succinic acid, fumaric acid and malic acid (i.e. C4 metabolites of the second half-cycle) (Fig.26h-j, see also Tab.1 in Paper IV). In addition, during sequential herbivory, MeJA increased the levels of cis-aconitic acid

(C6), while it further decreased the levels of succinic acid and malic acid (C4) (Fig. 25b, and Table 6).

These cumulative effects of MeJA on primary and secondary metabolism may have enhanced the plant resistance against herbivory, resulting in the decreased growth of young P. brassicae caterpillars (Fig.23d).

Caterpillars

Caterpillars + MeJA

Figure 25. MeJA enhances plant responses against caterpillars. Multivariate model for the effects of MeJA treatment on B.nigra metabolic response towards herbivory, relative to primary and secondary metabolism (OPLS-DA 1+1+0; R2Xcum = 28%, R2Ycum = 86%, Q2cum = 70%). (a) Score plot, with plants treated with herbivory by P. brassicae caterpillars (P) compared to same treatment on plant previously exposed to MeJA (MP). (b) Loading plot with metabolites (VIPs > 1.00).

Figure 26. Induced responses affect primary and secondary metabolism. Effects of single and sequential MeJA and herbivory on indivividual metabolites. MeJA- and herbivory- induced changes in abundances of defence-related secondary metabolites such as (a) sinigrin and (b) quercetin glucoside, and of growth-related primary metabolites, such as (c) fructose, (d) gentiobiose (or similar disaccharide), (e) maltotriose (trisaccharide), amino acids (f) phenylalanine and (g) tryprophan, and TCA cycle intermediates (h) citric acid, (i) α-ketoglutarate, and (j) fumaric acid. Sinigrin concentrations (µmol/g FW) were calculated comparing a calabiration curve of pure sinigrin standard. Abundance of other metabolites are reported as relative values from integrated chromatogram peak areas. Metabolites are colored according to the code system used in Fig.25. Significant changes in metabolite concentrations compared to control steady-state levels are reported as Student´s t-test, P-values < 0.05 (*), < 0.01 (**), and < 0.001 (***). Treatment abbreviations: controls (C), MeJA (M), P. brassicae herbivory (P), and sequential MeJA and herbivory (MP). See also Fig.2 and Tab.1 in Paper IV.

Table 6. MeJA strengthens herbivore-induced responses against caterpillars. Metabolite Class VIP (P vs. MP) Glucotropaeolin Glucosinolate 1.93 Gluconapin Glucosinolate 1.80 Maltotriose Sugar 1.78 ⬆ Sucrose Sugar 1.74 ⬆ Feruloylmalic acid Hydroxycinnamate 1.66 ⬆ cis-Aconitic acid TCA 1.63 ⬆ Flavanone (putative) Flavonol glucoside 1.60 ⬆ Sinigrin Glucosinolate 1.52 ⬆ Galactonic acid Organic acid 1.51 ⬆ 381 Unknown 1.51 ⬆ Neoglucobrassicin Glucosinolate 1.48 ⬆ Caffeoyl-glucoside (putative)(1) Hydroxycinnamate 1.39 ⬆ p-coumaroyl-D-glucose Hydroxycinnamate 1.37 ⬆ Glucoerucin Glucosinolate 1.32 ⬆ 3-methylbutyl-GSL Glucosinolate 1.22 ⬆ Caffeoyl-glucoside (putative (2) Hydroxycinnamate 1.17 ⬆ Disinapoylgentiobiose Hydroxycinnamate 1.15 ⬆ Inositol, myo- Sugar 1.14 ⬆ Dehydroascorbic acid Redox 1.09 ⬆ Sinalbin Glucosinolate 1.02 ⬆ -Linoleic acid Lipid 1.03 ⬆ Qn-3-sinapoylsophoroside-7-glucoside Flavonol glucoside 1.04 ⬇ Kmα 3-sinapoylsophoroside-7-glucoside Flavonol glucoside 1.04 ⬇ 4-hydroxyglucobrassicin Glucosinolate 1.04 ⬇ Succinic acid TCA 1.06 ⬇ Glucoraphanin Glucosinolate 1.06 ⬇ Km-3-hydroxyferuloylsophoroside-7-glucoside Flavonol glucoside 1.07 ⬇ Qn-glucoside (putative) Flavonol glucoside 1.07 ⬇ Sinapoylhydroxyferuloylgentiobiose Hydroxycinnamate 1.07 ⬇ Qn-3-sophorotrioside-7-glucoside Flavonol glucoside 1.11 ⬇ Trisingentiobiose Hydroxycinnamate 1.12 ⬇ Qn-7-sophoroside Flavonol glucoside 1.14 ⬇ Km 3-sinapoylsophoroside (putative) Flavonol glucoside 1.16 ⬇ Km-3-p-coumaroylsophoroside-7-glucoside Flavonol glucoside 1.16 ⬇ Qn-3-sinapoylsophoroside-7-glucoside (putative) Flavonol glucoside 1.17 ⬇ Km-3-hydroxyferuloyldiglucoside-7-glucoside Flavonol glucoside 1.17 ⬇ Malic acid TCA 1.21 ⬇ 421 Unknown 1.22 ⬇ Threonic acid Redox 1.28 ⬇ Km-glucoside (putative) Flavonol glucoside 1.32 ⬇ Tryptophan Amino acid 1.34 ⬇ 1-O-sinapoylglucose Hydroxycinnamate 1.47 ⬇ Fructose Sugar 1.53 ⬇ 1-caffeoyl- -D-glucose Hydroxycinnamate 1.64 ⬇ Glutathione (GSSG) Redox 1.66 ⬇ Oxalic acidβ Organic acid 1.67 ⬇ ⬇ Metabolite contribution to the OPLS-DA model (Fig.25), comparing the effect of MeJA treatment⬇ on B.nigra responses towards sequential herbivory by P. brassicae. Significant contribution to the model of each metabolite is reported for the respecitve variable importance in the projection (VIP) scores >1.00. Metabolites are ordered from higher to lower VIP values, for increased (blue ) and decreased (red ) abundance in plants pre-treated with MeJA (MP) compared to plants only treated with herbivory (P). See corresponding loading plot in Fig.25b. ⬆ ⬇

MeJA enhanced B. nigra defence responses and the production of defensive secondary metabolites, e.g. GLSs. At the same time, MeJA induced multiple effects in the plant chemistry and phenotype, such as early leaf senescence, purple stem coloration, differential accumulation of phenylpropanoids, and changes in primary metabolites, e.g. sugars and TCA cycle intermediates. MeJA application stimulates JA-responses via activity of MYC2, one of the main transcription factors involved in regulation of plant induced-defences (Lortzing & Steppuhn, 2016). Different phytohormones are involved in fine- tuning of MYC2 and JA-mediated defence responses (Erb et al., 2012). For instance, MYC2 interferes with ET responses via antagonistic interaction with EIN3 (ETHYLENE INSENSITIVE 3) (Song et al., 2014). Induced MYC2 activates regulation of GSL and phenylpropanoid metabolism (Dombrecht et al., 2007; Verhage et al., 2011; Schweizer et al., 2013), while it suppresses growth processes (Chen, et al., 2011; Huot et al., 2014; Campos et al., 2016). Thus, pleiotropic effects of MYC2 may explain the effects of MeJA on the phenotype of B. nigra which were observed in our study (Fig.23-26).

Using Arabidopsis mutant lines ein3-1 treated with MeJA, we tested the interaction of JA- and ET- responses, measuring MeJA-induced expressions of MYC2 and two downstream genes involved in defence against insects and pathogens, respectively controlled by MYC2 and EIN3: VSP2 (VEGETATIVE STORAGE PROTEIN 2) and PDF1.2 (PLANT DEFENSIN 1.2) (Manners et al., 1998; Liu et al., 2005; Verhage et al., 2011) (Fig.27a). This experiment confirmed JA and ET asymmetric cross-talk, with MeJA-induced expression of MYC2 and VSP2, but down-regulation of PDF1.2 in ein3-1 (Fig.27b).

Figure 27. JA/ET asymmetric regulation of herbivore-induced defences. (a) JA and ET antagonistic pathways involved in herbivore signalling and fine-tune of defence responses. (b) Relative expression levels (RTq-PCR) of MYC2, PDF1.2 and VSP2 in Arabidopsis, comparing wt (col-1) and ein3-1 mutants, after 72 hours induction with MeJA (1mM). Significance between the treatments was tested with ANOVA (N=3; Tuckey comparisons).

3.3.2 Ontogenic patterns of induced responses

To investigate herbivore-induced effects along the plant architecture, we examined the metabolic profiles of individual leaves (L1-L5), targeting for defence metabolites (Table 6) and following the responses throughout leaf ontogeny. As in the previous experiment, MeJA and herbivory induced a distinct shift from the steady-state metabolism of control plants, which increased during sequential treatments (PC2; Fig.28a). Herbivore-induced responses further increased towards younger leaves (PC1), and leaf ontogeny explained twice the total metabolic variation associated with the treatments (Fig.28b). Top young leaves (L1-L3) showed the strongest shifts after MeJA and herbivory, along with the highest levels of secondary metabolites, i.e. GSL, flavonols, and hydroxycinnamates (see Fig.4e,f in Paper IV).

We further tested the effect of MeJA on leaf ontogenic patterns targeting primary metabolism. Plants were divided into three ontogenetic groups of “top” (L1-L3), “mid” (L4-L6), and “low” (L7-L8) leaves, based on previous secondary metabolic profiles (Fig.4g and Fig.5a in Paper IV). The strongest responses were again observed in the youngest leaves (Fig.28c; also Fig.5 and Tab.3 in Paper IV), and largely reflected the effects of MeJA observed in our first experiment (Fig.24-26 and Table 6; also Tab.1 in Paper IV). MeJA affected primary metabolites directly involved in defence metabolism – including a decrease in α-linolenic acid (JA precursor) (Wasternack, 2015) and different effects on phenolics (caffeic acid, salicylic acid, and salicin) – and strongly induced changes in central carbon pathways, with increased levels of myo-inositol, maltose, and trehalose, and a decrease in fructose.

MeJA again increased citric acid and α-ketoglutaric acid levels (C5-C6), but also decreased those of fumaric acid (C4) (Fig.28c, and Tab.3 in Paper IV).

These shifts in primary metabolism suggests a reconfiguration of source- sink relationships, possibly as a strategy aimed at maximizing the plant fitness, balancing between plant growth and defence against insects (Bolton, 2009; Lortzing & Steppuhn, 2016). Regulation of TCA cycle integrates plant growth-defence processes, sustaining energy and photosynthetic demands (Noor et al., 2010; Sweetlove et al., 2010) and the biosynthesis of secondary metabolites. As carboxylic acids can fluctuate in response to environmental perturbations (Araujo et al., 2012), the asymmetric distribution of TCA intermediates observed in our study may correspond to an “open-flux” mode which regulates plant energy metabolism and redox balance (Gardeström et al., 2002; Igamberdiev & Eprintsev, 2016), as also suggested by the increase in the antioxidants metabolites ascorbic and dehydroascorbic acid (Noctor & Foyer, 1998; Smirnoff, 2011) (Fig.28c; and Tab.3 in Paper IV).

%) PC2 (12PC2

PC1 (25%)

LV1 (17%)

Figure 28. Leaf ontogeny shapes the plant herbivore-induced responses. Responses across leaf ontogeny relative to changes in secondary metabolism PCA model (five components; R2Xcum = 62%, Q2cum = 16%). Score plots colored according to (a) treatments and (b) leaf position (L1-L5). (c) Effect of MeJA on primary metabolism of young leaves. OPLS-DA loading plot (1+1+0; R2Xcum = 50%, R2Ycum = 88%, Q2cum = 75%), comparing shifts of primary metabolites in controls and MeJA-treated B. nigra for the “top” leaves (L1-L3). Plant leaves were divided in ontogenic groups, according to the patterns of secondary metabolism. See Fig.4g and Fig.5 in Paper IV. Treatments codes as in Fig.23-26.

Overall, we have shown how herbivore-induced responses regulate both primary and secondary metabolism, indicating an active reprogramming of growth-defence processes. Treatments with MeJA enhanced the plant resistance against caterpillars and resulted in increased accumulation of GSLs. The ontogenic patterns observed in leaves were also in agreement with the “optimal defence theory” (Mckey, 1974; Meldau et al., 2012), which predicts that plants prioritize defences in precious tissues with high fitness value, such as young leaves and reproductive organs (fruits and flowers).

Herbivory also induced VOC emissions (Tab.2 in Paper IV) mainly as GSL degradation products (e.g. ITCs and nitriles), but also terpenoids and GLVs. The plant VOC profile was not influenced by the MeJA application, which only increased the emissions of 1,1'-bicyclopentyl-2-one and 1-pentenol (JA metabolites; Schaller & Stintzi, 2009). Combined, MeJA and VOCs released from herbivore-attacked plants can indirectly protect the plants, functioning as airborne signals which attract herbivore parasitoids, or travel to neighbour plants and prime their JA-mediated defences (Farmer & Ryan, 1990; Karban et al., 2014; Balmer et al., 2015; Lortzing & Steppuhn, 2016). On the other hand, these same VOC emissions can also attract the herbivore butterflies.

Oviposition on leaves induces similar JA-mediated effects, which prime plant defences against future caterpillar herbivory (Lortzing & Steppuhn, 2016). P. brassicae butterflies usually oviposit on expanded mature leaves, but late instar caterpillars move upwards to feed on young leaves and flowers (Lucas-Barbosa et al., 2013; Bruinsma et al., 2014). Interestingly, in this study, herbivore-induced responses were increasingly prioritized towards younger leaves which induced large changes in secondary metabolites as well as in central energy pathways, such as sugar metabolism and TCA cycle. Young leaves also displayed increased levels of chlorophyll after 72 hours of MeJA application (see Fig.5b in Paper IV). Although this result sets in contrast with the commonly reported suppression of photosynthesis induced by herbivory (Zangerl et al., 2002; Bilgin et al., 2010; Halitschke et al., 2011; also observed in Paper II, relative to fully developed leaves), it also supports the hypothesis that herbivore-induced responses help the plant to prioritize both defence and growth processes in those young developing leaves that are quickly shifting from sink-tissues to future sources of photosynthesis.

4. Conclusions

4.1 The “Butterfly effect”

As a result of climate change, environmental factors are shifting very rapidly. The human impact on the planet is now so strong that we can speak about the Anthropocene as a new epoch in the Earth history (Steffen et al., 2011). By the end of this century, regional concentrations of O3 are predicted to exceed 40-70 ppb (Sitch et al., 2007), while O3 damage to agriculture today already accumulates to several billions of euros per year (Tiwari et al., 2016). Increasing temperatures are also affecting ecosystems world-wide, forcing plants and insects to migrate to new habitats (Morriën et al., 2010).

During plant-insect interactions, environmental pressures can have strong impact on both plants and insects (Edger et al., 2015) but induced metabolic changes in the plant are the major forces determining the success of either the host or the herbivores (Foyer et al., 2016). Although plant responses to single biotic and abiotic stresses usually induce unique metabolic signatures, responses to multiple stresses can be very difficult to predict (Mittler, 2006; Suzuki et al., 2014). Unpredictability is common to complex systems (Orrell 2010, Capra & Luisi 2014) where small perturbations in the initial conditions trigger vast downstream effects with unknown consequences (Lorenz, 2015).

In this thesis, I showed how multiple stresses alter plant metabolism and defence against insects, but also how herbivore-induced responses can affect the plant adaptation towards other concurrent stresses (Fig.29). Although single metabolite levels changed unexpectedly and in unique ways (Table 7), it was possible to identify more predictable response patterns which emerged from the shifts of the entire metabolome. Thus, the metabolomics approach appears to be especially helpful when the overall complexity of the system increases, for instance during concurrent O3 and herbivory (Paper I- II), or during plant interaction with multiple herbivore insects, such as aphids and caterpillars (Paper III). Particularly, O3 stress induced metabolic changes in the host-plant which affected the caterpillar development (Paper I). However, herbivory also interfered and redirected the initial plant response towards O3 stress (Paper II). Plant responses also differed according to the kind of herbivory between caterpillars and aphids, and were dependent on the density of the attacker (Paper III).

Metabolomics highlighted the link between growth and defence responses, which correlated to plant processes of photosynthesis, leaf development, and resistance against herbivory (Paper I-IV). The integration of multiple layers of biological information via metabolomics and transcriptomics enabled to predict actual physiological stress responses and to anticipate the occurrence of visible symptoms in leaves (Paper II). Combining metabolomics with manipulation of the plant-herbivore system using hormone elicitors (MeJA) or genetic mutants (Balmer et al., 2015) further provided insights into the mechanisms of central metabolic regulation during plant-insect interactions (Paper IV). I believe that when this systems approach will be applied and fully integrated to biotechnology and breeding programs (Weckwerth, 2011), metabolomics will rise as an essential element for studying plants and agro- ecosystems in constantly changing environments.

Figure 29. The “Butterfly effect”. Overview of how combination of abiotic and biotic stresses can shape B. nigra metabolism and its interaction with P. brassicae.

Table 7. Abiotic and biotic stress interaction on B. nigra metabolites.

Function Class Metabolite Stress interaction

O3 + caterpillars

Direct defence Glucosinolate Sinigrin Negative Direct defence Glucosinolate Glucoraphanin Positive Direct defence Glucosinolate Neoglucobrassicin Positive Direct defence Hydroxycinnamate 1-O-sinapoylglucose Positive Direct defence Flavonol glucoside Quercetin-3-sinapoylsophoros. Negative Central metabolism Sugar Maltotriose Positive Central metabolism Fatty acid Glycerol Negative Induced defence Fatty acid - Linolenic acid Positive Stress signalling Amino acid GABA Positive Stress tolerance Redox Threonicα acid Negative Indirect defence VOCs Furfural Negative Indirect defence VOCs Butenenitrile Positive Indirect defence VOCs Allyl-isothiocyanate Positive Indirect defence VOCs Dodecanal Neutral Indirect defence VOCs E-DMNT Neutral

Aphids + caterpillars

Direct defence Glucosinolate Sinigrin Neutral Direct defence Glucosinolate Glucoiberin Positive Direct defence Glucosinolate Glucobrassicin Positive Direct defence Glucosinolate 4-methoxy-glucobrassicin Positive Central metabolism Sugar Fructose Negative Central metabolism Sugar Maltose Positive Stress tolerance Sugar Trehalose Neutral Stress tolerance Redox Glutathione (GSH) Negative Induced defence Fatty acid - Linolenic acid Positive

α MeJA + caterpillars

Direct defence Glucosinolate Sinigrin Positive Direct defence Glucosinolate Gluconapin Positive Direct defence Glucosinolate Glucoerucin Positive Central metabolism Amino acid Phenylalanine Neutral Central metabolism Amino acid Tryptophan Positive Central metabolism Sugar Fructose Positive Central metabolism Sugar Maltotriose Positive Central metabolism TCA Citric acid Neutral Central metabolism TCA cis-Aconitic acid Positive Central metabolism TCA Malic acid Positive Central metabolism TCA Fumaric acid Neutral Indirect defence VOCs Allyl-isothiocyanate Neutral Indirect defence VOCs 1,1'-bicyclopentyl-2-one Positive Indirect defence VOCs 1-pentenol Positive

Potential interaction between herbivore-induced responses and other biotic and abiotic stress responses, estimated combining information from all studies presented this thesis. Sequential stresses may reinforce (positive), counteract (negative), or have no effect (neutral) on single metabolite levels in B. nigra.

5. Acknowledgements

I want to thank my supervisor Benedicte Albrectsen, for offering me this research project and introducing me to the chemical ecology community. Thank you for your support in these years, for the freedom and the trust. Thanks to my co-supervisor Thomas Moritz, for metabolomics mentorship and for being able to provide calm and stability in any difficult situation. Thanks to Göran (Slim) Samuelsson and Stefan Jansson for your wisdom.

Thanks to the Swedish Metabolomics Centre for the great organization and nice environment, especially: Jonas Gullberg, Annika Johansson, Maria Ahnlund, Jenny Hällqvist, and Inga-Britt Carlsson. Particularly, I want to thank Ilka Abreu for her help with LC-MS/Orbitrap, and Krister Lundgren for all the time spent teaching me LC-MS and GC-MS, and learning together the thermal desorption. May thanks also to Hans Stenlund and Rui Pinto, for help with Matlab and SIMCA, respectively. Thank you all for your patience!

Thanks to many people from the chemistry department, who provided help and support. Thank you especially to Mátyás Ripszam for your dedication to the VOCs analyses, and to Peter Haglund for the support, and for providing the TD-GC-TOF-MS. Thank you to Marcus Carlsson and Per-Anders Enquist for assisting in the lab during the identification of the unknown metabolites, a task which proved to require possibly other extra five years of PhD.

Thanks to Anastasia Matrosova and Zsofia Stangl for patiently teaching me how to use the Li-COR, and to Manuela Jurca, Maria Eriksson, and Vaughn Hurry for extra support and for borrowing the instruments. Thanks to Alex Makoveychuk for help with tissue cultures, to Rosìe Forsgren for the autoclaving and glassware work, and to Jenny Lönnebrink and the greenhouse staff, for their friendly manners and professionality.

Thank you to all my current and previous lab members for the discussions, technical and psychological support, and for all the good time, fika, lunches, beers, and fun. In order of featuring inside this five-year-long Phd movie: Franzi, Vicky, Ken, Tiggy, Abu, Tristan, Ylva, Johan, Barbara, Mona, Karen.

Thanks to the members of the Umeå Plant Science Centre, all of you are creating such a wonderful environment. There is no space for mentioning everyone, so I will especially thank the PhD students involved in teaching course lectures and laboratories with me: Jimmy, Daria, Abdel, Christoffer, Noemi, and particularly Vicky and Bernard and all the students that participated in course of “molecular ecology” and worked on B. nigra.

Thanks Katharina and Tomas, who strongly recommended me UPSC in 2011. Thanks to the European Science Foundation and EUROCORES Programme who funded my research within the A-BIO-VOC/EuroVOL and allowed me to have wonderful experiences in international conferences and meetings. I want to thank my collaborators for their help and for all the nice moments: James Blande, Tao Li, Timo Oksanen, Erik Poelman, Rieta Gols, Marcel Dicke, Philippe Reymond, Eliezer Khaling, Christelle Bonnet, Camille Ponzio, Foteini Pashalidou, and Holger Danner.

Sport and nature helped incredibly much in keeping my mind stable. Thank you Benjamin who told me about Tomtebo kolonilotten, and to Bernard & Gosia for spending precious time with me, mud, and mosquitos. Thanks to Daniel M., Daniel E., Josefin, and Fra, for bouldering with me. Thanks to IKSU simning, Stöcke TS Järne, and Umeå Parkourförening, and the UPSC bandy group, and especially Sacha and Thomas for organizing it.

Thank you to all my nice flat mates for all the funny moments spent together. Gaia, Andreas, Rogier, Victor, Isa, Daniel, and Ioulia. Thanks to all the nice people that created very special moments and atmospheres: Ewa & Fede, Mattias, Elina, Anders, and the Folketsbio. Søren & Leyla, Noemi & Daniel, Karen & Marielle, Gosia & Bernard, Anne & Martin, Claudia & Mátyás, Veronica & Audis, Anne-Sofie & Alex, Viveca & Szilard, Noalie, Eric, Hermann, Bianca, and LiquidSky, Thanks also to Andreas R. for a proper Italian carbonara with fruity Tasmanian black pepper.

Thanks to my family and friends, I miss you constantly: Fra, Leo, Marci, Ari, Eva, Laura, Tommi, Lupo, Marco, Marco (Flip), and Lorenzo. Thanks Piero for your time in Milano and for sharing passion for science. Thank you Hossein for your help, and for your hospitality in San Francisco. Thank you Dan and Holger for the great times in L.A., Joshua Tree park and Yosemite. Thanks Jorge and Stefano in CPH for your hospitality, great times and great beers. Thanks to Kirstin & Tommi and children, from Perth to Karlskoga. Thanks to Mark & Iris and little Vasco.

A special thanks to my parents. You are always there for me. Grazie. Per avermi sostenuto in lunghi anni di scuola e universitá, per la vostra gentilezza e il vostro amore, il tempo dedicato al lavoro e alla famiglia, e per avermi trasmesso amore per la cultura e rispetto per il prossimo. Non so come possa ripagarvi, e vi quindi vi dedico questa tesi 

Sist men inte minst, tack Ann-Grett & Douglas, och hela Josefins familj, för er varma gästfrihet och vänlighet; och tack Josefin, för att du finns, du är den viktigaste personen i mitt liv. Jag älskar dig ♥

6. Literature

Adhikari S, Russell FL. 2014. Effects of Apical Meristem Mining on Plant Fitness, Architecture, and Flowering Phenology in Cirsium Altissimum (Asteraceae). American Journal of Botany 101(12): 2079-2087. Agrawal AA, Kurashige NS. 2003. A role for isothiocyanates in plant resistance against the specialist herbivore Pieris rapae. Journal of Chemical Ecology 29(6): 1403-1415. Ali JG, Agrawal AA. 2012. Specialist versus generalist insect herbivores and plant defense. Trends in plant science 17(5): 293-302. Ansari MI, Lee RH, Chen SCG. 2005. A novel senescence-associated gene encoding gamma-aminobutyric acid (GABA): pyruvate transaminase is upregulated during rice leaf senescence. Physiologia Plantarum 123(1): 1-8. Araujo WL, Nunes-Nesi A, Nikoloski Z, Sweetlove LJ, Fernie AR. 2012. Metabolic control and regulation of the tricarboxylic acid cycle in photosynthetic and heterotrophic plant tissues. Plant Cell Environ 35(1): 1- 21. Ashmore MR. 2005. Assessing the future global impacts of ozone on vegetation. Plant Cell and Environment 28(8): 949-964. Atkinson NJ, Urwin PE. 2012. The interaction of plant biotic and abiotic stresses: from genes to the field. J Exp Bot 63(10): 3523-3543. Ayasse M, Stokl J, Francke W. 2011. Chemical ecology and pollinator-driven speciation in sexually deceptive orchids. Phytochemistry 72(13): 1667-1677. Bagard M, Le Thiec D, Delacote E, Hasenfratz-Sauder MP, Banvoy J, Gerard J, Dizengremel P, Jolivet Y. 2008. Ozone-induced changes in photosynthesis and photorespiration of hybrid poplar in relation to the developmental stage of the leaves. Physiologia Plantarum 134(4): 559-574. Baldoni E, Genga A, Cominelli E. 2015. Plant MYB Transcription Factors: Their Role in Drought Response Mechanisms. Int J Mol Sci 16(7): 15811-15851. Balmer A, Pastor V, Gamir J, Flors V, Mauch-Mani B. 2015. The 'prime-ome': towards a holistic approach to priming. Trends in plant science 20(7): 443-452. Bebber DP, Ramotowski MAT, Gurr SJ. 2013. Crop pests and pathogens move polewards in a warming world. Nature Climate Change 3(11): 985-988. Beekwilder J, van Leeuwen W, van Dam NM, Bertossi M, Grandi V, Mizzi L, Soloviev M, Szabados L, Molthoff JW, Schipper B, et al. 2008. The impact of the absence of aliphatic glucosinolates on insect herbivory in Arabidopsis. PLoS One 3(4): e2068. Bennett R, Donald A, Dawson G, Hick A, Wallsgrove R. 1993. Aldoxime-Forming Microsomal-Enzyme Systems Involved in the Biosynthesis of Glucosinolates in Oilseed Rape (Brassica napus) Leaves. Plant physiology 102(4): 1307-1312. Bienert GP, Schjoerring JK, Jahn TP. 2006. Membrane transport of hydrogen peroxide. Biochim Biophys Acta 1758(8): 994-1003. Bilgin DD, Zavala JA, Zhu J, Clough SJ, Ort DR, DeLucia EH. 2010. Biotic stress globally downregulates photosynthesis genes. Plant Cell Environ 33(10): 1597-1613. Blanco NE, Guinea-Díaz M, Whelan J, Strand Å. 2014. Interaction between plastid and mitochondrial retrograde signalling pathways during changes to plastid redox status. Philos Trans R Soc Lond B Biol Sci. 369(1640): 20130231. Blande JD, Holopainen JK, Li T. 2010. Air pollution impedes plant-to-plant communication by volatiles. Ecol Lett 13(9): 1172-1181. Blatt SE, Smallegange RC, Hess L, Harvey JA, Dicke M, van Loon JJA. 2008. Tolerance of Brassica nigra to Pieris brassicae herbivory. Botany-Botanique 86(6): 641-648. Boex-Fontvieille E, Rustgi S, von Wettstein D, Reinbothe S, Reinbothe C. 2015. Water-soluble chlorophyll protein is involved in herbivore resistance activation during greening of Arabidopsis thaliana. PNAS 112(23): 7303-7308. Bolton. 2009. Primary metabolism and plant defense-fuel for the fire. Mol Plant Microbe Interact. 22(5): 487-97. Booker F, Muntifering R, McGrath M, Burkey K, Decoteau D, Fiscus E, Manning W, Krupa S, Chappelka A, Grantz D. 2009. The Ozone Component of Global Change: Potential Effects on Agricultural and Horticultural Plant Yield, Product Quality and Interactions with Invasive Species. Journal of Integrative Plant Biology 51(4): 337-351. Bouche N, Fromm H. 2004. GABA in plants: just a metabolite? Trends Plant Sci 9(3): 110-115.

Bourgaud F, Gravot A, Milesi S, Gontier E. 2001. Production of plant secondary metabolites: a historical perspective. Plant Science 161(5): 839-851. Broekgaarden C, Poelman EH, Steenhuis G, Voorrips RE, Dicke M, Vosman B. 2008. Responses of Brassica oleracea cultivars to infestation by the aphid Brevicoryne brassicae: an ecological and molecular approach. Plant Cell and Environment 31(11): 1592-1605. Bruce TJA. 2015. Interplay between insects and plants: dynamic and complex interactions that have coevolved over millions of years but act in milliseconds. J Exp Bot 66(2): 455-465. Bruessow F, Gouhier-Darimont C, Buchala A, Metraux JP, Reymond P. 2010. Insect eggs suppress plant defence against chewing herbivores. Plant Journal 62(5): 876-885. Bruinsma M, Lucas-Barbosa D, ten Broeke CJ, van Dam NM, van Beek TA, Dicke M, van Loon JJ. 2014a. Folivory affects composition of nectar, floral odor and modifies pollinator behavior. J Chem Ecol 40(1): 39-49. Bundy JG, Davey MP, Viant MR. 2008. Environmental metabolomics: a critical review and future perspectives. Metabolomics 5(1): 3-21. Burger H, Dotterl S, Ayasse M. 2010. Host-plant finding and recognition by visual and olfactory floral cues in an oligolectic bee. Functional Ecology 24(6): 1234-1240. Burow M, Losansky A, Muller R, Plock A, Kliebenstein DJ, Wittstock U. 2009. The Genetic Basis of Constitutive and Herbivore-Induced ESP-Independent Nitrile Formation in Arabidopsis. Plant physiology 149(1): 561-574. Burow M, Muller R, Gershenzon J, Wittstock U. 2006. Altered glucosinolate hydrolysis in genetically engineered Arabidopsis thaliana and its influence on the larval development of Spodoptera littoralis. Journal of Chemical Ecology 32(11): 2333-2349. Bussy A. 1840. Sur la formation de l’huile essentielle de moutarde. Jour Pharm 27: 464–471 Butts CT. 2009. Revisiting the Foundations of Network Analysis. Science 325(5939): 414-416. Caldana C, Degenkolbe T, Cuadros-Inostroza A, Klie S, Sulpice R, Leisse A, Steinhauser D, Fernie AR, Willmitzer L, Hannah MA.2011. High-density kinetic analysis of the metabolomic and transcriptomic response of Arabidopsis to eight environmental conditions. Plant Journ 67(5):869-84. Campos ML, Yoshida Y, Major IT, de Oliveira Ferreira D, Weraduwage SM, Froehlich JE, Johnson BF, Kramer DM, Jander G, Sharkey TD, et al. 2016. Rewiring of jasmonate and phytochrome B signalling uncouples plant growth-defense tradeoffs. Nat Commun 7: 12570. Capra F. 1996. The Web of Life: A New Scientific Understanding of Living Systems. Anchor Books. Capra F, Luisi PL. 2014. The Systems View of Life: A Unifying Vision. Cambridge University Press Castagna A, Ranieri A. 2009. Detoxification and repair process of ozone injury: from O3 uptake to gene expression adjustment. Environ Pollut 157(5): 1461-1469. Chauvin A, Caldelari D, Wolfender JL, Farmer EE. 2013. Four 13-lipoxygenases contribute to rapid jasmonate synthesis in wounded Arabidopsis thaliana leaves: a role for lipoxygenase 6 in responses to long- distance wound signals. New Phytologist 197(2): 566-575. Chen IC, Hill JK, Ohlemuller R, Roy DB, Thomas CD. 2011. Rapid range shifts of species associated with high levels of climate warming. Science 333(6045): 1024-1026. Chen Q, Sun J, Zhai Q, Zhou W, Qi L, Xu L, Wang B, Chen R, Jiang H, Qi J, et al. 2011. The basic helix-loop-helix transcription factor MYC2 directly represses PLETHORA expression during jasmonate-mediated modulation of the root stem cell niche in Arabidopsis. Plant Cell 23(9): 3335-3352. Chen Z, Zheng Z, Huang J, Lai Z, Fan B. 2009. Biosynthesis of salicylic acid in plants. Plant Signal Behav 4(6): 493-496. Chevalier A, Gheusi F, Delmas R, Ordonez C, Sarrat C, Zbinden R, Thouret V, Athier G, Cousin JM. 2007. Influence of altitude on ozone levels and variability in the lower troposphere: a ground-based study for western Europe over the period 2001-2004. Atmospheric Chemistry and Physics 7(16): 4311-4326. Chini A, Fonseca S, Fernandez G, Adie B, Chico JM, Lorenzo O, Garcia-Casado G, Lopez-Vidriero I, Lozano FM, Ponce MR, et al. 2007. The JAZ family of repressors is the missing link in jasmonate signalling. Nature 448(7154): 666-U664. Cipollini DF, Sipe ML. 2001. Jasmonic acid treatment and mammalian herbivory differentially affect chemical defenses and growth of wild mustard (Brassica kaber). Chemoecology 11(3): 137-143. Clarke DB. 2010. Glucosinolates, structures and analysis in food. Analytical Methods 2(4): 310-325. Conner JK, Neumeier R. 1995. Effects of Black Mustard Population-Size on the Taxonomic Composition of Pollinators. Oecologia 104(2): 218-224. Consales F, Schweizer F, Erb M, Gouhier-Darimont C, Bodenhausen N, Bruessow F, Sobhy I, Reymond P. 2012. Insect oral secretions suppress wound-induced responses in Arabidopsis. J Exp Bot 63(2): 727- 737. Crawley MJ. 1989. Insect Herbivores and Plant-Population Dynamics. Annual Review of Entomology 34: 531-564.

Das S, Lagercrantz U, Lascoux M. 2007. Black mustard. In Genome Mapping and Molecular Breeding in Plants, Volume 2: Oilseeds. Edited by Kole C. Springer-Verlag. Davis GK. 2012. Cyclical Parthenogenesis and Viviparity in Aphids as Evolutionary Novelties. Journal of Experimental Zoology Part B-Molecular and Developmental Evolution 318(6): 448-459. Dawkins R, 1976. The Selfish Gene. Oxford University Press. De Vos M, Jander G. 2009. Myzus persicae (green peach aphid) salivary components induce defence responses in Arabidopsis thaliana. Plant Cell Environ 32(11): 1548-1560. De Vos M, Van Oosten VR, Van Poecke RM, Van Pelt JA, Pozo MJ, Mueller MJ, Buchala AJ, Metraux JP, Van Loon LC, Dicke M, et al. 2005. Signal signature and transcriptome changes of Arabidopsis during pathogen and insect attack. Mol Plant Microbe Interact 18(9): 923-937. Desurmont GA, Harvey J, van Dam NM, Cristescu SM, Schiestl FP, Cozzolino S, Anderson P, Larsson MC, Kindlmann P, Danner H, et al. 2014. Alien interference: disruption of infochemical networks by invasive insect herbivores. Plant Cell and Environment 37(8): 1854-1865. Dicke M, Baldwin IT. 2010. The evolutionary context for herbivore-induced plant volatiles: beyond the 'cry for help'. Trends in plant science 15(3): 167-175. Dizengremel P, Le Thiec D, Hasenfratz-Sauder MP, Vaultier MN, Bagard M, Jolivet Y. 2009. Metabolic- dependent changes in plant cell redox power after ozone exposure. Plant Biol 11 (Suppl 1): 35-42. Dombrecht B, Xue GP, Sprague SJ, Kirkegaard JA, Ross JJ, Reid JB, Fitt GP, Sewelam N, Schenk PM, Manners JM, et al. 2007. MYC2 differentially modulates diverse jasmonate-dependent functions in Arabidopsis. Plant Cell 19(7): 2225-2245. Dudareva N, Klempien A, Muhlemann JK, Kaplan I. 2013. Biosynthesis, function and metabolic engineering of plant volatile organic compounds. New Phytologist 198(1): 16-32. Eastmond PJ. 2004. Glycerol-insensitive Arabidopsis mutants: gli1 seedlings lack glycerol kinase, accumulate glycerol and are more resistant to abiotic stress. Plant Journal 37(4): 617-625. Edger PP, Heidel-Fischer HM, Bekaert M, Rota J, Gloeckner G, Platts AE, Heckel DG, Der JP, Wafula EK, Tang M, et al. 2015. The butterfly plant arms-race escalated by gene and genome duplications. PNAS 112(27): 8362-8366. Ellis C, Karafyllidis L, Turner JG. 2002. Constitutive activation of jasmonate signaling in an Arabidopsis mutant correlates with enhanced resistance to Erysiphe cichoracearum, Pseudomonas syringae, and Myzus persicae. Molecular Plant-Microbe Interactions 15(10): 1025-1030. Ellis PR, Farrell JA. 1995. Resistance to Cabbage Aphid (Brevicoryne brassicae) in 6 Brassica Accessions in New- Zealand. New Zealand Journal of Crop and Horticultural Science 23(1): 25-29. Ellis PR, Kift NB, Pink DAC, Jukes PL, Lynn J, Tatchell GM. 2000. Variation in resistance to the cabbage aphid (Brevicoryne brassicae) between and within wild and cultivated Brassica species. Genetic Resources and Crop Evolution 47(4): 395-401. Erb M, Meldau S, Howe GA. 2012. Role of phytohormones in insect-specific plant reactions. Trends in plant science 17(5):250-259. Farmer EE, Ryan CA. 1990. Interplant communication: airborne methyl jasmonate induces synthesis of proteinase inhibitors in plant leaves. PNAS 87(19): 7713-7716. Fatouros NE, Dicke M, Mumm R, Meiners T, Hilker M. 2008. Foraging behavior of egg parasitoids exploiting chemical information. Behavioral Ecology 19(3): 677-689. Fatouros NE, Huigens ME, van Loon JJ, Dicke M, Hilker M. 2005. Chemical communication: butterfly anti- aphrodisiac lures parasitic wasps. Nature 433(7027): 704. Felton GW, Bi JL, Summers CB, Mueller AJ, Duffey SS. 1994. Potential Role of Lipoxygenases in Defense against Insect Herbivory. Journal of Chemical Ecology 20(3): 651-666. Fernandez O, Bethencourt L, Quero A, Sangwan RS, Clement C. 2010. Trehalose and plant stress responses: friend or foe? Trends in plant science 15(7): 409-417. Ferreres F, Fernandes F, Pereira DM, Pereira JA, Valentao P, Andrade PB. 2009. Phenolics Metabolism in Insects: Pieris brassicae-Brassica oleracea var. costata Ecological Duo. J Agric Food Chem 57(19): 9035-9043. Fiehn O. 2002. Metabolomics--the link between genotypes and phenotypes. Plant Molecular Biology 48(1-2): 155-171. Firn RD, Jones CG. 2009. A Darwinian view of metabolism: molecular properties determine fitness. J Exp Bot 60(3): 719-726. Foyer CH, Rasool B, Davey JW, Hancock RD. 2016. Cross-tolerance to biotic and abiotic stresses in plants: a focus on resistance to aphid infestation. J Exp Bot 67(7): 2025-2037. Fraser CM, Chapple C. 2011. The phenylpropanoid pathway in Arabidopsis. Arabidopsis Book 9: e0152. American Society of Plant Biologists.

Fritz VA, Justen VL, Bode AM, Schuster T, Wang M. 2010. Glucosinolate Enhancement in Cabbage Induced by Jasmonic Acid Application. Hortscience 45(8): 1188-1191. Fuhrer J. 2003. Agroecosystern responses to combinations of elevated CO2, ozone, and global climate change. Agriculture Ecosystems & Environment 97(1-3): 1-20. Gardeström P, Igamberdiev AU, Raghavendra AS. 2002. Mitochondrial functions in the light and significance to carbon-nitrogen interactions. In Photosynthetic Nitrogen Assimilation and Associated Carbon and Respiratory Metabolism. Volume 12 of the series Advances in Photosynthesis and Respiration, 151- 172. Springer Geervliet JBF, Brodeur J. 1992. Host Species Affecting the Performance of the Larval Parasitoids Cotesia glomerata and C. rubecula (Hymenoptera, Braconidae) .2. Comparative Suitability of 3 Pieris Species (Lepidoptera, Pieridae). Mededelingen Van De Faculteit Landbouwwetenschappen, Universiteit-Gent, 57: 547-550. Gielen B, Vandermeiren K, Horemans N, D'Haese D, Serneels R, Valcke R. 2006. Chlorophyll a fluorescence imaging of ozone-stressed Brassica napus L. plants differing in glucosinolate concentrations. Plant Biology 8(5): 698-705. Glauser G, Grata E, Dubugnon L, Rudaz S, Farmer EE, Wolfender JL. 2008. Spatial and temporal dynamics of jasmonate synthesis and accumulation in Arabidopsis in response to wounding. Journal of Biological Chemistry 283(24): 16400-16407. Glazebrook J. 2005. Contrasting mechanisms of defense against biotrophic and necrotrophic pathogens. Annu Rev Phytopathol 43: 205-227. Golovitchev VI, Yang J. 2009. Construction of combustion models for rapeseed methyl ester bio-diesel fuel for internal combustion engine applications. Biotechnol Adv 27(5): 641-655. Gols R, Roosjen M, Dijkman H, Dicke M. 2003. Induction of direct and indirect plant responses by jasmonic acid, low spider mite densities, or a combination of jasmonic acid treatment and spider mite infestation. Journal of Chemical Ecology 29(12): 2651-2666. Gouhier-Darimont C, Schmiesing A, Bonnet C, Lassueur S, Reymond P. 2013. Signalling of Arabidopsis thaliana response to Pieris brassicae eggs shares similarities with PAMP-triggered immunity. J Exp Bot 64(2): 665-674. Goumenaki E, Taybi T, Borland A, Barnes J. 2010. Mechanisms underlying the impacts of ozone on photosynthetic performance. Environmental and Experimental Botany 69(3): 259-266. Halitschke R, Baldwin IT. 2003. Antisense LOX expression increases herbivore performance by decreasing defense responses and inhibiting growth-related transcriptional reorganization in Nicotiana attenuata. Plant Journal 36(6): 794-807. Halitschke R, Hamilton JG, Kessler A. 2011. Herbivore-specific elicitation of photosynthesis by mirid bug salivary secretions in the wild tobacco Nicotiana attenuata. New Phytol 191(2): 528-535. Halkier B, Gershenzon J. 2006. Biology and biochemistry of glucosinolates. Annu Rev Plant Biol 57: 303-333. Hatfield JL, Walthall CL. 2015. Meeting Global Food Needs: Realizing the Potential via Genetics x Environment x Management Interactions. Agronomy Journal 107(4): 1215-1226. Heil M, Baldwin IT. 2002. Fitness costs of induced resistance: emerging experimental support for a slippery concept. Trends in plant science 7(2): 61-67. Himanen SJ, Nissinen A, Auriola S, Poppy GM, Stewart CN, Jr., Holopainen JK, Nerg AM. 2008. Constitutive and herbivore-inducible glucosinolate concentrations in oilseed rape (Brassica napus) leaves are not affected by Bt Cry1Ac insertion but change under elevated atmospheric CO2 and O3. Planta 227(2): 427-437. Hodge S, Ward JL, Beale MH, Bennett M, Mansfield JW, Powell G. 2013. Aphid-induced accumulation of trehalose in Arabidopsis thaliana is systemic and dependent upon aphid density. Planta 237(4): 1057- 1064. Hopkins RJ, van Dam NM, van Loon JJA. 2009. Role of Glucosinolates in Insect-Plant Relationships and Multitrophic Interactions. Annual Review of Entomology 54: 57-83. Howe GA, Jander G. 2008. Plant immunity to insect herbivores. Annu Rev Plant Biol 59: 41-66. Huang SB, Taylor NL, Stroher E, Fenske R, Millar AH. 2013. Succinate dehydrogenase assembly factor 2 is needed for assembly and activity of mitochondrial complex II and for normal root elongation in Arabidopsis. Plant Journal 73(3): 429-441. Huot B, Yao J, Montgomery BL, He SY. 2014. Growth-defense tradeoffs in plants: a balancing act to optimize fitness. Mol Plant 7(8): 1267-1287. Igamberdiev AU, Eprintsev AT. 2016. Organic Acids: The Pools of Fixed Carbon Involved in Redox Regulation and Energy Balance in Higher Plants. Front Plant Sci 7: 1042. Ings TC, Montoya JM, Bascompte J, Bluthgen N, Brown L, Dormann CF, Edwards F, Figueroa D, et al. 2009. Ecological networks - beyond food webs. Journal of Animal Ecology 78(1): 253-269.

International Panel for Climate Change (IPCC). 2014. Synthesis Report. Edited by Pachauri RK, Meyer LA. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. IPCC, Geneva, Switzerland; pp. 151. Irmisch S, McCormick AC, Gunther J, Schmidt A, Boeckler GA, Gershenzon J, Unsicker SB, Kollner TG. 2014. Herbivore-induced poplar cytochrome P450 enzymes of the CYP71 family convert aldoximes to nitriles which repel a generalist caterpillar. Plant Journal 80(6): 1095-1107. Jaradat MR, Feurtado JA, Huang DQ, Lu YQ, Cutler AJ. 2013. Multiple roles of the transcription factor AtMYBR1/AtMYB44 in ABA signaling, stress responses, and leaf senescence. BMC Plant Biology 13:192. Jøndrup PM, Barnes JD, Port GR. 2002. The effect of ozone fumigation and different Brassica rapa lines on the feeding behaviour of Pieris brassicae larvae. Entomologia Exper et Applicata 104(1): 143-151. Jones AME, Bridges M, Bones AM, Cole R, Rossiter JT. 2001. Purification and characterisation of a non-plant myrosinase from the cabbage aphid Brevicoryne brassicae (L.). Insect Biochemistry and Molecular Biology 31(1): 1-5. Jones CG, Coleman JS. 1988. Plant Stress and Insect Behavior - Cottonwood, Ozone and the Feeding and Oviposition Preference of a Beetle. Oecologia 76(1): 51-56. Joudoi T, Shichiri Y, Kamizono N, Akaike T, Sawa T, Yoshitake J, Yamada N, Iwai S. 2013. Nitrated Cyclic GMP Modulates Guard Cell Signaling in Arabidopsis. Plant Cell 25(2): 558-571. Karban R, Agrawal AA. 2002. Herbivore Offense. Annual Review of Ecology and Systematics 33(1): 641-664. Karban R, Wetzel WC, Shiojiri K, Ishizaki S, Ramirez SR, Blande JD. 2014. Deciphering the language of plant communication: volatile chemotypes of sagebrush. New Phytologist 204(2): 380-385. Kask K, Kaennaste A, Talts E, Copolovici L, Niinemets U. 2016. How specialized volatiles respond to chronic and short-term physiological and shock heat stress in Brassica nigra. Plant Cell and Environment 39(9): 2027-2042. Katsir L, Schilmiller AL, Staswick PE, He SY, Howe GA. 2008. COI1 is a critical component of a receptor for jasmonate and the bacterial virulence factor coronatine. PNAS 105(19): 7100-7105. Kazana E, Pope TW, Tibbles L, Bridges M, Pickett JA, Bones AM, Powell G, Rossiter JT. 2007. The cabbage aphid: a walking mustard oil bomb. Proc. of the Royal Society B-Biological Sciences 274(1623): 2271- 2277. Keurentjes JJB, Angenent GC, Dicke M, Dos Santos VAPM, Molenaar J, van der Putten WH, de Ruiter PC, Struik PC, Thomma BPHJ. 2011. Redefining plant systems biology: from cell to ecosystem. Trends in plant science 16(4): 183-190. Khaling E, Papazian S, Poelman EH, Holopainen JK, Albrectsen BR, Blande JD. 2015. Ozone affects growth and development of Pieris brassicae on the wild host plant Brassica nigra. Env Pollution 199: 119-129. Khan MAM, Ulrichs C, Mewis I. 2011. Water stress alters aphid-induced glucosinolate response in Brassica oleracea var. italica differently. Chemoecology 21(4): 235-242. Kim JH, Jander G. 2007. Myzus persicae (green peach aphid) feeding on Arabidopsis induces the formation of a deterrent indole glucosinolate. Plant Journal 49(6): 1008-1019. Kos M, Houshyani B, Wietsma R, Kabouw P, Vet LEM, van Loon JJA, Dicke M. 2012. Effects of glucosinolates on a generalist and specialist leaf-chewing herbivore and an associated parasitoid. Phytochemistry 77: 162-170. Kouzai Y, Kimura M, Yamanaka Y, Watanabe M, Matsui H, Yamamoto M, Ichinose Y, Toyoda K, Onda Y, Mochida K, et al. 2016. Expression profiling of marker genes responsive to the defence-associated phytohormones salicylic acid, jasmonic acid and ethylene in Brachypodium distachyon. BMC Plant Biology 16. Lamb KP. 1959. Composition of the Honeydew of the Aphid Brevicoryne brassicae (L.) Feeding on Swedes (Brassica napobrassica Dc). Journal of Insect Physiology 3(1): 1-13. Lankau RA, Strauss SY. 2007. Mutual Feedbacks Maintain Both Genetic and Species Diversity in a Plant Community. Science 317(5844): 1561-1563. Laothawornkitkul J, Taylor JE, Paul ND, Hewitt CN. 2009. Biogenic volatile organic compounds in the Earth system. New Phytologist 183(1): 27-51. Lemoine NP, Burkepile DE, Parker JD. 2017. Insect herbivores increase mortality and reduce tree seedling growth of some species in temperate forest canopy gaps. PeerJ 5: e3102. Leonard E, Runguphan W, O'Connor S, Prather KJ. 2009. Opportunities in metabolic engineering to facilitate scalable alkaloid production. Nature Chemical Biology 5(5): 292-300. Li DK, Li Y, Zhang L, Wang XY, Zhao Z, Tao ZW, Wang JM, Wang J, Lin M, Li XF, et al. 2014. Arabidopsis ABA Receptor RCAR1/PYL9 Interacts with an R2R3-Type MYB Transcription Factor, AtMYB44. Int J Mol Sci 15(5): 8473-8490.

Li Q, Zheng Q, Shen W, Cram D, Fowler DB, Wei Y, Zou J. 2015. Understanding the biochemical basis of temperature-induced lipid pathway adjustments in plants. Plant Cell 27(1): 86-103. Li ZH, Peng JY, Wen X, Guo HW. 2013. ETHYLENE-INSENSITIVE3 Is a Senescence-Associated Gene That Accelerates Age-Dependent Leaf Senescence by Directly Repressing miR164 Transcription in Arabidopsis. Plant Cell 25(9): 3311-3328. Lin LZ, Sun JH, Chen P, Harnly J. 2011. UHPLC-PDA-ESI/HRMS/MSn Analysis of Anthocyanins, Flavonol Glycosides, and Hydroxycinnamic Acid Derivatives in Red Mustard Greens (Brassica juncea Coss Variety). J Agric Food Chem 59(22): 12059-12072. Lin MY, Horowitz LW, Payton R, Fiore AM, Tonnesen G. 2017. US surface ozone trends and extremes from 1980 to 2014: quantifying the roles of rising Asian emissions, domestic controls, wildfires, and climate. Atmospheric Chemistry and Physics 17(4): 2943-2970. Liu Y, Ahn JE, Datta S, Salzman RA, Moon J, Huyghues-Despointes B, Pittendrigh B, Murdock LL, Koiwa H, Zhu-Salzman K. 2005. Arabidopsis vegetative storage protein is an anti-insect acid phosphatase. Plant Physiol 139(3): 1545-1556. Lorenz EU. 2015. Predictability: Does the Flap of a Butterfly's Wings in Brazil Set off a Tornado in Texas? Resonance-Journal of Science Education 20(3): 261-263. – reproduced from original paper, 1972. Lortzing T, Steppuhn A. 2016. Jasmonate signalling in plants shapes plant-insect interaction ecology. Current Opinion in Insect Science 14: 32-39. Lucas-Barbosa D, van Loon JJA, Gols R, van Beek TA, Dicke M. 2013. Reproductive escape: annual plant responds to butterfly eggs by accelerating seed production. Functional Ecology 27(1): 245-254. Lundborg L, Nordlander G, Bjorklund N, Nordenhem H, Borg-Karlson AK. 2016. Methyl Jasmonate-Induced Monoterpenes in Scots Pine and Norway Spruce Tissues Affect Pine Weevil Orientation. Journal of Chemical Ecology 42(12): 1237-1246. Maag D, Erb M, Glauser G. 2015. Metabolomics in plant-herbivore interactions: challenges and applications. Entomologia Experimentalis et Applicata 157(1): 18-29. Manners JM, Penninckx IA, Vermaere K, Kazan K, Brown RL, Morgan A, Maclean DJ, Curtis MD, Cammue BP, Broekaert WF. 1998. The promoter of the plant defensin gene PDF1.2 from Arabidopsis is systemically activated by fungal pathogens and responds to methyl jasmonate but not to salicylic acid. Plant Molecular Biology 38(6): 1071-1080. Martin MV, Fiol DF, Sundaresan V, Zabaleta EJ, Pagnussat GC. 2013. oiwa, a Female Gametophytic Mutant Impaired in a Mitochondrial Manganese-Superoxide Dismutase, Reveals Crucial Roles for Reactive Oxygen Species during Embryo Sac Development and Fertilization in Arabidopsis. Plant Cell 25(5): 1573-1591. Mattiacci L, Dicke M, Posthumus MA. 1995. beta-Glucosidase: an elicitor of herbivore-induced plant odor that attracts host-searching parasitic wasps. PNAS 92(6): 2036-2040. Mckey D. 1974. Adaptive Patterns in Alkaloid Physiology. American Naturalist 108(961): 305-320. Meldau S, Erb M, Baldwin IT. 2012. Defence on demand: mechanisms behind optimal defence patterns. Annals of Botany 110(8): 1503-1514. Mewis I, Tokuhisa JG, Schultz JC, Appel HM, Ulrichs C, Gershenzon J. 2006. Gene expression and glucosinolate accumulation in Arabidopsis thaliana in response to generalist and specialist herbivores of different feeding guilds and the role of defense signaling pathways. Phytochemistry 67(22): 2450-2462. Michaeli S, Fromm H. 2015. Closing the loop on the GABA shunt in plants: are GABA metabolism and signaling entwined? Front Plant Sci 6: 419. Milkowski C, Strack D. 2010. Sinapate esters in brassicaceous plants: biochemistry, molecular biology, evolution and metabolic engineering. Planta 232(1): 19-35. Mirabella R, Rauwerda H, Allmann S, Scala A, Spyropoulou EA, de Vries M, Boersma MR, Breit TM, Haring MA, Schuurink RC. 2015. WRKY40 and WRKY6 act downstream of the green leaf volatile E-2-hexenal in Arabidopsis. Plant Journal 83(6): 1082-1096. Misof B, Liu SL, Meusemann K, Peters RS, Donath A, Mayer C, Frandsen PB, Ware J, Flouri T, et al. 2014. Phylogenomics resolves the timing and pattern of insect evolution. Science 346(6210): 763-767. Mittler R. 2006. Abiotic stress, the field environment and stress combination. Trends Plant Sci 11(1): 15-19. Moore B, Andrew RL, Kulheim C, Foley WJ. 2014. Explaining intraspecific diversity in plant secondary metabolites in an ecological context. New Phytologist 201(3): 733-750. Morant AV, Bjarnholt N, Kragh ME, Kjaergaard CH, Jorgensen K, Paquette SM, Piotrowski M, et al. 2008. The beta-glucosidases responsible for bioactivation of hydroxynitrile glucosides in Lotus japonicus. Plant Physiol 147(3): 1072-1091. Morrisey JP. 2009. Biological Activity of Defence-Related Plant Secondary Metabolites. In Plant-derived Natural Products: Synthesis, Function, and Application (269-279). Edited by Osbourn, AE, Lanzotti, V. Springer.

Morriën E, Engelkes T, Macel M, Meisner A, Van der Putten WH. 2010. Climate change and invasion by intracontinental range-expanding exotic plants: the role of biotic interactions. Annals of Botany 105(6): 843-848. Mousavi SA, Chauvin A, Pascaud F, Kellenberger S, Farmer EE. 2013. GLUTAMATE RECEPTOR-LIKE genes mediate leaf-to-leaf wound signalling. Nature 500(7463):422-6. Mueller SP, Krause DM, Mueller MJ, Fekete A. 2015. Accumulation of extra-chloroplastic triacylglycerols in Arabidopsis seedlings during heat acclimation. J Exp Bot 66(15): 4517-4526. Mulaudzi T, Ludidi N, Ruzvidzo O, Morse M, Hendricks N, Iwuoha E, Gehring C. 2011. Identification of a novel Arabidopsis thaliana nitric oxide-binding molecule with guanylate cyclase activity in vitro. FEBS Lett 585(17): 2693-2697. Munguia-Rosas MA, Ollerton J, Parra-Tabla V, De-Nova JA. 2011. Meta-analysis of phenotypic selection on flowering phenology suggests that early flowering plants are favoured. Ecol Lett 14(5): 511-521. Mutwil M, Klie S, Tohge T, Giorgi FM, Wilkins O, Campbell MM, Fernie AR, Usadel B, Nikoloski Z, Persson S. 2011. PlaNet: Combined Sequence and Expression Comparisons across Plant Networks Derived from Seven Species. Plant Cell 23(3): 895-910. Nakabayashi R, Saito K. 2015. Integrated metabolomics for abiotic stress responses in plants. Curr Opin Plant Biol 24: 10-16. Neilson EH, Goodger JQ, Woodrow IE, Moller BL. 2013. Plant chemical defense: at what cost? Trends Plant Sci 18(5): 250-258. Nguyen D, Rieu I, Mariani C, van Dam NM. 2016. How plants handle multiple stresses: hormonal interactions underlying responses to abiotic stress and insect herbivory. Plant Molecular Biology 91(6): 727-740. Noctor G, Foyer CH. 1998. Ascorbate and glutathione: Keeping active oxygen under control. Annual Review of Plant Physiology and Plant Molecular Biology 49: 249-279. Noor E, Eden E, Milo R, Alon U. 2010. Central Carbon Metabolism as a Minimal Biochemical Walk between Precursors for Biomass and Energy. Mol Cell 39(5): 809-820. Nuccio ML, Wu J, Mowers R, Zhou HP, Meghji M, Primavesi LF, Paul MJ, Chen X, Gao Y, Haque E, et al. 2015. Expression of trehalose-6-phosphate phosphatase in maize ears improves yield in well-watered and drought conditions. Nature Biotechnology 33(8): 862. Oelschlägel B, Nuss M, von Tschirnhaus M, Patzold C, Neinhuis C, Dotterl S, Wanke S. 2015. The betrayed thief - the extraordinary strategy of Aristolochia rotunda to deceive its pollinators. New Phytol 206(1): 342-351. Olivas NHD, Kruijer W, Gort G, Wijnen CL, van Loon JJA, Dicke M. 2017. Genome-wide association analysis reveals distinct genetic architectures for single and combined stress responses in Arabidopsis thaliana. New Phytologist 213(2): 838-851. Orrell D. 2010. Economyths. Ten Ways Economics Gets It Wrong. Wiley. Pandey GK, Pandey A, Prasad M, Bohmer M. 2016. Editorial: Abiotic Stress Signaling in Plants: Functional Genomic Intervention. Front Plant Sci 7: 681. Papazian S, Khaling E, Bonnet C, Lassueur S, Reymond P, Moritz T, Blande JD, Albrectsen BR. 2016. Central Metabolic Responses to Ozone and Herbivory Affect Photosynthesis and Stomatal Closure. Plant physiology 172(3): 2057-2078. Parolo G, Rossi G. 2008. Upward migration of vascular plants following a climate warming trend in the Alps. Basic and Applied Ecology 9(2): 100-107. Pashalidou FG, Fatouros NE, Van Loon JJA, Dicke M, Gols R. 2015. Plant-mediated effects of butterfly egg deposition on subsequent caterpillar and pupal development, across different species of wild Brassicaceae. Ecological Entomology 40(4): 444-450. Petersen B, Chen S, Hansen C, Olsen C, Halkier B. 2002. Composition and content of glucosinolates in developing Arabidopsis thaliana. Planta 214(4): 562-571. Pichersky E. 2011. New Synthesis-Duplicated Genes in the Ecological Interactions of Plants with Their Environment. Journal of Chemical Ecology 37(9): 923-923. Pieterse CMJ, Dicke M. 2007. Plant interactions with microbes and insects: from molecular mechanisms to ecology. Trends in plant science 12(12): 564-569. Pieterse CMJ, Leon-Reyes A, Van der Ent S, Van Wees SCM. 2009. Networking by small-molecule hormones in plant immunity. Nature Chemical Biology 5(5): 308-316. Pimm SL, Joppa LN. 2015. How Many Plant Species Are There, Where Are They, and at What Rate Are They Going Extinct? Annals of the Missouri Botanical Garden 100(3): 170-176. Pinto DM, Blande JD, Souza SR, Nerg AM, Holopainen JK. 2010. Plant Volatile Organic Compounds (VOCs) in Ozone (O-3) Polluted Atmospheres: The Ecological Effects. Journal of Chemical Ecology 36(1): 22-34. Pitzschke A, Hirt H. 2009. Disentangling the complexity of mitogen-activated protein kinases and reactive oxygen species signaling. Plant Physiol 149(2): 606-615.

Poelman EH, Galiart RJFH, Raaijmakers CE, van Loon JJA, van Dam NM. 2008. Performance of specialist and generalist herbivores feeding on cabbage cultivars is not explained by glucosinolate profiles. Entomologia Experimentalis et Applicata 127(3): 218-228. Ponzio C, Cascone P, Cusumano A, Weldegergis BT, Fatouros NE, Guerrieri E, Dicke M, Gols R. 2016. Volatile-mediated foraging behaviour of three parasitoid species under conditions of dual insect herbivore attack. Animal Behaviour 111: 197-206. Ponzio C, Gols R, Weldegergis BT, Dicke M. 2014. Caterpillar-induced plant volatiles remain a reliable signal for foraging wasps during dual attack with a plant pathogen or non-host insect herbivore. Plant Cell and Environment 37(8): 1924-1935. Potuschak T, Lechner E, Parmentier Y, Yanagisawa S, Grava S, Koncz C, Genschik P. 2003. EIN3-dependent regulation of plant ethylene hormone signaling by two Arabidopsis F box proteins: EBF1 and EBF2. Cell 115(6): 679-689. Prakash S, Hinata K. 1980. Taxonomy, cytogenetics and origin of crop brassicas, a review. Opera Bot 55:1–57. Lund Botanical Society. Quettier AL, Shaw E, Eastmond PJ. 2008. SUGAR-DEPENDENT6 encodes a mitochondrial flavin adenine dinucleotide-dependent glycerol-3-p dehydrogenase, which is required for glycerol catabolism and post germinative seedling growth in Arabidopsis. Plant Physiol 148(1): 519-528. Rask L, Andreasson E, Ekbom B, Eriksson S, Pontoppidan B, Meijer J. 2000. Myrosinase: gene family evolution and herbivore defense in Brassicaceae. Plant Molecular Biology 42(1): 93-113. Rejeb IB, Pastor V, Mauch-Mani B. 2014. Plant Responses to Simultaneous Biotic and Abiotic Stress: Molecular Mechanisms. Plants (Basel) 3(4): 458-475. Renner SS. 2006. Rewardless flowers in the Angiosperms and the role of insect cognition in their evolution. In: Plant–pollinator interactions: from specialization to generalization. 123-144. Edited by Waser NM, Ollerton J. Chicago, IL, USA: University of Chicago Press, Reymond P, Bodenhausen N, Van Poecke RM, Krishnamurthy V, Dicke M, Farmer EE. 2004. A conserved transcript pattern in response to a specialist and a generalist herbivore. Plant Cell 16(11): 3132-3147. Reymond P, Weber H, Damond M, Farmer EE. 2000. Differential gene expression in response to mechanical wounding and insect feeding in Arabidopsis. Plant Cell 12(5): 707-720. Rizhsky L, Liang H, Mittler R. 2002. The combined effect of drought stress and heat shock on gene expression in tobacco. Plant Physiol 130(3): 1143-1151. Rizhsky L, Liang H, Shuman J, Shulaev V, Davletova S, Mittler R. 2004. When defense pathways collide. The response of Arabidopsis to a combination of drought and heat stress. Plant Physiol 134(4): 1683- 1696. Rodriguez-Saona C, Crafts-Brander SJ, Pare PW, Henneberry TJ. 2001. Exogenous methyl jasmonate induces volatile emissions in cotton plants. Journal of Chemical Ecology 27(4): 679-695. Sakaki T, Saito K, Kawaguchi A, Kondo N, Yamada M. 1990. Conversion of monogalactosyldiacylglycerols to triacylglycerols in ozone-fumigated spinach leaves. Plant Physiol 94(2): 766-772. Salvatori E, Fusaro L, Strasser RJ, Bussotti F, Manes F. 2015. Effects of acute O3 stress on PSII and PSI photochemistry of sensitive and resistant snap bean genotypes (Phaseolus vulgaris L.), probed by prompt chlorophyll "a" fluorescence and 820 nm modulated reflectance. Plant Physiol Biochem 97: 368-377. Sana TR, Fischer S, Wohlgemuth G, Katrekar A, Jung KH, Ronald PC, Fiehn O.2010. Metabolomic and transcriptomic analysis of the rice response to the bacterial blight pathogen Xanthomonas oryzae pv. oryzae. Metabolomics. 6(3):451-465. Sardans J, Peñuelas J, Rivas-Ubach A. 2011. Ecological metabolomics: overview of current developments and future challenges. Chemoecology 21(4): 191-225. Sauve AMC, Thebault E, Pocock MJO, Fontaine C. 2016. How plants connect pollination and herbivory networks and their contribution to community stability. Ecology 97(4): 908-917. Schaller A, Stintzi A. 2009. Enzymes in jasmonate biosynthesis - Structure, function, regulation. Phytochemistry 70(13-14): 1532-1538. Schön M, Töller A, Diezel C, Roth C, Westphal L, Wiermer M, Somssich IE. 2013. Analyses of wrky18 wrky40 Plants Reveal Critical Roles of SA/EDS1 Signaling and Indole-Glucosinolate Biosynthesis for Golovinomyces orontii Resistance and a Loss-of Resistance Towards Pseudomonas syringae pv. tomato AvrRPS4. Molecular Plant-Microbe Interactions 26(7): 758-767. Schwachtje J, Baldwin IT. 2008. Why does herbivore attack reconfigure primary metabolism? Plant physiology 146(3): 845-851. Schweizer F, Fernandez-Calvo P, Zander M, Diez-Diaz M, Fonseca S, Glauser G, Lewsey MG, Ecker JR, Solano R, Reymond P. 2013. Arabidopsis Basic Helix-Loop-Helix Transcription Factors MYC2, MYC3, and

MYC4 Regulate Glucosinolate Biosynthesis, Insect Performance, and Feeding Behavior. Plant Cell 25(8): 3117-3132. Sewelam N, Oshima Y, Mitsuda N, Ohme-Takagi M. 2014. A step towards understanding plant responses to multiple environmental stresses: a genome-wide study. Plant Cell and Environment 37(9): 2024-2035. Shen W, Wei Y, Dauk M, Tan Y, Taylor DC, Selvaraj G, Zou J. 2006. Involvement of a glycerol-3-phosphate dehydrogenase in modulating the NADH/NAD+ ratio provides evidence of a mitochondrial glycerol-3- phosphate shuttle in Arabidopsis. Plant Cell 18(2): 422-441. Singh V, Louis J, Ayre BG, Reese JC, Shah J. 2011. TREHALOSE PHOSPHATE SYNTHASE11-dependent trehalose metabolism promotes Arabidopsis thaliana defense against the phloem-feeding insect Myzus persicae. Plant Journal 67(1): 94-104. Singh V, Shah J. 2012. Tomato responds to green peach aphid infestation with the activation of trehalose metabolism and starch accumulation. Plant Signal Behav 7(6): 605-607. Sitch S, Cox PM, Collins WJ, Huntingford C. 2007. Indirect radiative forcing of climate change through ozone effects on the land-carbon sink. Nature 448(7155): 791-794. Slansky F, Feeny P. 1977. Stabilization of Rate of Nitrogen Accumulation by Larvae of Cabbage Butterfly on Wild and Cultivated Food Plants. Ecological Monographs 47(2): 209-228. Smallegange RC, van Loon JJA, Blatt SE, Harvey JA, Agerbirk N, Dicke M. 2007. Flower vs. leaf feeding by Pieris brassicae: Glucosinolate-rich flower tissues are preferred and sustain higher growth rate. Journal of Chemical Ecology 33(10): 1831-1844. Smirnoff N. 2011. Vitamin C: The Metabolism and Functions of Ascorbic Acid in Plants. Biosynthesis of Vitamins in Plants: Vitamins B6, B8, B9, C, E, K - Pt B 59: 107-177. Soler R, Badenes-Perez FR, Broekgaarden C, Zheng SJ, David A, Boland W, Dicke M. 2012. Plant-mediated facilitation between a leaf-feeding and a phloem-feeding insect in a brassicaceous plant: from insect performance to gene transcription. Functional Ecology 26(1): 156-166. Song S, Huang H, Gao H, Wang J, Wu D, Liu X, Yang S, Zhai Q, Li C, Qi T, et al. 2014. Interaction between MYC2 and ETHYLENE INSENSITIVE3 modulates antagonism between jasmonate and ethylene signaling in Arabidopsis. Plant Cell 26(1): 263-279. Stamp N. 2003. Out Of The Quagmire Of Plant Defense Hypotheses. The Quarterly Review of Biology 78(1): 23-55. Steffen W, Grinevald J, Crutzen P, McNeill J. 2011. The Anthropocene: conceptual and historical perspectives. Phil Transact of the Royal Society a-Mathem Phys and Eng Scie 369(1938): 842-867. Steinbrenner AD, Gomez S, Osorio S, Fernie AR, Orians CM. 2011. Herbivore-Induced Changes in Tomato (Solanum lycopersicum) Primary Metabolism: A Whole Plant Perspective. Journal of Chemical Ecology 37(12): 1294-1303. Stephanopoulos GN, Aristidou AA, Nielsen J.1998. Review of Cellular Metabolism. In Metabolic Engineering, Principles and Methodologies, 21-79. Elsevier. Stephens AEA, Westoby M. 2015. Effects of insect attack to stems on plant survival, growth, reproduction and photosynthesis. Oikos 124(3): 266-273. Stork NE, McBroom J, Gely C, Hamilton AJ. 2015. New approaches narrow global species estimates for beetles, insects, and terrestrial arthropods. PNAS 112(24): 7519-7523. Sun C, Gao X, Fu J, Zhou J, Wu X. 2015. Metabolic response of maize (Zea mays L.) plants to combined drought and salt stress. Plant and Soil, 388 (1): 99-117. Suzuki N, Rivero RM, Shulaev V, Blumwald E, Mittler R. 2014. Abiotic and biotic stress combinations. New Phytol 203(1): 32-43. Sweetlove LJ, Beard KFM, Nunes-Nesi A. 2010. Not just a circle: flux modes in the plant TCA cycle. Trends in Plant Science 15 (8): 462-70. Thaler JS, Humphrey PT, Whiteman NK. 2012. Evolution of jasmonate and salicylate signal crosstalk. Trends Plant Sci 17(5): 260-270. Tiwari S, Grote R, Churkina G, Butler T. 2016. Ozone damage, detoxification and the role of isoprenoids - new impetus for integrated models. Functional Plant Biology 43(4): 324-336. Traw MB, Dawson TE. 2002. Reduced performance of two specialist herbivores (Lepidoptera : Pieridae, Coleoptera : Chrysomelidae) on new leaves of damaged black mustard plants. Environmental Entomology 31(4): 714-722. Tweeddale H, Notley-McRobb L, Ferenci T. 1998. Effect of slow growth on metabolism of Escherichia coli, as revealed by global metabolite pool ("Metabolome") analysis. Journal of Bacteriology 180(19): 5109- 5116. Vahisalu T, Puzorjova I, Brosche M, Valk E, Lepiku M, Moldau H, Pechter P, Wang YS, et al. 2010. Ozone- triggered rapid stomatal response involves the production of reactive oxygen species, and is controlled by SLAC1 and OST1. Plant Journal 62(3): 442-453. Vainonen JP, Kangasjarvi J. 2015. Plant signalling in acute ozone exposure. Plant Cell Environ 38(2): 240-252.

van Dam NM, Oomen MW. 2008. Root and shoot jasmonic acid applications differentially affect leaf chemistry and herbivore growth. Plant Signal Behav 3(2): 91-98. van Loon JJA, Schoonhoven LM. 1999. Specialist deterrent chemoreceptors enable Pieris caterpillars to discriminate between chemically different deterrents. Entomologia Experimentalis et Applicata 91(1): 29-35. Verhage A, Vlaardingerbroek I, Raaymakers C, Van Dam NM, Dicke M, Van Wees SC, Pieterse CM. 2011. Rewiring of the Jasmonate Signaling Pathway in Arabidopsis during Insect Herbivory. Front Plant Sci 2: 47. Viant MR, Kurland IJ, Jones MR, Dunn WB. 2017. How close are we to complete annotation of metabolomes? Current Opinion in Chemical Biology 36: 64-69. Walling LL. 2008. Avoiding effective defenses: Strategies employed by phloem-feeding insects. Plant physiology 146(3): 859-866. Wasternack C. 2015. How Jasmonates Earned their Laurels: Past and Present. Journal of Plant Growth Regulation 34(4): 761-794. Weckwerth W. 2011. Green systems biology - From single genomes, proteomes and metabolomes to ecosystems research and biotechnology. J Proteomics 75(1): 284-305. Wei HR, Persson S, Mehta T, Srinivasasainagendra V, Chen L, Page GP, Somerville C, Loraine A. 2006. Transcriptional coordination of the metabolic network in Arabidopsis. Plant physiology 142(2): 762- 774. Weng JK. 2014. The evolutionary paths towards complexity: a metabolic perspective. New Phytologist 201(4): 1141-1149. Wennberg PO, Dabdub D. 2008. Atmospheric chemistry. Rethinking ozone production. Science 319(5870): 1624-1625. Westman AL,Kresovich S. 1999. Simple sequence repeat (SSR)-based marker variation in Brassica nigra genebank accessions and weed populations. Euphytica 109: 85–92. Wheeler GS, Halpern MD. 1999. Compensatory responses of Samea multiplicalis larvae when fed leaves of different fertilization levels of the aquatic weed Pistia stratiotes. Entomologia Experimentalis et Applicata 92(2): 205-216. Winde I, Wittstock U. 2011. Insect herbivore counteradaptations to the plant glucosinolate–myrosinase system. Phytochemistry 72(13): 1566-1575. Wink M. 1997. Compartmentation of secondary metabolites and xenobiotics in plant vacuoles. Advances in Botanical Research Incorporating Advances in Plant Pathology, Vol 25 25: 141-169. Xu CC, Fan JL, Cornish AJ, Benning C. 2008. Lipid trafficking between the endoplasmic reticulum and the plastid in Arabidopsis requires the extraplastidic TGD4 protein. Plant Cell 20(8): 2190-2204. Yao I, Akimoto S. 2001. Ant attendance changes the sugar composition of the honeydew of the drepanosiphid aphid Tuberculatus quercicola. Oecologia 128(1): 36-43. Yi GE, Robin AHK, Yang K, Park JI, Hwang BH, Nou IS. 2016. Exogenous Methyl Jasmonate and Salicylic Acid Induce Subspecies-Specific Patterns of Glucosinolate Accumulation and Gene Expression in Brassica oleracea L. Molecules 21(10). Yonekura-Sakakibara K, Tohge T, Matsuda F, Nakabayashi R, Takayama H, Niida R, Watanabe-Takahashi A, Inoue E, Saito K. 2008. Comprehensive flavonol profiling and transcriptome coexpression analysis leading to decoding gene-metabolite correlations in Arabidopsis. Plant Cell 20(8): 2160-2176. Yukimune Y, Hara Y, Nomura E, Seto H, Yoshida S. 2000. The configuration of methyl jasmonate affects paclitaxel and baccatin III production in Taxus cells. Phytochemistry 54(1): 13-17. Zandalinas SI, Mittler R, Balfagon D, Arbona V, Gomez-Cadenas A. 2017. Plant adaptations to the combination of drought and high temperatures. Physiol Plant. Zang YX, Zhang H, Huang LH, Wang F, Gao F, Lv XS, Yang J, Zhu B, Hong SB, Zhu ZJ. 2015. Glucosinolate enhancement in leaves and roots of pak choi (Brassica rapa ssp chinensis) by methyl jasmonate. Horticulture Environment and Biotechnology 56(6): 830-840. Zangerl AR, Hamilton JG, Miller TJ, Crofts AR, Oxborough K, Berenbaum MR, de Lucia EH. 2002. Impact of folivory on photosynthesis is greater than the sum of its holes. PNAS 99(2): 1088-1091. Zavala JA, Baldwin IT. 2004. Fitness benefits of trypsin proteinase inhibitor expression in Nicotiana attenuata are greater than their costs when plants are attacked. BMC Ecol 4: 11.

Websites and software

AraNet (v2). Database for probabilistic co-functional gene network association for the model species Arabidopsis thaliana. http://www.inetbio.org/aranet/

Cytoscape. Open platform for visualizationof complex networks. http://www.cytoscape.org/ ePlant. Visualization tool for multiple levels of plant data. https://bar.utoronto.ca/eplant/

GeneMANIA. Database for gene-sets functional association network analyses. http://genemania.org/

KaPPA-View4. Database for representation and pathway analysis of gene and metabolite correlation networks. http://kpv.kazusa.or.jp/

Kyoto Encyclopedia of Genes and Genomes (KEGG). Database for understanding of high-level functions in biological system, such as the cell, the organism and the ecosystem. http://www.genome.jp/kegg/

Mapman. Visualization of expression profiling data sets. http://mapman.gabipd.org/home

The Arabidopsis Information Resource (TAIR). Database of genetic and molecular biology data for the model plant species Arabidopsis thaliana. https://www.arabidopsis.org/