The Pennsylvania State University the Graduate School Intercollege

The Pennsylvania State University

The Graduate School

Intercollege Graduate Program in Plant Biology

MAIZE DEFENSE RESPONSES AND STRATEGIES AGAINST ABOVEGROUND AND

BELOWGROUND INSECTS

A Dissertation in

Plant Biology

Lina Castano-Duque

Submitted in Partial Fulfillment of the Requirements

for the Degree of

Doctor of Philosophy

May 2017

The dissertation of Lina Castano-Duque was reviewed and approved* by the following:

Dawn S. Luthe

Professor of Plant Stress Biology

Dissertation Adviser

Kathleen M. Brown

Professor of Plant Stress Biology

Réka Albert

Distinguished Professor of Physics and Biology

Mary Barbercheck

Professor of Entomology

Committee Chair

Teh-Hui Kao

Distinguished Professor in Biochemistry and Molecular Biology

Graduate Program Chair

*Signatures are on file in the Graduate School.

Abstract

This dissertation had three main projects. In the first project we studied insect resistance traits against the root herbivore, western corn rootworm (WCR, Diabrotica virgifera) in the insect resistant maize genotype, Mp708 and its susceptible parent, Tx601 using biomechanical, root anatomy, molecular and biochemical analyses. These two maize genotypes differ in susceptibility to fall armyworm (FAW, Spodoptera frugiperda), WCR, and corn leaf aphid (CLA,

Rhopalosiphum maidis). The data suggested that Mp708 is more resistant to WCR than Tx601 due to strong nodal roots, constant root growth under infestation, constitutive and induced expression of ribosomal protein 2 (rip2), terpene synthase 23 (tps23) and maize insect resistant cysteine protease-1 (mir1), as well high constitutive and inducible levels of MIR1-CP protein, jasmonic acid (JA) and caryophyllene. We concluded that Mp708 showed resistance and antibiosis against

WCR and could be used as a model to explore the wide variety of mechanisms and traits involved in plant defense responses.

In the second project we explored the use of innovative proteomic and network analyses to understand the global defenses deployed by maize plants infested with the root-feeder, WCR.

Using the relatively new technique that employs 10-plex tandem mass spectrometry tags (TMT), we measured protein abundance changes in roots and leaves in response to these pest because metabolic changes in response to herbivory include a wide variety of mechanisms that enable the plant to survive. TMT data were analyzed by using a protein co-abundance and protein-protein interaction (PPI) networks. We detected 4878 proteins of which 863 had significant changes under

WCR infestation. Proteins with higher abundance during WCR infestation were involved in the

JA pathway and included lipoxygenase 5 (lox5), allene oxidase synthase 1 (aos1), and 12- oxophytodienoate reductase 2 (opr2). Other high abundance proteins during infestations were part

iii of biosynthesis and signaling pathways for proteases of serine/cysteine type, abscisic acid, reactive oxygen species, and ethylene (ET). We validated changes in these pathways by analyzing the expression of key genes in each, total protease activity, cysteine and serine-type protease inhibition, JA and ET production. Our data showed promising differences between tissue-specific defense responses in leaves and roots of plants infested by WCR. Also, we suggested a dual insect and microbial plant stress response and postulated that roots infested with WCR are under constant microbial pressure due to the interactions in rhizosphere.

In the third project we studied Mp708 tissue-specific responses to FAW by using the same proteomics and analysis techniques used for the second project. In addition we merged our proteomics network data with QTL known regions of resistance in Mp708 against FAW. We detected 4675 proteins of which 794 had significant changes under FAW infestation. Our analyses suggested an increase in the JA biosynthesis, REDOX changes in the glutathione and ascorbate pathways, and increase in ABA and ET biosynthesis and signaling. Infested leaf tissues showed high abundance and enrichment of proteinase inhibitors, cysteine proteases and peroxidases, these are well known plant defense proteins. Validations of total peroxidase activity showed an early activity spike in leaves of Mp708 and no change in the roots during FAW infestation , suggesting an early ROS signaling event that might not involve peroxide or an increase in quinone production mediated by peroxidases. Gene expression and phytohormone analyses showed that JA biosynthesis was activated in leaves and ET production increased only in roots, furthermore, JA and ET appears to control local MIR1-CP and RIP2 accumulation, and

ET could be a key regulator of MIR1-CP systemic accumulation. These findings open up new routes in the hormonal control of local and systemic defense signaling that could be similar for other defense proteins and secondary compounds. Using the protein-correlation network with the

iv known QTL regions, we found some proteins that might be part of the defense responses and tolerance against FAW that are involved in plant defenses, development and growth. Growth analyses determined that the insect susceptible maize genotype, Tx601, grew less in height and total root length during FAW infestation. We concluded that Mp708 defense mechanisms could involve lower trade-offs between plant growth and defense responses.

We have established an analysis pipeline for proteomics data that includes network biology approaches that can be used with different types of “omics” data from a wide variety of organisms to detect tissue-specific defense responses.

TABLE OF CONTENTS

List of Tables……………………………………………………………………………………viii

List of Figures…………………………………………………………………………………. ...ix

List of multimedia items………………………………………………………………………….xi

Acknowledgements………………………………………………………………………………xii

CHAPTER I. Introduction...... …………………………………………………………………. 1

CHAPTER II. A maize inbred exhibits innate resistance to the western corn rootworm,

Diabrotica virgifera……………………………………………………….………………………5

Introduction…………………………………………………….………………………………….5

Materials and methods……………………………………………………….……………………9

Results……………………………………………………….…………………………………...15

Discussion……………………………………………………….……………………………….20

CHAPTER III. Plant Bio-Wars: Protein Networks Reveal Tissue-Specific Defense Strategies in

Maize in Response to a Root Herbivore…………………………………………………………37

Introduction…………………………………………………….………………………………...37

Materials and methods……………………………………………………….…………………..42

Results and Discussion………………………………………………….……………………….54

Conclusions……………………………………………………….……………………………71

CHAPTER IV. Protein Networks Reveal Tissue Specific Defense Strategies in Maize in response to an aboveground herbivore…………………………………………………………………….96

Introduction…………………………………………………….………………………………...96

Materials and methods……………………………………………………….…………………..99

Results and Discussion………………………………………………….……………………...111

Conclusions……………………………………………………….…………………………….129

CHAPTER V. Conclusions…………………………………….……………………………….163

Bibliography…………………………………….………………………………………………………167

vii

Lists of Tables

Table 3.1. Proteins of interest from correlation network hotspot………………………………83 Table S3.1. Proteins identified from WCR TMT……...……………………………………….95 Table S3.2. Proteins present in the hierarchical clustering…………………………………….95 Table S3.3. Proteins in the abundance correlation network……………………………………95 Table S3.4. Proteins in the protein-protein interaction network……………………………….95

Table 4.1. Proteins that are part of the gene ontology (GO) term: “enzyme activity” ……….136 Table 4.2. Proteins that are part of the gene ontology (GO) term: “response to stimuli” ……137 Table 4.3. Proteins that belong to the subnetwork created from the protein correlation network, that fall within a known QTL for FAW……...………………………………………………...144 Table 4.4. Proteins of interest from the Hub from the protein abundance correlation network.147 Table S4.1. Identified proteins present in FAW TMT…………………………………………161 Table S4.2. Proteins and gene ontology annotation from the hierarchical clustering analysis...161 Table S4.3. Proteins in the abundance correlation network……………………………………161 Table S4.4. Proteins in the protein-protein interaction network……………………………….162

viii

List of Figures

Figure 2.1. Root length in Mp708 and Tx601 infestated with WCR……………………………26

Figure 2.2. Laser ablation tomography (LAT) cross-sections…………………………………..28

Figure 2.3. Time course analysis of aos and opr7………………………………………………30

Figure 2.4. Analysis of constitutive jasmonic acid (JA)..………………………………………31

Figure 2.5. Time course of rip2 and mpi...... ………………………………………32

Figure 2.6. Time course analysis of fpps3 and tps23.....………………………………………...33

Figure 2.7. Time course of mir1 transcript and MIR1-CP protein……………………………...34

Figure S2.1. Percent survival of WCR fed on Mp708 and Tx601 maize lines………………….35

Figure S2.2. Maximum cutting strength in lateral roots………………………………………...36

Figure 3.1. Hierarchical clustering of protein abundance ratios of Mp708…………………….74 Figure 3.2. Gene ontology (GO) hierarchical maps……………………………………………76 Figure 3.3. Gene ontology annotation that belong to Group 5………………………………..78 Figure 3.4. Betweenness centrality distribution of the correlation network.…………………..80 Figure 3.5. Hotspots from protein correlation and PPI networks……………………………...81 Figure 3.6. Node degree distribution of the PPI network………………………………………84 Figure 3.7. Gene ontology of PPI network.…………………………………………………….85 Figure 3.8. Analysis of salicylic acid (SA) and ethylene (ET) levels…………………………..87 Figure 3.9. Analysis of protease and inhibitors activity in leaves and roots……………………89 Figure 3.10. Analysis of benzothiazol volatile production from roots………………………….91 Figure 3.11. Summary of defense related responses in the maize……………………………..92

Figure 4.1. Hierarchical clustering of protein abundance ratios of Mp708…………………132 Figure 4.2. Parametric analysis of gene set enrichment (PAGE) …...………………………134 Figure 4.3. Singular gene ontology (GO) enrichment analysis…...…………………………141

Figure 4.4. Gene ontology enrichment of the proteins in the correlation network..…………143 Figure 4.5. Protein subnetworks from the protein correlation network…………...…………146 Figure 4.6. Gene ontology enrichment of the protein-protein interaction (PPI) network……150 Figure 4.7. Protein subnetworks from the protein-protein interaction (PPI) network………..151 Figure 4.8. Plant physiological changes measured in Mp708 and Tx601……………………153 Figure 4.9. Percent of carbon in leaves and roots of Mp708 and Tx601……………………..155 Figure 4.10. lox3 gene expression and ET accumulation in Mp708………………………….156 Figure 4.11. Gene expression and immunoblot Mir1-CP in Mp708…………………………158

Figure 4.12. Time course of peroxidase activity assay………………………………………..160

Multimedia Items

Note S3.1. Protein abundance correlation network from WCR infested plants…………………95 Note S3.2. Protein-protein interaction network from WCR infested plants……………………..95

Note S4.1. Protein abundance correlation network from FAW infested plants………………..161 Note S4.2. Protein-protein interaction network from FAW infested plants……………………162

Acknowledgments

There is a lot of people I want to thank for all their help and support that allowed me to keep going despite all the hardships that I faced through my Ph.D. research. I would like to thank all the members of my committee for their constructive criticism through the last seven years of research and to the Plant Biology program at Penn State for their support. I would like to thank my family, for making me a strong person who believes that a woman should be entitle to have an opinion and a goal beyond conventional gender roles. I want to thank my friends for being there when I needed a hug or a reminder that I can do better. I want to thank all the people that through several circumstances led me to become stronger, more resilient and assertive than ever.

Por ultimo quiero agradecer a la vida, porque con cada leccion llena de dolor, tristesa y desilucion he aprendido que soy mas fuerte de lo que pensaba, y que puedo llegar tan alto como me sea fisicamente possible.

To Dr. K. Anderson, and Dr. House, I owe you my life.

To Dr. Luthe, I will never forget: “Courage doesn't always roar. Sometimes courage is the little voice at the end of the day that says I'll try again tomorrow.”― Mary Anne Radmacher

xii

CHAPTER I. Introduction

There is a war between plants and insects, and like any war there is an art to it that can be described by tactics and strategies where, “all men can see these tactics whereby I conquer, but what none can see is the strategy out of which victory is evolved” -Sun Tzu. To understand plant defense strategies against insects at a whole organism level it is necessary to consider plant local and long distance defense responses in a tissue-specific way by using multi-level data analyses.

In this bio-war there is an overall communication strategy, first, the plant recognizes that is under attack leading to a local response that results in the deployment of local defenses. Then the plant might start a long distance communication event called systemic signaling, leading to the deployment of defenses in those systemic tissues (Schilmiller & Howe, 2005). The plant might keep its defenses high in case of future attack or shut them down if they are no longer needed

(Van Dam & Bezemer, 2006). An example of a maize insect defense response that is produced in the roots and used in the leaves is the maize insect resistant cysteine protease (Mir1-CP) (Lopez et al., 2007). The Mir1-CP war story started with an analysis of maize inbred lines, which showed that certain lines deterred the feeding and growth of the leaf feeding insect, fall armyworm (FAW, Spodoptera frugiperda) (Williams et al., 1990) due to Mir1-CP damage to the peritropic matrix of the insect guts (Mohan et al., 2006). Taken together these studies strongly suggested that there is above to belowground communication in maize.

Besides Mir1-CP, there are other examples at physiological level that show how plants can alter and mobilize their cellular content to avoid insect damage. During aboveground insect infestation, carbon, nitrogen and nutrients can reallocate. Carbon and nitrogen are building blocks for plant growth and defense molecules such as insecticidal proteins, phenolics and alkaloids that act as insect deterrents (Lincoln, 1993; Frost & Hunter, 2008; Kant et al., 2015;

Zhou et al., 2015). For example, Nicotiana attenuata infested aboveground with Manduca sexta showed decreased leaf and root carbohydrate levels (Machado et al., 2013). In oak seedlings, aboveground infestation by Orgyia leucostigma reduced carbon allocation to fine roots by 63% and nitrogen by 39% (Frost & Hunter, 2008). These changes in C and N distribution could be associated with defense compound production and result in reduced plant growth (Kant et al.,

2015).

When roots are under attack, a systemic defense response takes place that potentially affects aboveground herbivores and plant growth due to reallocation of resources. Maize roots severely damaged by western corn rootworm (Diabrotica virgifera virgifera, WCR) have decreased shoot dry weight and a 13.5% decrease in grain yield (Kahler et al., 1985). WCR tends to feed on nodal roots, resulting in an increased number of adventitious roots, decreased CO2 assimilation and decreased leaf area (Riedell & Reese, 1999). Differences in root chemical composition can be associated to WCR preference for nodal roots, which could be related to root antibiosis traits such as lignin and fiber content, that enhances insect resistance (Meihls et al.,

2012). Changes in root system architecture (RSA) in response to environmental changes have implications in whole plant architecture, growth, yield, abiotic stress resistance, nutrient uptake and development. Ecophysiological models recognize that different plant roots vary in their vulnerability to herbivores (van Dam, 2009). Under insect infestations certain types of roots are vital to plant survival and their growth rates change, which will affect the RSA and consequently plant growth and yield. The characterization of genes involved in root development is emerging

(Jung & McCouch, 2013), but the developmental RSA changes and associated gene expression in response to insect infestation are unknown.

The research that will be presented in this dissertation focused on the plant side of the plant-insect war, using maize (Zea mays L.) as a model organism. For the experiments we used the insect resistant genotype derived from Antigua germplasm, Mp708, or its insect susceptible parental genotype Tx601. The insects used for aboveground and belowground infestations were

FAW and WCR, respectively. FAW is an aboveground whorl-feeder and generalist pest of maize in the US and throughout Latin America (Brooks et al., 2007; Eleazar Ruiz-Najera et al., 2007).

WCR is a belowground root-feeder and specialist pest in North America. Although WCR has been controlled by crop rotation with soybean or transgenic maize expressing Bt-Cry3Bb1,(Gray et al., 2009) these control methods are becoming less effective (Gassmann, 2012). Most of the breeding programs looking for WCR resistance search for the phenotype of reduced lodging, but this is an indirect way to assess resistance since root antibiosis or toxicity to WCR is not measured. Therefore, this type of screening leads to tolerant lines and not antibiosis. We investigated maize defense responses and strategies to above and belowground herbivory by measuring herbivory effects on growth, carbon allocation, root phenotypes, phytohormone content, proteome or gene expression changes, enzyme activity and specific volatile production.

There is little information regarding intra-plant communication, resource allocation and the morphological, anatomical, biochemical and molecular basis of plant-insect interactions

(Machado et al., 2013; Soler et al., 2013) at the whole plant scale during FAW or WCR infestations of maize genotypes with known levels of resistance to chewing insect FAW

(Williams et al., 1985; Williams et al., 1990), root-feeding insect WCR (Gill et al., 2011), and phloem-sucking insect, corn leaf aphid (CLA, Rhopalosiphum maidis, (Louis et al., 2015).

Research on intra-plant communication, resource allocation and the morphological, anatomical, biochemical and molecular basis of plant-insect interactions is indispensable for plant breeding

3 programs (Jogaiah et al., 2012) that focus on the development of environmentally sustainable resistance to these herbivores.

The following hypothesis was tested in this project: maize has both local and systemic response pathways in leaves and roots that respond, communicate and defend the plant from aboveground and belowground herbivore pests.

Objective 1.1 Determine morphological and biochemical differences in maize defense response against WCR-belowground infestation and their potential effects on insects.

Objective 2.1 Determine the key elements of maize resistance during WCR-belowground infestation in roots and leaves using high throughput proteomics and network biology analysis.

Objective 2.2 Confirm the potential role of the pathways and proteins of interest in plant systemic defenses using other molecular and biochemical tools during WCR-belowground infestation.

Objective 3.1 Determine morphological and biochemical differences in maize defenses and their potential effects on insects during FAW-aboveground infestation.

Objective 3.2 Determine the key elements of maize resistance during FAW-aboveground infestation in roots and leaves using high throughput proteomics and network biology analysis.

Objective 3.3 Confirm the potential role of the pathways and proteins of interest in plant systemic defenses using other molecular and biochemical tools during FAW-belowground infestation.

CHAPTER II. A maize inbred exhibits innate resistance to the western corn rootworm,

Diabrotica virgifera

Introduction

Plant defense responses against root-feeding insect pests are a critical research topic in the area of plant-herbivore interactions. However, due to the difficulty of studying herbivore resistance mechanisms in the rhizosphere, far less is known about belowground than aboveground defense responses, nevertheless, characterization of plant resistance against belowground pests is vital. In general, there are three mechanisms of plant resistance to herbivory, these include 1] non-preference, 2] antibiosis and 3] tolerance (Painter, 1951; Painter,

1958). Non-preference mechanisms deter insect feeding and oviposition, antibiosis impairs herbivore performance, and tolerance is the ability of the plant to withstand and recover from herbivory (Painter, 1951). Although it is difficult to distinguish among these mechanisms when studying roots, it is clear that plant resistance to root-feeding pests involves a wide range of defense responses that are most likely linked to a combination of physical traits, signaling pathways and deterrent biomolecules and the ability to tolerate herbivory and or regrow after insect attack (Rasmann & Agrawal, 2008).

One model system for studying defenses to root herbivory is the important and widely distributed crop, maize (Zea mays L.) that is attacked by the western corn rootworm (WCR,

Diabrotica virgifera). WCR is a specialist pest of maize in North America (Hummel, 2003;

Brooks et al., 2007; Eleazar Ruiz-Najera et al., 2007) that causes economic losses in maize production in North America and Europe (Hummel, 2007; Gray et al., 2009; Tinsley et al., 2013;

Flagel et al., 2015). It has been estimated WCR is responsible for more than $1 billion of losses and pest control expenses in maize annually (Gray et al., 2009; Tinsley et al., 2016). For a

5 number of years it has been possible to control this insect with insecticides, crop rotation with soybeans (Levine et al, 2002), and transgenic Bt-maize. Unfortunately, some WCR populations have become resistant to these management strategies (Meinke et al., 1998; Levine et al., 2002;

Gassmann, 2012; Flagel et al., 2015) and finding additional sources of sustainable resistance is essential.

Exploiting innate or native defense tactics of maize is another strategy for identifying sustainable resistance traits that could be incorporated into hybrid maize. One potential defense against WCR that has been investigated is tolerance or compensatory growth in response to herbivore attack (Prischmann et al., 2007; Robert et al., 2014; Robert et al., 2015; Qu et al.,

2016). The reallocation of photosynthate between above and belowground organs appears to play an important role in tolerance (Robert et al., 2014; Robert et al., 2015). Indirect defenses also assist in protecting maize from WCR attack. It has been shown that aboveground feeding by

Spodoptera littoralis enhances the accumulation of the sesquiterpene, (E)-β-caryophyllene, which attracts WCR larvae (Robert, Christelle A. M. et al., 2012; Robert, C. A. M. et al., 2012) and natural enemies of WCR to maize roots (Rasmann et al., 2005; Kollner et al., 2008). It has been reported that the ability to synthesize and release this volatile has been lost from many

American maize lines (Degenhardt et al., 2009).

Despite research efforts over 60 years, as of 2009 there were no commercial maize hybrids with innate host plant resistance (El Khishen et al., 2009; Ivezic et al., 2009). There have been some reports that elevated levels of hydroxamic acids cause antibiosis (Assabgui et al.,

1993; Assabgui et al., 1995), but it also has been reported that the benzoxazinoid, DIMBOA attracts WCR to their feeding site in the nodal roots (Robert, Christelle A. M. et al., 2012). An evaluation of several maize genotypes for WCR resistance identified two genotypes SUM2162

6 and SUM2068 that appear to cause antibiosis in WCR, but the specific mechanism of resistance is unknown (El Khishen et al., 2009). Consequently, there is little known about innate host plant resistance traits against WCR in non-transgenic maize inbreds (Branson & Krysan, 1981; Abel et al., 2000; Hummel, 2003; Sappington et al., 2006; Rasmann & Agrawal, 2008).

Maize insect resistance is a variable and multidimensional trait due to crop genetic and phenotypic diversity (Meihls et al., 2012). For example, maize genotypes could be more or less resistant to WCR due to differences in root architecture (Branson et al., 1982), biomechanical strength (Meihls et al., 2012) and biochemical composition (van Dam, 2009). In this study we compared responses to WCR infestation in the insect-resistant maize inbred Mp708 and its susceptible parent, Tx601 (Williams et al., 1985). Mp708 was developed by traditional plant breeding from an Antiguan landrace (Williams et al., 1985; Williams et al., 1987; Williams et al., 1990) and has demonstrated resistance to three distinct feeding guilds of insects that include the chewing insect fall armyworm (FAW, Spodoptera frugiperda) (Williams et al., 1985;

Williams et al., 1990), root-feeding insect WCR (Gill et al., 2011), and phloem-sucking insect, corn leaf aphid (CLA, Rhopalosiphum maidis (Harfouche et al., 2006; Ankala et al., 2013; Louis et al., 2015). Mp708 resistance to FAW and CLA has been linked to acumulation of an insecticidal protease, Maize Insect Resistance 1- Cysteine Protease (MIR1-CP) (Pechan et al.,

2000; Louis et al., 2015), but to date no studies with Mp708 have correlated WCR feeding with insecticidal protein accumulation even though expression of several genes correlated with root hebivory and plant defense responses have been reported for other maize lines (Lawrence et al.,

2012; Lawrence et al., 2013).

MIR1-CP is a protease localized in the phloem that also is abundant in the roots. It rapidly accumulates in the whorl when Mp708 is attacked by FAW (Pechan et al 2000; Lopez et

7 al 2007). However, root removal prior to herbivore infestation reduces its accumulation aboveground in response to FAW feeding (Lopez et al. 2007). MIR1-CP is highly toxic to FAW and induces ruptures in the caterpillar peritrophic matrix, thus preventing the insect from absorbing nutrients (Mohan et al., 2006). It is likely that aphids are exposed to MIR1-CP when they feed on the phloem (Louis et al 2015). Since aphids lack a peritrophic matrix (Silva et al.,

2004), the toxic effect of MIR1-CP on CLA has not been determined, but could be due to its robust proteolytic activity (Mohan et al., 2006). Another defensive protein, RIP2 accumulates at the FAW feeding site in all maize inbreds tested including Mp708 (Chuang et al., 2014). RIP2 is toxic to FAW, but its mode of action as insecticidal protein is unknown (Chuang et al., 2014).

Importantly, no studies to date have determined if MIR1-CP and RIP2 accumulate in maize roots in response to WCR feeding.

In addition to the accumulation of toxic proteins, it has been demonstrated that the aboveground tissues of Mp708 contain elevated levels of jasmonic acid (JA) prior to herbivore attack and it likely that constitutively high JA levels prime Mp708 for subsequent herbivory

(Shivaji et al., 2010). The JA-signaling pathway and JA-family of compounds have been shown to participate in plant defense responses to root herbivory (Ankala et al., 2009; Koo & Howe,

2009; Erb & Glauser, 2010). Activation of the JA pathway during herbivory through JA- isoleucine conjugate (JA-ile) binding to COI1 (coronatine insensitive 1) and the degradation of

JAZ (JA-Zim domain proteins) (Chini et al., 2007; Thines et al., 2007) initiates the synthesis of enzymes involved in JA biosynthesis and accumulation of defensive proteins such as MIR1-CP

(Ankala et al., 2013), RIP2 (Chuang et al., 2014), chitinases, protease inhibitors (Ballare, 2011;

War et al., 2012), along with production of volatiles such as (E)-β-caryophyllene that attracts

WCR as well as its natural enemies (Rasmann et al., 2005; Rasmann & Agrawal, 2008; Robert,

Christelle A. M. et al., 2012; Capra et al., 2015).

To better understand defense response traits of maize to root-feeding herbivores, we need to examine traits of insect-resistant maize genotypes. Thus, the main objective of this study was to characterize resistance in Mp708 by comparing its responses to WCR with an insect susceptible genotype, Tx601. Our results suggest that Mp708 is resistant to WCR due to a suite of defensive mechanisms that include strong nodal roots, compensatory root growth despite infestation, high constitutive and inducible levels of JA, MIR1-CP, RIP2 and the presence of (E)-

β-caryophyllene. We propose that Mp708 is a unique maize genotype that is an excellent model for investigating herbivore defense responses due to its remarkable range of resistance to insects with different feeding behaviors.

Materials and methods

Plants and insects

Seeds for the maize genotypes, insect-resistant Mp708 and its susceptible parent, Tx601

(Williams et al., 1985) were provided by Dr. Paul Williams, USDA-ARS, Mississippi State

University. Plants were grown in Hagerstown Loam in the Plant Science greenhouse at the

Pennsylvania State University in 8 cm x 9 cm pots. Supplemental lightening was used as needed and temperature was kept between 22 and 27ºC. Diapause WCR eggs were obtained from Dr.

Bryan French, USDA-ARS, Brookings, SD and reared for 10 to 12 days on damp paper towels at

25ºC in the dark until hatching.

Insect infestation and tissue collection

Time course experiments were done by infesting each Mp708 and Tx601 plant at the V3 stage

(Ritchie et al., 1998) with ~20 WCR, that have hatched within ~24 hr, for 2, 4, or 7 days. Control plants were not infested with insects (including the 0 hr time point). Roots were cleaned, collected, weighed (0.1g per biological replicate) and stored at -80 °C until further use. In both, gene expression and immunoblot experiments a minimum of three biological replicates per treatment were used; a biological replicate was root tips pooled from two plants.

WCR bioassays

Bioassays were done by using a small brush to placing 20 WCR (~24 h hatched) neonates in the root system, where they were allowed to feed for 4 days. At the end of day 4, the seedlings were individually placed in plastic funnels connected to a vial with 75% (v/v) ethanol. The funnels were placed at room temperature under a constant light source that dried the soil for 7 days after which the number of WCR recovered was determined and percent survival calculated and analyzed by ANOVA in R version 3.2.1.

Root length and anatomy

Mp708 and Tx601 roots were examined to determine total length and anatomical changes after

WCR infestation. After 3, 6 and 9 days of continuous WCR infestation, 10 of each control and infested root systems were collected, cleaned and stored in 75% (v/v) ethanol. Root length was measured by separating all the roots from the base of the mesocotyl and scanning in a flatbed scanner at a resolution of 400 dots per inch (Epson Expression 1680, Seiko Epson Corporation,

Suwa, Japan). The scanned images were analyzed using WinRhizo software (Regent

Instruments, 2008) to determine total root length and the root system was classified into nodal and seminal roots depending on root diameter. Seminal roots had a diameter between 0.5-1 mm and nodal roots between 1-6 mm (Arsenault et al., 1995; Burton et al., 2012). Statistical analysis was performed on the data collected from the WinRhizo software using R statistical software version 3.2.1 (Team, 2015).

To further characterize differences in injury from WCR on Mp708 and Tx601, we used laser ablation tomography (LAT) to determine the regions of the nodal roots that were attacked by WCR after 9 d of infestation.(Chimungu et al., 2014). For LAT, a pulsed laser beam (Avia

7000, 355 nm pulsed laser) was used to ablate root tissue in a camera focal plane as the root segment is advanced with an imaging stage. The cross-section images were taken using a Canon

T3i 399 (Canon Inc. Tokyo, Japan) camera with 5X micro lens (MP-E 65 mm). The images were analyzed using ImageJ software (Schneider et al., 2012) and the amount of cortex lost was determined as the percentage of tissue missing from the estimated undamaged area. Five biological replicates were used per treatment, and one LAT generated image was used per plant.

Root length and percentage of cortex lost were analyzed by doing data transformation until the

Shapiro-Wilk test (Shapiro & Wilk, 1965) confirmed normal distribution, then a multiple-factor analysis of variance (ANOVA) was performed for each day individually, followed by a significant difference (HSD) Tukey pairwise comparison test in R version 3.2.1.

Biomechanical analysis of nodal roots

To measure the ability of Tx601 and Mp708 roots to resist cutting, a mechanical injury similar to

WCR herbivory, the maximum cutting stress (mCF) that the nodal roots could withstand was measured. Cutting tests were performed on nodal roots as these roots are typically targeted by

WCR larvae (Oleson et al., 2005; Hummel, 2007; Kadlicko et al., 2010). Maize seeds were surface sterilized with 2% (v/v) hypochlorite solution and pre-germinated at 20ºC for 48 hr prior to planting. Plants were grown in a controlled environment with a 10/14 hr night/day cycle, at

20oC/24oC respectively, within plastic tubes 300 mm in height and 150 mm diameter. Within each core, 5.5 L of a soil: sand: vermiculite mix (50:25:25 respectively) was packed to a depth of

280 mm. Prior to planting, the cores were saturated with 850 ml of nutrient solution as described in Zhu et al., (2010). Four replicates of Tx601 and Mp708 were harvested 3 weeks after planting when plants reached the V3 developmental stage (Ritchie et al., 1998). Planting was staggered to ensure sufficient time for biomechanical testing at the V3 stage. Each replicate was tested within

48 hours of roots being washed form the growth media eliminating potential confounding errors associated with decomposition. Biomechanical testing was performed using a singled edged razor blade (Ang et al., 2008) with root samples installed over a slotted plastic block and secured in place using elastic bands. Cutting force was recorded using an Instron 5966 universal test frame with a 10N load cell accurate to ± 2.5gms at maximum load. The cutting force was measured during the extension (cutting) phase. Extension was at a rate of 2 mm/min with maximum cutting force calculated as peak load / root cross sectional area. For each 60mm long segment of root multiple cut tests were performed along the axis with a minimum 10 mm between each cut to minimize the risk of influence on the next. Due to root diameter being smaller than the length of the razor blade the following cut test was performed using an unused portion of the blade by moving the root segment. Effectively a new area of the razor blade was used for each test ensuring potential blade blunting did not affect results.

RNA extraction, and quantitative real-time PCR analysis

Leaf and root tips (up to 2cm from root tip) from all root types were ground using a ball-mill tissue grinder (Genogrinder 2000; SpexCentriprep Inc., Metuchen, NJ, U.S.A.) for 2 min at

2,000 strokes/min under liquid nitrogen conditions. RNA was extracted from all ground tissues using TRIzol®-chloroform protocol, and treated with DNase (New England Biosciences) following manufacturer’s instruction. RNA content was measured using a Nanodrop (Thermo

Scientific) and cDNA was made using High Capacity cDNA Reverse Transcription Kit (ABI,

Foster City, CA) following manufacturer instructions. Real-time PCR analyses were done for time course expression experiments by using primers for aos, opr7, mpi, fpps3, tps23, mir1 and rip2 and actin as endogenous control gene aos F: 5’-CAA ACC GAC GAA TTT GAG CA-3’, R:

5’-GGA GGC TCG CAA CAA GTT G -3’; opr7 F: 5’-CCC ATG GCT ACC TCA TCG AT-3’,

R: 5’-CGT CAG TCC GGT CGT TGA T-3’; rip2 F: 5’-GAG ATC CCC GAC ATG AAG GA-

3’, R: 5’-CTG CGC TGC TGC GTT TT-3’; mpi F: 5’-GCG GAT TAT CGC CCT AAC C-3’, R:

5’-CGT CTG GGC GAC GAT GTC-3’; fpps3 F: 5’-CCT GGC TAG TTG TGC AAG CT-3’, R:

5’-GAA AAC AGT TTG GAC TGC CT-3’; tps23 F: 5’-TCA CCC ATG AGT GCC TCA GA-

3’, R: 5’-GTT GAC CGC CCT CTC TAG AAG A-3’; mir1 F: 5’- GAG GGT CTT GTC GTG

TTG AAC TT-3’, R: 5’- GCC ACA CCA TAA CGG ATT AAC TT-3’; actin F: 5’- GGA GCT

CGA GAA TGC CAA GAG CAG-3’, R: 5’- GAC CTC AGG GCA TCT GAA CCT CTC-3’.

The primers were designed using Primer Express software for real-time PCR (version 3.0) (ABI,

Foster City, CA). The PCR conditions used were: step 1: 50°C for 2 min and 95°C for 10 min, step 2: 95°C for 15 sec and 60°C for 1 min for 40 cycles, Step 3: 72°C for 10 min, Step 4: dissociation stage. The relative quantification values were obtained by using ABI 7500 Fast SDS

Software (version 1.4) (ABI, Foster City, CA), and analyzed with the R statistical software

(Team, 2015). The data was analyzed by first doing data transformation until Shapiro-Wilk test

(Shapiro & Wilk, 1965) confirmed normal distribution, then a multiple-factor analysis of variances (ANOVA) was done, followed by a significant difference (HSD) Tukey pairwise comparison test in R version 3.2.1.

Jasmonic acid (JA) quantification

Root tissues were collected as previously described and placed in 2 ml screw-cap FastPrep tubes

(Qbiogene, Carlsbad, CA) containing Zirmil beads (1.1 mm; SEPR Ceramic Beads and Powders,

Mountainside, NJ). Dihydro-jasmonic acid (dhJA) was added to each vial as internal standard

(100 ng) followed by 400 µl of 1-propanol:water:hydrochloric acid (2:1:0.002, v/v) and shaken for 40 s in a FastPrep FP 120 tissue homogenizer. Dichloromethane (1 ml) was added to each sample, followed by shaking for 40s in the homogenizer, and centrifugation at 13,000 x g for 1 min. The bottom dichloromethane and 1-propanol layer was then transferred to a 4 ml glass screw-cap vial and dried under an air stream. Samples were reconstituted in methanol:diethyl ether solution (1:9, v/v) and 2.3 µl of trimethylsilyldiazomethane hexane (Aldrich) were added to each. The vials were then capped and allowed to sit at room temperature for 25 min. Excess trimethylsilyldiazomethane was destroyed by adding 2.3 µl of 2.0 M acetic acid in hexane to each sample (Schmelz et al., 2003b; Schmelz et al., 2004). Finally, the phytohormones were collected by using a vapor phase extractions protocol previously described by Schmelz et al.

(2004). The extracts were run in a gas chromatograph mass spectrometer with electron ionization and identity and quantity of the total JA was determined by comparing the retention times and spectra of the internal standard. The data was analyzed by first doing data transformation until the Shapiro-Wilk test (Shapiro & Wilk, 1965) confirmed normal distribution, then a multiple- factor analysis of variances (ANOVA) was done, followed by a significant difference (HSD)

Tukey pairwise comparison test in R version 3.2.1.

Protein extraction, quantification and immunoblot analysis

Root samples were ground as described above in the RNA extraction steps. Then 500 µl of 2X

Laemmli sodium dodecyl sulfate (SDS) sample buffer (Laemmli, 1970) was added and the tubes were vortexed. Samples were boiled in a heat block for 10 min, vortexed, and centrifuged at

13,000 × g for 5 min. The supernatant containing the dissolved proteins was collected and protein quantification analysis was done using a NI™ (Non-Interfering™) protein assay (G- biosciences, USA) following manufacturer’s instructions and samples were stored at –80°C.

Immunoblot analysis were done by pooling 5 µg of each biological replicate for a total of 15 µg per treatment. Pooled proteins were run in 12% (w/v) polyacrylamide gels using constant voltage

(10 min at 180 volts and 35 min at 200 volts), followed by a semi-dry transfer (Panther Semi-Dry

Electroblotter, Thermo Scientific Owl, MA) onto nitrocellulose at constant current (100 mA per membrane) for 160 min. Incubation in primary antibody (anti-MIR1) was used in a ratio of

1:30,000 (v/v) for root extracts. The secondary anti-rabbit HRP conjugated antibody (Thermo

Fisher Scientific, Rockford, IL) was used in a ratio of 1:30,000 (v/v). Immunodetection was done using chemiluminescence (West Femto Maximum Sensitivity Substrate, Thermo Scientific,

MA). To corroborate even protein loading, separate gels were run and stained with

SimplyBlueTM (Thermo Scientific, MA).

Results

Survival of WCR on Tx601 and Mp708

We evaluated performance of WCR on Tx601 and Mp708 by assessing their survival following

4 d of feeding on plants of each genotype. A significantly higher percentage (35%) of WCR survived when the insects fed upon Tx601 compared to Mp708 (22%, P<0.05) (Fig. S2.1). This

15 finding confirms previous research (Gill et al., 2011) and verifies our use of Mp708 as a model to investigate physiological and biochemical resistance traits.

Changes in root length and growth during WCR infestations

To characterize the root damage caused by WCR feeding, we measured total root length at 3, 6 and 9 d following WCR infestation. Mp708 and Tx601 root lengths were not significantly different in the controls, but when plants were infested with WCR, Tx601 showed 30-40% less total root length than Mp708 (Fig. 2.1a, P<0.05). When the lengths of nodal and lateral roots were measured there were no significant differences in nodal root length between genotypes

(P=0.281) or between infested and control plants (P=0.166). Mp708 had longer lateral roots than

Tx601 in control plants at 3 d and infested treatments throughout the 9 d period (Fig. 1b,

P<0.05). To determine if changes in total root length could be caused by WCR-induced reduction in root growth, we related root length changes with roothairless 3 (rth3) gene expression

(Hochholdinger et al., 2008) during WCR infestation. rth3 encodes a putative GPI-anchored, monocot-specific, COBRA-like protein that has been linked to root hair elongation, various types of cell expansion and cell wall biosynthesis in maize (Hochholdinger et al., 2008). Throughout 7 d of continuous WCR exposure, constitutive expression of rth3 remained unchanged in Mp708 while it significantly decreased by day 7 in Tx601 (Fig. 2.1c, P<0.05), suggesting that Mp708 maintained root growth in spite of WCR feeding.

Differences in root anatomy and strength between Tx601 and Mp708

Images captured through the use of the laser ablation technique showed normal, undamaged, tissue structure and also damage resulting from 9 days of WCR infestation (Fig 2.2a, b, c, d). In

16 general, nodal roots of Mp708 appeared less damaged than those of Tx601. Image analysis of damaged roots showed that WCR typically fed on the root cortex with a 50% higher cortex loss in Tx601 compared to Mp708 (Fig. 2.2e, P<0.05).

To assess if there were differences in the nodal root toughness that could explain the higher losses of cortex in Tx601, we used a single edge razor blade system (Ang et al., 2008) and measured the maximum cutting force (mCF) needed to sever nodal roots at various positions from the root tip to the base of the stem. Roots grow acropetally (from the root tip) allowing distance from the root tip to be used as a proxy for root age with tissue age increasing closer to the root base (Loades et al., 2015). Significant differences in mCF were observed between

Mp708 and Tx601 (P<0.001) with mCF increasing linearly with increasing root age in Mp708.

In Tx601 correlations between age and mCF showed a lower R2 (0.095) compared to the one observed in Mp708 (0.673) (Fig. 2.2f) meaning in Tx601 there was not a strong correlation between mCF and root age. In Tx601, mCF increased linearly with increasing distance from the root tip up to ~60 mm from the root tip, but beyond this point Tx601 was not observed to either increase or decrease indicating a threshold in mCF (Fig. 2.2f).

Expression of two JA biosynthetic pathway genes and JA accumulation in roots during

WCR infestation

Because herbivory by chewing insects activates JA biosynthesis (McConn et al., 1997; Koo &

Howe, 2009), the transcript levels of two genes in this hormonal biosynthetic pathway, allene oxidase synthase (aos) and oxo-phytodienoate reductase 7 (opr7) were measured in WCR- infested roots (McConn et al., 1997; Koo & Howe, 2009; Yan et al., 2012). Constitutive (day 0) levels of both aos and opr7 transcripts were significantly higher in Mp708 than Tx601. During

WCR infestation, aos transcripts in Mp708 accumulated significantly and peaked by day 4, while transcript levels in Tx601 remained low and did not change during this time (Fig. 2.3a; P<0.05).

In Mp708, constitutive levels of opr7 transcripts were significantly higher than in Tx601 and these levels remained high throughout the 7 d infestation. In Tx601, opr7 transcript abundance gradually increased during the infestation, but transcript levels were only significantly higher than the control on day 7 (Fig. 2.3b; P<0.05). Because these genes are part of the JA biosynthetic pathway, the results suggest that Mp708 has the capacity to increase production of

JA earlier than Tx601 ultimately leading to higher constitutive JA levels in this genotype.

Previous research has shown that Mp708 whorls had higher constitutive JA levels than

Tx601 and that these levels increased in response to fall armyworm feeding (Shivaji et al., 2010), which suggested that Mp708 was genetically “primed” to respond to herbivory (Shivaji et al.,

2010). To determine if roots showed similar responses, we measured JA in roots of Mp708 and

Tx601 under non-infested conditions. We found that Mp708 roots had an approximately 3-fold higher JA concentration than Tx601 prior to WCR-feeding (Fig. 2.4a, P<0.05). In addition, JA levels in Mp708 roots increased approximately 3-fold in response to WCR feeding after 4 d of infestation (Fig. 2.4b; P<0.05), which corresponded with the higher levels of aos expression only in Mp708 (Fig. 2.3a). Furthermore, JA levels in leaves did not increase in response to belowground WCR infestation (Fig. 2.4b).

Accumulation of defense genes and MIR1-CP protein in roots in response to WCR infestation

To better understand the downstream molecular differences between Tx601 and Mp708 caused by WCR feeding, we measured root transcript levels of five insect-defense related genes: rip2

(Chuang et al., 2014), maize proteinase inhibitor (mpi) (Tamayo et al., 2000; Vila et al., 2005), farnesyl diphosphate synthase 3 (fpps3) (Richter et al., 2015), terpene synthase 23 (tps23)

(Rasmann et al., 2005; Rasmann & Turlings, 2007; Degenhardt et al., 2009), and mir1. MIR1-

CP protein abundance was also determined by immunoblot analysis (Pechan et al., 2000; Mohan et al., 2008).

When infested with WCR, rip2 expression in Mp708 roots significantly increased and peaked at day 4, and its transcript levels were significantly higher than those of Tx601 at days 4 and 7. The expression of mpi in Mp708 and Tx601 increased dramatically after 2 days of infestation and remained high in both genotypes (Fig. 2.5b; P<0.05). Increased expression of rip2 and mpi genes corresponded with aos and opr7 transcript changes and indicated that rip2, but not mpi, transcripts accumulate faster and to higher levels in Mp708 than in Tx601 (Fig. 2.3a, b).

Since terpene-derived compounds appear to be involved in plant defenses (Richter et al.,

2015), we measured expression of fpps3, which encodes the enzyme involved in producing farnesyl diphosphate (FPP), a precursor of sesquiterpenes, polyphenols, squalene, triterpenes and ubiquinones (Sallaud et al., 2009; Richter et al., 2015). We also measured transcript levels of tps23, which functions down stream of fpps3 and facilitates production of the sesquiterepene,

(E)-β-caryophyllene (Kollner et al., 2008). fpps3 transcripts significantly increased by day 4 in

Mp708 and Tx601, but were no significant differences between the genotypes (Fig. 2.6a). In contrast, abundance of tps23 transcripts in Mp708 increased during WCR infestation and was significantly higher than in Tx601 at all time points. In Tx601, tps23 transcript levels were low and did not increase during infestation (Fig. 2.6b). These results imply that Mp708 is capable of producing (E)-β-caryophyllene that could indirectly contribute to WCR resistance by attracting

19 its natural enemies. In fact, a previous study (Smith et al., 2012) demonstrated that Mp708 plants constitutively produced 10-fold greater levels of (E)-β-caryophyllene than Tx601.Futhermore, there were no significant differences in the constitutive and induced (E)-β-caryophyllene levels in Mp708. This could send a “decoy” signal indicating that the plant is already infested with

WCR and attract entomopathogenic nematodes that are natural enemies of WCR (Robert, C. A.

M. et al., 2012).

We also examined the transcript profile of mir1, which is expressed in whorls of Mp708 and not Tx601 (Pechan et al., 2000; Mohan et al., 2008). mir1 transcript levels increased dramatically in Mp708 roots during WCR infestation and peaked at day 4 (Fig. 2.7a), furthermore, MIR1-CP protein abundance followed the same accumulation trend as the transcripts (Fig. 2.7b). Importantly, the mir1 transcript and protein accumulation coincided with aos, opr7 expression and JA accumulation profiles (Fig. 2.4b), suggesting that MIR1-CP insecticidal properties could be part of Mp708 resistance to WCR in addition to high constitutive and inducible JA levels, and rip2 and tps23 expression and the presence of (E)-β-caryophyllene.

Discussion

We present data suggesting that WCR resistance in Mp708 is due to a combination of traits that could contribute to non-preference, antibiosis and tolerance (Painter, 1951; Painter, 1958).

These traits include root biomechanical resistance to cutting, compensatory root growth, constitutively elevated JA levels, synthesis of insecticidal proteins such as MIR1-CP and RIP2 and (E)-β-caryophyllene production. To the best of our knowledge, characterization of innate insect resistance to WCR in non-transgenic maize inbred lines has not been previously reported.

Bioassays showed that fewer WCR larvae survived if fed upon Mp708 roots compared to Tx601,

20 validating previous bioassays done with multiple maize lines at the V8 developmental stage that showed fewer larvae were recovered from Mp708 plants compared to Tx601 and B73 lines (Gill et al., 2011). These results open the possibility that Mp708 has desirable traits for plant breeding programs targeting WCR resistance (Jogaiah et al., 2012).

The results show that Mp708 and Tx601 roots were differentially damaged by WCR feeding. Both Mp708 and Tx601 roots tended to increase in length over time, but following

WCR infestation, the total root length of Tx601 was lower than that of Mp708 because fewer lateral roots were measured in Tx601. Lateral roots from Tx601 showed no correlation between resistance to cutting and root age, while Mp708 nodal and lateral roots were more resistant to cutting, which could be one reason for lower WCR feeding and survival on Mp708. Lateral roots from Tx601 showed no correlation between resistance to cutting and root age while Mp708 showed the same trend as nodal roots (Fig. S2.2), implying that Tx601 lateral roots are more prone to damage by organisms with piercing or chewing feeding strategies. Loss of lateral root length could contribute to the poorer performance of plants under WCR infestation, since lateral roots are responsible for most of nutrient and water uptake contributing to overall plant fitness

((Paez-Garcia et al., 2015). In Tx601, the reduction in root length could be the result of both greater root consumption by WCR and reduced root growth. Furthermore, in the field damaged roots can become more susceptible to other diseases that can gain entry through the wounds. We related root length changes with rth3 gene expression (Hochholdinger et al., 2008) during WCR infestation. This gene is a marker of growth in the root apical meristem (Rost & Bryant, 1996;

Bassani et al., 2004; Hochholdinger et al., 2008) and its expression profiles showed that Mp708 was able to maintain rth3 transcript levels during infestation, whereas in Tx601 its expression decreased significantly by day 7. Compensatory maintenance of root growth during WCR

21 infestation could result in similar shoot biomass and CO2 assimilation as uninfested plants

(Riedell & Reese, 1999), leading to unaffected yields (Branson et al., 1982). Taken together, these results suggest that Mp708 has a root system more resistant to WCR feeding than Tx601 and is able to maintain root growth during WCR infestation.

The site of WCR feeding within the nodal roots was determined using laser ablation tomography. These images showed that WCR tended to feed on the nodal root cortex and caused more damage in Tx601 than Mp708. Furthermore, we determined that Mp708 nodal and lateral roots were more resistant to cutting than Tx601, which could be one reason for lower

WCR feeding and survival on Mp708. Tougher roots could make it more difficult for WCR to feed on the roots and access nutrients. In tobacco, decreased root toughness and lignin accumulation has been linked to low tolerance to root wireworms due to weaker root tension following a fracture toughness test (Johnson et al., 2010). The study however only observed significant differences in fracture toughness between tobacco lines, and not in cutting resistance

(Johnson et al., 2010).

These findings combined with the production of potentially toxic proteins like MIR1-CP and RIP2 suggests that Mp708 has a suite of robust defense traits in its roots. These root traits in

WCR-resistant maize should lead to higher yields and fewer WCR adults in the field compared to maize genotypes with smaller roots and compromised root growth during belowground infestation (Branson et al., 1982).

Many studies have shown that feeding by chewing insects increases the expression of genes involved in JA biosynthesis and accumulation (Koo & Howe, 2009), but the expression of these genes and accumulation of JA in roots during root herbivore attack has not been studied extensively in maize (Erb et al., 2009a; Erb et al., 2012). We showed that aos transcript levels in

Mp708 roots increased up to day 4 of WCR infestation while in Tx601 they remained lower and relatively constant. Transcript levels for orp7 were higher in Mp708 and remained high throughout the infestation whereas those in Tx601 were initially low and only increased slightly during the infestation. Notably, the constitutive expression of these two genes was significantly higher in Mp708 than Tx601 roots, similar to high expression of these two genes in leaves of older Mp708 plants (Shivaji et al., 2010).

The constitutive and induced expression of aos and opr7 may contribute to the higher constitutive and inducible levels of JA in Mp708 than in Tx601. These results support the finding that Mp708 plants are under constant alert against herbivory (Shivaji et al., 2010). Our results appear to be consistent with studies that linked JA accumulation with high constitutive and inducible gene expression and accumulation of insecticidal proteins (Zhu, 2010; Ankala et al.,

2013). Studies in insect-susceptible maize (cultivar Delprim) have provided conflicting results on inducible levels of JA in roots with WCR infestation. One study showed low accumulation of this phytohormone in roots during WCR infestation (Erb et al., 2012) while another showed 50% increase in JA levels (Erb et al., 2009a). Mp708 roots have high constitutive and inducible levels of JA that suggests it plays a key role in activating downstream defenses to resist WCR attack.

To understand the downstream molecular changes involved in Mp708 and Tx601 defense responses to WCR, we examined accumulation of four genes, rip2, mpi, fpps3, and tps23. mpi and fpps3 significantly increase in both maize lines while rip2 and tps23 transcripts showed high endogenous and induced levels only in Mp708. Because fpps3 produces FPP, a precursor of sesquiterpenes, polyphenols, squalene, triterpenes and ubiquinones (Sallaud et al., 2009; Richter et al., 2015), it is possible that both inbreds can increase production of FPP-derived compounds related to plant defenses. Downstream from fpps3 is tps23, a herbivore-induced gene that leads to

23 the production of (E)-β-caryophyllene (Kollner et al., 2008), a volatile that attracts entomopathogenic nematodes that are natural enemies of WCR (Kollner et al., 2008). Mp708, but not Tx601, expressed high constitutive and inducible levels of tps23 transcripts in the roots

(Fig. 6b). This coupled with prior data showing that Mp708 has much higher (E)-β- caryophyllene than Tx601 that repelled FAW larvae (Smith et al., 2012a) further supports its role in WCR defense. Thus the presence of (E)-β-caryophyllene could play two important roles in

Mp708: deterring FAW feeding and attracting the natural enemies of WCR.

Because diet-based bioassays with WCR are problematic, we were not able to directly determine the effect of the insecticidal protein MIR1-CP on WCR performance. However, mir1 transcript levels and MIR1-CP protein increased during WCR infestation, implicating MIR1-CP in defense. Diabrotica species have a peritrophic matrix (Silva et al., 2004), therefore it is possible that consumption of MIR1-CP by WCR could damage this structure as it does in FAW

(Pechan et al., 2000). These results indicate that both Mp708 and Tx601 use the products of mpi and fpps3 to defend against herbivory, whereas rip2, tps23 and mir1 are only inducible in Mp708 and could be key players in its resistance.

It appears that Mp708 has multiple traits that enable it perform well during WCR infestation in addition to being resistant to insects with different feeding behaviors like FAW

(Williams et al., 1985; Williams et al., 1990) and CLA (Louis et al., 2015). Mp708 was developed from landraces of maize that most likely originated in Mesoamerica (Williams et al.,

1987), where many phytophagous maize pests including Diabrotica sp., FAW and CLA may also have originated (de Lange et al., 2014). One could speculate that multiple generations of selection for adequate yield despite intense insect pressure led to the incorporation of multiple resistance traits into these landraces, which ultimately were incorporated into Mp708 by

24 selective breeding. Hence, Mp708 displays a suite of resistance traits to three types of insect pests: a whorl feeder (FAW), phloem feeder (CLA) and root feeder (WCR) that encompass both constitutive and inducible defense responses. Because populations of WCR are developing resistance to Bt-transgenes (Gassmann, 2012; Flagel et al., 2015), soil-applied insecticide and persist despite crop rotation with soybean (Bigger, 1932; Gray et al., 2009), the availability of a non-transgenic genotype with this remarkable range of native host plant resistance will be especially useful for discovering new resistance traits that can be implemented in plant breeding and pest management programs against a highly adaptable insect like WCR.

(a) 3 d 6 d 9 d

Length of lateral roots after 6d WCR infestation

(b) 200 3 d 6 d 9 d a a 150

Corn line a Mp708 100 Tx601

Lateral roots length (cm)

0 (c) Control WCR Roots rth3 WCR

Treatment

Figure 2.1. oot length in Mp708 and Tx601 at days 3, 6 and 9 after continuous infestation with

WCR. (a) Total root length, (b) lateral root length in Mp708 and Tx601 at days 3, 6 and 9 after continuous infestation with WCR. (c) Time course analysis of rth3 transcript accumulation in

Mp708 and Tx601 in response to continuous infestation with WCR. Relative expression (RQ) of rth3 was measured by qRT-PCR. Gene expression levels were normalized to actin. Length and

RQ data were normalized and analyzed using analyzed using multiple-factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test. Letters represent results of the HSD (P<0.05) and error bars show the standard error.

(a) Tx601 Control (c) Mp708 Control

(b) Tx601 WCR (d) Mp708 WCR

(e) (f)

Figure 2.2. Laser ablation tomography (LAT) cross-sections from nodal roots of Tx601 and Mp708 after 9 d of continuous infestation with WCR. Root cross-sections from control (a) Tx601

28 or (c) Mp708 and (b) WCR-infested Tx601 and (d) Mp708. (e) Percentage of cortex lost in Mp708 and Tx601 in non-infested control plants or those infested with WCR for 9 days. (f) Maximum cutting strength in nodal roots. Linear regression analysis shows a significant difference between lines (P<0.001) as a function of distance from stem base. For the LAT images (e) the percentage loss was determined from images analyzed with ImageJ software, normalized and analyzed using multiple-factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test. Letters represent results of the HSD (P<0.05) and error bars show the standard error.

(a) Roots aos

(b) Roots opr7

Figure 2.3. Time course analysis of aos (a) and opr7 (b) transcript accumulation in roots of

Mp708 and Tx601 in response to WCR infestation. Gene expression levels were determined in roots of V3 stage plants 0, 2, 4 and 7 d after belowground infestation with WCR. Relative expression (RQ) of aos and opr7 was measured by qRT-PCR. Gene expression levels were normalized to actin. RQ data were normalized and analyzed using multiple-factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test. Results of the

HSD are represented by letters and error bars represent the standard error.

(a) Constitutive levels of JA in roots

(b) Constitutive and induced levels of JA

Figure 2.4. Analysis of constitutive jasmonic acid (JA) levels in root from Mp708 and Tx601

(a) and JA accumulation in root tips and leaves of Mp708 infested with WCR (b). Control plants were not infested with WCR. JA data were normalized and analyzed using multiple-factor

ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test.

Letters represent results of the HSD and error bars show the standard error.

(a) Roots rip2

(b) Roots mpi

Figure 2.5. Time course of maize defense genes in response to WCR. (a) rip2 and (b) mpi transcript accumulation in roots of Mp708 and Tx601. Gene expression levels were determined in root of V3 stage plants 0, 2, 4 and 7 d after belowground infestation with WCR. Relative expression (RQ) of rip2 and mpi were measured by qRT-PCR. Gene expression levels were normalized to actin. RQ data were normalized and analyzed using multiple-factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test. Letters represent results of the HSD (P<0.05) and error bars show the standard error.

(a) Roots fpps3

(b) Roots tps23

Figure 2.6. Time course of maize genes involved in volatile production. (a) fpps3 and (b) tps23 transcript accumulation in roots of Mp708 and Tx601 in response to continuous WCR infestation. Gene expression levels were determined in V3 stage plants 0, 2, 4 and 7 d after belowground infestation with WCR. Relative expression (RQ) of fpps3 and tps23 were measured by qRT-PCR. Gene expression levels were normalized to actin. RQ data were normalized and analyzed using multiple-factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test. Letters represent results of the HSD (P<0.05) and error bars show the standard error.

(a) Roots mir1

(b) 2 d 3 d 4 d

control WCR control WCR control WCR

Root

Figure 2.7. Time course of maize insecticidal protein accumulation. (a) mir1 transcript and (b)

MIR1-CP protein accumulation in roots in roots of Mp708 in response to WCR infestation. Gene expression levels were determined in V3 stage plants 0, 2, 4 and 7 d after belowground infestation with WCR. Relative expression (RQ) of mir1 was measured by qRT-PCR. Gene expression levels were normalized to actin. RQ data were normalized and analyzed using multiple-factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test. Letters represent results of the HSD (P<0.05) and error bars show the standard error. For the western blots each lane represents a total of 15 µg of root proteins pooled from

34 three biological replicates. Simply Blue staining of the gels shows equal protein loading for the samples.

Figure S2.1. Percent survival of WCR fed on Mp708 and Tx601 maize lines. Percent survival was normalized and analyzed using multiple-factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test, the results of the HSD are represented by the letters and the error bars represent the standard error.

Figure S2.2. Maximum cutting strength in lateral roots. Linear regression analysis shows a significant difference between Mp708 (P<0.001) as a function of distance from stem base.

CHAPTER III. Plant Bio-Wars: Protein Networks Reveal Tissue-Specific Defense

Strategies in Maize in Response to a Root Herbivore

Introduction

Humans live within social networks (Guimerà et al., 2005; Ahn et al., 2010; Sekara et al., 2016); each individual can be represented as a node and each social relationship as a link (Kitano, 2002;

Sekara et al., 2016). A person is a biological network composed by a plethora of nodes such as molecules, metabolites, proteins, and organs connected by metabolic processes, regulation or physical interacting events (Oltvai & Barabasi, 2002; Barabasi & Bonabeau, 2003; Albert, 2005;

Ahn et al., 2010; Weckwerth, 2011; Walley et al., 2016). Plants are the same; they are biological networks composed of molecules linked by physical, regulatory or relative abundance mechanisms (Weckwerth, 2011; Walley et al., 2016). These networks can be perturbed by exposing the organism to external stimuli such as insect herbivory. One way for researchers to understand plant-insect interactions is to investigate global plant defense responses as a perturbed living network that is complex, robust, dynamic and versatile (Kitano, 2002;

Weckwerth, 2011). Nevertheless, a whole system analysis of maize defense responses against insects that uses network biology where nodes are proteins that have patterns in their connections has not yet been done.

Plant networks can serve as representations of the biological events that take place within the organism, for example, when a plant is attacked by an insect, the attacked organs deploy local direct defenses that become network nodes. Typical local defenses include accumulation of secondary metabolites such as glucosinolates, 2,4-dihydroxy-7-methoxy-1,4-benzoxazin-3-one

(DIMBOA), alkaloids, phenolics or insecticidal proteins, and enzymes (Schilmiller & Howe,

2005; Howe, 2008; Zhu-Salzman et al., 2008; Koo & Howe, 2009; War et al., 2012). The

37 attacked organ then communicates to other organs that in turn respond by deploying their defense mechanisms in which each activated gene, metabolite, protein, and enzyme becomes another node in the defense network (Schilmiller & Howe, 2005; Van Dam & Bezemer, 2006;

Chen et al., 2009; Erb et al., 2009c; War et al., 2012). This communication is called systemic signaling and involves the plant’s recognition of its attacker, followed by local and systemic responses. The communication events among organs assume the identity of network links, these links can be physical interactions between proteins, similarity in expression patterns, or gene regulatory events (Walley et al., 2016). In a dynamic network where time is a factor, spread of information such as responses to external stimuli can travel through nodes and links as ripples.

Paths in the network that are used for transferring information become important for processing of stimuli and nodes within paths are highly influential on network dynamics (Barabasi &

Bonabeau, 2003; Albert, 2005). If the network is static, other topology characteristics become diagnostic of nodes or groups of nodes that are vital for the organism to respond to external stimuli, such as network connectivity distribution, node connectivity or number of paths that go through the node and these characteristics are called network properties (Alm & Arkin, 2003;

Doncheva et al., 2012; Hazem Radwan & Janice, 2014; Zhang et al., 2016).

Plant systemic defense responses against herbivores must involve changes in metabolic processes and molecules capable of communicating the alert response (Schilmiller & Howe,

2005). For these molecules to act as mediators of communication they must be induced by insect feeding, have mobility and regulate resistance. Phytohormones are examples of potential mediators (Soler et al., 2013). Recently it was discovered that a family of plant elicitor peptides

(PEPs) might mediate and amplify defense signaling in maize in response to Spodoptera exigua oral secretions by increasing jasmonic acid (JA) and ethylene (ET) content and activating

38 expression of genes involved proteinase inhibitor accumulation, volatile terpene synthesis and the benzoxazinoid pathway (Huffaker et al., 2013). Local and systemic defense responses in plants vary depending on the type of insect feeding on them, implying that multiple signaling molecules function in systemic cross-talk during plant defense (Erb & Lu, 2013; Soler et al.,

2013). An approach to detect communication signals, phytohormones and defense pathways involved in local and systemic defenses is to determine activation or repression of plant metabolic pathways in those organs under infested conditions using network biology.

Root herbivores are ubiquitous in nature, cause severe damage to plants, and their interactions with the roots are influenced by other soil dwelling organisms. The root-insect- microbe tritropic interaction affects the way roots respond to insects compared to other plant organs (Johnson et al., 2016). One approach for studying and contrasting defense responses in roots and leaves is to conduct a comprehensive “omics” analysis of plant defense responses using a network biology approach. This type of approach has not previously been used to study aboveground or belowground plant defense responses to insects. The final goal of the network biology approach is to model biological phenomena at one or many levels. Levels that can be modeled are molecular signaling at the sub-cellular level, networks of physiological processes at the cellular and tissue level, plant growth and development at the individual level, genetic variation among individuals at the population level, plant performance, competition and trophic interaction at the community level, and the trade-off dynamics at the ecosystem level (Keurentjes et al., 2011). This approach is challenging due to the large amount of data that must be acquired to generate models within levels or to integrate the levels. However the use of high-throughput technologies, and computer analysis such as systems biology can help to integrate genotyping, phenotyping/ morphotyping and molecular phenotype analyses (Weckwerth, 2011; Bassel et al.,

2012; Studham & Macintosh, 2012) to achieve the goal of understanding and producing insect resistant plants (Barah & Bones, 2015).

The majority of “omics” analyses in plant-insect interactions involve hierarchical clustering to determine groups of genes, proteins or metabolites with similar abundance patterns.

One of these analyses used Z. mays infested with corn stalk borer (Ostinia furnacalis) or treated with methyl jasmonate (me-JA). The results showed 39,636 genes with differential expression in response to all treatments against control plants. These genes were involved in metabolic processes, catalytic activity as a response to stimuli, and transfer of molecules. During insect infestation genes in the JA pathway such as lipoxygenases, allene oxide synthase, and other defensive proteins as terpene synthase and serine protease inhibitors, were up regulated (Yang et al., 2015). Another “omics” analysis used hierarchical clustering of plant-insect interactions under drought stress. In this analysis Arabidopsis was subjected to drought or infected with a pathogen (Botrytis cinerea) followed by infestation with Pieris rape caterpillars. Transcriptional changes induced by insect feeding had different activation times that were dependent on the plant’s prior exposure to pathogen or drought stress. When plants were exposed to pathogen infection and drought conditions before caterpillar infestation, there was a significant downregulation of pathogen defenses and plant growth. Overall, insect attack reprogrammed plant responses from previous stresses to herbivore defenses (Olivas et al., 2016). Finally, an analysis that integrated proteomics and transcriptomics using hierarchical clustering indicated that oral secretions (OS) from Mythimna separata elicited a long-term defense response in maize that included transcription factors, JA pathway and terpene-derived VOCs (Qi et al.).

“Omics” data can be combined with network analysis, and this approach has been applied in other fields. For example, research on hepatitis C virus infection used a protein correlation

40 network built from proteomics data to determine target proteins and genes vital to the response to this virus. This network was created by linking proteins with similar expression patterns and used degree and betweenness topological network measurements to identify pathogen interactors and find proteins of interest (McDermott et al., 2012). In maize, there are “omics” databases that have been the basis for building a protein-protein interaction network (PPI) that uses interacting orthologs (interologs) of 13 different species (Oryza sativa, Arabidopsis thaliana, Homo sapiens, Mus musculus, Rattus norvegicus, Drosophila melanogaster, Caenorhabditis elegans, Saccharomyces cerevisiae, Schizosaccharomyces pombe, Escherichia coli, Bacillus subtilis, Helicobacter pylori, Campylobacter jejuni, and Synechococcus). These protein-protein interactions are predicted in silico to occur in maize and are based on experimentally confirmed interactions available on four interactome databases [BioGRID (Stark et al., 2006); DIP

(Salwinski et al., 2004); IntAct (Aranda et al., 2010); and MINT (Chatr-aryamontri et al., 2007).

The basis of this network was that evolutionary conserved proteins keep their conserved interacting partners if the protein function is the same (Musungu et al., 2015). Both protein correlation networks and maize PPI can be used as tools to analyze any type of “omics” data to identify proteins and pathways of interest (Hazem Radwan & Janice, 2014).

To understand plant defenses against insect attack at whole organism level it is necessary to consider both local and systemic defense responses (Schilmiller & Howe, 2005; Perkins et al.,

2013). We studied local and systemic defense responses in an insect resistant maize genotype,

Mp708, during infestation with specialist root herbivore, Diabrotica virgifera virgifera (Western corn rootworm, WCR), an economically devastating pest in North America and Europe

(Hummel, 2007; Gray et al., 2009; Tinsley et al., 2013; Flagel et al., 2015). Mp708 has been shown to be resistant to fall armyworm (Williams et al., 1985), corn leaf aphid (Louis et al.,

2015) and WCR (Castano-Duque manuscript submitted). We generated a quantitative and qualitative protein database using tandem mass spectrometry tags (TMT), and measured protein abundance changes in roots and leaves in response to WCR infestation. These data were then used for network analyses.

If we can understand local and systemic defenses in Mp708 during WCR infestation, it will help us understand other plant-herbivore interactions because large and complex biological systems have common properties (Barabasi & Bonabeau, 2003). The main goals of this study were to generate protein correlation and protein-protein interaction networks to determine which modules in the maize metabolic network are activated and can be considered as insect resistant modules (Hazem Radwan & Janice, 2014). Another goal was to create a data analysis pipeline based on proteomics and systems biology approaches that could be used as a toolbox in the plant-insect signaling field. Determining changes in metabolic and signaling pathways and differences in protein accumulation that are important for plant defenses can help us design environmentally sustainable strategies for protecting crops from insect damage. This will be an advantage in the massive production of food that is needed in the world, especially in areas where insects are a menace to maize yield.

Materials and methods

Plants and insects

Seed for the maize genotypes, insect-resistant Mp708 and its susceptible parent, Tx601

(Williams et al., 1985) were provided by Dr. Paul Williams, USDA-ARS, Mississippi State

University. Plants were grown in Hagerstown Loam in the Plant Science greenhouse at the

Pennsylvania State University in 8 cm x 9 cm pots. Supplemental lightening was used as needed.

Diapause WCR eggs were obtained from Dr. Bryan French, USDA-ARS, Brookings, SD and incubated for 10 to 12 days on damp paper towels at 25ºC in the dark until hatching.

Insect infestation and tissue collection

Time course experiments were done by infesting each Mp708 and Tx601 plant at the V3 stage

(Ritchie et al., 1998) with ~20 WCR (~24 h hatched) for 2, 4, or 7 days. Control plants were not infested with insects (including the 0 hr time point). Roots were cleaned, collected, weighed

(0.1g per biological replicate) and stored at -80 °C until further use. In both, gene expression and immunoblot experiments a minimum of three biological replicates per treatment were used; a biological replicate was root tips (up to 2cm from root tip) pooled from two plants. For gene expression analysis three biological replicates were used and a replicate was pooled root tips from two plants. For ethylene experiments three biological replicates were used and each replicate was the root system of one plant.

Protein extraction for proteomic analysis

For the proteomics experiment root tips and leaves (0.5g per biological replicate) were collected from control and 4-day WCR-infested Mp708 seedlings. The tissues were ground using a ball- mill tissue grinder (Genogrinder 2000; SpexCentriprep Inc., Metuchen, NJ, U.S.A.) for 2 min at

2,000 strokes/min under chilled conditions. Extraction buffer (1.8 M sucrose, 0.5 M Tris-HCl pH

8.5, 0.1 M EDTA, 4% β-mercaptoethanol and protease inhibitor) was added to each sample (1:5 ratio) vortexed and left on ice for 5 min, then Tris-saturated phenol pH 7.5 was added (1:1 ratio) vortexed and centrifuged at 6,000xg for 10 min at 4°C. Supernatant was collected then five volumes of cold precipitation solution (0.1M ammonium acetate in methanol) were added and

43 left overnight at -20°C. Samples were centrifuged at 6,000xg for 10 min at 4°C, the pellet was washed three times with 100% cold acetone followed by centrifugation at 6,000xg for 5 min at

4°C. Pellets were air dried then rehydrated with 50 µl of buffer (0.1 M ammonium bicarbonate and 4 M urea) and vortexed for 2 h at room temperature, supernatant was recovered and 50 µl of rehydration buffer was added to the pellet which was vortexed again for 2 h at room temperature, then both supernatants were mixed. Protein quantification was performed on the supernatant mix by using NITM protein assay kit (G BIOSCIENCES, St. Louis, MO). The mixed supernatants

(50µg) were reduced with 250 mM dithithreitol (DTT) then alkylated with 250 mM of iodoacetate in the dark, and the alkylation reaction was stopped by incubating with 5 mM DTT at room temperature. The urea content in samples was diluted to 1.2 M by adding 50 mM Tris-HCl pH 8.0 and 5 mM calcium chloride.

Enzyme digestion

Enzyme digestion was performed by adding trypsin (Sigma, St. Louis) in 1:50 (enzyme: protein) ratio and lysyl endopeptidase (Wako Pure Chemicals, Japan) in 1:90 (Enzyme : protein) ratio and incubated overnight at 37°C. Then desalting of the digested sample was performed using Bond

Elut OMIX C18 pipette tips (Agilent, Santa Clara, CA) following manufacturer instructions.

Samples were dried in a speedvac system (Thermo Fisher Scientific, Waltham, MA) at room temperature.

Tandem mass tags (TMT) labeling

Samples were rehydrated in 100 µl of 100 mM triethylammonium bicarbonate buffer (TEAB).

The10-plex TMT (ThermoFisher Scientific, Grand Island, NY) tags were suspended in 41µl of

100% (V/V) acetonitrile and each one added to the rehydrated sample for labeling. Samples were labeled in the following order: infested leaves bioreplicates 1 (126), 2 (127N) and 3 (130C); control leaves bioreplicates 1 (127C) and 2 (128N): infested roots bioreplicates 1 (128C),

2(129N) and 3(131); control roots bioreplicates 1(129C) and 2 (130N). Labeled samples were shaken for 1 h at room temperature and the reaction was quenched by adding 0.8 µl of 50%

(V/V) hydroxylamine. All samples were combined, dried in a vacuum concentrator and sent to the proteomics core facility at the University of Wisconsin, Madison for further analysis.

High pH sample fractionation

Dried TMT samples were resolubilized in 250 µL of 9.5 mM ammonium formate in water, pH

10. The TMT labeled mixture was fractionated by pH using a Phenomenex Gemini 5µm C18

110Å, 4.6mm x 250mm column at a flow rate of 1.0 ml/min with 9.5 mM ammonium formate in

80% acetonitrile in water, pH 10 solvent. Sample detection was conducted with a Waters 2996 photodiode array detector monitoring at 214 nm and 280 nm. Fractions were collected every 1 min from the time of injection to the end of column re-equilibration. Recovered fractions were pooled into one mixed fraction every 6th fraction to generate five total pooled fractions. Pooled fractions were frozen and lyophilized. Each fraction was resolubilized in 200 µl of 0.1% formic acid in water and the pH was checked. For fractions with pH >3, formic acid was added to bring the pH down to 3. Aliquots were then placed in autosampler vials for high-pressure liquid chromatography with mass spectrometry (HPLC/MS/MS).

Mass spectrometry

HPLC/MS/MS was performed using an Agilent 1100 Nanopump system with MicroWPS autosampler. For each sample, 3 µL was loaded onto a Thermo Easyspray C18 column, 75µm x

15cm packed with PepMap 3µm, 100Å C18. Gradient elution was performed using 0.1% formic acid in water and 0.1% formic acid in acetonitrile. Electrospray ionization was at 1.9 kV and mass spectrometry data was acquired over the m/z range 350-2000 using a top ten data- dependent acquisition approach. MS1 spectra were collected in the Orbitrap at 120,000 resolving power and MS/MS spectra were acquired in the Orbitrap analyzer at 30,000 resolving power.

MS/MS fragmentation was achieved by HCD at normalized collision energy of 37% and isolation width of 2.00. The AGC target value for MS/MS was adjusted to 75,000. Dynamic exclusion was enabled with a repeat count of one and exclusion duration of 45 sec. Singly- charged ions and ions of unknown charge state were rejected from selection as precursors for

MS/MS.

Protein identification and quantitative analysis

Protein identification analysis was performed using Proteome Discoverer (Thermo) version

1.4.1.14. The database used was a protein database downloaded from Ensemble Genomes and with date stamp of 4-Dec-2014. To this database was appended a set of common contaminant proteins such as trypsin and human keratin. Search parameters were as follows: fully tryptic terminus specificity (both sides), up to 2 missed cleavages within a peptide sequence, fixed carbamidomethylation of cysteine residues, fixed modification of peptide N-termini and lysine residues with TMT reagent mass, variable oxidation of methionine, and variable deamidation of asparagine and glutamine residues, 10 ppm maximum mass error for precursors, 0.02 Da mass error for fragment ions. Peptide assignment confidence was determined using a reverse-decoy

46 strategy. A false-discovery rate (FDR) of 95% (“medium confidence”) was accepted for passing peptide identifications to the quantitation method.

Quantitative analysis using the TMT reporter ions was performed by Proteome

Discoverer following database searching. Peptides were rejected from quantitation if they were not unique to a protein, MS/MS spectra were rejected if they had a co-isolation of >75% (i.e.

>75% of the ion current within an MS/MS isolation window was due to a ions not related to the identified peptide.) For instances where a given peptide had reporter ion channels with no ion current, these missing values were replaced by the minimum value detected for any reporter ion in any peptide. In these data, that corresponded to a signal of approximately 100 counts. Further processing was performed using an in-house developed script to adjust and normalize reporter ion abundances and collect peptide results into proteins. Reporter ion abundances were scaled by multiplying each abundance by the ion injection time for that spectrum. This reduces the effect of single, highly-abundant peptides from dominating the quantitative value. The abundance for each reporter ion channel was then normalized to the median value for that channel, correcting for deviations in protein inputs, digestion efficiency, and labeling.

Analysis of variance

Protein abundance data was transformed into ratios by taking each bioreplicate of infested leaves or infested roots (Lin and Rin) and dividing them by the average of their respective control bioreplicates (Lc and Rc). The control ratios were determined dividing each control bioreplicate with the average of the control bioreplicate (for example, Rc:Rcaverge). Ratios were transformed using logarithm base 10 and analyzed by ANOVA using R (Team, 2015).

Heatmap generation

The heatmap was built using by taking the logarithmic base 10 ratios of roots infested:roots control (Rin/Rc) and leaves infested:leaves control (Lin/Lc) to do hierarchical clustering in R

(Supplemental data R code). We used only the proteins that had significant differences in abundance (p-value= or < than 0.05). The similarity dendogram in the heatmap map was built by using the Euclidean distance method in R and we determined abundance patterns using a maximum of 10 k-means cut off. These 10 k-means groups became proteins families that were individually annotated by using Biomart maize RefGen_V4 and gene ontology. Singular gene ontology enrichment analysis was performed using agriGO online tools with the TMT protein list as input list and the maize genome locus from maizesequence.org as background.

Protein correlation networks

Protein abundance adjacency matrixes generated by Pearson correlation analyses were visualized using Cytoscape (Cline et al., 2007). The adjacency matrix was built using Pearson correlation on the log transformed ratios, a filter correlation of 0.99 and filter abundance value of 0.001 from an in-house code in R (Supplemental R code) (McDermott et al., 2012). The matrix was imported to Cytoscape version 3.2.1 (Cline et al., 2007) to build the visualization and annotations were added to each protein by importing the gene ontology data from gramene.org using the ensembl and biomart tools for Z. mays. Topological analysis of the network was done using the NetworkAnalyzer tool from Cytoscape (Doncheva et al., 2012). The betweenness centrality distributions were fit to the power law (y=axb) where y was the betweenness centrality and x was the number of neighbors. We used jActiveModules to determine the subnetworks of highly activated proteins using a scored p-value to determine “hotspots” (Ideker et al., 2002).

The scored p-value was developed to rank the highly connected subnetworks that had significant differences in abundance under certain stimuli using a scoring system of protein p-values (Ideker et al., 2002). Then a singular gene ontology enrichment analysis was performed using agriGO online tools with the TMT protein list as input list and the maize genome locus from maizesequence.org as background.

Protein-protein interaction networks

The protein-protein interaction subnetwork was built using identified proteins and the log transformed ratios from the TMT analysis. The subnetwork was extracted from predicted protein-protein interaction network (PPI) from maize (Musungu et al., 2015) using Cytoscape version 3.2.1 (Cline et al., 2007). We used jActiveModules to determine the subnetworks of highly activated proteins by using a scored p-value to determine “hotspots” (Ideker et al., 2002).

Hotspots were found using the jActiveModules (Ideker et al., 2002) application in Cytoscape followed by singular gene ontology enrichment analysis that was performed using agriGO online tools with the TMT protein list as input list and the maize genome locus from maizesequence.org as background.

Ethylene (ET) measurements

Whole root systems and leaves were excised after 2, 3 and 4 days of WCR infestation were cleaned and placed in 180 ml glass flasks and sealed for 24 h, non-infested samples also were collected. The tissue fresh weight was recorded for each sample. Gas (100 µl) was taken from the flask with a gas tight syringe and directly injected into an Agilent 6890 gas chromatograph, equipped with a flame ionization detector and a Rt-QS-BOND column (30 m x 0.53 mm ID x 20

µm film thickness, Restek, Bellefonte, PA). The sample was run using helium carrier gas at an initial flow rate of 9 mL min-1 with an average velocity of 69 cm s-1, injector temperature of 150°C, and conducted isothermally at 70°C with the detector at 200°C. The amount of ET was quantified by comparing the chromatograph of the samples against a standard curve made by using a 9 µg mL-

1 ET in helium standard (Restek, PA) and analyzed with the R statistical software. The data was analyzed by first doing data transformation until the Shapiro-Wilk test (Shapiro and Wilk, 1965) confirmed normal distribution, then a multiple-factor analysis of variances (ANOVA) was done, followed by a significant difference (HSD) Tukey pairwise comparison test.

Salicylic acid (SA) quantification

Root tissues were collected as previously described and placed in 2 ml screw-cap FastPrep tubes

(Qbiogene, Carlsbad, CA) containing Zirmil beads (1.1 mm; SEPR Ceramic Beads and Powders,

Mountainside, NJ). Dihydro-jasmonic acid (dhJA) was added to each vial as internal standard

(100 ng) followed by 400 µl of 1-propanol:water:hydrochloric acid (2:1:0.002, v/v) and shaken for 40 s in a FastPrep FP 120 tissue homogenizer. Dichloromethane (1 ml) was added to each sample, followed by shaking for 40 s in the homogenizer, and centrifugation at 13,000 x g for 1 min. The bottom dichloromethane and 1-propanol layer was then transferred to a 4 ml glass screw-cap vial and dried under an air stream. Samples were reconstituted in methanol:diethyl ether solution (1:9, v/v) and 2.3 µl of trimethylsilyldiazomethane hexane (Aldrich) were added to each. The vials were then capped and allowed to sit at room temperature for 25 min. Excess trimethylsilyldiazomethane was destroyed by adding 2.3 µl of 2.0 M acetic acid in hexane to each sample (Schmelz et al., 2003; Schmelz et al., 2004). Finally, phytohormones were collected

50 by using a vapor phase extraction protocol previously described by Schmelz et al. (2004). The extracts were run in a gas chromatograph mass spectrometer with electron ionization and identity and quantity of the total SA was determined by comparing the retention times and spectra of the internal standard. The data were analyzed by first doing data transformation until the Shapiro-

Wilk test (Shapiro and Wilk, 1965) confirmed normal distribution, then a multiple-factor analysis of variances (ANOVA) was done, followed by a significant difference (HSD) Tukey pairwise comparison test in R version 3.2.1.

Root volatiles and bioassays

Mp708 seedlings were infested or non-infested with 35 WCR neonates for volatile collection or 20 WCR neonates for bioassays for 18 d. Volatiles were collected for 24 h from roots by using root chambers as described in (Ali et al., 2010) at a rate of 0.5 L per hour. Super-Q® traps were rinsed with 150 μl of dichloromethane; n-nonyl-acetate (20 ng) was added as an internal standard. Samples were injected and analyzed by gas chromatography mass spectrometry by using electron ionization (6890 gas chromatograph interfaced to a Hewlett-

Packard 5973N mass selective detector). The column (30 m × 0.25 mm i.d., 0.25 μm film thickness, HP-1MS) was maintained at 40°C for 0.5 min and then increased by 10°C per min to

280°C. Identifications of all volatile compounds were confirmed by comparing mass spectra from those in the library database and quantified by comparing the chromatograph of the samples against the standard. Amount of each volatile was normalized by the fresh weight of the root sample that was measured 24 h after the volatile collection was finished.

Bioassays were performed by collecting the worms from the soil and root system after VOC collection. The worms were counted, weighted immediately and the percentage of survival was calculated. Data was analyzed with the R statistical software version 3.2.1 by using logarithmic, square root, inverse power or box-cox transformations until the Shapiro-Wilk test (Shapiro and

Wilk, 1965) confirmed normal distribution then a multiple-factor analysis of variance (ANOVA) was performed for each day individually, followed by a significant difference (HSD) Tukey pairwise comparison test. If the data did not achieve normality a Wilcoxon test was performed.

Protease activity and protease inhibition activity with serine or cysteine-type inhibitors

Leaf and root samples (0.1g) from Mp708 were ground using a ball-mill tissue grinder

(Genogrinder 2000; SpexCentriprep Inc., Metuchen, NJ, U.S.A.) for 2 min at 2,000 strokes/min under liquid nitrogen conditions. Soluble proteins were extracted by adding 1 ml of 0.046 M Tris buffer, pH 8.1 and 0.0115 M CaCl2 and 5% (weight/volume) insoluble, cross-linked polyvinylpolyrrolidone (Afla Aesar, Haverhill, MA). Samples were vortexed, kept on ice for 10 min and centrifuged at 11,000 x g for 10 min at 4°C. Supernatants were aliquoted and used for determining protease activity, which was assayed by incubating 100 µl of leaf or root extract with 100 µl of the substrate, 2% (w/v) azocasein (Sigma, St. Louis) for 4 hr at 47°C. The reaction was terminated by adding 5% (w/v) TCA and incubated for 15 min at room temperature, followed by centrifugation at 10,000 x g for 5 min. The supernatant was removed and mixed with 0.5 N NaOH in a 1:1 ratio and absorbance at 440 nm was determined. Total protease activity was defined as 1 unit per 1 unit of absorbance. Specific activity was calculated by dividing the activity by the protein concentration as determined by the Bradford assay. Cysteine protease activity was determined by adding 1 mM dithiotheitol (DTT) to the reaction mixture.

To determine the relative proportion of the protease activity that was due to serine proteases, root extracts were preincubated with 50 nM of aprotinin (Sigma, St. Louis) at room temperature for

30 min prior to the addition of the azocasein substrate. The proportion of cysteine protease activity was determined by preincubating the root extract with 0.5mM N-[N-(L-3-trans- carboxyirane-2carbonyl)-L-leucyl]-agmatine (E64) (Sigma, St. Louis) and 1 mM DTT before the enzyme assay was initiated Protease inhibition activity with serine and cysteine inhibitors was calculated as [1 – (activity with inhibitor / activity without inhibitor)] * 100. . Data was analyzed with the R statistical software version 3.2.1 by first doing data transformation until the Shapiro-

Wilk test (Shapiro and Wilk, 1965) confirmed normal distribution, then a t-test analysis was done.

RNA extraction, and quantitative real-time PCR analysis

Leaf and root samples were ground using a ball-mill tissue grinder (Genogrinder 2000;

SpexCentriprep Inc., Metuchen, NJ, U.S.A.) for 2 min at 2,000 strokes/min under liquid nitrogen conditions. RNA was extracted the samples using TRIzol®-chloroform protocol, and treated with DNase (New England Biosciences) following manufacturer’s instruction. RNA content was measured using a Nanodrop (Thermo Scientific) and cDNA was made using High Capacity cDNA Reverse Transcription Kit (ABI, Foster City, CA) following manufacturer instructions.

Quantitative real-time PCR (qRT-PCR) analyses were done for time course expression experiments using primers for acs2 and actin as endogenous control genes, acs2 F: 5’- CTC TTC

TCG TGG ATG GAC CT-3’, R: 5’- CGT TGA GCT TCA CCT TGT GT-3’; actin F: 5’- GGA

GCT CGA GAA TGC CAA GAG CAG-3’, R: 5’- GAC CTC AGG GCA TCT GAA CCT CTC-

3’. The primers were designed with Primer Express software for qRT-PCR (version 3.0) (ABI,

Software (version 1.4) (ABI, Foster City, CA), and analyzed with the R statistical software

(Team, 2015). Data were analyzed by first doing data transformation until Shapiro-Wilk test

(Shapiro and Wilk, 1965) confirmed normal distribution, then a multiple-factor analysis of variances (ANOVA) was done, followed by a significant difference (HSD) Tukey pairwise comparison test.

Results and discussion

We identified total 4878 proteins present in both leaves and roots of control and WCR-infested plants with a 95% FDR. Analysis of variance indicated that 863 proteins with p ≤ 0.05 or less when comparing log ratios of infested and control leaves and roots (Table S3.1). Several types of analyses were performed on the protein abundance ratios of the identified proteins including hierarchical clustering with metabolic annotations, and systems analyses of two protein networks, one built based on protein abundance correlations and the other based on protein- protein physical interactions. These analyses showed many proteins and pathways of interest that were validated by analyzing phytohormone content, gene expression and protease activity.

Heatmap

Clustering analysis was performed on the 863 proteins with significant differences in protein abundance caused by WCR infestation. Proteins were clustered in groups based on their patterns of abundance ratios by using the Euclidian similarity distance method that divided the proteins in

10 family groups (Fig. 3.1; Table S3.2). The protein families and their members that are relevant to plant defenses are described below.

Group 1.

Proteins in this group had abundance ratios (Rin/Rc) that increased in roots during WCR- infestation. It included 211 proteins with root ratios varying from 1.1 to 2.9 while leaf ratios varied from 1.1 to 0.8 (Fig. 3.1, Group 1). Gene ontology analysis of the molecular function of these proteins showed significant enrichment of proteins with oxidoreductase, peroxidase, lipoxygenase, and electron carrier activities (Fig. 3.2a). Peroxidases produce reactive oxygen species (ROS) that serve as signaling molecules that activate plant defenses against insect and pathogen attack (Howe & Jander, 2008; Almagro et al., 2009; Chen et al., 2009). Peroxidases have been linked to direct herbivore defenses because their oxidation of plant phenolic compounds produces quinones that impair insect digestion (War et al., 2012). Some of these peroxidases (AC197758.3_FGP004, GRMZM2G107228_P01, GRMZM2G061088_P01,

GRMZM2G450233_P01, GRMZM2G120517_P02) function in the ascorbate pathway that is associated with the response to fungi or microbes, suggesting that the high abundance of peroxidases in roots could be linked to the interaction with soil fungi or microbial populations in addition to herbivore defenses (Table S3.1). Because peroxidases can be related to plant-microbe interactions (Almagro et al., 2009), it is possible that soil microbes could interact with wounded roots and active plant antimicrobial defense responses. We found several lipoxygenases in this group (GRMZM2G104843; GRMZM2G102760_P01; GRMZM2G156861;

GRMZM2G040095_P02; GRMZM2G109130_P01), which was not unexpected because they function in the JA biosynthetic pathway. We found additional proteins involved in JA

55 biosynthesis including allene oxidase (GRMZM2G077316_P01) and 12-oxophytodienoate reductase 2 (GRMZM2G000236_P01) (Table S3.1).

There were other proteins with several molecular functions potentially involved in defense responses such as lipases (GRMZM2G320298_P01; GDXG) and phospholipases

(GRMZM2G129238_P01; phospholipase C). Phospholipases are well-characterized enzymes that hydrolyze phospholipids and have many roles in cellular regulation, lipid metabolism, membrane remodeling and signaling in response to stimuli (Yoshida, 1979a; Wang et al., 2001), including herbivory and JA accumulation (Yoshida, 1979a; Yoshida, 1979b; Yoshida, 1979c;

Creelman & Mullet, 1997; Wasternack & Hause, 2013). Ribosome inactivating proteins (RIPs)

(GRMZM2G035890_P01, GRMZM2G173809_P01) were upregulated in infested roots. The molecular function of RIPs is to depurinate ribosomal RNA residues and to block translational elongation translation (Endo et al., 1987, Endo and Tsurugi, 1987, Nielsen and Boston, 2001).

Because they often contain lectin-binding domains, they can bind to gut proteins and impair herbivore digestion (Michiels et al., 2010). The maize protein RIP2 (GRMZM2G063536) has been shown to inhibit fall armyworm growth by 20% in bioassays (Chuang et al., 2014).

We found a protein with a putative lectin-domain receptor-like protein kinase family member (GRMZM2G465165_P01), and another RIP protein with a lectin-binding domain

(GRMZM2G173809_P01). The lectin-binding domain has been linked to plant perception of insect attack because it could act as a receptor of carbohydrate fragments from either the herbivore or damaged cell walls (Chen et al., 2009; Vandenborre et al., 2011; Lannoo & Van

Damme, 2014; Macedo et al., 2015). Consequently the kinase protein could help the maize plant perceive and transduce signals that the plant is under WCR attack. Other proteins that could potentially function in perception of insect attack included leucine-rich repeat proteins

(GRMZM2G004572_P01, GRMZM2G121573_P02), signal perception protein with a PAS domain (GRMZM2G032351_P02) (Dunham et al., 2003), and inositol monophosphate synthase

(GRMZM2G051276_P01).

Other types of proteins found in Group 1 are serine protease inhibitors

(GRMZM2G029780_P01, GRMZM2G031572_P01 S28-type, GRMZM2G029641_P01 S10- type). Serine protease inhibitors have been linked to plant defense responses because they inhibit digestive proteases in the insect gut and inhibit their ability to obtain essential amino acids from dietary proteins (Green & Ryan, 1972; Chen et al., 2009). In addition to protease inhibitors, proteins with putative protease activity were identified that included a cysteine peptidase

(GRMZM2G108849_P02 C1-type); and serine type carboxypeptidases

(GRMZM2G339091_P01, GRMZM2G128056_P01 S10-type, GRMZM2G157156_P01 S1C- type). Cysteine peptidases have been linked to plant defenses against pathogens (Bozkurt et al.,

2011; Pogány et al., 2015; Misas-Villamil et al., 2016) and insects (Pechan et al., 2000; Chen et al., 2007; Chen et al., 2009; Louis et al., 2015; Medeiros et al., 2016). In fact, it is known that

Maize Insect Resistance 1-Cysteine Protease (MIR1-CP, GRMZM2G150276) increases in

Mp708 roots in response to WCR infestation (Castano-Duque manuscript submitted). Although the TMT analysis indicated that MIR1-CP abundance in roots increased 1.77-fold in response to

WCR infestation, the increase was not significant (P=0.165).

We found proteins involved in ascorbate biosynthesis (GRMZM2G112792_P01) and glutathione metabolism (GRMZM2G162486_P01). Both ascorbate and glutathione are key elements in the REDOX hub that integrate metabolism and environmental stimuli to modulate plant responses (Foyer & Noctor, 2011). The reduced forms of both ascorbate and glutathione serve as antioxidants in maize, and high levels of reduced glutathione have been linked to maize

57 resistance against aphids (Rhopalosiphum padi L. and Sitobion avenae F.) (Sytykiewicz, 2016).

These results suggest a potential correlation between insect infestation and pools of reduced ascorbate and glutathione.

Finally, other proteins within Group 1 that might have a role in plant defenses are proteins involved cinnamaldehyde metabolism (GRMZM2G179981_P01,

GRMZM2G016836_P02). This compound is capable of inhibiting the biosynthesis of fungal cell walls (Bang et al., 2000) and chitin (Kang et al., 2007) and it attracts southern corn rootworms (Metcalf & Lampman, 1989; Whitworth et al., 2002). The dynamics of cinnamaldehyde as an attractant and a potential deterrent of insects is unknown. Also, higher abundance in infested roots was 4-hydroxy-3-methylbyt-2-en-1-yl diphosphate synthase (HMB-

PP synthase, GRMZM2G137409_P01) an enzyme involved of the non-mevalonate terpenoid biosynthetic pathway. This enzyme produces (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate

(HMB-PP), which is subsequently converted to isopentenyl pyrophosphate (IPP). IPP is a precursor for volatiles such as caryophyllene that are known to participate in plant defenses

(Richter et al., 2015).

The herbicide safener-binding protein of maize (GRMZM2G085924_P01) that protects against injury from chloroacetanilide and thiocarbamate herbicides (Scott-Craig et al., 1998) also increased in abundance, but its role in herbivore defense responses is unknown. The abundance of several additional proteins increased in roots in response to WCR infestation. These included a protein with a saposin-domain (GRMZM2G123029_P03) that might be involved in activation of lysosomal lipid-degrading enzymes (Munford et al., 1995); cupin 1 (GRMZM2G094328_P01) that plays a role in disease (Carrillo et al., 2009) and insect resistance in rice (Zhang et al.,

2013); and spermidine synthase involved in growth and development (GRMZM2G047867_P02)

(Table S3.2).

Group 2 and 3

Two groups in the heatmap had Rin/Rc ratios were less than 1, while Lin/Lc ratios were greater than 1 (Fig. 3.1, Group 2 and 3). These represent proteins that were repressed in the roots, but increased in leaves during WCR infestation. The first group had 270 proteins with Rin/Rc ratios from 0.7 to 1 (Fig. 1, group 2) while the second group had 88 proteins with Rin/Rc ratios less than 0.7 (Fig. 1 group 3). For the first group the GO analysis of molecular function showed enrichment in proteins involved in translation and nucleic acid or RNA binding (Fig. 3.2b).

Furthermore, GO analysis for cellular localization showed significance in proteins that localized in the ribonucleoprotein complex (Fig. 3.2c). All of these GO terms were linked with proteins involved in controlling transcription and translation. Some of the proteins found were: translation elongation factors (GRMZM2G033283_P01; GRMZM2G101859_P01;

GRMZM2G439201_P02; GRMZM2G028313_P02; GRMZM2G082974_P01;

GRMZM2G129804_P01; GRMZM2G113696_P01) and ribosomal proteins. Down regulation of ribosomal proteins in the roots implies that the plant is reducing protein production. Because many of the proteins that increase in abundance in Group 1 are related to herbivore defenses, it is possible that decreased protein synthesis affects proteins and enzymes involved in house-keeping functions (Wasternack & Hause, 2013; Robert et al., 2015) (Table S3.2).

The second group had 88 proteins with Rin/Rc ratios below 0.7. GO analysis showed no significantly enriched terms, nevertheless some proteins in this group are involved in nucleotide binding, ribosomal complexes, regulation of transcription and translation (Table S3.2). The repressed proteome showed decreased levels of transcription and translation and suggests that the

59 plant is diminishing protein production and switching pathways from development and growth to defense.

Group 4.

In the fourth group there were 47 proteins that showed Rin/Rc ratios higher than 1.0 and Lin/Lc ratios lower than 0.8 (Fig. 3.1, group 4). There were no significantly enriched GO terms, but some of the proteins in these group are involved in ABA biosynthesis

(GRMZM2G155889_P04); DNA processing (GRMZM2G008728_P03); nucleosome assembly

(GRMZM2G069911_P02); and formation of flavonoids (GRMZM2G058292_P01) (Table S3.2).

This group was similar to Group 1 due to the high Rin/Rc ratios, the difference was that the systemic organ showed low protein abundance, thus these proteins were associated with local defense responses.

Group 5.

In fifth group included proteins did not have a marked difference in abundance profiles between the two organs, which suggests that they were involved in maintaining plant metabolism. This group had 210 proteins, with Rin/Rc ratios from 1.25 to 0.61 and Lin/Lc ratios from 1.0 to 0.52

(Fig. 3.1, group 5). GO analysis of the molecular function of these proteins showed that only 146 of them had annotations with 46 significantly enriched terms. Among the enriched terms were nitrogen metabolism, amino acid metabolism, carboxylic acid metabolism, redox homeostasis, protein transport, phosphorylation, biological regulation, post-translational protein modification, oxidoreductase activity and ketone metabolism (Fig. 3.3). There were a few proteins in this group that had ratios greater than 1 in roots such as an aldehyde dehydrogenase involved in valine degradation (GRMZM2G001898_P01), small heat shock protein HSP20

(GRMZM2G037146_P01), a cytochrome that transfers electrons between the two reaction

60 center complexes of oxygenic photosynthetic membranes (GRMZM2G448174_P01), lipid transfer protein (GRMZM2G176347_P01), 50S ribosomal protein L24

(GRMZM2G076087_P01) and UDP-glucose-1-phosphate uridylyltransferase activity

(GRMZM2G098370_P02) (Table S3.2). These proteins potentially could be involved in defense responses but further validations should be done to confirm this.

Group 6.

The sixth group consisted of 6 proteins that were repressed during infestation in both organs with ratios below 1 for Rin/Rc and Lin/Lc (Fig. 3.1, group 6). These proteins were a 60S ribosomal protein L35 (AC206642.4_FGP001), chloroplast associated protein (GRMZM2G030240_P01), protein with GH3 and JAR1 domains that is auxin-induced (GRMZM2G162413_P01), histidine kinase-like ATPase domain (GRMZM2G084819_P03), and non-specific serine/threonine kinase

(GRMZM2G429940_P01) (Table S3.2). It seemed that proteins associated with auxin-related responses are repressed in both roots and leaves.

Group 7.

The seventh group of proteins were highly upregulated in leaves and downregulated in roots.

This group consisted of 13 proteins with ratios below 0.8 for Rin/Rc and above 1.9 for Lin/Lc

(Fig. 3.1, group 7). This group contained proteins with the following putative functions: a protein with a signal transduction response regulator receiver domain and cheY-like domain that receives a signal from the sensor domain partner (GRMZM2G341405_P04), a protein with a malectin-like binding domain (AC207019.3_FGP003), a protein with a phospholipase C-domain

(GRMZM2G060194_P02), vingain, a cysteine-type peptidase enzyme similar to papain

(GRMZM2G070011_P01), a protein involved in translation regulation of double stranded RNA

(GRMZM2G094526_P03), a protein with a rare lipoprotein A domain and a double psi-beta

61 barrel (DPBB) similar to pollen allergen (GRMZM2G099092_P01), a MYC/MYB transcription factor (GRMZM2G136831_P02), and a glutharedoxin from group II (GRMZM5G880300_P02)

(Table S3.2). These data suggested that leaves have a higher abundance of signal perception and transduction proteins that could function in transmitting systemic signals from the attacked roots to the leaves.

Group 8 and 9

Groups 8 and 9 had protein abundances that were highly upregulated in response to WCR infestation in only in the roots. This group had 17 proteins with very high ratios of Rin/Rc (2.9 to

5.3) while Lin/Lc ratios are close to or less than 1 (Fig. 3.1, groups 8 and 9). These proteins included a substilin /chymotrypsin-like inhibitor pis7 (GRMZM2G028393_P01), Bowman-Birk type wound-induced proteinase inhibitor (WIP1, GRMZM2G156632_P01) and pathogenesis protein related protein 10 (PR10, GRMZM2G112488_P01). Other proteins involved in signaling and transcription were the calcium binding EF-hand protein (GRMZM2G075456_P01, caleosin- domain), a protein involved in posttranscriptional gene regulation (GRMZM2G582965_P01, dsRBD domain), an unknown protein (GRMZM5G882399_P01) and a transporter

(GRMZM2G061303_P01) (Table S3.2). Proteins involved in direct defenses against insect herbivores and signal transduction pathways were the most highly upregulated proteins in this group. Pathogenesis response genes (PR genes) generally are upregulated in response to microbes (Dixon et al., 1990; Niki et al., 1998; McGee et al., 2001; Takahashi et al., 2004), and it is possible that it functions in WCR-infested plants to protect the roots microbes in the rhizosphere.

Group 10.

This group had two proteins that were highly downregulated in the roots, but upregulated in leaves. Both proteins had Rin/Rc ratios below 0.17 and Lin/Lc ratios near 1.5 (Fig. 3.1, Group

10). These proteins were an uncharacterized protein associated with mitochondria

(GRMZM5G826972_P01) and a protein containing an armadillo domain that is involved in signal transduction (GRMZM2G086648_P04) (Table S3.2).

Protein correlation networks

In addition to the heatmap analysis, a protein correlation network was developed. In this network, each node represented a protein and each edge was a positive protein abundance correlation with a score of 0.99 or higher, meaning that two proteins were linked if their patterns of abundance were similar enough to pass the stringent Pearson correlation parameters (Note

S3.1). The idea behind this network is that similar protein abundance patterns under specific stressors could lead to proteins that have a similar functions during plant herbivore defense responses (Provart, 2012). The network properties were evaluated by using the NetworkAnalyzer

(Doncheva et al., 2012) tool from Cytoscape. Edges were combined to treat the network as undirected since the edges were pattern similarities instead of directed metabolic processes or regulations. NetworkAnalyzer showed that the network had 526 nodes and 6666 edges (Note

S3.1).

The betweenness centrality distribution was used as an indicator of the network connectivity (Yu et al., 2007; McDermott et al., 2012). Betweenness is a network property that measures the number of shortest paths going through a certain node in the network. Higher betweenness values indicate that a particular protein is more critical because the paths of more proteins with similar abundance patterns go through that protein (Yu et al., 2007). In general, it

63 is known that proteins with high betweenness scores tend to play an important role in the organism’s responses to stresses (Azuaje, 2014). We fit the betweenness centrality distribution to the power law (y=axb; a=0.850, b=-1.578, R2=0.77 and correlation=0.754) (Fig. 3.4). This fit showed that there are few proteins with a high betweenness centrality score, and these proteins could be important for plant defense responses. Some of the proteins with the highest betweenness score were involved in translation and transcription such as RNA binding proteins

(GRMZM2G167356_P01, GRMZM5G872147_P02), aminoacyl-tRNA synthetase

(GRMZM2G146589_P01) and translation elongation factor Tu (GRMZM2G057535_P01).

Terpene synthesis (GRMZM2G038153_P01), proteins with oxidoreductase activity

(AC234163.1_FGP002, GRMZM2G178415_P01, GRMZM2G162158_P01), and cysteine protease activity (GRMZM2G070011_P01) (Table 3.1, Table S3.3 and Note S3.1) were also identified. Most of these proteins were similar to the proteins found in the heatmap analysis that showed high abundance in the roots during WCR infestation (Fig. 3.1, Group 1), demonstrating that both correlation network and heatmap analyses revealed similar proteins that potentially function in plant defense responses to root herbivores.

We determined the “hotspots” or hubs in the network using scored p-values of protein hubs where higher scores indicated hubs with more activation (Ideker et al., 2002). Five hubs were found, Hub 1 had 41 proteins and a scored p-value of 10.173, Hub 2 had 7 proteins and a scored p-value of 5.936, Hub 3 had 7 proteins and a scored p-value of 3.564, Hub 4 had 10 proteins and a scored p-value of 3.513, Hub 5 has 12 proteins and a scored p-value of 3.392

(Note S3.1). Proteins in Hub 1 showed higher abundances in infested roots compared to leaves, thus these proteins were highly abundant in the attacked organ during WCR infestation (Fig. 3.5a and Note S3.1). In this hub the Rin/Rc protein ratios varied from 5.26 (GRMZM2G008093_P01,

64 tubulin) to 1.36 (GRMZM2G102183_P02, malate synthase in the glycoxylate cycle) and the

Lin/Lc ratios varied from 1.43 (GRMZM5G882399_P01, uncharacterized protein) to 0.72

(GRMZM2G085924_P01, herbicide safener binding protein).

Among the proteins with high Rin/Rc ratios in Hub 1 was a protein with a 2-oxoglutarate dioxygenase domain that is involved in ethylene and anthocyanin biosynthesis (Table 3.1). Both of these compounds have been linked to plant defenses: ethylene has been proposed as a regulator of defenses (O'Donnell et al., 1996; Harfouche et al., 2006) and anthocyanin accumulation increases the coloration of plant tissues that repels some insects (Schaefer &

Rolshausen, 2006). Other protein was a pathogenesis-related protein 10 (PR10), the PR protein expression could be linked to the presence of microbes in the soil that are interacting with the wounded roots. A calcium binding EF-hand protein was found in Hub 1 (Table 3.1). This protein is involved in the metabolism of oxylipins that are compounds involved in plant defense signaling against insect wounding (Somerville et al., 2000). Several of the proteins in Hub 1, lipoxygenase 3, lipoxygenase 5, lipoxygenase 1, lipoxygenase 6 and a hydroperoxide dehydratase (aos1) are involved in JA biosynthesis (Table 3.1). Several other proteins in this group are involved in plant defense responses such as Bowman-Birk type trypsin inhibitor, clathrin adaptor protein, protein with a Von Willebrand factor type A domain, RIPs, a putative polyphenol oxidase (PPO), a glutathione S transferase and a UDP-glycosyltransfarase (Table

3.1). Both, 1-aminocyclopropane-1-carboxylate oxidase (also called 2-oxoglutarate-dependent dioxygenase) and UDP-glucoside transferase are involved in DIMBOA and DIBOA biosynthesis, which are molecules with insecticidal properties found in maize (Jonczyk et al.,

2008; Erb et al., 2009a; Erb et al., 2009b). Many of these proteins are involved plant defense responses against insects such as JA and DIMBOA biosynthesis pathway, peroxidases, RIPs and

65 polyphenol oxidases (PPO) (Chen, 2008; Howe, 2008; Chen et al., 2009; Kerchev et al., 2012;

War et al., 2012) (Table 3.1, Table S3.3 and Note S3.1).

Protein-protein interaction networks

In addition to the a protein correlation network, we extracted a subnetwork from a layout protein- protein interaction network created by Musungu et al. (2015) based on interacting maize orthologs (interologs). The interologs are predicted in silico to occur in maize and are based on experimentally confirmed interactions in 13 species. These protein interaction events are available on four interactome on-line databases (Musungu et al., 2015). In the subnetwork, we only kept proteins and their interaction partners if they were detected in the TMT dataset. There were 1218 nodes and 3919 edges that represented predicted physical interactions in this subnetwork (Notes S3.2). We fit the betweenness distribution to the power law (y=axb) where y was the betweenness centrality and x was the number of neighbors (a=0, b=1.355, R2=0.325 in logarithmic value and correlation=0.042). Due to the low R2 value obtained using the betweenness distribution, we fit the node degree distribution to the power law (a=423.26, b=-

1.395, R2=0.914 in logarithmic value and correlation=0.96) (Fig. 3.6). Because this had a higher

R2, we concluded that the node degree distribution was a better fit to describe the network overall connectivity. The degree distribution showed that our protein-protein interaction network had a scale-free behavior due to the low number of highly connected proteins and high number of proteins with one or two interactive partners that is common behavior in the World Wide Web and several other biological networks (Albert, 2005; Sekara et al., 2016).

We determined the hubs of highly activated proteins by using a scored p-value to determine the “hotspots” (Ideker et al., 2002). Five hubs were found, the Hub 1 had 175 proteins

66 with a score of 4.93, Hub 2 had 179 proteins with a score of 4.62, Hub 3 had 187 with a score of

4.43, Hub 4 had 200 proteins with a score of 4.43 and Hub 5 had 227 proteins with a score of

4.08. Hub 1 had a wide range of Rin/Rc ratios, from 1.86 (GRMZM5G870932_P01, phosphoenolpyruvate carboxylase) to 0.53 (GRMZM2G481755_P02, a protein part of the membrane bound o-acetyltransferase family involved in phosphatidylcholine acyl editing) (Notes

S3.2). Proteins in Hub 1 had Lin/Lc ratios varying from 1.48 (GRMZM2G040247_P01, a protein involved in vesicle-mediated transport) to 0.37 (AC206642.4_FGP001, a 60S ribosomal L35 protein) (Supplemental Cytoscape file 2). The annotation of the proteins in Hub 1 showed proteins in various GO classes: “response to other organisms” (7 proteins ), “ROS response” (3 proteins), “JA biosynthesis” (1 protein), “ethylene signaling” (1 protein), “ethylene biosynthesis” (1 protein) and “ABA biosynthesis” (1 protein) (Fig. 3.7, Table S3.2 and Notes

S3.2).

The other 161 proteins from Hub 1 not linked to a pathway were evaluated using GO tools in AgriGO. We found that 130 had GO annotations and 34 GO terms were significantly enriched, some of these terms were protein folding, oxidoreductase activity, lyase activity, amino acid metabolism, translation and RNA binding (Fig. 3.7). Many of the GO enriched terms found in this Hub 1 also were found by the correlation network and heatmap analyses. These included oxidoreductases that might be involved in REDOX reactions.

In all the five hubs there were common pathways represented by the proteins in each hub, these common pathways were, ethylene biosynthesis and signaling, ROS metabolism and response, response to other organisms, and JA biosynthesis. As previously discussed in the clustering analysis, all of these pathways have been linked to plant defense responses. It seems that the protein-protein interaction network was able to confirm the activation of already known

67 defense pathways in the plant. These results might be due to the nature of how the layout network used to build our subnetwork was created. A protein will be part of the layout network built by Musungu et al. (2015) if they are maize interologs, thus, interactions that might take place in maize but have not been experimentally documented before will not be included in the network, leaving out possible novel defense proteins and pathways.

Validations of the heatmap and network analyses

Phytohormones

All three of the analysis methods used identified proteins and pathways that showed significant abundance changes, enrichment, and location in a particular hub. We chose to measure phytohormone levels as a way to determine if there were changes in phytohormone pathways as showed by the network analyses. As previously described in chapter II, the JA levels in roots infested by WCR increases significantly while there are no changes in the leaf JA content. This confirms that the proteins present in the heatmap, and network analyses that are involved in JA biosynthesis contributed to increased JA accumulation in roots but not leaves. It was expected that WCR herbivory would suppress SA levels and enhance JA accumulation in roots due to the antagonism between these two phytohormones (Niki et al., 1998; Thaler et al., 2002; Takahashi et al., 2004; Schweiger et al., 2014). Nevertheless, the network analysis predicted an increase in

SA accumulation in roots due to increase in protein abundance of PR10 and peroxidases associated with response to microbes in roots. SA levels in control leaves were approximately

10-fold higher than control roots (Fig. 3.8a). During WCR infestation leaf SA levels decreased by 40%, SA levels in the roots did not change. It is possible that the similar SA levels in control

68 and WCR-infested roots is due to the physical interaction between roots and soil microbes that could preventing the plant from deceasing the SA pathway in this organ.

The heatmap analysis and both network analyses predicted an increase in ethylene biosynthesis in roots infested with WCR. To test this, we examined the response of acs2 that encodes the ET biosynthetic enzyme, aminocyclopropane-1-carboxylic acid synthase (Wang et al., 2002). Previous studies showed that both JA and ET are involved in Mp708 defenses against

FAW and corn leaf aphid (Harfouche et al Ankala et al., 2009, Ankala et al., 2013, Louis et al.), but little is known about the ET pathway in response to WCR attack. When plants were infested with WCR, acs2 expression in roots increased dramatically by day 2 and remained elevated until day 4 (Fig. 3.8b). Expression of acs2 in leaves gradually increased between 2 to 7 days suggesting a delay in acs2 expression in leaves compared to roots (Fig. 3.8b). This behavior of acs2 expression in response to WCR infestation suggests that immediately attacked organs

(roots) might activate ET production first, followed by a delayed response in distal organs

(leaves). The roots of plants infested with WCR for 4 days also showed significant increase in

ET levels compared to 2 days (Fig. 3.8c). This suggests that activating the ET biosynthesis pathway and ET release may be temporally different.

Protease activity

The heatmap and network analyses showed that the levels of several proteases increased in abundance in WCR-infested roots (Table S3.2, Note S3.1, S3.2). To test this finding, total protease activity of WCR-infested roots and leaves was measured and the results showed a 2-fold increase in protease activity in roots, but not leaves of WCR infested plants (Fig. 3.9, p ≤ 0.005).

We then examined the effects of specific serine and cysteine protease inhibitors on total protease

69 activity (Note S1). In WCR-infested roots, the serine protease inhibitor significantly decreased the protease activity (Fig. 3.9b, P=0.01) while cysteine inhibitor did not (Fig. 3.9c, P=0.06).

Previous work has shown that the cysteine protease MIR1-CP is induced above and belowground in response to herbivory in Mp708 (Pechan et al., 2000), but it could be one of a suite of proteases that accumulate in roots in response to WCR. This suggests that serine proteases are the prominent proteases induced in roots by WCR infestation. Serine proteases have been linked to plant defense strategies against coleopterans (Medeiros et al., 2016), but their mode of action as insecticidal proteins has not been elucidated. In other organisms like fungi the mode of action of serine peptidases has been linked to degradation of the insect cuticle and regulation of the pH in the fungi-insect contact interphase (Barelli et al., 2016), perhaps these defense mechanisms could be deployed to protect maize roots from WCR.

Root volatiles and bioassays

Because the abundance of hydroxymethylbutenyl 4-diphosphate synthase (HMB-PP), an enzyme involved in volatile biosynthesis, increased approximately 1.4-fold in infected roots, we measured the release of volatiles from this organ. The levels of benzothiazole (BZO) increase approximately 10-fold in WCR-infested roots (Figure 3.10). BZO is a volatile compound that has antimicrobial, antiviral (Ali & Siddiqui 2013) and antifungal properties (Yadav et al., 2011).

BZO also exerts insecticidal properties against Bradysia odoriphaga (Diptera: Sciaridae) by decreasing the insect capacity to acquire nutrients and reducing its digestive enzymatic activities

(Zou et al., 2007). In maize, BZO has been identified in root exudates as a compound with antimicrobial properties against Phytophthora capsici (Yang et al., 2014). Also, BZO has been detected in rice plants infested with Tibraca limbativentris Stal. (Hemiptera: Pentatomidae)

(Melo Machado et al., 2014) and in mechanically damaged poplar cuttings (Hu et al., 2009). Due to the increase in proteins involved in microbial and insect defense pathways in WCR-infested roots, perhaps the plant is producing volatiles with antimicrobial and insecticidal properties to defend against the insect and other organisms that are interacting with wounded roots. Because proteins involved in microbial and insect defense pathways increase in WCR-infested roots, it is possible that BZO is produced to defend against WCR and opportunistic pathogens that might interact with vulnerable wounded roots. To determine if Mp708 showed resistance and antixenosis responces after 15 d of infestation with WCR, we performed bioassays that showed fewer rootworms recovered from Mp708 compared to its insect-susceptible parental genotype,

Tx601 (P=0.001; Fig. 3.11a, b) and WCR average weight was lower in Mp708 (P=0.003; Fig.

3.11c). These result suggest that there are defense responses against WCR that coexist with defense responses against bacteria and fungi due to the interaction between roots and soil microbiota.

Conclusions

We have successfully evaluated maize defense responses to WCR infestation in roots and leaves using proteomic data by implementing hierarchical clustering and two types of network analyses.

The three analysis methods were used to allow us to determine proteins and pathways of interest at whole plant and at individual organ levels. The methods showed similarities in the pathways of interest, nevertheless the PPI network served as diagnostic to corroborate the involvement of already known defense pathways in the plant, while hierarchical clustering and protein correlation networks allowed discovery of proteins that could have novel roles in plant defense responses. These analyses revealed proteins with higher abundance during WCR infestation that

71 involved JA, ABA and ET biosynthesis and signaling pathways. The three analysis methods showed that proteases of serine/cysteine type are involved in plant defenses in infested roots. We determined that total protease activity increased in roots with WCR treatments, and that serine- type proteases are highly contributing to this activity. We do not know the physiological damage that this type of proteases causes to WCR, but we hypothesized that they could be breaking down cuticle on the outer layers of the larvae. Activation of ET biosynthetic pathway was found in both network analyses and volatile analysis showed that ET production goes up in infested roots.

We propose that this phytohormone might be involved in short distance communication due to its high diffusion from plant tissues into the soil and surrounding areas (Zechmeister-Boltenstern

& Nikodim, 1999) or might be an intra-plant communication signal produced by the roots that could be transported to the leaves through aerenchyma in the roots (Colmer, 2003). The three analyses indicated that roots infested with WCR also showed responses against other biotic stresses like fungi and pathogens. This idea can be correlated to plant physiological responses like the production of the volatile BZO an antimicrobial, antifungal (Yadav et al., 2011) and insecticidal (Zhao et al., 2016) compound. We postulate that the plant is constantly balancing its defenses against root herbivores and other biotic stressors because the roots are in constant contact with soil microorganisms. Future research directions could focus on understanding the plant-insect-microbe interaction balance and how it benefits the plant.

In terms of individual plant organ responses, we observed a distinct difference between the local and systemic responses to WCR. The roots responded by increasing JA, ET and ABA biosynthesis and production of insect defenses such as proteases, serine-type proteases,

Bowman-Birk type trypsin inhibitor, RIPs, peroxidases, lectin-type proteins, cinnamaldehyde, terpenoids and benzothiazol (Fig. 3.12). On the other hand, leaves tended to increase production

72 of proteins that control transcription and translation, and proteins with signaling domains (Fig.

3.11). Upregulation of signal perception proteins in leaves could imply that systemic tissues are reading themselves to receive a signal from roots that could lead to increasing organ specific responses and change the metabolic focus from growth to controlling protein production.

Finally, we have established an analysis pipeline for proteomics data that includes network biology approaches that can be used with different types of “omics” data from a wide variety of organisms. Hierarchical clustering as well as protein correlation networks analysis can serve as discovery tools to find novel defense protein or pathways. Several hormonal biosynthetic and signaling pathways of JA and ET were activated in the roots, while in leaves there was an increase in signaling perception proteins that could be related to systemic defense activation. In roots there was evidence for the coexistence of plant defense responses to microbes and fungi in addition to those involved in herbivore defense.

Figure 3.1. Hierarchical clustering of protein abundance ratios of Mp708 maize infested with

WCR for 4 days. The heatmap was built using the proteins ratios from Lin:Lc and Rin:Rc from proteins that showed statistically significant differences in abundance (P ≤0.05). The dendogram in the heatmap was generated by applying Euclidian similarity distance method to determine abundance patterns using a maximum of 10 k-means cut off. The 10 k-means are labeled as protein families, they are color coded in the bar next to the dendogram and each family was enclosed by a red box, 1 (□), 2 (□), 3 (□), 4 (□), 5 (□), 6 (□), 7 (□), 8 (□), 9 (□), and 10 (□).

(a)

(b) (c)

Figure 3.2. Gene ontology (GO) hierarchical maps of (a) the molecular function for proteins from Group 1 in the clustering analysis. (b) The molecular function and (c) cellular component

76 for the proteins from group 2 in the clustering analysis. GO annotations were performed using

AgriGO on line tools, in boxes: color represent statistical significant terms, the darker and redder the color the significance is higher, the numbers in the parenthesis are the enrichment multiple- test adjusted p-value, first ratio is the number of GO terms of proteins that were searched from protein groups 1 or 2 and background (Maize Genome Locus from maizesequence.org), and the second ratio is the total number proteins queued and background (Maize Genome Locus from maizesequence.org).

Figure 3.3. Gene ontology annotation enrichment of the proteins that belong to Group 5. These are 46 GO terms significantly enriched in Group 5 from the heatmap analysis. Blue bars are the queue list of the proteins that had GO annotations from the 210 member of Group 5 and green

78 bars are the background list of proteins associated to a specific GO terms in the whole maize genome (Maize genome locus from maizesequence.org)

Figure 3.4. Betweenness centrality distribution of the protein abundance correlation network fitted to the power law (y=axb) where y was betweenness centrality and x is number of neighbors

(a=0.850, b=-1.578, R2=0.77 in logarithmic value and correlation=0.754).

Figure 3.5. Hotspots or modules with the highest scored p-value from (a) protein abundance correlation subnetwork, inner colors represent the Rin : Rc ratios and outer colors the Lin : Lc.

(b) hotspot from the protein-protein interaction network with their pathway annotations. These hotspots were created using the jActiveModules tool in Cytoscape.

Table 3.1. Proteins of interest from the hotspot with the highest scored p-value generated from the protein abundance correlation network.

GRMZM Identification Name Rin/Rc Lin/Lc number ratio ratio AC212219.3_FGP005 protein with 2-oxoglutarate dioxygenase 5.07 1.11 domain GRMZM2G112488_P01 pathogenesis-related protein 3.95 1.10 GRMZM5G892627_P01 UDP-glucoside transferase 3.78 1.29 GRMZM2G075456_P01 calcium binding EF-hand protein 3.45 1.11 GRMZM5G815098_P01 bowman-Birk type trypsin inhibitor 3.4 1.13 GRMZM5G883149_P03 clathrin adaptor protein 2.93 0.93 GRMZM2G110567_P01 protein with a von willebrand factor type A 2.73 0.85 domain GRMZM2G156861_P02 lipoxygenase 3 2.26 1.05 GRMZM2G102760_P01 lipoxygenase 5 2.17 0.96 GRMZM2G156861_P01 lipoxygenase 1 2.02 0.98 GRMZM2G040095_P02 lipoxygenase 6 1.87 0.95 GRMZM2G067225_P01 hydroperoxide dehydratase (aos1) 1.65 1.08 GRMZM2G035890_P01 ribosome inactivated protein 1 1.95 0.98 GRMZM5G851266_P01 putative polyphenol oxidase 1.95 0.98 GRMZM2G335618_P01 glutathione S transferase 1.88 0.91 GRMZM2G173809_P01 ribosome inactivated protein with ricin 2 1.76 1.11 domain GRMZM2G304712_P01 UDP-glycosyltransfarase 1.69 0.93 AC148152.3_FGP005 1-aminocyclopropane-1-carboxylate oxidase 1.63 0.9 GRMZM2G079348_P01 catalase isoenzyme 3 1.54 0.92

Figure 3.6. Node degree distribution of the protein-protein interaction network fitted to the power law (y=axb) where y was number of nodes and x is node degree (a=423.26, b=-1.395,

R2=0.914 in logarithmic value and correlation=0.96).

Figure 3.7. Gene ontology of the molecular function for the hotspots with the highest scored p- value from protein-protein interaction network. This are the statistical significantly enriched terms of the secondary GO level. Blue bars are the protein from the hotspot and green bars are

85 the background used for the enrichment analysis (Maize Genome Locus from maizesequence.org).

SA from roots and leaves (a)

(b acs2 (c) ET from roots )

Figure 3.8. Analysis of salicylic acid (SA) and ethylene (ET) levels in Mp708 maize plants infested with WCR. (a) SA levels were determined in roots and leaves after 4 days of infestation

87 with WCR. (b) Relative expression (RQ) of acs2, which is involved in ET synthesis was measured by qRT-PCR. (c) Gene expression levels were normalized to actin. ET production was measured in roots at various intervals from 2 to 7 days of WCR infestation. Control plants were not infested with WCR. RQ and ET data were normalized and analyzed using multiple-factor

ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test.

Letters represent results of the HSD and error bars show the standard error.

(a)

(b (c) )

Figure 3.9. Analysis of (a) protease activity in leaves and roots of Mp708 genotype infested with

WCR at V3 stage after 4 days of infestation with WCR. Analysis of percentage of protease

89 activity inhibited by using (b) serine protease inhibitors or (c) cysteine protease inhibitors.

Control plants were not infested with WCR. Protease activity is measure in terms of absorbance at 440 nm using azocasine as substrate, specific activity was used by measuring total protein content in samples. Inhibition of cysteine proteases was done by using 0.5mM E64 and of serine proteases by using 50nM aprotinin. Specific activity was analyzed using multiple-factor

ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test.

Letters represent results of the HSD and error bars show the standard error.

Figure 3.10. Analysis of benzothiazol volatile production from roots of Mp708 genotype after

18 days of infestation with WCR. Control plants were not infested with WCR. Volatile abundance data were normalized and analyzed using t-test. Letters represent results of the t-test and error bars show the standard error.

(a) Mp708-WCR Tx601-WCR

(b) (c)

Figure 3.11. WCR bioassay performed in Mp708 and Tx601 maize genotypes at 18 d. (a) Picture of WCR recovered from Mp708 and Tx601 roots. (b) Percent survival of WCR fed on Mp708 and Tx601 maize genotypes. (c) Average worm weight from the WCR recovered. Percent survival and average worm weight were and analyzed by using Wilcoxon test, the results

92 of the are represented by the letters (P<0.05) and the error bars represent the standard error. White bar indicates 1 cm.

(a) (b)

Up regulated pathways: Up regulated proteins: JA biosynthesis Ribosomal proteins ABA biosynthesis Translational and transcriptional ET biosynthesis/Perception machinery Proteins involved in responses to other Control of transcription and translation organisms Protein containing an armadillo domain

Up regulated direct defenses: Up regulated direct defenses: RIPs production Vingain, a cysteine-type peptidase Proteases (Cysteine/Serine type) enzyme Peroxidases DIMBOA biosynthesis Up regulated signaling: Serine protease inhibitors Proteins with sensor-domains Cinnamaldehyde biosynthesis Protein with a phospholipase C-domain Cupin-1 Protein with a rare lipoprotein A domain Up regulated indirect defences: Protein with a malectin-like binding Terpenoid metabolism domain Cinnamaldehyde biosynthesis DIMBOA biosynthesis Down regulated Defense related proteins that were up Up regulated signaling: regulated in roots Lectin binding proteins Persoxidases Oxylipin biosynthesis

Down regulated Transcription and translational machinery

Figure 3.12. Summary of defense related responses in the maize (a) roots and (b) leaves under

WCR infestation determined by using the clustering and network analysis. Color in boxes represent orange for roots responses and green for leaves responses.

Table S3.1 Proteins identified in leaves and roots of control and infested plants at 95 % false discovery rate. Logarithm base 10 of protein abundance ratios were analyzed using multiple- factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test and letters were assigned to each treatment detonating significant differences.

Table S3.2. Proteins present in the hierarchical clustering analysis generated by applying an

Euclidian similarity distance method to determine abundance patterns using a maximum of 10 k- means cut off. The 10 k-means are labeled as protein families and represent 10 different protein abundance patterns.

Note S3.1. Protein abundance correlation network built with Cytoscape (562 nodes and 13332 edges).

Table S3.3. Proteins and their annotations present in the abundance correlation network from the

Note S2.

Note S3.2. Protein-protein interaction network built with Cytoscape (1218 nodes and 3919 edges).

Table S3.4. Proteins and their annotations present in the protein-protein interaction network from the Note S3.

CHAPTER IV. Protein Networks Reveal Tissue-Specific Defense Strategies in Maize in

response to an aboveground herbivore

Introduction

Constitutive and induced herbivore defenses in plants vary in the leaves and roots depending on the site of insect attack, this is the case for insecticidal proteins (Ankala et al., 2013), phytohormones and metabolites (Erb et al., 2012; Gulati et al., 2014). Some of these defenses could be elucidated by conducting “omics” and network biology analysis of plant responses to herbivory in leaves and roots. A better understanding of these complex plant defense responses to insect attack, will facilitate the design of environmentally sustainable strategies for protecting crops from herbivore pests that include breeding for more insect resistant plants.

The maize genotype Mp708 was developed for resistance to fall armyworm (FAW,

Spodoptera frugiperda) and other caterpillar pests (Williams et al., 1990) by traditional plant breeding. The FAW is a serious pest of maize in the US, Latin America (Brooks et al., 2007;

Eleazar Ruiz-Najera et al., 2007) and Africa (Goergen et al., 2016). Recently, however, it has been shown to have enhanced resistance to the corn leaf aphid (CLA, Rhopalosiphum maidis)

(Louis et al., 2015) and western corn root worm (WCR, Diabrotica virgifera virgifera L.) (Gill et al., 2011). Mp708 was developed from Antiguan landrace that was most likely selected by farmers because it yielded well under high insect pressure (Williams et al., 1990). Mp708 has constitutively elevated levels of jasmonic acid (JA) that further increase in response to FAW feeding (Shivaji et al., 2010). This maize genotype rapidly mobilizes a toxic cysteine protease,

Mir1-CP (maize insect resistant cysteine protease-1) (Pechan et al., 2000), from roots to leaves in response to FAW attack (Lopez et al., 2007). One of the parents of Mp708 is the susceptible

96 genotype, Tx601 that does not accumulate Mir1-CP or has elevated JA levels (Jiang et al., 1995;

Shivaji et al., 2010). The jasmonic acid (JA) signaling pathway enhances resistance to chewing insects such as FAW (McConn et al., 1997; Fonseca et al., 2009; Galis et al., 2009) and the JA family of compounds activates local and systemic plant responses to foliar herbivory (Ankala et al., 2009; Koo & Howe, 2009; Erb & Glauser, 2010) that leads to biosynthesis and accumulation of defensive proteins such as Mir1-CP (Ankala et al., 2013), RIP2 (ribosomal inactivating protein 2), (Chuang et al., 2014), chitinases and protease inhibitors (Ballare, 2011; War et al.,

2012).

In some cases, the phytohormone ethylene (ET) interacts with the JA pathway to either enhance or suppress its activity (Schmelz et al., 2003; Erb & Glauser, 2010; Broekgaarden et al.,

2015). ET has been studied extensively in relation to abiotic stresses (Schellingen et al., 2015),

(Sarazin et al., 2015) (Arraes et al., 2015) and herbivory in maize (Schmelz et al., 2003; Ankala et al., 2009; Zhu et al., 2011; Louis et al., 2015). Studies indicate that both JA and ET are involved in defenses against FAW in Mp708 (Ankala et al., 2009; Ankala et al., 2013), but there is evidence of JA and ET antagonism in Arabidopsis (Song et al., 2014) and tomato (Tian et al.,

2014) in response to herbivory. The impact of JA and ET on induction of maize defenses in response to insects like FAW (Ankala et al., 2009) and CLA (Louis et al., 2015) has been shown.

Neverthless, an organ-specific study examining ET production, signaling and downstream control over defense protein accumulation in maize has not been done.

In this study we used multidimensional analysis of plant defenses against FAW to understand organ-specificity responses in leaves and roots. Multidimensional analysis integrates different sources of data, such as physiological, biochemical, and high-throughput OMICS to generate a comprehensive picture of plant responses to stimuli (Weckwerth, 2011; Barah &

Bones, 2015). For example, a metabolomics study on maize seedling leaves and roots found several significant metabolite changes in these tissues during infestation with Spodoptera littoralis (Marti et al., 2012). The study showed that leaves had a high abundance of HDMBOA-

Glc and HDM2BOA-Glc, benzoxazinoid derivatives reported to deter insect attack. In roots, as well as in leaves, the only significant change in metabolites was increased tryptophan levels, which is a precursor of benzoxazinoid synthesis (Marti et al., 2012). Other high-throughput data studies that used hierarchical clustering showed that oral secretions from Mythimna separata elicited a long-term defense response in maize that included transcription factors, JA pathway metabolites and terpene-derived VOCs (Qi et al., 2016). To our knowledge there are no plant- insect “omics” studies that used network biology analysis to integrate multiple levels of data to explain the plant phenotype (Weckwerth, 2011; Bassel et al., 2012; Studham & Macintosh,

2012). This type of analysis could find novel pathways, proteins of interest and differentiate organ-specific defense responses.

We studied the responses Mp708 to FAW feeding using high throughput tandem mass tagged (TMT) proteomics with network biology analysis to discover defense responses and mechanisms in leaves and roots. Our research focused on elucidating organ-specific defense mechanisms against FAW in a maize genotype with known insect resistance. The proteomic data were used for hierarchical clustering and network biology analysis using correlation and protein physical interaction networks. These networks also were combined with previously identified

QTL regions associated with resistance (Williams et al., 1990; Brooks et al., 2005; Brooks et al.,

2007). Finally we performed validations of proteins and pathways of interest using gene expression analysis, phytohormone content, immunoblots, leaf and root growth analysis, carbon content and peroxidase activity assays. Ideally the proteomics data and analysis pipelines from

98 this study could facilitate the development of future network biology analyses analyzing plant insect interactions.

Materials and methods

Plants and insects

Seed for the maize genotypes, Mp708 (insect-resistant) and its susceptible parent, Tx601

(Williams et al., 1985) were provided by Dr. Paul Williams, USDA-ARS, Mississippi State

University. Plants were grown in Hagerstown Loam in the Plant Science glasshouse at The

Pennsylvania State University in 8 cm x 9 cm pots until they reached the V3 (three fully expanded leaves, (Ritchie et al., 1998)). Temperatures in the greenhouse were kept between 22 to 27ºC and supplemental lightening was used. FAW eggs were obtained from Benzon Research or Dr. Williams’ laboratory and caterpillars were reared on artificial diet (Ali et al., 1990) until they were ready for infestation (3rd larval instar).

Insect infestation and tissue collection

Infestation experiments were performed when the plants reached V3 stage (Ritchie et al., 1998) by placing one FAW larva on the whorl of each Mp708 and Tx601 plant. Insects were allowed to feed up to 8 h and control plants were not infested with insects (including the 0 h time point).

Leaves and roots were cleaned, weighed and stored at -80°C until further use. Pooled tissues of root tips or leaves were collected using three plants per biological replicate for proteomic analyses and two plants per biological replicate for the other experiments. In proteomic, gene expression and immunoblot experiments a minimum of three biological replicates per treatment were used.

Protein extraction for proteomic analysis

For the proteomics experiment root tips and leaves (0.5g per biological replicate) were collected from control and 24 h FAW-infested Mp708 seedlings. The tissues were ground using a ball- mill tissue grinder (Genogrinder 2000; SpexCentriprep Inc., Metuchen, NJ, U.S.A.) for 2 min at

2,000 strokes min-1 under chilled conditions. Extraction buffer (1.8 M sucrose, 0.5 M Tris-HCl pH 8.5, 0.1 M EDTA, 4% β-mercaptoethanol and protease inhibitor) was added to each sample

(1:5 ratio) vortexed and left on ice for 5 min, then Tris-saturated phenol pH 7.5 was added (1:1 ratio) vortexed and centrifuged at 6,000xg for 10 min at 4°C. Supernatant was collected, then five volumes of cold precipitation solution (0.1M ammonium acetate in methanol) were added and left overnight at -20°C. Samples were centrifuged at 6,000xg for 10 min at 4°C, the pellet was washed three times with 100% cold acetone followed by centrifugation at 6,000xg for 5 min at 4°C. Pellets were air dried then rehydrated with 50 µl of buffer (0.1 M ammonium bicarbonate and 4 M urea) and vortexed for 2 h at room temperature, then the supernatant was recovered and

50 µl of rehydration buffer was added to the pellet which was vortexed again for 2 h at room temperature, and finally both supernatants were mixed. Protein quantification was performed on the mixed supernatants using the NITM protein assay kit (G BIOSCIENCES, St. Louis, MO).

The supernatant (50 µg of protein) was reduced with 250 mM dithithreitol (DTT) then alkylated with 250 mM of iodoacetate in the dark, and the alkylation reaction was stopped by incubating with 5 mM DTT at room temperature. The urea content in samples was diluted to 1.2 M by adding 50 mM Tris-HCl pH 8.0 and 5 mM calcium chloride.

Enzyme digestion

100

Enzyme digestion was performed by adding trypsin (Sigma, St. Louis) in 1:50 (enzyme:protein) ratio and lysyl endopeptidase (Wako Pure Chemicals, Japan) in 1:90 (enzyme:protein) ratio and incubated overnight at 37°C. Then desalting of the digested sample was performed using Bond

Elut OMIX C18 pipette tips (Agilent, Santa Clara, CA) following manufacturer instructions.

Samples were dried in a Speedvac system (Thermo Fisher Scientific, Waltham, MA) at room temperature.

Tandem mass tags (TMT) labeling

Samples were rehydrated in 100 µl of 100 mM triethylammonium bicarbonate buffer (TEAB).

The10-plex TMT (ThermoFisher Scientific, Grand Island, NY) tags were suspended in 41 µl of

100% acetonitrile and each one added to the rehydrated sample for labeling. Samples were labeled in the following order: infested leaf bioreplicates 1 (126), 2 (127N) and 3 (130C); control leaf bioreplicates 1 (127C) and 2 (128N): infested root bioreplicates 1 (128C), 2(129N) and

3(131); control root bioreplicates 1(129C) and 2 (130N). Labeled samples were shaken for 1 h at room temperature and the reaction was quenched by adding 0.8 µl of 50% hydroxylamine. All samples were combined, dried in a vacuum concentrator and sent to the proteomics core facility at the University of Wisconsin, Madison for further analysis.

High pH sample fractionation

Dried TMT samples were resolubilized in 250 µL of 9.5 mM ammonium formate in water, pH

10. The TMT labeled mixture was fractionated by pH using a Phenomenex Gemini 5µm C18

110Å, 4.6mm x 250mm column at a flow rate of 1.0 ml/min with 9.5 mM ammonium formate in

80% acetonitrile in water, pH 10 solvent. Sample detection was conducted with a Waters 2996

101 photodiode array detector monitoring at 214 nm and 280 nm. Fractions were collected every 1 min from the time of injection to the end of column re-equilibration. Recovered fractions were pooled into one mixed fraction every 6th fraction to generate five total pooled fractions. Pooled fractions were frozen and lyophilized. Each fraction was resolubilized in 200 µl of 0.1% formic acid in water and the pH was checked. For fractions with pH >3, formic acid was added to bring the pH down to 3. Aliquots were then placed in autosampler vials for high-pressure liquid chromatography with mass spectrometry (HPLC/MS/MS).

Mass spectrometry

HPLC/MS/MS was performed using an Agilent 1100 Nanopump system with MicroWPS autosampler. For each sample, 3 µL was loaded onto a Thermo Easyspray C18 column, 75µm x

MS/MS.

Protein identification and quantitative analysis

102

Protein identification analysis was performed using Proteome Discoverer (Thermo) version

Quantitative analysis using the TMT reporter ions was performed by Proteome

Discoverer following database searching. Peptides were rejected from quantitation if they were not unique to a protein, MS/MS spectra were rejected if they had a co-isolation of >75% (i.e.

103 abundance for each reporter ion channel was then normalized to the median value for that channel, correcting for deviations in protein inputs, digestion efficiency, and labeling.

Analysis of variance

Protein abundance data was transformed into ratios by taking each bioreplicate of infested leaves or infested roots (Lin and Rin) and dividing them by the average of their respective control bioreplicates (Lc and Rc). The control ratios were determined dividing each control bioreplicate with the average of the control bioreplicate (for example, Rc:Rcaverge). Ratios were transformed using Log10 and analyzed by ANOVA using R (Team, 2015) then a honest significant difference

(HSD) Tukey pairwise comparison test was performed.

Heatmap generation

The heatmap was built by taking the Log10 ratios of leaves infested:leaves control (Lin:Lc) and roots infested:roots control (Rin:Rc) to do hierarchical clustering in R. We used only the proteins that had significant differences in abundance (p≤0.05). The similarity dendogram in the heatmap was built using the Euclidean distance method in R and we determined abundance patterns using a maximum of 10 k-means cut off. These 10 k-means groups became proteins families that were individually annotated by using Biomart maize RefGen_V4. Parametric analysis of gene set enrichment (PAGE) was done on the protein list and ratios from the hierarchical clustering map by using agriGO online tools. The PAGE method we used takes into account the expression levels when determining gene ontology (GO) enrichment (Kim & Volsky, 2005). Also, within each clustered group singular gene ontology enrichment analysis was performed using agriGO online tools (Du et al., 2010) with the TMT protein list from the group as input list and the maize

104 genome locus from maizesequence.org as background. The significant GO terms for molecular function were visualized using REVIGO online tools (Supek et al., 2011) and Cytoscape.

Protein correlation networks

Protein abundance adjacency matrixes generated by Pearson correlation analyses were visualized using Cytoscape (Cline et al., 2007). The adjacency matrix was built using Pearson correlation on the Log10 transformed ratios, a filter correlation of 0.99 and filter abundance value of 0.001 from an in-house code in R (McDermott et al., 2012). The matrix was imported to Cytoscape version 3.2.1 (Cline et al., 2007) to build the visualization and annotations were added to each protein by importing the gene ontology data from gramene.org using the ensembl and biomart tools for Z. mays. Topological analysis of the network was done using the NetworkAnalyzer tool from Cytoscape (Doncheva et al., 2012). The betweenness centrality distributions were fit to the power law (y=axb) where y was the betweenness centrality and x was the number of neighbors.

Gene ontology enrichment was performed on all the proteins in the network by using ClueGO tool from Cytoscape (Bindea et al., 2009) that created a functionally grouped annotation network for biological process and molecular function ontologies. A subnetwork was created keeping only the proteins that fall within a known QTL for FAW resistance. We used jActiveModules to determine the subnetworks of highly activated proteins using a scored p-value to determine

“hotspots” (Ideker et al., 2002). The scored p-value was developed to rank the highly connected subnetworks that had significant differences in abundance under certain stimuli using a scoring system of protein p-values (Ideker et al., 2002).

Protein-protein interaction (PPI) networks

105

The protein-protein interaction (PPI) subnetwork was built using identified proteins and the log transformed ratios from the TMT analysis. The subnetwork was extracted from predicted protein-protein interaction network (PPI) from maize (Musungu et al., 2015) using Cytoscape version 3.2.1 (Cline et al., 2007). This available PPI used interacting orthologs (interologs) of 13 different species (Oryza sativa, Arabidopsis thaliana, Homo sapiens, Mus musculus, Rattus norvegicus, Drosophila melanogaster, Caenorhabditis elegans, Saccharomyces cerevisiae, Schizosaccharomyces pombe, Escherichia coli, Bacillus subtilis, Helicobacter pylori, Campylobacter jejuni, and Synechococcus). These protein-protein interactions were predicted in silico to occur in maize and are based on experimentally confirmed interactions available on four interactome databases [BioGRID (Stark et al., 2006); DIP (Salwinski et al.,

2004); IntAct (Aranda et al., 2010); and MINT (Chatr-aryamontri et al., 2007). We used jActiveModules to determine the subnetworks of highly activated proteins and highly activated component by using a scored p-value to determine “hotspots” (Ideker et al., 2002). Gene ontology enrichment was performed on all the proteins in the network by using ClueGO tool from Cytoscape (Bindea et al., 2009).

Plant growth, root architecture and carbon content

Plant height of Mp708 and Tx601 was measured to determine if growth changed in response to insect infestation. Plant height was measured at 6 and 9 days followed by cleaning and collecting roots and whole leaves. The samples were dried for 5 days at 55ºC and five root systems from each treatment were stored in 75% ethanol to measure root length.

Root lengths of Mp708 and Tx601 were examined to determine if there were changes after FAW infestation. Root samples used were removed from the 75% ethanol and root types

106 were separated from the base of the mesocotyl; all the root segments were scanned using a flatbed scanner at a resolution of 400 dots per inch (Epson Expression 1680, Seiko Epson

Corporation, Suwa, Japan). Root types were separated as thin s (>0mm and <=0.2mm), seminal- like 1 (>0.2 and (=0.5mm), seminal-like 2 (>0.5mm and <=1mm) and nodal (>1 and <=6mm).

Total root lengths of each root type were analyzed by WinRhizo software (Regent Instruments,

2008) (Burton et al., 2012). ANOVA was performed on the root length data using R.

Dried leaves and roots were ground using a ball-mill tissue grinder (Genogrinder 2000;

SpexCentriprep Inc., Metuchen, NJ, U.S.A.) for 5-10 min at 2,000 strokes min-1 under liquid nitrogen conditions. Samples were analyzed for carbon content using a CE Instruments EA 1110

CHNS-O elemental Analyzer by complete and instantaneous oxidation of the sample by "flash combustion" (Matejovic, 1996). The combustion products are separated by a chromatographic column and detected by the thermal conductivity detector (T.C.D.). ANOVA was performed on the carbon content data using R.

RNA extraction, and real-time PCR analysis

Fresh leaf and root samples were ground using a ball-mill tissue grinder (Genogrinder 2000;

SpexCentriprep Inc., Metuchen, NJ, U.S.A.) for 2 min at 2,000 strokes min-1 under liquid nitrogen conditions. RNA was extracted from all ground tissues using TRIzol®-chloroform protocol, and treated with DNase (New England Biosiences) following manufacturer instructions. RNA content was measured using a Nanodrop (Thermo Scientific) and cDNA was made by using High Capacity cDNA Reverse Transcription Kit (ABI, Foster City, CA) following manufacturer instructions. Real-time PCR analyses were done for time course expression experiments using primers for lox3, mir1 and rip2. actin was used as endogenous

107 because previous studies have shown that it expression is not affected by herbivory (Chuang et al., 2014). The primers follow: lox3 F: 5’- GCT ACG TAC GAG CTG GTA CAT GAA-3’, R:

5’- GCC GCT CTC TTC CCG TTT-3’; acs2 F: 5’- CTC TTC TCG TGG ATG GAC CT-3’, R:

5’- CGT TGA GCT TCA CCT TGT GT-3’; mir1 F: 5’- GAG GGT CTT GTC GTG TTG AAC

TT-3’, R: 5’- GCC ACA CCA TAA CGG ATT AAC TT-3’; rip2 F: 5’-GAG ATC CCC GAC

ATG AAG GA-3’, R: 5’-CTG CGC TGC TGC GTT TT-3’; actin F: 5’- GGA GCT CGA GAA

TGC CAA GAG CAG-3’, R: 5’- GAC CTC AGG GCA TCT GAA CCT CTC-3’. Primers were designed using Primer Express software for real-time PCR (version 3.0) (ABI, Foster City, CA).

The PCR conditions used were: step 1: 50°C for 2 min and 95°C for 10 min, step 2: 95°C for 15 sec and 60°C for 1 min for 40 cycles, Step 3: 72°C for 10 min, Step 4: dissociation stage. The relative quantification values were obtained by using ABI 7500 Fast SDS Software (version 1.4)

(ABI, Foster City, CA), and analyzed with R statistical software (Team, 2015). The data were analyzed by first doing Log10, square root, inverse power or box-cox data transformation until the Shapiro-Wilk test (Shapiro and Wilk, 1965) confirmed normal distribution, then a multiple- factor analysis of variances (ANOVA) was done, followed by a significant difference (HSD)

Tukey pairwise comparison test in R version 3.2.1.HSD) Tukey pairwise comparison test.

Ethylene (ET) measurements

Plants infested with FAW as previously described were used. The whole root system and leaf tissues were cleaned and placed in 180 ml glass flasks and sealed for 24 h at 25°C. The tissue

FW was recorded for each sample. After the incubation, an air sample (200 µl) was taken from the flasks (200 µl) using a gas tight syringe and directly injected into an Agilent 6890 gas chromatograph, equipped with a flame ionization detector and Rt-QS-BOND column (30 m x

108

0.53 mm ID x 20 micron film thickness, Restek, Bellefonte, PA). The gas sample was analyzed using helium carrier gas at an initial flow rate of 9 mL min-1 with an average velocity of 69 cm s-

1, injector temperature of 150°C. The run was conducted isothermally at 70°C with the detector at 200°C. The amount of ET was quantified by comparing the chromatograph of the samples against a standard curve made by using a 9µg mL-1 ET in helium standard (Restek, PA) and data were analyzed with R statistical software. Data were analyzed by first doing Log10, square root, inverse power or box-cox data transformation until the Shapiro-Wilk test (Shapiro and Wilk,

1965) confirmed normal distribution, then a multiple-factor analysis of variances (ANOVA) was done, followed by a significant difference (HSD) Tukey pairwise comparison test.

Hormonal inhibition analysis

Production of JA was inhibited by treating plants with ibuprofen [α-methyl-4-(2-methylpropyl) benzeneacetic acid] (Sigma-Aldrich) a lipoxygenase inhibitor (Hu et al., 2006). An ibuprofen solution (5 mM in water) was used to water the soil (10 ml) and spray leaves of Mp708 plants at

6 and 24 h before FAW infestations. Controls were non-infested plants treated with IBU, or plants that were untreated (NS treatment). ET perception was blocked using 1- methylcyclopropene (1-MCP), commercially available as Ethylbloc (1-MCP) (Floralife,

Waterloo, SC, U.S.A.). The Ethylbloc solution was prepared at a final concentration of 5 mg L-1 following the manufacturer’s protocol (Harfouche et al., 2006) and previous Ethylbloc experiments (Ankala et al., 2009). A beaker containing this solution was placed within a

Plexiglass chamber that enclosed up to 12 plants and each ETB treatment lasted 6 h. Plants were placed in the chamber 24 h before FAW infestation. Controls were non-infested plants treated

109 with ETB or untreated plants (NS treatment) placed in Plexiglass chambers, but not treated with

ETB.

Protein extraction, quantification and immunoblot analysis

Root samples were ground as described in the RNA extraction steps. Then 500 µl of 2X

Laemmli sodium dodecyl sulfate (SDS) sample buffer (Laemmli, 1970) was added and tubes were vortexed. Samples were boiled in a heat block for 10 min, vortexed, and centrifuged at

13,000×g for 5 min. The supernatant containing the dissolved proteins was collected and protein quantification analysis was done using a NI™ (Non-Interfering™) protein assay (G-biosciences,

USA) following manufacturer instructions and samples were stored at –80°C. Immunoblot analysis were done by pooling 5 µg of each biological replicate for a total of 15 µg per treatment.

Pooled proteins were run in 12% polyacrylamide gels using constant voltage (10 min at 180 volts and 35 min at 200 volts), followed by a semi-dry transfer (Panther Semi-Dry Electroblotter,

Thermo Scientific Owl, MA) onto nitrocellulose at constant current (100 mA per membrane) for

160 min. Incubation in primary antibody (anti-MIR1-CP or anti-RIP2) was used in a ratio of

1:30,000 for root extracts. The secondary anti-rabbit HRP conjugated antibody (Thermo Fisher

Scientific, Rockford, IL) was used in a ratio of 1:30,000. Immunodetection was done using chemiluminescence (West Femto Maximum Sensitivity Substrate, Thermo Scientific, MA). To corroborate even protein loading, separate gels were run and stained with SimplyBlueTM

(Thermo Scientific, MA).

Peroxidase and peroxidase assays

110

Leaf and root samples collected at 0, 2, 6, 12, and 24 h were ground as described above. For peroxidase activity assays, each sample was mixed with 1.25 ml of 0.1M potassium phosphate buffer (pH 7.0) with 5% cross-linked polyvinylpolyrrolidone (Afla Aesar, Haverhill, MA) and incubated on ice for 5 min, then centrifuged at 11,000xg for 10 min. The supernatants (5 µl) were assayed for peroxidase activity (change in absorbance at λ=450) using 3 mM guaiacol (190

µl) and 3% hydrogen peroxide as substrates. Peroxidase enzyme activity was determined as ∆

Absorbance m-1 µg-1 of protein. Hydrogen peroxide concentrations were measured at 0, 2, 6 and

24 h, 500 µl of phosphate buffer (pH 6.0) was added to each sample and centrifuged for 30 min.

Supernatants from each sample (50 µl) were assayed using a hydrogen peroxide colorimetric kit following manufacturer protocol (Enzo Life Siences, NY). The hydrogen peroxide concentration

-1 was determined as µM mg of fresh weight. The data was analyzed by first doing Log10, square root, inverse power or box-cox data transformation until the Shapiro-Wilk test (Shapiro and

Wilk, 1965) confirmed normal distribution, then a multiple-factor analysis of variances

(ANOVA) was done, followed by a significant difference (HSD) Tukey pairwise comparison test by using R.

Results and discussion

A total of 4675 proteins were identified in both leaves and roots of control and FAW-infested plants with a 95% FDR. ANOVA showed 794 proteins with a p≤0.05 when comparing Log10 ratios of infested and control leaves and roots (Table S4.1). To analyze the protein data collected with the TMT, we performed hierarchical clustering with metabolic annotations, and network biology analysis of a protein abundance correlation network and a protein-protein interactions

111

(PPI) network. These analyses revealed many proteins and pathways of interest that were subsequently validated

Hierarchical clustering

Hierarchical clustering was performed on the 794 proteins with significant differences in protein abundance ratios. Proteins were clustered in groups based on their patterns of abundance ratios using the Euclidian similarity distance method that divided the proteins in 10 families (Fig. 1;

Table S4.2). Parametric analysis of gene set enrichment (PAGE) performed on all 794 proteins showed that there were 35 significantly enriched gene ontology terms for the Lin:Lc abundance ratios and none for the Rin:Rc ratios. PAGE took into account protein abundance levels when determining GO enrichment (Kim & Volsky, 2005). Of the 35 enriched GO terms, only enzyme regulator activity and response to stimuli showed up regulation in Lin:Lc ratios (Fig. 4.2). Eleven proteins were part of the “enzyme regulator activity” GO term and all of them showed Lin:Lc protein ratios higher than 1 indicating that these proteins increased in abundance in response to

FAW (Fig. 4.2, Table 4.1). Several of them were serine and cysteine protease inhibitors or proteins with adenyl-nucleotide exchange factor activity (Fig. 4.2, Table 4.1). Protease inhibitors have long been linked to plant defense strategies against herbivores (Green & Ryan, 1972). For example, a maize proteinase inhibitor has been shown to decrease digestive protease activity in

Spodoptera littoralis larvae (Tamayo et al., 2000). Consequently, protease inhibitors increasing in abundance in response to FAW could be acting as direct defenses against FAW attack.

Fifty six proteins were part of the “response to stimuli” GO term, eleven were peroxidases that had Lin:Lc ratios between 0.6 and 2 and Rin:Rc ratios between 0.8 and 1.4

(Fig. 4.2, Table 4.2). Peroxidases participate in a wide variety of biological roles such as such as

112 lignin and suberin formation, cross-linking of cell wall components, and synthesis of phytoalexins (Almagro et al., 2009). Also they are key elements in the ROS metabolism and programmed host cell death associated with pathogens (Almagro et al., 2009) and defense response signaling (Hu et al., 2006; War et al., 2012). Plants oxidize phenolic compounds into quinones as direct defense metabolites against insect. These compounds are produced and oxidized by peroxidases and/or polyphenol oxidases (PPO) and damage the insect by inhibiting protein digestion (War et al., 2012).

Other three proteins part of the response to stimuli GO term were cysteine peptidases

(Fig. 4.2, Table 4.2), a protease (Mir1-CP) has been described as part of the defense response mechanism that Mp708 deploys against FAW (Jiang et al., 1995; Pechan et al., 2000) by damaging the peritropic matrix of the caterpillar gut, leading to lower acquisition of nutrients and possible death (Mohan et al., 2006). Although none of these proteases were identified as Mir1-

CP, this protein was in our TMT data set (GRMZM2G150276_P01, Table S4.1) with a Lin:Lc ratio of 1.3, a Rin:Rc ratio of 0.95 and a p=0.4. It is possible that the other proteases with p≤0.05 also could be attacking the FAW gut.

The hierarchical clustering analysis generated 10 protein families. The proteins in each family potentially relevant to caterpillar defenses are described below.

Group 1

In this group there were 332 proteins that have Lin:Lc ratios between 1.7 and 0.9 while the

Rin:Rc ratios were between 0.5 and 1.1, indicating that these proteins increased in abundance in leaves infested with FAW and decreased or remained the same in roots (Fig. 4.1, Table S4.2).

Single gene ontology enrichment analysis of the molecular function showed that 26 of the

113 proteins in this group are involved in oxido-reductase activities (Fig. 4.1, 4.3a, Table S4.2).

Changes in REDOX state play a crucial role in JA signaling pathway in plants (Walter et al.,

2007; War et al., 2012) and caterpillar induced responses (Paudel et al., 2013). In tobacco,

REDOX changes are linked to responses to pathogens, and phytohormone levels that control mitogen activated protein kinases (MAPKs) and SA-induced protein kinases (SIPK) involved in control of downstream pathogen-induced gene expression (Matern et al., 2015). It is possible that

REDOX changes in leaves can lead to activation of plant defense responses against insects that differentially affect downstream MAPKs signaling pathway in leaves and roots. Also, the

“electron carrier activity” and “transporter activity” GO terms are significantly enriched in this group (Fig. 4.3a, Tale S4.2), which supports the idea that REDOX changes caused by changes in ion concentrations (Kant et al., 2015) could be involved in leaf defense responses against caterpillars. These REDOX related proteins were different and had high abundance in leafs in response to herbivory compared to the ones that showed down regulation in the parametric analysis of gene set enrichment (PAGE) (Fig. 4.2a). Perhaps multiple REDOX-related proteins have different functions in the plant defense response causing their abundance profiles to be different, further molecular studies could help understand their roles in plant defenses.

Group 2

In this group there were 249 proteins that had Lin:Lc ratios between 0.9 and 0.5 while the Rin:Rc ratios between 1.8 and 0.7, meaning that these proteins generally increased in abundance in roots of plants infested with FAW (Fig. 4.1, Table S4.2). Single gene ontology enrichment analysis of the molecular function showed that 132 proteins in this group were involved in binding activities

(Fig. 4.1, 4.3b, Table S4.2). Some of the possible binding activities were cofactor, nucleotide,

114

GTP and RNA binding. Considering that these proteins were activated in systemic tissues, some of the biding activities could be involved in receptor complexes crucial for perceiving insect infestation. Also, nucleotide binding could reflect control of gene expression in systemic tissues.

Furthermore, there were 34 proteins involved in oxido-reductase related activities, as mentioned for Group 1. This suggests that REDOX changes could be linked to the control of phytohormone abundance and gene expression in systemic defense responses. Because these REDOX changes occurred in systemic tissues, it is possible that there are difference in downstream responses such as MAPK activation that are different from REDOX-induced responses in leaves.

Group 3

In this group there were 119 proteins that had Lin:Lc ratios between 2.1 and 1.4 while the Rin:Rc ratios between 1.4 and 0.8. These proteins accumulated to a greater extent in leaves than roots of

FAW- infested (Fig. 4.1, Table S4.2). Single gene ontology enrichment analysis of the molecular function showed 17 proteins were involved in oxido-reductase activities and nine in electron carrier activity (Fig. 4.1, 4.3b, Table S4.2). These proteins appeared to accumulate in greater abundance than those in Group 1 and could be involved in modulating the leaf REDOX status during FAW infestation.

Group 4

In this group there were 52 proteins that had Lin:Lc ratios between 1.5 and 0.7 while the Rin:Rc ratios between 2.6 and 1.0, meaning that these proteins increased in abundance only in roots of

FAW-infested plants (Fig. 4.1 , Table S4.2). Single gene ontology enrichment analysis of the molecular function showed 22 proteins involved in diverse catalytic activities and 11 were

115 involved in oxido-reductase activities (Fig. 4.1, 4.3b, Table S4.2). These proteins with GO annotations involved in REDOX changes are different from the ones in the other groups. Taken together the data from Groups 1 to 3, it seems like different sets of enzyme involved in REDOX conditions increase in leaves and roots during FAW infestation, implying that there is a level of tissue-specificity over REDOX changes.

Group 5

In this group there were 16 proteins that had Lin:Lc ratios between 1.2 and 0.7 while the Rin:Rc ratios between 0.7 and 0.3, meaning that these proteins increased slightly in abundance in leaves and were highly repressed in roots of FAW-infested plants (Fig. 4.1 , Table S4.2). One of these proteins is ubiquitin (GRMZM2G127101_P01) and its high abundance in leaves was expected considering that JA regulation of gene expression in response to herbivory uses the SCFCOI1

(SKIP, CULLIN, F-BOX complex and CORONATINE INSESITIVE 1) ubiquitin ligase complex (Thines et al., 2007). Other protein in this group was isopentenyl pyrophosphate isomerase (IPPS; GRMZM2G108285_P01) that is involved in the mevalonate pathway that lead to production of isoprenoids and volatiles such as caryophyllene (Pare & Tumlinson, 1999) that herbivore defense responses.

Group 6

In this group there were nine proteins that had Lin:Lc ratios between 4.2 and 2.4 while the

Rin:Rc ratios between 1.9 and 0.9, indicating that the abundance of these proteins increased highly in leaves of FAW-infested plants (Fig. 4.1 , Table S4.2). One of these highly abundant

116 proteins was a proteinase inhibitor (GRMZM2G028393_P01) that increases in response to wounding in potato (Lee et al., 1986). Other protein belongs to the calmodulin-binding motif (IQ motif) family protein (GRMZM2G041765_P01). These proteins tend to be part of signaling pathways responding to abiotic or biotic stimuli (War et al., 2012; Cheval et al., 2013) such as insect herbivory (Scholz et al., 2016). Perhaps this protein is involved in the signaling responses against FAW in Mp708. Also, a peroxidase (GRMZM2G116902_P01) was part of this group that could, play a similar role defensive as the peroxidases described in the GO PAGE analysis.

Group 7

In this group there were eight proteins that had Lin:Lc ratios between 0.5 and 0.3 while the

Rin:Rc ratios between 1.3 and 0.5, meaning that they were repressed in leaves of plants infested with FAW (Fig.4.1 , Table S4.2). In this group there was a protein involved in flavonoid biosynthesis (GRMZM2G422750_P01). Flavonoids generally have strong antioxidative properties that could help roots cope with oxidative stress (Mierziak et al., 2014).

Group 8

In this group there was one protein with a Lin:Lc ratio of 1.3 and Rin:Rc ratio of 0.2, meaning that this protein slightly increase in abundance in leaves, but was highly repressed in roots of plants infested with FAW. This protein (GRMZM2G063729) is involved in RNA processing

(Fig. 4.1, Table S4.2).

Group 9

117

In this group there were four proteins that had Lin:Lc ratios between 3.0 and 2.0 while the

Rin:Rc ratios between 0.7 and 0.4, meaning that these proteins increased in abundance only in leaves and were repressed in roots of FAW-infested plants (Fig. 4.1 , Table S4.2). Some of these proteins were, a cysteine peptidase named vignain (GRMZM2G070011_P01), protease could be performing an insecticidal role similar to that of Mir1-CP in Mp708. Also, an uncharacterized protein (GRMZM2G114680_P03) that is similar to Scarecrow-like protein 18 (SCL18) from

Arabidopsis, and could be acting as a transcription factor whose response is highly correlated with insect defense in maize.

Group 10

In this group there were four proteins that had Lin:Lc ratios between 1.9 and 1.3 while the

Rin:Rc ratios between 2.9 and 1.9, meaning that these proteins increased in abundance in both leaves and roots of plants infested with FAW (Fig. 4.1 , Table S4.2). Some of these proteins were pathogenesis-related protein 10 (PR10, GRMZM2G112488_P01), this protein could be induced by microorganisms in the wound site or in the soil that are in contact with the damaged leaves or roots. A Bowman-Birk type proteinase inhibitor WIP1 (GRMZM2G156632_P01) was present in this group. The transcripts of this protein have been shown to be induced locally and systemically in response to wounding in maize (Rohrmeier & Lehle, 1993). This also appears to accumulate in Mp708 leaves and roots in response to FAW feeding.

Protein correlation networks

In addition to the heatmap analysis, a protein correlation network was developed. In this network, each node represented a protein and each edge was a protein abundance correlation

118 with a score of 0.99 or higher, meaning that two proteins were linked if their patterns of abundance were sufficiently similar to pass the stringent Pearson correlation parameters (Note

S4.2). The idea behind this network is that similar protein abundance patterns under specific stressors could lead to proteins that have a similar functions during herbivore defense responses

(Provart, 2012). In this type of network analysis proteins with similar abundance tend to be highly correlated under specific biotic stressors like herbivory. These proteins tend to cluster together in the network and could have a similar functions during plant herbivore defense responses (Provart, 2012). The network properties were evaluated by using the NetworkAnalyzer

(Doncheva et al., 2012) tool from Cytoscape, which showed that the network had 813 nodes and

8598 non-directed/multi-edges node pairs (Note S4.1, Table S4.3). Gene ontology enrichment performed on the whole protein correlation network showed a wide variety of enriched GO terms. Enriched biological process GO terms were “reactive oxygen species”, metabolic process such as “glycosyl compound metabolic process” (Fig. 4.4a), which are both processes that have been linked to plant defenses responses to caterpillars (War et al., 2012). Enriched molecular function GO terms were oxidoreductase activity, translation elongation and adenyl-nucleotide exchange factor (Fig. 4.4b). As seen with the hierarchical cluster analysis, REDOX changes in the plant seem to play an important role in the response to herbivores.

Because the QTLs for FAW resistance in Mp708 have been characterized (Williams et al., 1990; Brooks et al., 2005), we created a subnetwork (62 proteins, Table S4.3, Note S4.1) of the proteins present in the correlation network that fall within the QTL bin locations (Fig. 4.5a).

Within this subnetwork 24 proteins had significant protein abundance changes (Table 4.3). The protein with the highest abundance in the leaves was a senescence protein

(GRMZM2G046186_P01; Table 4.3). Developmental changes could be due to plant trade-offs

119 between defense and growth (War et al., 2012).. Other protein with a Lin:Lc ratio of 1.6 and a

Rin: Rc ratio of 1.0 had an ankyrin repeat-containing domain (GRMZM2G150791_P01; Table

4.3). In rice, overexpression of a protein with an ankyrin domain showed enhanced resistance to

Xanthomonas oryzae pv oryzae (Xoo) and Magnaporthe oryzae, possibly because this protein acts as a positive regulator in rice basal defense mediated by SA- and JA-signaling pathways

(Mou et al., 2013). In maize this protein with an ankyrin domain could be part of JA defense responses against FAW. Another protein in this network had a FAS1 (fasciclin-like) domain

(GRMZM2G174799_P01; Table 4.3) that has been linked to proteins that are anchored in the membrane, in Arabidopsis. These proteins function as cell adhesion molecules and are linked to shoot development (Johnson et al., 2003). Other protein in this network was a Pi starvation- induced protein (GRMZM2G118037_P01; Table 4.3). This family of proteins perform proteolysis and are involved in hormone synthesis, viral protein processing and receptor maturation (Bergeron et al., 2000). Therefore, this protein might be involved phytohormone synthesis or signaling in response to FAW.

We determined the activated Hubs in the network by scoring the p-values of protein

Hubs, where a high score indicated more activation in terms of p-values (Ideker et al., 2002).

Five hubs were found, Hub 1 had 44 proteins and a score of 7.2, Hub 2 had 31 proteins and a score of 7.1, Hub 3 had 37 proteins and a score of 7.0, Hub 4 had 42 proteins and a score of 5.3,

Hub 5 had 9 proteins and a score of 5.2 (Note S4.1). Proteins in Hub 1 showed higher abundances in infested leaves compared to roots, thus these proteins were highly abundant in the attacked organ during FAW infestation (Fig. 4.5b, Table 4.4, Note S4.1). In this Hub the Lin:Lc protein ratios varied from 1.72 (GRMZM2G059556_P01, NADH ubiquinone oxidoreductase) to

1.17 (GRMZM2G157332_P07, uncharacterized protein) and the Rin:Rc ratios varied from 0.97

120

(GRMZM2G320307_P02, glucosinolate biosynthetic process) to 0.64 (GRMZM2G116689_P01, uncharacterized protein) (Table 4; Note S1). Among the proteins in Hub 1 were proteins fell within a QTL in the correlation network (Fig. 4.5b, Table 4.4). They were Pi starvation-induced protein (GRMZM2G118037_P01), FAS1 (fasciclin-like) domain (GRMZM2G174799_P01), nascent polypeptide-associated complex (GRMZM2G110116_P01), and acyl carrier protein/fatty acid biosynthetic process (GRMZM2G181542_P01, GRMZM2G062481_P01; Table 4.4).

Protein-protein interaction (PPI) networks

In addition to the a protein correlation network, we extracted a subnetwork by only keeping proteins and their interaction partners if they were detected in the TMT dataset and were present in a layout protein-protein interaction network created by Musungu et al. (2015) based on interacting maize orthologs (interologs). The interologs are predicted in silico to occur in maize and are based on experimentally confirmed interactions in 13 species available on four interactome on-line databases (Musungu et al., 2015). The basis of this network was that evolutionary conserved proteins keep their conserved interacting partners if the protein function is the same (Musungu et al., 2015). The network properties were evaluated using the

NetworkAnalyzer (Doncheva et al., 2012) tool from Cytoscape, which showed that the network had 1181 nodes with 316 isolated nodes, 339 self-loops and 4142 edges (Note S4.2, Table S4.4).

The protein with the highest Lin:Lc ratio (8.5) was auxin response factor 27

(GRMZM2G160005_P01) with a Rin:Rc ratio of 1.1 and p=0.8. The protein with the highest

Rin:Rc ratio (2.8) was actin-depolymerizing factor (GRMZM2G463471_P01), the Lin:Lc ratio was 1.1 and p=0.9. GO analysis showed biological process enrichment of proteins involved in oxidoreductase activities (Fig. 4.6).

121

From this PPI network, a subnetwork was created with the proteins that showed significant protein abundance changes, it had 228 nodes and 181 edges. The protein with the highest Lin:Lc ratio (2.0) was thioredoxin, and its root ratio was 1.0; the protein with the highest

Rin:Rc ratio (1.7) was a UDP-glucuronosyl transferase with a Lin:Lc ratio of 0.9. The highest connected group of nodes from this network had 74 proteins. The protein with the highest Lin:Lc ratio (1.7) and Rin:Rc ratio of 1.1 was a ribosomal protein S19 and the one with the highest

Rin:Rc ratio (1.5) with a Lin:Lc ratio of 0.9 was a polyubiquitin-like protein isoform 1 (Fig.

4.7a).

We determined the activated Hubs in the network by scoring the p-values of protein

Hubs, a high score indicated more activation in terms of p-values (Ideker et al., 2002). Five hubs were found, Hub 1 had 125 proteins and a score of 5.10, Hub 2 had 131 proteins and a score of

4.48, Hub 3 had 138 proteins and a score of 4.46, Hub 4 had 152 proteins and a score of 4.42,

Hub 5 has 128 proteins and a score of 4.39 (Fig. 4.7b, Note S4.2). In Hub 1 there were proteins assigned to well known pathways such as response to other organisms that have oxidoreductase activity (GRMZM2G036921_P01, GRMZM2G118610_P01, GRMZM2G139680_P01) or are involved in a defense response signaling pathway (AC203535.4_FGP001; mRNA splicing protein CDC5 (Myb superfamily). Also proteins in ROS responses or metabolism such as superoxide dismutase (GRMZM2G106928_P01) and catalase isozyme 1

(GRMZM2G088212_P01) and involved in ABA signaling (GRMZM2G152757_P01) were found in this Hub. The proteins in Hubs 2 and 3 were very similar to those in Hub 1, but Hub 2 contained a protein involved in JA biosynthesis (GRMZM2G000236_P01; 12-oxophytodienoate reductase 2). In Hubs 4 and 5 there were proteins that belonged to similar pathways as the other hubs, however Hub 4 contained proteins involved in ethylene biosynthesis (GRMZM2G061135,

122 methionine adenosyltransferase and GRMZM2G067265, methionine gamma-lyase). Hub 5 contained a protein involved in JA signaling (GRMZM2G033219_P01, 6,7-dimethyl-8- ribityllumazine synthase).

Validations of pathways and proteins of interest

Phenomic changes in two maize genotypes after 24h of FAW infestation

TMT data showed that there were proteins with significant changes that are involved in plant growth and development (Table S4.2). We looked at phenomic changes between insect resistant

(Mp708) and insect susceptible (Tx601) genotypes to determine if the TMT results could be revealing differences in growth and defenses trade-offs. Mp708 plants tended to be taller and have longer thin and seminal-like roots than Tx601 when they were infested with FAW (Figure

4.8). Mp708 showed positive height change from day 6 to 9 in controls and FAW-infested plants, while Tx601 showed no changes in height when infested with FAW (p≤0.05, Fig. 4.8a) and at 9 days FAW-infested Mp708 plants were significantly taller than infested Tx601(p≤0.05, Fig.

4.8a). At day six, Tx601 roots were significantly shorter than control and FAW-Mp708 (p≤0.05,

Fig. 4.8b) and by day nine Tx601 roots were only shorter than FAW-infested Mp708 (p≤0.05,

Fig. 4.8b). We evaluated length of different root types and found that Mp708 had longer thin and seminal-like roots on days 6 and 9 during FAW infestation than Tx601 (p≤0.05, Fig. 4.8c). These results imply that Mp708 produces the same amounts of thin and seminal-like roots in control and infested conditions while Tx601 has less of these roots under infestations. Thin roots are important for water acquisition in young roots, and root penetration into the soil (Paez-Garcia et al., 2015), also, root growth and maintenance under stress is metabolically very costly for the plant (over 50% of daily photosynthate production) (Lambers et al., 1996), thus, maintaining

123 root growth under FAW infestation allows Mp708 plants to thrive better than susceptible genotypes. This result points at a phenotypic difference in plant growth responses from insect resistant and susceptible genotypes infested with FAW.

Because there were differences in plant height and root length between Mp708 and

Tx601 during FAW infestation we looked at changes in percent carbon in leaves and roots. The percent carbon in leaves was ~2% higher in Mp708 than in Tx601 at days 6 and 9 when plants were infested with FAW (p≤0.05, Fig. 4.9a). Also the percent carbon in roots was ~2% higher in

Mp708 than in Tx601 only at day nine when plants were infested with FAW (p≤0.05, Fig. 4.9b).

The difference in carbon content could be associated with the longer length of thin and seminal like roots in Mp708 compared to Tx601 during FAW infestation. Difference in carbon content allocation could explain maintenance of root growth in insect resistant genotype. These results suggest that growth and defense trade-offs in Mp708 are lower than in Tx601 under caterpillar infestation, and it is possible that this carbon/cost differences could affect other biological process in Mp708.

Expression of a JA biosynthetic pathway gene and ET accumulation in leaves and roots during

FAW infestation

One of the key genes in the JA biosynthesis in maize is lox3, because it encodes an enzyme that catalyzes an early reaction in the pathway (McConn et al., 1997, Koo and Howe, 2009, Yan et al., 2013). We measured expression of lox3 to determine how FAW feeding might alter gene expression and possibly affect JA biosynthesis in Mp708 leaves and roots. In leaves of FAW- infested plants lox3 transcript levels significantly increased compared to control, while transcript levels in roots were not different than those in non-infested plants (Fig. 4.10a). These results

124 indicated that lox3 expression significantly increases in the organ attacked by the herbivore, and remains unaltered in other parts of the plant.

Due to the activation of ET pathway seen in the network biology analysis we determined

ET levels in leaves of plants infested with FAW. ET production in leaves did not change during the 24 h infestation and there were no significant differences between infested and control plants

(Fig. 4.10b). ET levels in roots of plants infested aboveground with FAW increased after 24 h compared to levels at 2 and 6 h and controls (Fig. 4.10c). FAW infestation altered ET levels relative to the controls in tissues distal to the feeding sites. This result may be interpreted as long-distance communication between infested and non-infested tissues that involves ET.

Effects of blocking JA and ET pathways on MIR1-CP and RIP2

TMT analysis showed that proteases, JA and ET pathways are involved in the plant defense responses to FAW. To understand some downstream molecular plant defense changes, first we examined accumulation of two proteins that have insecticidal properties, MIR1-CP and RIP2

(Pechan et al., 2000; Bass et al., 2004; Brooks et al., 2005; Brooks et al., 2007; Chuang et al.,

2014) in leaves and roots during FAW infestation. We found MIR1-CP in our TMT data set but its p-value was 0.4 and RIP2 was not in our database, nevertheless due to the amount of evidence of MIR1-CP and RIP2 importance as insecticidal proteins present in Mp708 we decided to use them as protein markers. There was no increase in mir1 transcripts in leaves or roots and rip2 transcripts increased ~60% only in leaves of plants attacked by FAW but was not a significant change (Fig. 4.11a, b). These results are consistent with previous studies that show relatively high constitutive mir1 expression in Mp708 at the V7 stage with no changes in transcript abundance at 24 hours after herbivory (Zhu, 2010).

125

To link the JA and ET pathways to our two defense proteins markers we evaluated downstream abundance changes of MIR1 and RIP2 while impairing JA synthesis and ET perception in leaves and roots. Previous studies have indicated that JA and ET enhanced foliar accumulation mir1 transcripts and MIR1-CP protein in plants infested with FAW (Ankala et al.,

2013), but only ET was involved during corn leaf aphid (CLA, Rhopalosiphum maidis) infestation (Louis et al., 2015). Inhibition of JA accumulation by IBU treatment decreased

MIR1-CP protein abundance in tissues locally attacked by FAW, but had no effect in root tissue

(Fig. 4.11c). To determine if the IBU concentration was insufficient, a 3-fold higher concentration was tested, but it led to necrotic lesions in the leaves of control and infested plants.

Since the JA pathway controls many plant defenses such as production of reactive oxygen species, polyphenols, tannins, proteinase inhibitors (War et al., 2012), repression of other plant defenses besides MIR1-CP in the roots could contribute to plant defenses.

When plants were treated with ETB and infested with FAW, MIR1-CP abundance decreased slightly in leaves and dramatically in roots (Fig. 4.11c). These results showed that during FAW infestation JA could alter the MIR1-CP-mediated leaf defenses, while ET might be crucial for these defenses in leaves and roots. We conclude that ET perception is a positive regulator of leaf and root defenses involving MIR1-CP during FAW infestation. ET might act distally from the site of infestation towards the systemic organs (Harfouche et al., 2006; Zhu et al., 2011) due to its volatility. A recent review of ET and hormonal crosstalk during plant defenses indicates that the ET signaling network can directly affect insect attackers or influence plant defenses in distal tissues (Broekgaarden et al., 2015).

Treating FAW infested plants with IBU or ETB decreased the accumulation of RIP2 in the leaves and had no effects in the roots, where the protein was not detected (Fig. 4.11d). JA and

126

ET seem to control RIP2 abundance only in tissues immediately attacked by FAW. These results indicate RIP2 accumulation appears to be restricted to the tissue attacked by the herbivore and it is not systemically expressed (Chuang et al., 2014). These results differ from previously published work that showed a significant increase in rip2 transcripts in the leaves in response to

FAW feeding (Chuang et al., 2014), but those plants were older (V8 stage) than the seedlings

(V3 stage) used in this study. It appears that during FAW infestation JA can regulate MIR1-CP and RIP2 accumulation in leaves. These results show different hormonal regulation of MIR1 and

RIP2 in response to leaf or root-chewing insects. We propose a defense model that links JA and

ET with defense responses in leaves and roots infested by a leaf-chewing insects. In the leaves,

FAW feeding increases JA and does not affect ET accumulation, but it induces both JA and ET accumulation in roots. In leaves of plants infested with FAW, JA biosynthesis and ET perception control MIR1 and RIP2 abundance, while in roots only ET perception controls MIR1-CP accumulation.

Peroxidase and peroxide changes in leaves and roots under FAW infestation

The hierarchical clustering and the two different network analyses showed that peroxidases are key defense players in both leaves and roots of FAW-infested plants. Total peroxidase activity was determined in leaves and roots of control and infested Mp708 and Tx601 plants because there were several peroxidases with variable protein abundances in Mp708 leaf and root tissues.

In Mp708 leaf peroxidase activity significantly increased at 2h and by 24h it decreased to constitutive 0h levels. In Tx601 the enzyme activity was at constitutive levels from 0-6h then significantly decreased (Fig. 4.12a). Peroxidases catalyze the biochemical reaction that consumes peroxide and transforms a reduced electron donor to an oxidized compound. We measured

127 peroxide content in leaves and found that this compound significantly increased only in Tx601 after 24h of FAW infestation (Fig. 4.12b).

Increase in peroxide content correlates with a decrease of peroxidase activity because peroxide is the substrate used by this enzyme. In roots, the peroxidase activity decreases in

Tx601 after 24h, with no change in Mp708 or in peroxide content (Fig. 4.12c, d). Activity in the leaves goes down significantly between 6h and 12 h of infestation. In general, leaves have higher peroxide levels than roots of both maize genotypes (Figure 4.12b, c, d) while the peroxidase levels are higher in roots than in leaves (Fig. 4.12a, c). In Mp708, earlier changes of peroxidase activity in leaves could indicate that this is an early defense response against FAW. Further enzyme activity analysis focused on polyphenol oxidase (PPO) showed that basal levels in leaves are higher in Tx601 compared to Mp708, but only Mp708 genotype decreased the enzyme activity after 6h of infestation then went back to basal levels (Fig. 4.12e). PPO activity was below measureable range in the roots of both genotypes.

The profiles of peroxidase activity in leaves of both genotypes showed an early increase followed by a decrease that is only significantly lower at 24h compared to 0h in Tx601.

Peroxidases seem to be involved in lignification (Saathoff et al., 2013), programmed cell death, membrane protection against oxidative stress through ROS removal and production (Saathoff et al., 2013; Zhou et al., 2015), and production of quinones (Kant et al., 2015; Zhou et al., 2015).

All of these roles have been linked to plant defense responses against insects (Kant et al., 2015;

Zhou et al., 2015). Studies performed in buffalograss (Nuttall, Buchloë dactyloides) using insect susceptible (378) and resistant (Prestige) genotypes against western chinch bug Blissus occiduus showed an early increase of peroxidase activity in the resistant genotype (Hoang, 2010). In wild- type tomato (Solanum lycopersicum) peroxidase gene expression increases during infestation by

128

Helicoverpa zea (corn earworm), furthermore, peroxidase overexpression led to increase of genes encoding protease inhibitors and pathogenesis-related (PR) proteins (Suzuki et al., 2012).

Our data suggests that Mp708 increases activity of peroxidases in leaves without affecting the levels of peroxide, thus the enzyme could be using other substrate such as lipid peroxide or could be affecting other reactions that do not involve hydrogen peroxide.

Conclusions

1. Parametric analysis of gene set enrichment (PAGE) showed several proteinase inhibitors and cysteine proteases were highly enriched in Lin:Lc ratios and it is likely that they contribute to

FAW defense in Mp708. Also, this analysis revealed that peroxidases might be playing a role in the defense strategies against this pest.

2. Hierarchical clustering and network analyses of proteomics data showed promising results in unraveling potential metabolic pathways, signaling events and proteins with insecticidal properties. All the analyses showed an increase in the JA biosynthetic pathway, changes in oxido-reductase status and peroxidase abundance. Hierarchical clustering and protein-correlation networks indicated REDOX changes in the glutathione and ascorbate pathways, while the PPI network revealed the involvement of ABA and ET biosynthesis and signaling.

3. Using the protein-correlation network with the known QTL regions in Mp708 that explain the resistant against FAW, we found some proteins that might be part of the defense mechanisms of this genotype. These proteins are, Pi starvation-induced protein (GRMZM2G118037_P01),

FAS1 (fasciclin-like) domain (GRMZM2G174799_P01), nascent polypeptide-associated

129 complex (GRMZM2G110116_P01), and acyl carrier protein/fatty acid biosynthetic process

(GRMZM2G181542_P01, GRMZM2G062481_P01; Table 4).

4. The insect susceptible maize genotype, Tx601, grew less in height and total root length during

FAW infestation. Thin and seminal like roots significantly decreased during FAW infestation in

Tx601. Also, the carbon content in leaves and roots was lower in Tx601 compared to Mp708 only in FAW-infested plants after 9 days of infestation. This could be linked to the decrease in plant height and root length in Tx601 compared to Mp708. Mp708 defense mechanisms could involve lower trade-off between plant growth and defense responses.

5. ET accumulation vary among leaves and roots, and these hormones elicit differential control over local and systemic accumulation of MIR1-CP. During FAW infestation, JA and ET appears to control local MIR1-CP and RIP2 accumulation, ET could be a key regulator of MIR1-CP systemic accumulation. These findings open up new routes in the hormonal control of local and systemic defense signaling that could be similar for other defense proteins and secondary compounds. Furthermore, the question prevails as to why there is a different phytohormone control over defenses in an organ-specific matter.

6. Peroxidase activity increases early in leaves of Mp708 while it does not change in the roots during FAW infestation.

7. We have established an analysis pipeline for proteomics data that includes network biology approaches that can be used with different types of “omics” data from a wide variety of

130 organisms to detect tissue-specific defense responses to be able to understand the organism strategies to survive or diminish the external stress.

131

Figure 4.1. Hierarchical clustering of protein abundance ratios that showed statistically significant differences in abundance (P≤ 0.05) in Mp708 infested with FAW for 24 h. The

132 heatmap was built using the proteins ratios from Lin:Lc and Rin:Rc from proteins that showed statistically significant differences in abundance (P ≤0.05). The dendogram in the heatmap was generated by applying Euclidian similarity distance method to determine abundance patterns using a maximum of 10 k-means cut off. The 10 k-means are labeled as protein families, they are color coded in the bar next to the dendogram and each family was enclosed by a red box, 1 (□), 2

(□), 3 (□), 4 (□), 5 (□), 6 (□), 7 (□), 8 (□), 9 (□), and 10 (□).

133

Figure 4.2. Gene ontology (GO) enrichment of the proteins present in the hierarchical map. (a)

GO enrichment of the molecular function. (b) GO enrichment of the biological process. GO

134 annotations were performed using AgriGO on line tools, in boxes: color represent statistical significant terms, the blue color represents significant downregulation in Lin:Lc ratios. Red color represents significant enrichment and upregulation Lin:Lc ratios. There were no Rin:Rc GO enriched terms. The C1 number is the enrichment multiple-test adjusted P-value on the Lin:Lc ratios.

135

Table 4.1. Proteins that are part of the gene ontology (GO) term: “enzyme activity” that was significantly enriched and up regulated in the Lin/Lc ratio. Columns represent protein id, protein ratios, p-value, groups are the 10 k-mean families clusters, chromosome location of the proteins and a description of the protein domains and GO annotation.

Protein id Lin:Lc Rin:Rc PVAL groups Chr Description GRMZM2G028393 3.4 2.0 0.008251 6 8 serine-type endopeptidase inhibitor activity GRMZM2G007928 1.3 0.9 0.014686 1 3 serine-type endopeptidase inhibitor activity GRMZM2G156632 2.0 2.0 0.007129 10 8 serine-type endopeptidase inhibitor activity GRMZM2G044684 1.5 1.0 0.04468 3 4 adenyl-nucleotide exchange factor activity GRMZM2G013461 1.2 0.9 0.024556 1 8 cysteine-type endopeptidase inhibitor activity GRMZM2G440968 1.0 0.6 0.048186 5 6 cysteine-type endopeptidase inhibitor activity GRMZM2G102000 1.4 1.0 0.031196 3 2 enzyme regulator activity GRMZM2G002440 1.4 0.9 0.00403 3 4 adenyl-nucleotide exchange factor activity GRMZM2G042789 2.9 1.3 0.018048 6 8 serine-type endopeptidase inhibitor activity GRMZM2G035948 1.4 0.9 0.006085 1 2 adenyl-nucleotide exchange factor activity GRMZM2G447785 1.5 1.0 0.000452 3 2 serine-type endopeptidase inhibitor activity

136

Table 4.2. Proteins that are part of the gene ontology (GO) term: “response to stimuli” that was significant enriched and up regulated in the Lin/Lc ratio. Columns represent protein id, protein ratios, p-value, groups are the 10 k-mean families clusters, chromosome location of the proteins and a description of the protein domains and GO annotation.

Protein id Lin:Lc Rin:Rc PVAL groups Chr Description GRMZM2G075828 1.1 0.7 0.044155 1 7 antiporter activity/drug transmembrane transporter activity GRMZM2G120563 0.7 1.2 0.041159 2 7 antiporter activity/drug transmembrane transporter activity GRMZM2G123922 0.7 1 0.002035 2 10 ATP binding/ nucleotide binding Bifunctional inhibitor/plant lipid transfer protein/seed storage GRMZM2G078876 1.8 1 0.030671 3 1 helical domain GRMZM2G032280 0.9 1.3 0.007835 4 5 biotin-[acetyl-CoA-carboxylase] ligase activity Bundle sheath cell specific protein 1/ response to ABA, water GRMZM2G168552 1.4 0.7 0.037615 1 8 deficit and ripening GRMZM2G079348 0.8 1.2 0.013474 2 4 catalase activity/ heme binding GRMZM2G025078 1.3 0.8 0.01928 1 1 copper ion binding GRMZM2G107562 1.5 0.8 0.000891 1 7 Copper ion binding protein/ electron carrier activity GRMZM2G134738 1.4 0.9 0.017999 1 8 cytochrome-c oxidase activity/ mitochondrion GRMZM2G112488 1.6 2.1 0.021391 10 1 defense response/ Pathogenesis-related protein 10 GRMZM2G105682 1.2 0.8 0.042294 1 7 electron carrier activity/ Cupredoxins superfamily fatty acid elongation/ acetyl-CoA carboxylase activity/ biotin GRMZM2G377341 3.1 1 0.036529 6 5 carboxylase activity GRMZM2G101287 1.3 0.9 0.000556 1 5 galactolipid biosynthetic process/ Pathogen induced protein 2-4

137

GRMZM2G032110 0.6 1 0.019582 2 4 gene silencing by RNA GRMZM2G009326 0.7 1 0.0068 2 1 GRAM domain GRMZM2G028313 0.7 1 0.033131 2 2 GTPase activity GRMZM2G055276 1.3 0.9 0.00041 1 3 HABP4 protein domain/ PAI-1 mRNA-binding domain GRMZM2G170044 0.8 1.1 0.005935 2 7 lipid transport/ Plant lipid transfer protein/Par allergen GRMZM2G161673 0.6 1.5 0.029636 2 6 magnesium protoporphyrin IX methyltransferase activity GRMZM2G102786 1.6 0.6 0.029843 1 7 metal ion binding GRMZM2G173186 0.6 0.7 0.029121 2 2 mismatched DNA binding GRMZM2G064753 1.4 0.9 0.044483 3 3 mitochondrial electron transport, cytochrome c to oxydase GRMZM2G023346 1.6 0.7 0.028248 1 7 NPH3 domain GRMZM2G393671 1.3 0.9 0.02486 1 7 nucleotide binding/ Stress responsive alpha-beta barrel GRMZM2G135893 1.3 0.8 0.001566 1 4 oxidation-reduction process/ Glutathione peroxidase GRMZM2G370815 1.3 0.9 0.023519 1 4 PB1 (Phox and Bem1) domain GRMZM2G133475 1.2 0.8 0.012994 1 2 Peroxidase 66/ metal ion binding GRMZM2G035506 1.3 0.8 0.013411 1 5 peroxidase activity GRMZM2G116846 0.7 0.9 0.00406 2 1 peroxidase activity GRMZM2G126261 1.2 0.9 0.019516 1 7 peroxidase activity GRMZM2G427937 1.3 0.8 0.026661 1 7 peroxidase activity GRMZM2G103342 0.8 1.2 0.010343 2 3 Peroxidase/ heme binding GRMZM2G120517 1.4 0.9 0.003361 3 2 Peroxidase/ heme binding GRMZM2G150893 0.6 1 0.00017 2 5 Peroxidase/ metal ion binding GRMZM2G427815 1.9 0.9 0.045986 3 7 Peroxidase/ response to oxidative stress

138

GRMZM2G061088 1 1.4 0.024315 4 8 Peroxidase/oxidoreductase GRMZM2G087590 1.9 1.1 0.01532 3 3 Photosynthesis and Mog1/PsbP, alpha/beta/alpha sandwich Plasminogen activator inhibitor 1 RNA-binding protein/ GRMZM2G116282 1.2 1 0.017273 1 3 plasma membrane positive regulation of catalytic activity/ Mrp/NBP35 ATP- GRMZM2G129208 0.8 1 0.005983 2 2 binding protein GRMZM2G053939 0.6 1 0.011987 2 7 PREDICTED: alanine aminotransferase 2 GRMZM2G041060 1.4 0.6 0.01148 1 4 protein folding protein transport by the Tat complex/ integral component of GRMZM2G472651 1.6 0.9 0.021584 3 1 membrane GRMZM2G124963 0.6 1.2 0.002965 2 2 pyridoxal phosphate binding GRMZM2G112538 2 1.4 0.011065 3 1 response to biotic stimulus/ Pathogenesis-related protein 10 GRMZM2G044132 1.3 0.8 0.040668 1 2 response to stress (Water deficit, ABA and ripening) GRMZM2G133718 1.7 1.1 0.038534 3 3 serine-type carboxypeptidase activity GRMZM2G028393 3.4 2 0.008251 6 8 serine-type endopeptidase inhibitor activity GRMZM2G042789 2.9 1.3 0.018048 6 8 serine-type endopeptidase inhibitor activity SKP1-like protein 1A/ ubiquitin-dependent protein catabolic GRMZM2G101446 1.1 0.9 0.040247 1 4 process GRMZM2G104179 1.3 0.9 0.019384 1 8 Tetratricopeptide repeat protein 1 GRMZM2G136522 0.8 1.3 0.001346 4 7 Thioredoxin reductase GRMZM2G005024 0.7 1.6 0.000953 2 10 tryptophan synthase activity GRMZM2G168849 0.6 1.1 0.000747 2 1 UVR domain

139

GRMZM2G158976 1.3 0.7 0.001244 1 1 VQ domain/ modulating plastid sigma factors GRMZM2G377168 2.1 1.1 0.010795 3 5

140

(a)

(b)

Figure 4.3. Singular gene ontology (GO) enrichment analysis of molecular function performed on hierarchical cluster families (a) one, (b) two, (c) three, and (d) four by using agriGO online

141 tools and visualized using REVIGO online tools. [In circles: color represents the GO enrichment

P-value and the border color represents the number of proteins associated with that term]. Both color descriptors are continuous color mapping with the darker red indicating a lower P-value and darker green indicating a higher number of proteins associated with that GO term. Similar

GO terms are linked in the graph.

142

(a)

(b)

Figure 4.4. Gene ontology enrichment of the proteins in the correlation network. (a) Biological process GO and (b) molecular function GO for all proteins within the correlation network. All the GO terms are significantly enriched with a P≤ 0.05. **Significantly enriched terms that belong to GO groups.

143

Table 4.3. Proteins that belong to the subnetwork created from the protein correlation network, that fall within a known QTL for FAW resistance and have proteins abundance p≤ 0.05. Columns represent protein id, protein ratios, p-value, groups are the 10 k-mean families clusters, chromosome location of the proteins and a description of the protein domains and GO annotation. Protein id Lin:Lc Rin:Rc pval Chr groups Description GRMZM2G046186_P01 2 1 0.016523 7 3 leaf senescence/ Kinesin like protein Inorganic diphosphatase activity, inorganic GRMZM2G090718_P01 1.6 1 0.002121 5 3 pyrophosphatase GRMZM2G150791_P01 1.6 1 0.011301 7 3 Ankyrin repeat-containing domain GRMZM2G090271_P01 1.6 1 0.031745 2 3 Plastid-specific 30S ribosomal protein 2 GRMZM2G093574_P02 1.6 0.7 0.049721 6 1 40S ribosomal protein S21/ribosome GRMZM2G073504_P01 1.5 1 0.013082 7 3 UBX domain GRMZM2G118037_P01 1.5 0.9 0.018359 6 1 Pi starvation-induced protein/ Peptidase S8 domain GRMZM2G174799_P01 1.4 0.9 1.94E-04 7 3 FAS1 (fasciclin-like) domain GRMZM2G070199_P01 1.4 0.9 0.001019 6 3 oxidation-reduction process, Cytochrome c GRMZM2G062481_P01 1.4 0.9 0.024026 6 1 Acyl carrier protein/fatty acid biosynthetic process GRMZM2G068496_P01 1.4 1 0.039763 6 1 Ribosome/Ribosomal protein L29e PREDICTED: LOW QUALITY PROTEIN: GRMZM2G134711_P01 1.4 0.8 0.041169 7 1 probable ADP-ribosylation factor GTPase- activating protein AGD11 GRMZM2G005973_P01 1.3 0.9 0.007915 6 1 Ribosomal_L1 Nascent polypeptide-associated complex subunit GRMZM2G110116_P01 1.3 0.9 0.023103 9 1 beta/regulation of transcription, DNA-templated

144

GRMZM2G181542_P01 1.3 0.8 0.040492 9 1 Acyl carrier protein/fatty acid biosynthetic process NADH-ubiquinone oxidoreductase 19 kDa subunit, GRMZM2G006085_P01 1.2 0.7 0.00345 1 1 mitochondrial respiratory chain complex I GRMZM2G077711_P01 0.9 1.5 0.031896 9 4 DnaJ domain or J-domain IMP cyclohydrolase GRMZM2G306374_P01 0.8 1 0.035871 2 2 activity/phosphoribosylaminoimidazolecarboxamide formyltransferase activity GRMZM2G105644_P01 0.7 1.1 1.74E-04 5 2 geranylgeranyl reductase activity hydrolase activity, CDC48 N-terminal domain and GRMZM2G063060_P01 0.7 0.9 0.00339 9 2 AAA+ ATPase domain Transaminase/transferase/pyridoxal phosphate GRMZM2G010328_P01 0.7 1 0.032148 1 2 binding UTP:glucose-1-phosphate uridylyltransferase GRMZM2G032003_P02 0.6 1 0.001526 7 2 activity pyruvate, phosphate dikinase activity, PEP-utilising GRMZM2G306345_P04 0.5 1.2 3.69E-04 6 2 enzyme GRMZM2G306345_P01 0.5 1.4 0.015216 6 2 pyruvate, phosphate dikinase activity

145

Figure 4.5. Protein subnetworks from the protein correlation network. (a) Subnetwork of proteins that fall within a known QTL for FAW resistance, highlighted in yellow are proteins with P≤ 0.05. (b) Hub with the highest scored P-value inner colors represent the Lin:Lc ratios and outer colors the Rin:Rc. These hotspots were created by using the jActiveModules tool in Cytoscape. Node shapes represent the protein family from the hierarchical clustering analysis, diamond is one, triangle is two, hexagon is three, rectangle is four, v-shape is six, parallelogram is nine and ellipse is no-family.

146

Table 4.4. Proteins of interest from the hotspot with the highest scored p-value generated from the protein abundance correlation network. Columns represent protein id, protein ratios, p-value, groups are the 10 k-mean families clusters, chromosome location of the proteins and a description of the protein domains and GO. In Name Lin:Lc Rin:Rc pval Chr groups Description QTL GRMZM2G059556_P01 1.7 0.7 8.42E-04 4 1 NADH ubiquinone oxidoreductase 13kD-like subunit transferase activity, transferring acyl groups other than amino- GRMZM2G325001_P01 1.7 0.7 0.002639 5 1 acyl groups glucosinolate biosynthetic process/Pyridine nucleotide- GRMZM2G320307_P02 1.7 1 0.040327 5 3 disulphide oxidoreductase GRMZM2G044498_P01 1.6 0.9 0.001377 7 3 PREDICTED: cylicin-2-like GRMZM2G122793_P01 1.6 0.9 1.43E-04 9 3 Ferredoxin-thioredoxin reductase catalytic chain, chloroplastic GRMZM2G010929_P01 1.6 0.9 0.001632 8 3 DNA binding/HB transcription factor GRMZM2G116689_P01 1.5 0.6 0.006919 7 1 Polyubiquitin 2 GRMZM2G136262_P01 1.5 0.9 0.012279 3 3 fatty acid biosynthetic process/phosphopantetheine binding GRMZM2G447785_P01 1.5 1 4.52E-04 2 3 serine-type endopeptidase inhibitor activity GRMZM2G094928_P01 1.5 0.7 0.032114 7 1 vacuolar proton-transporting V-type ATPase complex GRMZM2G144653_P02 1.5 0.8 0.016862 7 1 Thioredoxin/glycerol ether metabolic process GRMZM2G025992_P03 1.5 0.9 0.010792 7 1 Superoxide dismutase [Cu-Zn] 2 GRMZM2G118037_P01 1.5 0.9 0.018359 6 1 1 Pi starvation-induced protein GRMZM2G028379_P02 1.5 0.9 0.009423 1 3 CHCH motif/ Cysteine alpha-hairpin motif superfamily GRMZM2G144180_P01 1.4 0.7 0.01685 2 1

147

GrpE protein homolog/adenyl-nucleotide exchange factor GRMZM2G002440_P02 1.4 0.9 0.00403 4 3 activity GRMZM2G416388_P01 1.4 0.8 0.023232 1 1 CBS domain protein GRMZM2G062481_P01 1.4 0.9 0.024026 6 1 1 Acyl carrier protein/fatty acid biosynthetic process Zinc finger C-x8-C-x5-C-x3-H type family protein/RNA GRMZM2G124476_P02 1.4 0.8 0.021218 10 1 binding GRMZM2G055936_P01 1.4 0.8 0.008559 8 1 Peroxiredoxin-5/oxidoreductase activity GRMZM2G120517_P02 1.4 0.9 0.003361 2 3 peroxidase activity GRMZM2G073401_P01 1.4 0.7 0.012893 1 1 Chaperonin/protein folding GRMZM2G174799_P01 1.4 0.9 1.94E-04 7 3 1 FAS1 (fasciclin-like) domain GRMZM2G119691_P01 1.4 0.9 0.003834 5 3 EMB12/translation initiation factor activity GRMZM2G138523_P01 1.4 0.9 0.002883 5 1 huntingtin interacting protein K GRMZM2G113696_P01 1.4 0.9 0.011489 2 1 Eukaryotic translation initiation factor 5A/ribosome binding GRMZM2G134738_P01 1.4 0.9 0.017999 8 1 cytochrome-c oxidase activity GRMZM2G033283_P01 1.4 0.9 0.012429 8 1 Translation machinery-associated protein 22 GRMZM2G135893_P01 1.3 0.8 0.001566 4 1 glutathione peroxidase activity GRMZM2G101480_P01 1.3 0.8 0.001302 4 1 MBF1 transcription factor GRMZM2G110116_P01 1.3 0.9 0.023103 9 1 1 Nascent polypeptide-associated complex subunit beta Putative FKBP-type peptidyl-prolyl cis-trans isomerase family GRMZM2G031204_P01 1.3 0.8 0.024439 8 1 protein/peptidyl-prolyl cis-trans isomerase activity GRMZM2G181542_P01 1.3 0.8 0.040492 9 1 1 Acyl carrier protein/fatty acid biosynthetic process GRMZM2G308689_P02 1.3 1 4.33E-04 6 1 mRNA splicing factor SYF2

148

NADH ubiquinone oxidoreductase B14 subunit/aerobic GRMZM2G014382_P01 1.3 0.8 0.007387 1 1 respiration GRMZM2G049390_P01 1.3 0.9 0.017266 5 1 prefoldin complex/Prefoldin (PFD) GRMZM2G393671_P01 1.3 0.9 0.02486 7 1 Stress responsive alpha-beta barrel GRMZM2G055276_P01 1.3 0.9 4.10E-04 3 1 Hyaluronan/mRNA-binding protein GRMZM2G439201_P02 1.3 0.9 0.02847 7 1 eukaryotic translation elongation factor 1 complex GRMZM2G145449_P01 1.2 0.9 0.009989 3 1 Peroxiredoxin-5/aerobic respiration GRMZM2G157018_P01 1.2 0.8 0.019925 1 1 ATP synthase D chain, mitochondrial GRMZM2G169182_P01 1.2 0.8 0.033196 3 1 PB1 (Phox and Bem1) domain PREDICTED: LOW QUALITY PROTEIN: uncharacterized GRMZM2G329300_P01 1.2 0.9 0.0253 1 1 threonine-rich GPI-anchored glycoprotein PJ4664.02-like GRMZM2G157332_P07 1.2 0.8 0.028867 8 1

149

(a)

(b)

Figure 4.6. Gene ontology enrichment of the protein-protein interaction (PPI) network. GO of biological process (a) and molecular function (b) from all the proteins within the PPI created using our TMT data set. All the terms are significantly enriched with a P≤ 0.05. **Significantly enriched terms that belong to GO groups.

150

Figure 4.7. Protein subnetworks from the protein-protein interaction (PPI) network. (a) Hub with the highest scored P-value and (b) subnetwork of the highest connected component. Hub was

151 created by using the jActiveModules tool in Cytoscape. Node shapes represent the protein family from the hierarchical clustering analysis, diamond is one, hexagon is two, triangle is three, v- shape is four, rectangle is five and elipse is no-family. Color in nodes represent pathway GO term, green is ABA signaling, pink is ET biosynthesis, red is ET signaling, gray is JA biosynthesis, purple is JA signaling, light yellow is ROS metabolism, dark yellow is ROS response, aqua is response to other organism and blue is no-pathway.

152

d d (a)

(b) d d d d

(c)

Figure 4.8. Plant physiological changes measured in Mp708 and Tx601 at days 6 or 9 after 24h 153

FAW infestation. (a) Height, (b) total root length and (c) total length of different root types. Bar graphs represent Mp708 (grey) and Tx601 (white), data were normalized and analyzed using multiple-factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test. Letters represent results of the HSD (P<0.05) and error bars show the standard error.

154

(a) Leaves d d

(b) Root d d

Figure 4.9. Percent of carbon in leaves and roots of Mp708 and Tx601 at days 6 or 9 after 24 hours of FAW infestation. (a) Percentage of carbon in leaves and (b) roots. Bar graphs represent Mp708 (grey) and Tx601 (white), data were normalized and was analyzed using multiple-factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test. Letters represent results of the HSD (P<0.05) and error bars show the standard error.

155

(a) lox3

(b) Leaves

Figure 4.10. lox3 gene expression and ET accumulation in Mp708 after 24 hours of FAW 156 infestation. (a) lox3 transcript abundance. Time course analysis of ethylene (ET) accumulation in (b) leaves and (c) roots. Control plants were not infested with the herbivores. Relative expression (RQ) was measured by qRT-PCR and gene expression levels were normalized to actin. Ethylene levels were determined as described in Materials and Methods. Data were normalized and analyzed using multiple-factor ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test. Letters represent results of the HSD (P<0.05) and error bars show the standard error.

157

Figure 4.11. Gene (a) mir1 (c) Ibuprofen (JA) 1-MCP (ET) expression and immunoblot Control FAW Control FAW Mir1-CP in Mp708 after 24 NS IBU NS IBU NS 1-MCP NS 1-MCP hours of FAW infestation. Leaf (a) mir1 transcript Root abundance after 24 h of Leaf infestation. Relative Root expression (RQ) was measured by qRT-PCR and gene expression levels were (b) rip2 (d) normalized to actin. RQ Ibuprofen (JA) 1-MCP (ET) Control FAW Control FAW data were normalized and 1-MCP IBU NS IBU NS NS 1-MCP analyzed using multiple- Leaf factor ANOVA followed by Root honest significant difference Leaf (HSD) Tukey pairwise Root comparison test. Results of the HSD are represented by letters and error bars represent the standard error. Immunoblot analysis of Mir1-CP accumulation in response to FAW in the presence of jasmonic acid (JA) synthesis inhibitors and ethylene (ET) perception inhibitor. NS refers to plants that were not treated with the inhibitors. Each

158 lane contains 30 µg of leaf protein or 15 µg of root protein pooled from three biological replicates. Ponceau S staining of the blots shows equal protein loading.

159

(a) Peroxidase in leaves (b) Peroxide in leaves

(e) PPO in leaves

Figure 4.12. Time course of peroxidase activity assay in (a) leaves, (c) roots and (e) PPO

160 activity in leaves at 0, 2, 6, 12 and 24 hours after FAW. Also, time course peroxide concentration assay in (b) leaves and (e) roots at 0, 2, 6, 24 h after FAW. Time course Bar graphs represent

Mp708 (black) and Tx601 (white), data were normalized and analyzed using multiple-factor

ANOVA followed by honest significant difference (HSD) Tukey pairwise comparison test.

Results of the HSD are represented by letters and error bars represent the standard error.

Table S4.1. Identified proteins present in leaves and roots of control and FAW infested plants

(95 % false discovery rate) with their protein abundance ratios. Logarithm of protein abundance ratios were analyzed using multiple-factor analysis of variance (ANOVA) followed by honest significant difference (HSD) Tukey pairwise comparison test. Letters were assigned to each group treatment column denotating significant differences.

Table S4.2. Proteins and gene ontology annotation from the hierarchical clustering analysis. The

10 k-mean groups are labeled as protein families and represent 10 different protein abundance patterns determined by using an Euclidian similarity distance method to determine abundance patterns using a maximum of 10 k-means cut off.

Note S4.1. Protein abundance correlation network built with Cytoscape (813 nodes and 17197 edges).

Table S4.3. Proteins and their annotations present in the abundance correlation network from the

Supplemental Cytoscape file 1.

161

Note S4.2. Protein-protein interaction network built with Cytoscape, using TMT data (1181 nodes and 4142 edges).

Table S4.4. Proteins and their annotations present in the protein-protein interaction network from the Supplemental Cytoscape file 2.

162

CHAPTER V. Conclusions

Insect resistant maize genotype, Mp708, has multi-trait that enable it perform well during WCR infestation in addition to being resistant to insects with different feeding behaviors like FAW

(Williams et al., 1985; Williams et al., 1990) and CLA (Louis et al., 2015). Among those traits

Mp708 has a root system more resistant to WCR feeding than Tx601 and is able to maintain root growth during WCR infestation. Mp708 has tough roots that could make it more difficult for

WCR to feed on the roots and access nutrients compared to Tx601, and high constitutive and inducible levels of JA in Mp708 suggests that it plays a key role in activating downstream defenses to resist WCR attack. Mp708 and Tx601 use the products of mpi and fpps3 to defend against herbivory, whereas rip2, tps23 and mir1 are only inducible in Mp708 and could be key players in its resistance. Finally, we conclude that Mp708 has a suite of robust defense traits in its roots and the presence of (E)-β-caryophyllene could play two important roles, deterring FAW feeding and attracting the natural enemies of WCR.

We were able to further characterize Mp708 defense responses to WCR in roots and leaves using proteomic data by implementing hierarchical clustering and two types of network analyses, protein co-relation and protein-protein interaction (PPI) networks. These analyses revealed proteins with higher abundance during WCR infestation that involved JA, ABA and ET biosynthesis and signaling pathways. Also, they showed that proteases of serine/cysteine type are involved in plant defenses in infested roots, further characterization of protease activity allowed us to determine that total protease activity increased in roots with WCR treatments, and that serine-type proteases are highly contributing to this activity. We do not know the physiological damage that this type of proteases are doing on WCR, but we hypothesized that they could be breaking down cuticle on the outer layers of the larvae. Activation of ET biosynthetic pathway

163 was found in both network analyses and volatile analysis showed that ET production goes up in infested roots. We proposed that this phytohormone might be involved in short distance communication due to its high diffusion from plant tissues into the soil and surrounding areas

(Zechmeister-Boltenstern & Nikodim, 1999) or might be an intra-plant communication signal produced by the roots that could be transported to the leaves through aerenchyma in the roots

(Colmer, 2003).

In terms of individual plant organ responses to WCR, we observed a distinct difference between the local and systemic responses. The roots responded by increasing JA, ET and ABA biosynthesis and production of insect defenses such as proteases, serine-type proteases,

Bowman-Birk type trypsin inhibitor, RIPs, peroxidases, lectin-type proteins, cinnamaldehyde, terpenoids and benzothiazol (BZO). On the other hand, leaves tended to increase production of proteins that control transcription and translation, and proteins with signaling domains, implying that systemic tissues are reading themselves to receive a signal from roots that could lead to increasing organ specific responses and change the metabolic focus from growth to controlling protein production. Several hormonal biosynthetic and signaling pathways of JA and ET were activated in the roots, while in leaves there was an increase in signaling perception proteins that could be related to systemic defense activation. In roots there was evidence for the coexistence of plant defense responses to microbes and fungi in addition to those involved in herbivore defense.

The three analyses indicated that roots infested with WCR also showed responses against other biotic stresses like fungi and pathogens. This idea was correlated to root production of the volatile, BZO, an antimicrobial, antifungal (Yadav et al., 2011) and insecticidal (Zhao et al.,

2016) compound. We postulated that the plant is constantly balancing its defenses against root herbivores and other biotic stressors because the roots are in constant contact with soil

164 microorganisms. Future research directions could focus on understanding the plant-insect- microbe interaction balance and how it benefits the plant.

Hierarchical clustering and network analyses of proteomics data from FAW infested plants showed promising results in unraveling potential metabolic pathways, signaling events and proteins with insecticidal properties. All the analyses showed an increase in the JA biosynthetic pathway, changes in oxido-reductase status, peroxidase abundance, REDOX changes in the glutathione and ascorbate pathways, and increase in ABA/ET biosynthesis and signaling. Also, infested leaves show high abundance and enrichment of proteinase inhibitors, cysteine proteases and peroxidases. Peroxidase activity increases early in leaves of Mp708 while it does not change in the roots during FAW infestation, implying that an early ROS-related signal event that does not correlate with increase in peroxide changes or a spike in quinone production could be happening in the leaves.

Using the protein-correlation network with the known QTL regions in Mp708 that explain the resistance against FAW, we found some proteins that might be part of the defense mechanisms of this genotype. Due to the association of some of these proteins with plant development and growth we determined that the insect susceptible maize genotype, Tx601, grew less in height and total root length during FAW infestation. Thin and seminal like roots significantly decreased during FAW infestation in Tx601. Also, the carbon content in leaves and roots was lower in Tx601 compared to Mp708 only in FAW-infested plants after 9 days of infestation. This could be linked to the decrease in plant height and root length in Tx601 compared to Mp708 that could involve lower trade-off in Mp708 between plant growth and defense responses. ET accumulation increases only in roots during FAW infestation. Both JA and ET appears to control local MIR1-CP and RIP2 accumulation whereas ET could be a key

165 regulator of MIR1-CP systemic accumulation. These findings open up new routes in the hormonal control of local and systemic defense signaling that could be similar for other defense proteins and secondary compounds. Furthermore, the question prevails as to why there is a different phytohormone control over defenses in an organ-specific matter.

Mp708 was developed from landraces of maize that most likely originated in

Mesoamerica (Williams et al., 1987), where many phytophagous maize pests including

Diabrotica sp., FAW and CLA may also have originated (de Lange et al., 2014). We have established an analysis pipeline for proteomics data that includes network biology approaches that can be used with different types of “omics” data from a wide variety of organisms to detect tissue-specific defense responses to be able to understand the organism strategies to survive or diminish the external stress.

166

BIBLIOGRAPHY

Abel CA, Berhow MA, Wilson RL, Binder BF, Hibbard BE. 2000. Evaluation of

conventional resistance to European corn borer (Lepidoptera : Crambidae) and western

corn rootworm (Coleoptera : Chrysomelidae) in experimental maize lines developed from

a backcross breeding program. Journal of Economic Entomology 93(6): 1814-1821.

Ahn Y-Y, Bagrow JP, Lehmann S. 2010. Link communities reveal multiscale complexity in

networks. Nature 466(7307): 761-764.

Albert R. 2005. Scale-free networks in cell biology. Journal of Cell Science 118(21): 4947.

Ali A, Luttrell RG, Schneider JC. 1990. Effects of temperature and larval diet on development

of the Fall armyworm (Lepidoptera: Noctuidae). Annals of the Entomological Society of

America 83(4): 725-733.

Ali JG, Alborn HT, Stelinski LL. 2010. Subterranean herbivore-induced volatiles released by

citrus roots upon feeding by Diaprepes abbreviatus recruit entomopathogenic

Nematodes. Journal of Chemical Ecology 36(4): 361-368.

Ali R, Siddiqui N. 2013. Biological aspects of emerging benzothiazoles: A Short Review.

Journal of Chemistry 2013: 1-12.

Alm E, Arkin AP. 2003. Biological networks. Current Opinion in Structural Biology 13(2):

193-202.

Almagro L, Ros LVG, Belchi-Navarro S, Bru R, Barcelo AR, Pedreno MA. 2009. Class III

peroxidases in plant defence reactions. Journal of Experimental Botany 60(2): 377-390.

Ang KY, Lucas PW, Tan HTW. 2008. Novel way of measuring the fracture toughness of

leaves and other thin films using a single inclined razor blade. New Phytologist 177(3):

830-837.

167

Ankala A, Kelley RY, Rowe DE, Williams WP, Luthe DS. 2013. Foliar herbivory triggers

local and long distance defense responses in maize. Plant Science 199: 103-112.

Ankala A, Luthe DS, Williams WP, Wilkinson JR. 2009. Integration of ethylene and jasmonic

acid signaling pathways in the expression of maize defense protein Mir1-CP. Molecular

Plant-Microbe Interactions 22(12): 1555-1564.

Aranda B, Achuthan P, Alam-Faruque Y, Armean I, Bridge A, Derow C, Feuermann M,

Ghanbarian AT, Kerrien S, Khadake J, et al. 2010. The IntAct molecular interaction

database in 2010. Nucleic Acids Research 38(Database issue): D525-D531.

Arraes FBM, Beneventi MA, de Sa MEL, Paixao JFR, Albuquerque EVS, Marin SRR,

Purgatto E, Nepomuceno AL, Grossi-de-Sa MF. 2015. Implications of ethylene

biosynthesis and signaling in soybean drought stress tolerance. Bmc Plant Biology 15:

213.

Arsenault J-L, Pouleur S, Messier C, Guay R. 1995. WinRHIZO, a root-measuring system

with a unique overlap correction method. HortScience 30: 906.

Assabgui RA, Arnason JT, Hamilton RI. 1993. Hydroxamic acid content in maize (Zea mays)

roots of 18 Ontario recommended hybrids and prediction of antibiosis to the western corn

rootworm, Diabrotica virgifera virgifera LeConte [Coleoptera: Chrysomelidae].

Canadian Journal of Plant Science 73(1): 359-363.

Assabgui RA, Arnason TJ, Hamilton RI. 1995. Field evaluations of hydroxamic acids as

antibiosis factors in elite maize inbreds to the Western corn root worm (Coleoptera:

Chrysomelidae). Journal of Economic Entomology 88(5): 1482-1493.

Azuaje FJ. 2014. Selecting biologically informative genes in co-expression networks with a

centrality score. Biology Direct 9.

168

Ballare CL. 2011. Jasmonate-induced defenses: a tale of intelligence, collaborators and rascals.

Trends in Plant Science 16(5): 249-257.

Bang KH, Lee DW, Park HM, Rhee YH. 2000. Inhibition of fungal cell wall synthesizing

enzymes by trans-cinnamaldehyde. Bioscience Biotechnology and Biochemistry 64(5):

1061-1063.

Barabasi AL, Bonabeau E. 2003. Scale-free networks. Scientific American 288(5): 60-69.

Barah P, Bones AM. 2015. Multidimensional approaches for studying plant defence against

insects: from ecology to omics and synthetic biology. Journal of Experimental Botany

66(2): 479-493.

Barelli L, Moonjely S, Behie SW, Bidochka MJ. 2016. Fungi with multifunctional lifestyles:

endophytic insect pathogenic fungi. Plant Molecular Biology 90(6): 657-664.

Bass HW, Krawetz JE, Obrian GR, Zinselmeier C, Habben JE, Boston RS. 2004. Maize

ribosome-inactivating proteins (RIPs) with distinct expression patterns have similar

requirements for proenzyme activation. Journal of Experimental Botany 55(406): 2219-

2233.

Bassani M, Neumann PM, Gepstein S. 2004. Differential expression profiles of growth-related

genes in the elongation zone of maize primary roots. Plant Molecular Biology 56(3): 367-

380.

Bassel GW, Gudinier A, Brady SM, Hennig L, Rhee S, Smet ID. 2012. Systems analysis of

plant functional, transcriptional, physicial interaction, and metabolic netwoks. The Plant

Cell 24(10):3859-3875.

Bergeron F, Leduc R, Day R. 2000. Subtilase-like pro-protein convertases: from molecular

specificity to therapeutic applications. Journal of Molecular Endocrinology 24(1): 1-22.

169

Bigger JH. 1932. Short rotation fails to prevent attack of Diabrotica longicornis Say. Journal of

Economic Entomology 25: 196-199.

Bindea G, Mlecnik B, Hackl H, Charoentong P, Tosolini M, Kirilovsky A, Fridman W-H,

Pages F, Trajanoski Z, Galon J. 2009. ClueGO: a Cytoscape plug-in to decipher

functionally grouped gene ontology and pathway annotation networks. Bioinformatics

25(8).

Borrego EJ, Kolomiets MV. 2016. Synthesis and functions of jasmonates in maize. Plants 5(4):

41.

Bozkurt TO, Schornack S, Win J, Shindo T, Ilyas M, Oliva R, Cano LM, Jones AME,

Huitema E, van der Hoorn RAL, et al. 2011. Phytophthora infestans effector AVRblb2

prevents secretion of a plant immune protease at the haustorial interface. Proceedings of

the National Academy of Sciences 108(51): 20832-20837.

Branson TF, Krysan JL. 1981. Feeding and oviposition behavior and life-cycle strategies of

Diabrotica - An evolutionary view with implications for pest-management (Coleoptera,

Chrysomelidae). Environmental Entomology 10(6): 826-831.

Branson TF, Sutter GR, Fisher JR. 1982. Comparison of the tolerant and susceptible maize

inbred under artificial infestations of Diabrotica virgifera virgifera (Coleoptera,

Chrysomelidae) –Yield and adult emergence. Environmental Entomology 11(2): 371-372.

Broekgaarden C, Caarls L, Vos IA, Pieterse CMJ, Van Wees SCM. 2015. Ethylene: Traffic

Controller on Hormonal Crossroads to Defense. Plant Physiology 169(4): 2371-2379.

Brooks TD, Bushman BS, Williams WP, McMullen MD, Buckley PM. 2007. Genetic basis of

resistance to Fall armyworm (Lepidoptera : Noctuidae) and Southwestern corn borer

170

(Lepidoptera : Crambidae) leaf-feeding damage in maize. Journal of Economic

Entomology 100(4): 1470-1475.

Brooks TD, Willcox MC, Williams WP, Buckley PM. 2005. Quantitative trait loci conferring

resistance to fall armyworm and southwestern corn borer leaf feeding damage. Crop

Science 45(6): 2430-2434.

Brown AE, Smith WEC, Shivaji R, Williams WP, Luthe DS, Sandoya GV, Smith C. 2011.

Mp708 a maize (Zea mays) line resistant to herbivory constitutively releases (E)-b-

caryophyllene. Abstracts of Papers of the American Chemical Society 242.

Burg SP, Burg EA. 1966. Interaction between auxin and ethylene and its role in plant growth.

Proceedings of the National Academy of Sciences of the United States of America 55(2):

262-&.

Burton AL, Williams M, Lynch JP, Brown KM. 2012. RootScan: Software for high-

throughput analysis of root anatomical traits. Plant and Soil 357(1-2): 189-203.

Capra E, Colombi C, De Poli P, Nocito FF, Cocucci M, Vecchietti A, Marocco A, Stile MR,

Rossini L. 2015. Protein profiling and tps23 induction in different maize lines in

response to methyl jasmonate treatment and Diabrotica virgifera infestation. Journal of

Plant Physiology 175: 68-77.

Carrillo MGC, Goodwin PH, Leach JE, Leung H, Cruz CMV. 2009. Phylogenomic

relationships of rice Oxalate oxidases to the cupin superfamily and their association with

disease resistance QTL. Rice 2(1): 67-79.

Chang KN, Zhong S, Weirauch MT, Hon G, Pelizzola M, Li H, Huang S-sC, Schmitz RJ,

Urich MA, Kuo D, et al. 2013. Temporal transcriptional response to ethylene gas drives

growth hormone cross-regulation in Arabidopsis. Elife DOI: 10.7554/eLife.00675

171

Chatr-aryamontri A, Ceol A, Palazzi LM, Nardelli G, Schneider MV, Castagnoli L,

Cesareni G. 2007. MINT: the Molecular INTeraction database. Nucleic Acids Research

35(Database issue): D572-D574.

Chen H, Gonzales-Vigil E, Wilkerson CG, Howe GA. 2007. Stability of plant defense proteins

in the gut of insect herbivores. Plant Physiology 143(4): 1954-1967.

Chen M-S, Wu J-X, Zhang G-H. 2009. Inducible direct plant defense against insect herbivores.

Chinese Bulletin of Entomology 46(2): 175-186.

Chen MS. 2008. Inducible direct plant defense against insect herbivores: A review. Insect

Science 15(2): 101-114.

Cheval C, Aldon D, Galaud JP, Ranty B. 2013. Calcium/calmodulin-mediated regulation of

plant immunity. Biochimica Et Biophysica Acta-Molecular Cell Research 1833(7): 1766-

1771.

Chimungu JG, Brown KM, Lynch JP. 2014. Large root cortical cell size improves drought

tolerance in maize. Plant Physiology 166(4): 2166-2178.

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): 666-671.

Chuang W-P, Herde M, Ray S, Castano-Duque L, Howe GA, Luthe DS. 2014. Caterpillar

attack triggers accumulation of the toxic maize protein RIP2. New Phytologist 201(3):

928-939.

Cline MS, Smoot M, Cerami E, Kuchinsky A, Landys N, Workman C, Christmas R, Avila-

Campilo I, Creech M, Gross B, et al. 2007. Integration of biological networks and gene

expression data using Cytoscape. Nature Protocols 2(10): 2366-2386.

172

Colmer TD. 2003. Long-distance transport of gases in plants: a perspective on internal aeration

and radial oxygen loss from roots. Plant, Cell & Environment 26(1): 17-36.

Creelman RA, Mullet JE. 1997. Biosynthesis and action of jasmonates in plants. Annual

Review of Plant Physiology and Plant Molecular Biology 48: 355-381. de Lange ES, Balmer D, Mauch-Mani B, Turlings TCJ. 2014. Insect and pathogen attack and

resistance in maize and its wild ancestors, the teosintes. New Phytologist 204(2): 329-

341.

Degenhardt J, Hiltpold I, Koellner TG, Frey M, Gierl A, Gershenzon J, Hibbard BE,

Ellersieck MR, Turlings TCJ. 2009. Restoring a maize root signal that attracts insect-

killing nematodes to control a major pest. Proceedings of the National Academy of

Sciences of the United States of America 106(32): 13213-13218.

Dixon RA, Harrison MJ, John GS 1990. Activation, structure, and organization of genes

involved in microbial defense in plants. Advances in Genetics: Academic Press, 165-234.

Doncheva NT, Assenov Y, Domingues FS, Albrecht M. 2012. Topological analysis and

interactive visualization of biological networks and protein structures. Nature Protocols

7(4): 670-685.

Du Z, Zhou X, Ling Y, Zhang Z, Su Z. 2010. agriGO: a GO analysis toolkit for the agricultural

community. Nucleic Acids Res 38 (Web Server issue):W64-70.

Dunham CM, Dioum EM, Tuckerman JR, Gonzalez G, Scott WG, Gilles-Gonzalez MA.

2003. A distal arginine in oxygen-sensing Heme-PAS domains is essential to ligand

binding, signal transduction, and structure. Biochemistry 42(25): 7701-7708.

El Khishen AA, Bohn MO, Prischmann-Voldseth DA, Dashiell KE, Wade French B,

Hibbard BE. 2009. Native resistance to western corn rootworm (Coleoptera:

173

Chrysomelidae) larval feeding: characterization and mechanisms. Journal of Economic

Entomology 102(6): 2350-2359.

Eleazar Ruiz-Najera R, Molina-Ochoa J, Carpenter JE, Espinosa-Moreno JA, Alfredo

Ruiz-Najera J, Lezama-Gutierrez R, Foster JE. 2007. Survey for Hymenopteran and

Dipteran parasitoids of the Fall armyworm (Lepidoptera: Noctuidae) in Chiapas, Mexico.

Journal of Agricultural and Urban Entomology 24(1): 35-42.

Erb M, Flors V, Karlen D, de Lange E, Planchamp C, D'Alessandro M, Turlings TCJ, Ton

J. 2009a. Signal signature of aboveground-induced resistance upon belowground

herbivory in maize. Plant Journal 59(2): 292-302.

Erb M, Glauser G. 2010. Family business: multiple members of major phytohormone classes

orchestrate plant stress responses. Chemistry-a European Journal 16(34): 10280-10289.

Erb M, Glauser G, Robert CAM. 2012. Induced immunity against belowground insect

herbivores- Activation of defenses in the absence of a jasmonate burst. Journal of

Chemical Ecology 38(6): 629-240.

Erb M, Gordon-Weeks R, Flors V, Camañes G, Turlings TC, Ton J. 2009b. Belowground

ABA boosts aboveground production of DIMBOA and primes induction of chlorogenic

acid in maize. Plant signaling and behavior 4(7): 636-638.

Erb M, Lenk C, Degenhardt J, Turlings TCJ. 2009c. The underestimated role of roots in

defense against leaf attackers. Trends in plant science 14(12): 653-659.

Erb M, Lu J. 2013. Soil abiotic factors influence interactions between belowground herbivores

and plant roots. Journal of Experimental Botany 64(5): 1295-1303.

174

Flagel LE, Swarup S, Chen M, Bauer C, Wanjugi H, Carroll M, Hill P, Tuscan M, Bansal

R, Flannagan R, et al. 2015. Genetic markers for western corn rootworm resistance to

Bt Toxin. G3-Genes Genomes Genetics 5(3): 399-405.

Fonseca S, Chini A, Hamberg M, Adie B, Porzel A, Kramell R, Miersch O, Wasternack C,

Solano R. 2009. (+)-7-iso-Jasmonoyl-L-isoleucine is the endogenous bioactive

jasmonate. Nature Chemical Biology 5(5): 344-350.

Foyer CH, Noctor G. 2011. Ascorbate and glutathione: The heart of the redox hub. Plant

Physiology 155(1): 2-18.

Frost CJ, Hunter MD. 2008. Herbivore-induced shifts in carbon and nitrogen allocation in red

oak seedlings. New Phytologist 178(4): 835-845.

Fuchs Y, Lieberma.M. 1968. Effects of kineting IAA and gibberellin on ethylene production

and their interactions in growth of seedlings. Plant Physiology 43(12): 2029-2036.

Galis I, Gaquerel E, Pandey SP, Baldwin IT. 2009. Molecular mechanisms underlying plant

memory in JA-mediated defence responses. Plant Cell and Environment 32(6): 617-627.

Gallie DR, Geisler-Lee J, Chen J, Jolley B. 2009. Tissue-specific expression of the ethylene

biosynthetic machinery regulates root growth in maize. Plant Molecular Biology 69(1-2):

195-211.

Gassmann AJ. 2012. Field-evolved resistance to Bt maize by western corn rootworm:

Predictions from the laboratory and effects in the field. Journal of Invertebrate Pathology

110(3): 287-293.

Gill TA, Sandoya G, Williams P, Luthe DS. 2011. Belowground resistance to western corn

rootworm in Lepidopteran-resistant maize genotypes. Journal of Economic Entomology

104(1): 299-307.

175

Gray ME, Sappington TW, Miller NJ, Moeser J, Bohn MO. 2009. Adaptation and

invasiveness of western corn rootworm: Intensifying research on a worsening pest.

Annual Review of Entomology 54: 303-321.

Green TR, Ryan CA. 1972. Wound-induced proteinase inhibitor in plant leaves – Possible

defense mechanism against insets. Science 175(4023): 776-777.

Guimerà R, Mossa S, Turtschi A, Amaral LAN. 2005. The worldwide air transportation

network: Anomalous centrality, community structure, and cities' global roles.

Proceedings of the National Academy of Sciences 102(22): 7794-7799.

Gulati J, Baldwin IT, Gaquerel E. 2014. The roots of plant defenses: Integrative multivariate

analyses uncover dynamic behaviors of roots’ gene and metabolic networks elicited by

leaf herbivory. The Plant journal : for cell and molecular biology 77(6): 880-892.

Harfouche AL, Shivaji R, Stocker R, Williams PW, Luthe DS. 2006. Ethylene signaling

mediates a maize defense response to insect herbivory. Molecular Plant-Microbe

Interactions 19(2): 189-199.

Hazem Radwan A, Janice IG. 2014. Pattern discovery in protein networks reveals high-

confidence predictions of novel Interactions. Innovative Applications of Artificial

Intelligence; Twenty-Sixth IAAI Conference.

Hoang A. 2010. The role of peroxidase in the defense response of buffalograss to chinch bugs.

Entomology, University of Nebraska, Lincoln Digital commons.

Hochholdinger F, Wen T-J, Zimmermann R, Chimot-Marolle P, da Costa e Silva O, Bruce

W, Lamkey KR, Wienand U, Schnable PS. 2008. The maize (Zea mays L.)

roothairless3 gene encodes a putative GPI-anchored, monocot-specific, COBRA-like

protein that significantly affects grain yield. Plant Journal 54(5): 888-898.

176

Howe GA, & Schaller, A. 2008. Direct defenses in plants and their induction by wounding and

insect herbivores. Induced Plant Resistance to Herbivory. Netherlands Springer 7-29.

Howe GA, Jander G. 2008. Plant immunity to insect herbivores. Annual Review of Plant

Biology 59: 41-66.

Hu F, Huang J, Xu Y, Gian X, Zhong J-J. 2006. Responses of defense signals, biosynthetic

gene transcription and taxoid biosynthesis to elicitation by a novel synthetic jasmonate in

cell cultures of Taxus chinensis. Biotechnology and Bioengineering 94(6): 1064-1071.

Hu Z, Shen Y, Shen F, Luo Y, Su X. 2009. Evidence for the signaling role of methyl

jasmonate, methyl salicylate and benzothiazole between poplar (Populus simonii × P.

pyramidalis ‘Opera 8277’) cuttings. Trees 23(5): 1003-1011.

Huffaker A, Pearce G, Veyrat N, Erb M, Turlings TCJ, Sartor R, Shen Z, Briggs SP,

Vaughan MM, Alborn HT, et al. 2013. Plant elicitor peptides are conserved signals

regulating direct and indirect antiherbivore defense Proceedengs of national academy of

science 110(14): 5707-5712.

Hummel HE. 2003. Introduction of Diabrotica virgifera virgifera into the Old World and its

consequences: A recently acquired invasive alien pest species on Zea mays from North

America. Communications in Agricultural and Applied Biological Sciences 68(4a): 45-

57.

Hummel HE. 2007. Diabrotica virgifera virgifera LeConte: inconspicuous leaf beetle--

formidable challenges to agriculture. Communications in agricultural and applied

biological sciences 72(2): 7-32.

177

Ideker T, Ozier O, Schwikowski B, Siegel AF. 2002. Discovering regulatory and signalling

circuits in molecular interaction networks. Bioinformatics (Oxford, England) 18 Suppl 1:

S233-240.

Ivezic M, Raspudic E, Brmez M, Majic I, Brkic I, Tollefson JJ, Bohn M, Hibbard BE,

Simic D. 2009. A review of resistance breeding options targeting western corn rootworm

(Diabrotica virgifera virgifera LeConte). Agricultural and Forest Entomology 11(3).

Jiang BH, Siregar U, Willeford KO, Luthe DS, Williams WP. 1995. Association of a 33-

kilodalton cysteine proteinase found in corn callus with the inhibition of fall armyworm

larval growth. Plant physiology 108(4): 1631-1640.

Jogaiah S, Govind SR, Tran L-SP. 2012. Systems biology-based approaches toward

understanding drought tolerance in food crops. Critical Reviews in Biotechnology 33: 23-

39.

Johnson KL, Jones BJ, Bacic A, Schultz CJ. 2003. The fasciclin-like arabinogalactan proteins

of arabidopsis. A multigene family of putative cell adhesion molecules. Plant Physiology

133(4): 1911-1925.

Johnson SN, Erb M, Hartley SE. 2016. Roots under attack: contrasting plant responses to

below- and aboveground insect herbivory. New Phytologist 210(2): 413-418.

Johnson SN, Hallett PD, Gillespie TL, Halpin C. 2010. Below-ground herbivory and root

toughness: a potential model system using lignin-modified tobacco. Physiological

Entomology 35(2): 186-191.

Jonczyk R, Schmidt H, Osterrieder A, Fiesselmann A, Schullehner K, Haslbeck M, Sicker

D, Hofmann D, Yalpani N, Simmons C, et al. 2008. Elucidation of the final reactions

178

of DIMBOA-glucoside biosynthesis in maize: Characterization of Bx6 and Bx7. Plant

Physiology 146(3): 1053-1063.

Jung JKH, McCouch S. 2013. Getting to the roots of it: Genetic and hormonal control of root

architecture Getting to the roots of it: Genetic and hormonal control of root

architecture Frontiers in plant science 4: e186

Kadlicko SR, Tollefson JJ, Prasifka JR, Baca F, Stankovic G, Delic N. 2010. Evaluation of

Serbian commercial maize hybrid tolerance to feeding by larval western corn rootworm

(Diabrotica virgifera virgifera LeConte) using ME novel “difference approach”. Maydica

55(3-4): 179-185.

Kahler AL, Olness AE, Sutter GR, Dybing CD, Devine OJ. 1985. Root damage by western

corn-rootworm and nutrient content in maize. Agronomy Journal 77(5): 769-774.

Kang TH, Hwang EI, Yun BS, Park KD, Kwon YM, Shin CS, Kim SU. 2007. Inhibition of

chitin synthases and antifungal activities by 2 '-benzoyloxycinnamaldehyde from

Pleuropterus ciliinervis and its derivatives. Biological & Pharmaceutical Bulletin 30(3):

598-602.

Kant MR, Jonckheere W, Knegt B, Lemos F, Liu J, Schimmel BCJ, Villarroel CA, Ataide

LMS, Dermauw W, Glas JJ, et al. 2015. Mechanisms and ecological consequences of

plant defence induction and suppression in herbivore communities. Annals of Botany

115(7): 1015-1051.

Kerchev PI, Fenton B, Foyer CH, Hancock RD. 2012. Plant responses to insect herbivory:

interactions between photosynthesis, reactive oxygen species and hormonal signalling

pathways. Plant Cell and Environment 35(2): 441-453.

179

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.

Kim S-Y, Volsky DJ. 2005. PAGE: Parametric Analysis of Gene Set Enrichment. BMC

Bioinformatics 6(1): 144.

Kitano H. 2002. Systems biology: A brief overview. Science 295(5560): 1662-1664.

Kollner TG, Held M, Lenk C, Hiltpold I, Turlings TCJ, Gershenzon J, Degenhardt J. 2008.

A maize (E)-beta-caryophyllene synthase implicated in indirect defense responses against

herbivores is not expressed in most American maize varieties. Plant Cell 20(2): 482-494.

Koo AJK, Howe GA. 2009. The wound hormone jasmonate. Phytochemistry 70(13-14): 1571-

1580.

Laemmli UK. 1970. Cleavage of structural Ppoteins during the assembly of the head of

bacteriophage T4. Nature 227(5259): 680-685.

Lambers H, Atkin OK, Millenaar FF. 1996. Respiratory patterns in roots in relation to their

functioning. Plant Roots. The Hidden Half: 323-362.

Lannoo N, Van Damme EJM. 2014. Lectin domains at the frontiers of plant defense. Frontiers

in Plant Science 5: 397.

Lawrence SD, Novak NG, El Kayal W, Ju CJT, Cooke JEK. 2012. Root herbivory: molecular

analysis of the maize transcriptome upon infestation by Southern corn rootworm,

Diabrotica undecimpunctata howardi. Physiologia Plantarum 144(4).

Lawrence SD, Novak NG, Xu H, Cooke JEK. 2013. Herbivory of maize by southern corn

rootworm induces expression of the major intrinsic protein ZmNIP1;1 and leads to the

discovery of a novel aquaporin ZmPIP2. Plant signaling & behavior 8(8).

180

Lee JS, Brown WE, Graham JS, Pearce G, Fox EA, Dreher TW, Ahern KG, Pearson GD,

Ryan CA. 1986. Molecular characterization and phylogenetic studies of a wound-

induvible protease inhibitor gene in Lycopersicon speciaes. Proceedings of the National

Academy of Sciences of the United States of America 83(19): 7277-7281.

Levine E, Spencer JL, Isard SA, Onstad DW, Gray ME. 2002. Adaptation of the western

corn rootworm to crop rotation: evolution of a new strain in response to a management

practice. American Entomologist 48(2): 94-117.

Lincoln DE. 1993. The influence of plant carbon-dioxide and nutrient supply on susceptibility to

insect herbivores. Vegetatio 104: 273-280.

Loades KW, Bengough AG, Bransby MF, Hallett PD. 2015. Effect of root age on the

biomechanics of seminal and nodal roots of barley (Hordeum vulgare L.) in contrasting

soil environments. Plant and Soil 395(1): 253-261.

Lopez L, Camas A, Shivaji R, Ankala A, Williams P, Luthe D. 2007. Mir1-CP, a novel

defense cysteine protease accumulates in maize vascular tissues in response to herbivory.

Planta 226(2): 517-527.

Louis J, Basu S, Varsani S, Castano-Duque L, Jiang V, Williams WP, Felton GW, Luthe

DS. 2015. Ethylene Contributes to maize insect resistance1-Mediated Maize Defense

against the Phloem Sap-Sucking Corn Leaf Aphid. Plant physiology 169(1): 313-324.

Macedo MLR, Oliveira CFR, Oliveira CT. 2015. Insecticidal activity of plant lectins and

potential application in crop protection. Molecules 20(2): 2014-2033.

Machado RAR, Ferrieri AP, Robert CA, Glauser G, Kallenbach M, Baldwin IT, Erb M.

2013. Leaf-herbivore attack reduces carbon reserves and regrowth from the roots via

jasmonate and auxin signaling. New Phytologist 200: 1234-1246.

181

Marti G, Erb M, Boccard J, Glauser G, Doyen GR, Villard N, Robert CAM, Turlings TCJ,

Rudaz S, Wolfender JL. 2012. Metabolomics reveals herbivore-induced metabolites of

resistance and susceptibility in maize leaves and roots. Plant cell environment 36(3): 621-

623.

Matejovic I. 1996. The application of Dumas method for determination of carbon, nitrogen, and

sulphur in plant samples. Rostlinna Vyroba 42(7): 313-316.

Matern S, Peskan-Berghoefer T, Gromes R, Kiesel RV, Rausch T. 2015. Imposed

glutathione-mediated redox switch modulates the tobacco wound-induced protein kinase

and salicylic acid-induced protein kinase activation state and impacts on defence against

Pseudomonas syringae. Journal of Experimental Botany 66(7): 1935-1950.

McConn M, Creelman RA, Bell E, Mullet JE, Browse J. 1997. Jasmonate is essential for

insect defense Arabidopsis. Proceedings of the National Academy of Sciences of the

United States of America 94(10): 5473–5477.

McDermott JE, Diamond DL, Corley C, Rasmussen AL, Katze MG, Waters KM. 2012.

Topological analysis of protein co-abundance networks identifies novel host targets

important for HCV infection and pathogenesis. Bmc Systems Biology 6.

McGee JD, Hamer JE, Hodges TK. 2001. Characterization of a PR-10 pathogenesis-related

gene family induced in rice during infection with Magnaporthe grisea. Molecular Plant-

Microbe Interactions 14(7): 877-886.

Medeiros HA, Mingossi BF, Dias OR, Franco PF, Vicentini R, Mello OM, Moura SD, Silva-

Filho CM. 2016. Sugarcane serine peptidase inhibitors, serine peptidases, and Clp

protease system subunits associated with sugarcane borer (Diatraea saccharalis)

herbivory and wounding. International Journal of Molecular Sciences 17(9): 1444.

182

Meihls LN, Kaur H, Jander G 2012. Natural variation in maize defense against insect

herbivores. Cold spring harbor symposium in quantitative biology. 269-283.

Meinke LJ, Siegfried BD, Wright RJ, Chandler LD. 1998. Adult susceptibility of Nebraska

western corn rootworm (Coleoptera: Chrysomelidae) populations to selected insecticides.

Journal of Economic Entomology 91(3): 594-600.

Melo Machado RC, Sant'Ana J, Blassioli-Moraes MC, Laumann RA, Borges M. 2014.

Herbivory-induced plant volatiles from Oryza sativa and their influence on chemotaxis

behaviour of Tibraca limbativentris stal. (Hemiptera: Pentatomidae) and egg parasitoids.

Bulletin of Entomological Research 104(3): 347-356.

Metcalf RL, Lampman RL. 1989. Estragole analogs as attractants for corn rootworms

(Coleoptera, Chrysomelidae). Journal of Economic Entomology 82(1): 123-129.

Michener HD. 1938. The action of ethylene on plant growth. American Journal of Botany 25(9):

711-720.

Michiels K, Van Damme EJM, Smagghe G. 2010. Plant-insect interactions: what can we learn

from plant lectins? Archives of Insect Biochemistry and Physiology 73(4): 193-212.

Mierziak J, Kostyn K, Kulma A. 2014. Flavonoids as important molecules of plant interactions

with the environment. Molecules 19(10): 16240-16265.

Misas-Villamil JC, van der Hoorn RAL, Doehlemann G. 2016. Papain-like cysteine proteases

as hubs in plant immunity. New Phytologist 212(4):902-907.

Mohan S, Ma PWK, Pechan T, Bassford ER, Williams WP, Luthe DS. 2006. Degradation of

the S. frugiperda peritrophic matrix by an inducible maize cysteine protease. Journal of

Insect Physiology 52(1): 21-28.

183

Mohan S, Ma PWK, Williams WP, Luthe DS. 2008. A naturally occurring plant cysteine

protease possesses remarkable toxicity against insect pests and synergizes Bacillus

thuringiensis toxin. PLoS ONE 3(3): e1786.

Mou SL, Liu ZQ, Guan DY, Qiu AL, Lai Y, He SL. 2013. Functional analysis and

expressional characterization of rice ankyrin repeat-containing protein, OsPIANK1, in

basal defense against Magnaporthe oryzae attack. Plos One 8(3).

Munford RS, Sheppard PO, Ohara PJ. 1995. Saposin-like proteins (SAPLIP) carry out

diverse funtions on a common backbone structure. Journal of Lipid Research 36(8):

1653-1663.

Musungu B, Bhatnagar D, Brown RL, Fakhoury AM, Geisler M. 2015. A predicted protein

interactome identifies conserved global networks and disease resistance subnetworks in

maize. Frontiers in Genetics 6: 201-Article No.: 201.

Niki T, Mitsuhara I, Seo S, Ohtsubo N, Ohashi Y. 1998. Antagonistic effect of salicylic acid

and jasmonic acid on the expression of pathogenesis-related (PR) protein genes in

wounded mature tobacco leaves. Plant and Cell Physiology 39(5): 500-507.

O'Donnell PJ, Calvert C, Atzorn R, Wasternack C, Leyser HMO, Bowles DJ. 1996.

Ethylene as a signal mediating the wound response of tomato plants. Science 274(5294):

1914-1917.

Oleson JD, Park YL, Nowatzki TM, Tollefson JJ. 2005. Node-injury scale to evaluate root

injury by corn rootworms (Coleoptera : Chrysomelidae). Journal of Economic

Entomology 98(1): 1-8.

Olivas NHD, Coolen S, Huang PP, Severing E, van Verk MC, Hickman R, Wittenberg

AHJ, de Vos M, Prins M, van Loon JJA, et al. 2016. Effect of prior drought and

184

pathogen stress on Arabidopsis transcriptome changes to caterpillar herbivory. New

Phytologist 210(4): 1344-1356.

Oltvai ZN, Barabasi AL. 2002. Life's complexity pyramid. Science 298(5594): 763-764.

Paez-Garcia A, Motes CM, Scheible W-R, Chen R, Blancaflor EB, Monteros MJ. 2015.

Root traits and phenotyping strategies for plant improvement. Plants-Basel 4(2): 334-

355.

Painter RH. 1951. Insect Resistance in Crop Plants: Macmillan.

Painter RH. 1958. Resistance of Plants to Insects. Annual Review of Entomology 3(1): 267-290.

Pare PW, Tumlinson JH. 1999. Plant volatiles as a defense against insect herbivores. Plant

Physiology 121(2): 325-331.

Paudel J, Copley T, Amirizian A, Prado A, Bede JC. 2013. Arabidopsis redox status in

response to caterpillar herbivory. Frontiers in Plant Science 4: 113.

Pechan T, Ye LJ, Chang YM, Mitra A, Lin L, Davis FM, Williams WP, Luthe DS. 2000. A

unique 33-kD cysteine proteinase accumulates in response to larval feeding in maize

genotypes resistant to fall armyworm and other lepidoptera. Plant cell 12(7): 1031–1041.

Perkins LE, Cribb BW, Brewer PB, Hanan J, Grant M, de Torres M, Zalucki MP. 2013.

Generalist insects behave in a jasmonate-dependent manner on their host plants, leaving

induced areas quickly and staying longer on distant parts. Proceedings of the Royal

Society B-Biological Sciences 280(1756): e20122646.

Pogány M, Dankó T, Kámán-Tóth E, Schwarczinger I, Bozsó Z. 2015. Regulatory

proteolysis in Arabidopsis-pathogen interactions. International Journal of Molecular

Sciences 16(10): 23177-23194.

185

Prischmann DA, Dashiell KE, Schneider DJ, Hibbard BE. 2007. Field screening maize

germplasm for resistance and tolerance to western corn rootworms (Col.:

Chrysomelidae)†. Journal of Applied Entomology 131(6): 406-415.

Provart N. 2012. Correlation networks visualization. Frontiers in Plant Science 3: 240.

Qi J, Sun G, Wang L, Zhao C, Hettenhausen C, Schuman MC, Baldwin IT, Li J, Song J,

Liu Z, et al. 2016. Oral secretions from Mythimna separata insects specifically induce

defence responses in maize as revealed by high-dimensional biological data. Plant, Cell

& Environment 39(8): 1749-1766.

Qu W, Robert CAM, Erb M, Hibbard BE, Paven M, Gleede T, Riehl B, Kersting L,

Cankaya AS, Kunert AT, et al. 2016. Dynamic Precision Phenotyping Reveals

Mechanism of Crop Tolerance to Root Herbivory. Plant Physiology.

Rasmann S, Agrawal AA. 2008. In defense of roots: A research agenda for studying plant

resistance to belowground herbivory. Plant Physiology 146(3): 875-880.

Rasmann S, Kollner TG, Degenhardt J, Hiltpold I, Toepfer S, Kuhlmann U, Gershenzon J,

Turlings TCJ. 2005. Recruitment of entomopathogenic nematodes by insect-damaged

maize roots. Nature 434(7034): 732-737.

Rasmann S, Turlings TCJ. 2007. Simultaneous feeding by aboveground and belowground

herbivores attenuates plant-mediated attraction of their respective natural enemies.

Ecology Letters 10(10): 926-936.

Richter A, Seidl-Adams I, Koellner TG, Schaff C, Tumlinson JH, Degenhardt J. 2015. A

small, differentially regulated family of farnesyl diphosphate synthases in maize (Zea

mays) provides farnesyl diphosphate for the biosynthesis of herbivore-induced

sesquiterpenes. Planta 241(6): 1351-1361.

186

Riedell WE, Reese RN. 1999. Maize morphology and shoot CO2 assimilation after root damage

by western corn rootworm larvae. Crop Science 39(5): 1332-1340.

Ritchie JT, Singh U, Godwin DC 1998. Cereal growth, development and yield. In: Tsuji GY,

Hoogenboom G, Thornton PK eds. Understanding Options for Agricultural Production:

Kluwer academic publishers, 78-98.

Robert CAM, Erb M, Duployer M, Zwahlen C, Doyen GR, Turlings TCJ. 2012. Herbivore-

induced plant volatiles mediate host selection by a root herbivore. New Phytologist

194(4): 1061-1069.

Robert CAM, Erb M, Hibbard BE, French BW, Zwahlen C, Turlings TCJ. 2012. A

specialist root herbivore reduces plant resistance and uses an induced plant volatile to

aggregate in a density-dependent manner. Functional Ecology 26(6): 1429-1440.

Robert CAM, Ferrieri RA, Schirmer S, Babst BA, Schueller MJ, Machado RAR, Arce

CCM, Hibbard BE, Gershenzon J, Turlings TCJ, et al. 2014. Induced carbon

reallocation and compensatory growth as root herbivore tolerance mechanisms. Plant,

Cell & Environment 37(11): 2613-2622.

Robert CAM, Schirmer S, Barry J, Wade French B, Hibbard BE, Gershenzon J. 2015.

Belowground herbivore tolerance involves delayed overcompensatory root regrowth in

maize. Entomologia Experimentalis et Applicata 157(1): 113-120.

Rohrmeier T, Lehle L. 1993. WIP1, a wound-inducible gene from maize with homology to

Bowman-Birk proteinase inhibitors. Plant molecular biology 22(5): 783-792.

Rost TL, Bryant JA. 1996. Root organization and gene expression patterns. Journal of

Experimental Botany 47(304): 1613-1628.

187

Saathoff AJ, Donze T, Palmer NA, Bradshaw J, Heng-Moss T, Twigg P, Tobias CM,

Lagrimini M, Sarath G. 2013. Towards uncovering the roles of switchgrass peroxidases

in plant processes. Frontiers in Plant Science 4: 202.

Sallaud C, Rontein D, Onillon S, Jabes F, Duffe P, Giacalone C, Thoraval S, Escoffier C,

Herbette G, Leonhardt N, et al. 2009. A Novel pathway for sesquiterpene biosynthesis

from Z,Z-farnesyl pyrophosphate in the wild tomato Solanum habrochaites. Plant Cell

21(1): 301-317.

Salwinski L, Miller CS, Smith AJ, Pettit FK, Bowie JU, Eisenberg D. 2004. The Database of

Interacting Proteins: 2004 update. Nucleic Acids Research 32(Database issue): D449-

D451.

Sappington TW, Siegfried BD, Guillemaud T. 2006. Coordinated diabrotica genetics research:

accelerating progress on an urgent insect pest problem. American Entomologist 52(2): 90-

97.

Sarazin V, Duclercq J, Mendou B, Aubanelle L, Nicolas V, Aono M, Pilard S, Guerineau F,

Sangwan-Norreel B, Sangwan RS. 2015. Arabidopsis BNT1, an atypical TIR-NBS-

LRR gene, acting as a regulator of the hormonal response to stress. Plant Science 239:

216-229.

Schaefer HM, Rolshausen G. 2006. Plants on red alert: do insects pay attention? BioEssays

28(1): 65-71.

Schellingen K, Van der Straeten D, Remans T, Loix C, Vangronsveld J, Cuypers A. 2015.

Ethylene biosynthesis is involved in the early oxidative challenge induced by moderate

Cd exposure in Arabidopsis thaliana. Environmental and Experimental Botany 117: 1-11.

188

Schilmiller AL, Howe GA. 2005. Systemic signaling in the wound response. Current Opinion

in Plant Biology 8(4): 369-377.

Schmelz EA, Alborn HT, Tumlinson JH. 2003a. Synergistic interactions between volicitin,

jasmonic acid and ethylene mediate insect-induced volatile emission in Zea mays.

Physiologia Plantarum 117(3): 403-412.

Schmelz EA, Engelberth J, Alborn HT, O'Donnell P, Sammons M, Toshima H, Tumlinson

JH. 2003b. Simultaneous analysis of phytohormones, phytotoxins, and volatile organic

compounds in plants. Proceedings of the National Academy of Sciences of the United

States of America 100(18): 10552-10557.

Schmelz EA, Engelberth J, Tumlinson JH, Block A, Alborn HT. 2004. The use of vapor

phase extraction in metabolic profiling of phytohormones and other metabolites. Plant

Journal 39(5): 790-808.

Schneider CA, Rasband WS, Eliceiri KW. 2012. NIH Image to ImageJ: 25 years of image

analysis. Nature Methods 9(7): 671-675.

Scholz SS, Heyer M, Vadassery J, Mithöfer A. 2016. A role for calmodulin-like proteins in

herbivore defense pathways in plants. Journal of Endocytobiosis and Cell Research

27(1): 1-12.

Schweiger R, Heise AM, Persicke M, MÜLler C. 2014. Interactions between the jasmonic and

salicylic acid pathway modulate the plant metabolome and affect herbivores of different

feeding types. Plant, Cell & Environment 37(7): 1574-1585.

Scott-Craig JS, Casida JE, Poduje L, Walton JD. 1998. Herbicide safener-binding protein of

maize - Purification, cloning, and expression of an encoding cDNA. Plant Physiology

116(3): 1083-1089.

189

Sekara V, Stopczynski A, Lehmann S. 2016. Fundamental structures of dynamic social

networks. Proceedings of the National Academy of Sciences. 113(36): 9977-9982.

Shapiro SS, Wilk MB. 1965. An analysis of variance test for normality (Complete samples).

Biometrika(52): 591-611.

Shivaji R, Camas A, Ankala A, Engelberth J, Tumlinson JH, Williams WP, Wilkinson JR,

Luthe DS. 2010. Plants on constant alert: elevated levels of jasmonic acid and

jasmonate-induced transcripts in caterpillar-resistant maize. Journal of Chemical Ecology

36(2): 179-191.

Silva CP, Silva JR, Vasconcelos FF, Petretski MlDA, DaMatta RA, Ribeiro AF, Terra WR.

2004. Occurrence of midgut perimicrovillar membranes in paraneopteran insect orders

with comments on their function and evolutionary significance. Arthropod Structure &

Development 33(2): 139-148.

Smith WEC, Shivaji R, Williams WP, Luthe DS, Sandoya GV, Smith CL, Sparks DL,

Brown AE. 2012. A maize line resistant to herbivory constitutively releases (E)-beta-

caryophyllene. Journal of Economic Entomology 105(1): 120-128.

Soler R, Erb M, Kaplan I. 2013. Long distance root-shoot signalling in plant-insect community

interactions. Trends in Plant Science 18(3): 149-156.

Somerville C, Browse J, Jaworski JG, Ohlrogge J 2000. Lipids: chapter 10. In: BB Buchanan

WG, RL Jones ed. Biochemistry and Molecular Biology of Plants. Rockville, MD:

American Society of Plant Physiologists, 456–527.

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.

190

Stark C, Breitkreutz B-J, Reguly T, Boucher L, Breitkreutz A, Tyers M. 2006. BioGRID: a

general repository for interaction datasets. Nucleic Acids Research 34(Database issue):

D535-D539.

Studham ME, Macintosh G. 2012. Phytohormone signaling pathway analysis method for

comparing hormone responses in plant-pest interactions. BMC Research Notes 5: 392.

Supek F, Bosnjak M, Skunca N, Smuc T. 2011. REVIGO summarizes and visualizes long lists

of gene ontology terms. Plos One 6(7).

Suzuki H, Dowd PF, Johnson ET, Hum-Musser SM, Musser RO. 2012. Effects of elevated

peroxidase levels and corn earworm feeding on gene expression in tomato. Journal of

Chemical Ecology 38(10): 1247-1263.

Sytykiewicz H. 2016. Expression patterns of genes involved in ascorbate-glutathione cycle in

aphid-infested Maize (Zea mays L.) seedlings. International Journal of Molecular

Sciences 17(3): 268.

Takahashi H, Kanayama Y, Zheng MS, Kusano T, Hase S, Ikegami M, Shah J. 2004.

Antagonistic interactions between the SA and JA signaling pathways in Arabidopsis

modulate expression of defense genes and gene-for-gene resistance to cucumber mosaic

virus. Plant and Cell Physiology 45(6): 803-809.

Tamayo MC, Rufat M, Bravo JM, San Segundo B. 2000. Accumulation of a maize proteinase

inhibitor in response to wounding and insect feeding, and characterization of its activity

toward digestive proteinases of Spodoptera littoralis larvae. Planta 211(1): 62-71.

Team RC 2015. R: a lenguage and environment for statistical computing. Vienna, Austria: R

Foundation for Statistical Computing.

191

Thaler JS, Fidantsef AL, Bostock RM. 2002. Antagonism between jasmonate- and salicylate-

mediated induced plant resistance: effects of concentration and timing of elicitation on

defense-related proteins, herbivore, and pathogen performance in tomato. Journal of

Chemical Ecology 28(6): 1131-1159.

Thines B, Katsir L, Melotto M, Niu Y, Mandaokar A, Liu G, Nomura K, He SY, Howe GA,

Browse J. 2007. JAZ repressor proteins are targets of the SCFCO11 complex during

jasmonate signalling. Nature 448(7154): 661-665.

Tian DL, Peiffer M, De Moraes CM, Felton GW. 2014. Roles of ethylene and jasmonic acid

in systemic induced defense in tomato (Solanum lycopersicum) against Helicoverpa zea.

Planta 239(3): 577-589.

Tinsley NA, Estes RE, Gray ME. 2013. Validation of a nested error component model to

estimate damage caused by corn rootworm larvae. Journal of Applied Entomology

137(3): 161-169.

Tinsley NA, Mitchell PD, Wright RJ, Meinke LJ, Estes RE, Gray ME. 2016. Estimation of

efficacy functions for products used to manage corn rootworm larval injury. Journal of

Applied Entomology 140(6): 414-425. van Dam NM. 2009. Belowground Herbivory and Plant Defenses. Annual Review of Ecology

Evolution and Systematics 40: 373-391.

Van Dam NM, Bezemer TM. 2006. Chemical communication between roots and shoots -

Towards an integration of aboveground and belowground induced responses in plants.

Chemical ecology 16: 127-143.

Vandenborre G, Smagghe G, Van Damme EJM. 2011. Plant lectins as defense proteins

against phytophagous insects. Phytochemistry 72(13): 1538-1550.

192

Vila L, Quilis J, Meynard D, Breitler JC, Marfa V, Murillo I, Vassal JM, Messeguer J,

Guiderdoni E, San Segundo B. 2005. Expression of the maize proteinase inhibitor (mpi)

gene in rice plants enhances resistance against the striped stem borer (Chilo

suppressalis): effects on larval growth and insect gut proteinases. Plant Biotechnology

Journal 3(2): 187-202.

Walley JW, Sartor RC, Shen Z, Schmitz RJ, Wu KJ, Urich MA, Nery JR, Smith LG,

Schnable JC, Ecker JR, et al. 2016. Integration of omic networks in a developmental

atlas of maize. Science 353(6301): 814.

Walter A, Mazars C, Maitrejean M, Hopke J, Ranjeva R, Boland W, Mithöfer A. 2007.

Structural requirements of jasmonates and synthetic analogues as inducers of Ca2+

signals in the nucleus and the cytosol of plant cells. Angewandte Chemie International

Edition 46(25): 4783-4785.

Wang XQ, Ullah H, Jones AM, Assmann SM. 2001. G protein regulation of ion channels and

abscisic acid signaling in Arabidopsis guard cells. Science 292(5524): 2070-2072.

War AR, Paulraj MG, Ahmad T, Buhroo AA, Hussain B, Ignacimuthu S, Sharma HC.

2012. Mechanisms of plant defense against insect herbivores. Plant signaling & behavior

7(10): 1306-1320.

Wasternack C, Hause B. 2013. Jasmonates: biosynthesis, perception, signal transduction and

action in plant stress response, growth and development. An update to the 2007 review in

Annals of Botany. Annals of Botany 111(6): 1021-1058.

Weckwerth W. 2011. Green systems biology - From single genomes, proteomes and

metabolomes to ecosystems research and biotechnology. Journal of Proteomics 75(1):

284-305.

193

Whitworth RJ, Wilde GE, Shufran RA, Milliken GA. 2002. Comparison of adult corn

rootworm (Coleoptera : Chrysomelidae) sampling methods. Journal of Economic

Entomology 95(1): 96-105.

Williams WP, Buckley PM, Davis FM. 1985. Larval growth and behaviour of the fall

armyworm (Lepidoptera, Noctuidae) on callus initiated from suceptible and resistant corn

hybrids. Journal of Economic Entomology 78(4): 951-954.

Williams WP, Buckley PM, Davis FM. 1987. Feeding response of corn-earworm (Lepidoptera,

Noctuidae) to callus and extracts of corn in the laboratory. Environmental Entomology

16(2): 532-534.

Williams WP, Davis FM, Windham GL. 1990. Registrationof Mp708 germplasm line of

maize. Crop science 30(3): 757.

Yadav PS, Devprakash, Senthilkumar GP. 2011. Benzothiazole: different methods of

synthesis and diverse biological activities. International Journal of Pharmaceutical

Sciences and Drug Research 3(1): 1-7.

Yan Y, Christensen S, Isakeit T, Engelberth J, Meeley R, Hayward A, Emery RJN,

Kolomiets MV. 2012. Disruption of OPR7 and OPR8 reveals the versatile functions of

jasmonic acid in maize development and defense. Plant Cell 24(4): 1420-1436.

Yang FS, Zhang YL, Huang QX, Yin GH, Pennerman KK, Yu JJ, Liu ZX, Li DF, Guo AP.

2015. Analysis of key genes of jasmonic acid mediated signal pathway for defense

against insect damages by comparative transcriptome sequencing. Scientific Reports 5.

Yang FS, Zhang YL, Huang QX, Yin GH, Pennerman KK, Yu JJ, Liu ZX, Li DF, Guo AP.

2015. Analysis of key genes of jasmonic acid mediated signal pathway for defense

194

against insect damages by comparative transcriptome sequencing. Scientific Reports 5:

16500.

Yoshida S. 1979a. Freezing-injury and phospholipid degradation in vivo in woody plant cells. 1.

Subcellular localization of phospholipase D in living bark tisues of the black locust tree

(Robina pseudoacaia L.). Plant Physiology 64(2): 241-246.

Yoshida S. 1979b. Freezing-injury and phospholipid degradation in vivo in woody plant cells. 2.

Regulatory effects of divalent cations on activiy of membrane bound phospholipase D.

Plant Physiology 64(2): 247-251.

Yoshida S. 1979c. Freezing-injury and phospholipid degradation in vivo in woody plant cells. 3.

Effets of freezing on activity of membrane bound phospholipase D in microsome

enriched membranes. Plant Physiology 64(2): 252-256.

Yu HY, Kim PM, Sprecher E, Trifonov V, Gerstein M. 2007. The importance of bottlenecks

in protein networks: Correlation with gene essentiality and expression dynamics. Plos

Computational Biology 3(4): 713-720.

Zechmeister-Boltenstern S, Nikodim L. 1999. Effect of water tension on ethylene production

and consumption in montane and lowland soils in Austria. European Journal of Soil

Science 50(3): 425-432.

Zhang SJ, Yang WZ, Zhao QQ, Zhou XJ, Jiang L, Ma S, Liu XQ, Li Y, Zhang CY, Fan

YL, et al. 2016. Analysis of weighted co-regulatory networks in maize provides insights

into new genes and regulatory mechanisms related to inositol phosphate metabolism. Bmc

Genomics 17: 129.

Zhang YX, Gao P, Xing Z, Jin SM, Chen ZD, Liu LT, Constantino N, Wang XW, Shi WB,

Yuan JS, et al. 2013. Application of an improved proteomics method for abundant

195

protein cleanup: molecular and genomic mechanisms study in plant defense. Molecular &

Cellular Proteomics 12(11): 3431-3442.

Zhao Y, Xu C, Wang Q, Wei Y, Liu F, Xu S, Zhang Z, Mu W. 2016. Effects of the microbial

secondary metabolite benzothiazole on the nutritional physiology and enzyme activities

of Bradysia odoriphaga (Diptera: Sciaridae). Pesticide Biochemistry and Physiology 129:

49-55.

Zhou S, Lou Y-R, Tzin V, Jander G. 2015. Alteration of plant primary metabolism in response

to insect herbivory. Plant Physiology 169(3): 1488-1498.

Zhu CH, Gan LJ, Shen ZG, Xia K. 2006. Interactions between jasmonates and ethylene in the

regulation of root hair development in Arabidopsis. Journal of Experimental Botany

57(6): 1299-1308.

Zhu L. 2010. Study of Defense Genes Expression Induced by Leaf Herbivory in Roots of Insect-

resistant Maize. Masters in Agronomy, The Pennsylvania State University The

Pennsylvania State University.

Zhu Z, An F, Feng Y, Li P, Xue L, Mu A, Jiang Z, Kim J-M, To TK, Li W, et al. 2011.

Derepression of ethylene-stabilized transcription factors (EIN3/EIL1) mediates jasmonate

and ethylene signaling synergy in Arabidopsis. Proceedings of the National Academy of

Sciences of the United States of America 108(30): 12539-12544.

Zhu-Salzman K, Luthe DS, Felton GW. 2008. Arthropod-inducible proteins: Broad spectrum

defenses against multiple herbivores. Plant Physiology 146(3): 852-858.

Zou C-S, Mo M-H, Gu Y-Q, Zhou J-P, Zhang K-Q. 2007. Possible contributions of volatile-

producing bacteria to soil fungistasis. Soil Biology and Biochemistry 39(9): 2371-2379.

196

VITA: Lina M. Castano-Duque

PROFESSIONAL PREPARATION University of Nevada, Reno Biochemistry and Molecular Biology B.A. 2010 Pennsylvania State Plant Biology Ph.D 2017 University

APPOINTMENTS NSF GRFP Fellow, 2011-2014 Bunton-Waller Fellow, 2010 McNair Scholar, 2009-2010 Research assistant, Nevada Agricultural Experimental Station, University of Nevada, 2008-2009

PUBLICATIONS -Lina Castano-Duque, and Dawn S. Luthe. Protein Networks Reveal Tissue Specific Defense Strategies in Maize in response to an aboveground herbivore. -Lina Castano-Duque, Anjel Helms, Jared Ali and Dawn S. Luthe. Plant Bio-Wars: Maize Protein Networks Reveal Tissue-Specific Defense Strategies in Response to a Root Herbivore. -Lina Castano-Duque, Kenneth Loades, John F. Tooker, Kathleen Brown, W. Paul Williams and Dawn S. Luthe. A maize inbred exhibits resistance traits against western corn rootworm, Diabrotica virgifera virgifera. -Louis J, Basu S, Varsani S, Castano-Duque L, Jiang V, Williams WP, Felton GW, Luthe DS, 2015, Ethylene Contributes to maize insect resistance1-Mediated Maize Defense against the Phloem Sap-Sucking Corn Leaf Aphid. Plant physiology, 169(1):313-324 -Chuang, W.-P, M Herde, S. Ray, L. Castano-Duque, G.A. Howe, and D.S. Luthe, 2014, Caterpillar attack triggers accumulation of the toxic maize protein RIP2: New Phytologist, 201: 928-939. -Fong, C., F. Diaz, L. Jurado, K. Castillo, F. Gonzalez, L. Castano Duque, J. Osorio, H. Cardenas, 2008, Effect Of Eggs Population Density On Viability And Time Of Development Of Drosophila melanogaster(Drosophilidae): Colombia Acta Biologica Colombiana, 13:123-132

EMPLOYEMNT HISTORY -Graduate student researcher: Advisor: Dr. Dawn Luthe Pennsylvania State University Research experience, From September 2011 to present -Undergraduate student researcher: Dr. John C. Cushman University of Nevada, Reno Senior thesis project, December 2008 to May 2010

VOLUNTEER EXPERIENCE -Co-Chair for Graduate Woman in Science Voices annual conference (2014 and 2015) -Mentor in PlantingScience Master Plant Science Team (MPST) program from ASPB (October 2012- on going).

197