University of Florida Thesis Or Dissertation Formatting

INVESTIGATING PLANT-PHYTOPHTHORA INTERACTIONS USING TRANSCRIPTOMICS AND DETERMINING THE SIGNIFICANCE OF Ca2+ SIGNALING IN HOST DEFENSE

ZUNAIRA AFZAL

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2019

To my beloved son, Muhammad and daughter, Hannah

ACKNOWLEDGMENTS

I would like to take this opportunity to express my sincere gratitude towards my advisor, Dr. Gul Shad Ali, for the guidance, the friendship and support during all these years of my doctoral studies. Things I learned from him will surely guide me throughout my professional and personal life. I also would like to give a big thanks to Dr. Rosemary

Loria, who supported me and talked me through the hard times. She is surely the one who has the best intentions to help students under any circumstances. I would like to thank Dr. Roger Kjelgren, who always helped and managed the resources to support me throughout my degree. I would like to extend my appreciation to my committee members Dr. Jeff Rollins, Dr. Karen Garrett and Dr. Chen Jianjun, for their support, valid suggestions, specially Dr. Chen for his kind advice and career counseling. I would also like to thank, Jessica Ulloa, Pamela Hicks and all the Department of Plant

Pathology and MREC staff for their great support throughout my degree. I also want to mention Mary Brennan who, although got retired shortly after I started working in the laboratory, but she helped and taught me a lot during that period. I will always remember her kindness and compassion. I want to thank Matthew Creech for always helping me when I didn’t know where to go and what to do. I also like to acknowledge

Dr. Dave Norman and his lab members specifically Dr. Ana Bocsanczy and Jeanne

Yuen for lending me the lab supplies and letting me use equipment always with a gentle smile. I highly oblige Rubina Ali for her help, support, friendship and all the memories we created together.

My deepest appreciation goes to my husband Naveed Asghar for his support, help, patience and love. I want to acknowledge my kids Muhammad Dawood and

Hannah Naveed who bared my absence and attention even when I was physically

present with them. I would like to thank my parents and brothers, who even being far away, always made a way to be present and show their support and love, mainly my mother who always motivates me to keep going and fulfil my dreams. I thank Irum Iqrar,

Stephanie Suarez and Khansa Jamil for their friendship for life that started here during my Ph.D. I also want to thank Dr. Brantlee Spakes- Richter for developing and polishing my teaching skills.

I would like to thank the Institute of International Education (IIE), United States

Education Foundation in Pakistan (USEFP) and Department of Plant Pathology for sponsoring my stay and studies in the United States. Finally, my deepest appreciation is towards Senator J. William Fulbright (Late) who started Fulbright scholarship program that is surely an amazing life changing experience for people like me.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 8

LIST OF FIGURES ...... 9

ABSTRACT ...... 11

CHAPTER

1 LITERATURE REVIEW ...... 13

Important Phytophthora Species ...... 14 Plant Defense ...... 17 Phytophthora-Plant Interaction ...... 22 Recognition of Phytophthora spp. by Plants and Induction of PTI ...... 23 Supression of PTI by Phytophthora Effectors ...... 28 Phytophthora Effectors Targeting Host Susceptibility Factors ...... 30 Perspective on Developing Phytophthora Resitant Plants ...... 32 Research Objectives ...... 33

2 COMPARATIVE TRANSCRIPTOME ANALYSIS BETWEEN A RESISTANT AND A SUSCEPTIBLE WILD TOMATO ACCESSION IN RESPONSE TO PHYTOPHTHORA PARASITICA ...... 35

Materials and Methods...... 37 P. parasitica Inoculation, RNA Extraction and Sequencing ...... 37 RNA-Seq Analysis ...... 38 Validation of RNA-seq Results ...... 39 Results ...... 40 P. parasitica Infection on S. pimpinellifolium Accessions ...... 40 Analysis of RNA-seq Data ...... 41 Differential Expression Analysis ...... 42 Analysis GO Enrichment Analysis of DEGs ...... 44 Parametric Gene Set Enrichment Analysis Based on GO ...... 45 KEGG Annotations ...... 48 Visualization of DETs in Plant Pathways ...... 49 Validation of RNA-seq Results with qRT-PCR ...... 51 Discussion ...... 51 Antimicrobial Genes in Relevance to Resistance Response Against P. parasitica ...... 52 Contrasting Expression of Downy Mildew Susceptibility Factor in Response to P. parasitica ...... 54 Induction of Plant Heat Stress Tolerance Related Genes ...... 55

R Genes ...... 56

3 TRANSCRIPTOME PROFILE OF CARRIZO CITRANGE ROOTS IN RESPONSE TO PHYTOPHTHORA PARASITICA INFECTION ...... 76

Materials and Methods...... 79 P. parasitica Inoculation, Sample Collection and Sequencing ...... 79 RNA-Seq Quality Assessment ...... 79 Reference Genome Mapping ...... 79 De novo Assembly and Differential Expression Analysis ...... 80 Functional Annotation of Transcripts ...... 81 Validation of DETs by Quantitative Real Time PCR ...... 82 Discussion ...... 92

4 THE PHYTOPHTHORA RXLR EFFECTOR AVRblb2 MODULATES PLANT IMMUNITY BY INTERFERING WITH Ca2+ SIGNALING PATHWAY ...... 134

Materials and Methods...... 139 Plasmid Construction and Transient Expression ...... 139 Protein Extraction and Western Blotting ...... 140 In planta Co-immunoprecipitation ...... 141 Hypersensitive Response Analysis in N. benthamiana Plants ...... 142 Results ...... 143 Calmodulin-binding to Avrblb2 is not Required for Localization of Avrblb2 .... 143 Calmodulin-binding to AVRblb2 is not Required for AVRblb2 binding with C14 ...... 143 Calmodulin-binding to AVRblb2 is Required for Rpi-blb2-Dependent Cell Death ...... 144 Discussion ...... 145

5 SUMMARY AND CONCLUSIONS ...... 158

LIST OF REFERENCES ...... 163

BIOGRAPHICAL SKETCH ...... 172

LIST OF TABLES

Table page

2-1 qRT-PCR primers for selected genes...... 58

2-2 Summary statistics of RNA-seq data and mapping results...... 59

2-3 Number of DEGs found among all comparisons...... 59

2-4 Comparison of GO term enrichment analysis...... 60

3-1 List of genomes that were used to make custom BLAST database...... 102

3-2 The statistics of Trinity generated de novo assembly...... 103

3-3 qRT-PCR primers for selected genes...... 104

3-4 Mapping coverage of RNA-Seq samples to Citrus and P. parasitica genomes...... 104

3-5 Top KEGG Pathways for DETs...... 105

3-6 R gene DETs induced upon P. parasitica infection in citrus roots...... 106

LIST OF FIGURES

Figure page

2-1 Disease Symptoms and HR response...... 61

2-2 Venn diagrams presenting the overlap of expressed genes...... 62

2-3 Principle component analysis (PCA) plot...... 63

2-4 Venn diagram analysis of DEGs...... 64

2-5 Heat map showing hierarchical cluster analysis ...... 65

2-6 Overview of GO annotations assignment of DEGs ...... 66

2-7 Gene set enrichment analysis by PAGE tool of AgriGO...... 67

2-8 Heatmap showing expression profiles of core defense genes...... 69

2-9 Visualization of DEGs involved in MAPK signaling pathway...... 70

2-10 DEGs mapped on plant hormone signal transduction pathway...... 71

2-11 Visualization of differentially expressed genes in biotic stress...... 73

2-12 Heatmap showing expression profiles of DEGs encoding DUF26 RLKs ...... 74

2-13 qRT-PCR based validation of plant defense related DEGs ...... 75

3-1 Plot showing ExN50 statistics...... 115

3-2 Pipeline of RNA-Seq data analysis...... 116

3-3 Infection dynamics of Phytophthora parasitica in citrus roots ...... 117

3-4 Sample correlation and Principle component analysis (PCA)...... 118

3-5 Volcano plots showing pairwise comparisons ...... 119

3-6 Hierarchical cluster analysis of differential expressed transcripts ...... 120

3-7 Venn diagrams shown number of citrus DETs...... 121

3-8 Functional annotation statistics of citrus DETs...... 122

3-9 Top hit species distribution graph...... 123

3-10 Distribution of GO categories assigned to DETs ...... 124

3-11 Detailed GO term annotations of DETs ...... 125

3-12 Interactive graphs for summary visualization of most abundant GOs ...... 126

3-13 Visualization of DETs involved in MAPK signaling pathway ...... 127

3-14 Mapping of DETs on Plant-pathogen interaction pathway...... 128

3-15 DETs mapped on Plant hormone signaling pathway...... 129

3-16 KEGG enzyme enrichment analysis ...... 130

3-17 DETs mapped on Flavanoid biosynthesis pathway...... 131

3-18 Visualization of DETs involved in biotic stress pathway...... 132

3-19 qRT-PCR based validation of top plant defense related DETs ...... 133

4-1 The Phytophthora Avrblb2 effectors interact with calmodulin...... 149

4-2 The P. infestans and P. parasitica Avrblb2 interact with calmodulin in a Ca2+– dependent manner...... 150

4-3 Most of the divergent Avrblb2 homologs interact with calmodulin in a Ca2+- dependent manner...... 151

4-4 Calmodulin-binding site mapping of the P. infestans Avrblb2...... 152

4-5 Identification of amino acids in Avrblb2, which are critical for interaction to calmodulin...... 153

4-6 Protein expression confirmation of Avrblb2 mutants and C14 cloned in plant expression vectors...... 154

4-7 Visualization of YFP tagged WT and NCB in Nb/WT and Nb/Rpi leaves...... 155

4-8 The Avrblb2/Rpi-blb2-induced cell death is dependent on calmodulin binding to Avrblb2...... 156

4-9 In planta co-immunoprecipitation of Avrblb2 mutants with C14...... 157

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

INVESTIGATING PLANT-PHYTOPHTHORA INTERACTIONS USING TRANSCRIPTOMICS AND DETERMINING THE SIGNIFICANCE OF Ca2+ SIGNALING IN HOST DEFENSE

Zunaira Afzal

December 2019

Chair: Gul Shad Ali Major: Plant Pathology

Oomycetes in the genus Phytophthora are responsible for causing destructive diseases in several economically important crops. Being unique among other phytopathogens, control strategies against Phytophthora diseases are very limited.

Understanding Plant-Phytophthora interactions and plant genetic resistance is very important to manage Phytophthora diseases.

Genetic resistance among wild and cultivated plants against Phytophthora spp. are the best resources that can be utilized to develop Phytophthora resistant plants.

Transcriptome analysis of a host in response to a pathogen is a powerful technique to narrow down the list of potential genes involved in resistance development.

Transcriptomes of a Phytophthora parasitica resistant citrus rootstock and two wild tomato accessions (resistant and susceptible) were analyzed. Substantial transcriptional reprogramming of biotic stress components like high induction of oxidoreductase activity, modulation of phytohormone signaling, transcription factors, proteolysis and calcium signaling indicated activation of multiple defense mechanisms in both plants leading to resistant outcome against P. parasitica. Core plant defense genes like

protease inhibitors, chitinases, defensin, PR-1, a downy mildew susceptibility factor etc. were highly induced whereas, WRKY transcriptions factors, most of the R genes including a powdery mildew susceptibility gene (Mlo) were highly repressed in tomato exclusively during the resistance outcome. In citrus, unlike tomato, significant induction of most of the R genes including Mlo was observed. Induction of several G-type lectin

S-receptor-like serine threonine- kinases (an R genes class) is being reported for the first time during Plant-Phytophthora interaction. Functional significance of defense related genes identified here requires further investigation.

Spatio-temporal oscillations in cellular Ca2+concentration is a key early event in plant response to pathogens including Phytophthora spp. Interaction of Avrblb2, a core Phytophthora effector, with host proteins pertaining to Ca2+ homeostasis was found. A key discovery was that Avrblb2 interacts with calmodulin (CaM), a primary calcium sensor only in the presence of Ca2+whereas the absence of Ca2+leads to

Avrblb2 interaction with the C14 protease to promote infection. In planta functional analyses revealed that Avrblb2-CaM binding is required for its recognition by Rpi-blb2 to incite a hypersensitive response. In light of these findings we suggest that

Ca2+ signaling plays a critical role in determining virulence or avirulence activity of

Avrblb2 in host cells.

CHAPTER 1 LITERATURE REVIEW

Oomycetes are a unique class of eukaryotes classified in the kingdom Protoctista along with algae, diatoms and other planktons under the sub kingdom Stramenopila

(Kamoun et al., 2015). Morphologically they are quite similar to filamentous fungi but physiologically, genetically and biochemically they are very different from fungi. For example, unlike true fungi (Eumycota) oomycetes have aseptate multinucleated hyphae and are diploid instead of haploid like most Eumycota fungi. Oomycete and fungi also greatly differ in their cell wall composition. Fungal cell walls contain chitin as the primary carbohydrate polymer whereas oomycete have cellulose and beta glucans as major structural carbohydrate of their cell walls. Unlike fungi, most of the oomycetes lack sterol biosynthesis (Latijnhouwers et al., 2003).

Phytopathogenic oomycetes are associated with destructive diseases in many economically important crops. They are mainly distributed in the genus Phytophthora and Pythium whereas obligate biotrophic oomycete pathogens (generally categorized as downy mildews) are scattered among various genera. The genus Phytophthora is comprised of 160 known species (Orlikowski et al., 2010). Phytophthora pathogens are accountable for enormous crop losses worldwide and have also been considered as an emerging threat to forest ecosystems (Hansen, 2015).

Control strategies for Phytophthora diseases are very limited (Attard et al., 2014).

Despite their differences, fungi and oomycetes use quite similar strategies to infect and colonize their hosts, but most of the chemicals that effectively control fungal diseases are ineffective in controlling oomycete diseases. This lack of efficacy is due in part to the fact that many of the fungicides are formulated to target sterol synthesis and chitin in

fungal cell walls and both these things are missing among oomycetes (Latijnhouwers et al., 2003). To efficiently control Phytophthora diseases a deeper understanding about their physiology, pathogenicity and interaction strategies with their hosts is required.

Important Phytophthora Species

Currently, all known species within the genus Phytophthora are plant pathogens with varied life styles, infection mode and host range (Le Berre et al., 2008). In a recent ranking of top ten Oomycete plant pathogens based upon their economic and scientific importance, six positions are occupied by Phytophthora spp. and Phytophthora infestans ranked as the top most oomycete pathogen (Kamoun et al., 2015). P. infestans that is an airborne foliar pathogen with a host range primarily limited to solanaceous species. Late blight of potato caused by P. infestans, was responsible for

Irish potato famine in 1800s that tremendously shaped human history by causing more than a million deaths and displacement of another million (Nowicki et al., 2012). This disease has been considered to lay the foundation of the science of Plant Pathology. P. infestans is still considered the deadliest potato pathogen causing about $ 6.7 billion crop losses worldwide (Haas et al., 2009). Late blight of tomato is another major socioeconomically important disease caused by P. infestans. In addition to potato and tomato, P. infestans is associated with the diseases of some other important solanaceous species including nightshade (Solanum nigrum), petunia, Calibrachoa and

Nicotiana benthamiana (Becktell et al., 2006; Nelson, 2008). Being so devastating as a pathogen for nearly two centuries, P. infestans has been widely studied and has been developed into the model oomycete pathogen(Haas et al., 2009). Molecular biology research on P. infestans has provided the basics to understand Phytophthora-plant interactions (Kamoun et al., 2015).

Despite of its relatively recent discovery (1995); Phytophthora ramorum, the causal agent of sudden oak death, sudden larch death and ramorum blight; holds second place in the list of top ten oomycete phytopathogens (Kamoun et al., 2015).

Currently, it is considered the most damaging pathogen of oaks worldwide, posing a major threat to the timber industry (Kamoun et al., 2015). (Kliejunas, 2010). Like P. infestans, P. ramorum is an airborne pathogen. But unlike P. infestans, P. ramorum has a wide host range and is considered to have the widest host range of all known

Phytophthora spp. Thus, it can serve as an excellent model to study pathogen adaptations for infecting multiple host species (Kamoun et al., 2015).

Ranked fourth on the list, Phytophthora sojae is the major pathogen of soybean that accounts for $1-2 billion crop losses worldwide by causing root rot in older plants and damping off in the seedlings (Tyler, 2007). It is a soilborne pathogen with a very narrow host range and economic damage limited to only soybean. Along with P. ramorum, the P. sojae genome was the first sequenced oomycete genome. Molecular mechanisms underlying pathogenicity of P. sojae have been studied to understand its interaction with its host and to determine the genetic basis of resistance against this pathogen (Kamoun et al., 2015).

The devastating vegetable pathogen, Phytophthora capsici ranked fifth on the list. It can infect several different plant families including cucurbits, legumes and solanaceous vegetables including tomato and pepper. P. capsici is considered to be the most genetically diverse eukaryotic organism yet described. A genome sequence and single nucleotide polymorphism based linkage map is available and transcriptomic and

effetoromics studies have been done to understand infection strategies of this pathogen in different hosts (Kamoun et al., 2015).

Phytophthora cinnamomi is another high impact Phytophthora pathogen, associated with the devastating diseases of numerous plants in horticulture, nursery and forest industry worldwide. This pathogen is considered among the most invasive pathogens with no effective control strategies yet available. Once invaded it causes damage to various plant species up to their extinction. In Australia, it was declared a key threat to biodiversity in the Commonwealth Environment Protection and Biodiversity

Conservation Act 1999 (Shearer et al., 2004). Major diseases of P. cinnamomi has been reported in avocado (Messenger et al., 2000), eucalyptus (Podger, 1978) and oaks

(Bergot et al., 2004). Due to its high impact, P. cinnamomi has been widely studied using numerous approaches (Kamoun et al., 2015).

Holding eighth place on the list of top ten oomycetes, Phytophthora parasitica is a soilborne pathogen that can infect roots and stems of up to sixty different plant families including economically important crops (Goodwin et al., 1989). Black shank in tobacco and root rot and gummosis in citrus are major diseases caused by P. parasitica

(Kamoun et al., 2015). In addition to late blight in tomato caused by P. infestans, two other tomato diseases; Phytophthora root rot and buck eye rot of fruit are caused by P. parasitica (Neher and Duniway, 1992). P. parasitica is also responsible for root rot and leaf spot in peace lily (Spathiphyllum) that is an economically important ornamental plant grown around the globe (Mounika et al., 2017). In cotton, boll rot and damping off in seedlings had also been reported to be caused by P. parasitica (Mitra, 1929)

(Pinckard and Guidroz, 1973). It is also found causing fruit rot in guva (Mitra, 1929).

Apart from cultivated plants P. parasitica can also infect the model plant Arabidopsis thaliana, the most exhaustively studied plant with significant amounts of genomic transcriptomics and proteomics data available to support a deeper understanding of plant biology and to explore the molecular basis of plant-pathogen interactions

(Koornneef and Meinke, 2010). The model oomycete pathogen P. infestans is airborne with a narrow host range so it cannot represent soilborne Phytophthora spp. with broad host ranges. P. sojae, although a widely studied soilborne oomycete, is just a major pathogen of soybean only. So, due to its broader host range and ability to infect model plants, P. parasitica has now emerged as a model of soilborne oomycete pathogens

(Meng et al., 2014).

Despite of their wide-spread and large economic impact, control strategies for

Phytophthora diseases are still very limited. All these Phytophthora spp. are being studied widely by using different omics approaches to understand their pathogenicity and molecular mechanisms used to bypass host plant defense responses, with an ultimate objective of developing long lasting resistance in host plants against these noxious pathogens.

Plant Defense

Plants are constantly attacked by numerous microbes, insects and herbivores.

Being sessile, plants can’t escape from these invaders like other motile organisms. So, they have evolved strong defense strategies to counter these attackers.

Plant defense comprises of both constitutive preformed and induced defense responses that offers several layers of protection against pathogen invasion (Anderson et al., 2010; Doughari, 2015). Induced defense of plants comprises of two major layers, the first one is known as pathogen associated molecular pattern (PAMP) triggered

immunity (PTI) in which some molecules or structures of pathogen are perceived by plant pattern recognition receptors (PRRs), followed by an array of defense responses that combat pathogen invasion. If the pathogen manages to survive PTI then the second layer of immunity called effector triggered immunity (ETI) comes in action. Due to these sophisticated multilayer defense mechanisms, most of the plants are nonhosts to most of the microbes. Few microbes are capable of overcoming defense in certain plant hosts and thus are adapted pathogens of those hosts.

Effectors are proteins secreted by pathogens inside hosts to modulate defense and, in the case of incompatible interactions, these modulations are perceived by plants through resistance genes (R genes) leading to activation of downstream components of defense ultimately leading to local cell death or hypersensitive response (HR) to prevent spread of the pathogen. Most characterized R genes encode intracellular receptor proteins, containing nucleotide binding (NB) and leucine rich repeat (LRR) domains

(Jones and Dangl, 2006). However, some membrane associated R genes that function as PRRs have also been reported (Zipfel et al., 2006). Based on their functional domains R genes are broadly categorized into five classes: the first R gene class named as CNL (CC-NB-LRR) comprise of at least one coiled-coil domain in addition to the typical NB-LRR domain, the second class called TNL (TIR-NB-LRR) contains a Toll- interleukin receptor (TIR)-like domain in addition to NB-LRR domain, RLPs (receptor like proteins) comprised of a receptor serine threonine kinase and an extracellular LRR domain (ser/thr-LRR) constitutes the third class, the fourth class RLK (Kin-LRR) contains a kinase and an extracellular leucin rich repeat, and the last class (Others) comprise of all other R genes without any defined conserved domains (Sanseverino et

al., 2009). Several P. infestans resistance genes (Rpi) have been identified from tomato, potato and other plants and are being used to develop Phytophthora resistant plant (Rodewald and Trognitz, 2013) (Zhu et al., 2012). Most of these Rpi genes are very specific in recognizing their Phytophthora counter parts (avr genes), however, there is significant allelic diversty not only among different Phytophthora species but also among different races and strains of the same species (Rodewald and Trognitz,

2013). Two R genes, a TIR-NB-LRR and RSP4 were shown to be involved in conferring resistance against P. parasitica in a citrus cultivar (Poncirus trifoliata cv.

Rubidoux) (Boava et al., 2011). In addition to local cell death (HR), R genes are known to activate prolonged resistance by inducing phytohormones and pathogenicity related genes (PR-genes) that cumulatively give rise to broad spectrum systemic acquired resistance (SAR) to fight against future infections (Jones and Dangl, 2006). Induced systemic resistance (ISR) and SAR incited by soil-borne infections in resistant rootstocks can prevent infections in foliar parts of the plant (Vernooij et al., 1994; Park et al., 2007).

Calcium Signaling in Plant-Pathogen Interactions

Calcium signaling has been widely reported to regulate a wide range of regular physiological functions as well as responses to environmental stresses in eukaryotes and prokaryotes (Marchadier et al., 2016) (Bruni et al., 2017). The calcium ion (Ca2+) is considered as a key secondary messenger. Spatio-temporal oscillation in intra and extracellular Ca2+ concentration (calcium signatures) is among the earliest molecular event in plant responses to microbes including Phytophthora spp. (Aldon et al., 2018)

(Atkinson et al., 1990). The generation of diverse calcium signatures in response to different pathogens or even the closely related variants within the same pathogenic

species have been observed. (Tavernier et al., 1995; Zimmermann et al., 1997; Xu and

Heath, 1998; Blume et al., 2000; Grant et al., 2000; Lecourieux et al., 2002; Gust et al.,

2007; Ranf et al., 2008). These Ca2+-signatures are decoded by combinations of various Ca2+-binding proteins (calcium sensors) and their downstream targets (Ranty et al., 2016). Specificity of pathogen-induced calcium signatures, their decoding and relaying by combinations of calcium sensors with their downstream partners have been suggested to shape the plant immune responses apart from physiological or other stress responses (Lecourieux et al., 2006; Ranty et al., 2016).

Plants carry large number of diverse calcium sensor proteins, divided into four major families: the Ca2+ binding proteins or calmodulins (CaM), calmodulin like proteins

(CML), Ca2+ dependent protein kinases (CDPK) and the calcineurin B like proteins

(CBL). CaM is a ubiquitous calcium sensor, highly conserved among eukaryotes whereas CML, CDPK and CBL are only found in plants (Ranty et al., 2016). Among all calcium sensors, calmodulin is relatively well studied in numerous plant microbe interactions (Harding et al., 1997; Do Heo et al., 1999; Kim et al., 2002; Takabatake et al., 2007; Reddy et al., 2011; Du et al., 2015a; Choi et al., 2009; Do Heo et al., 1999;

Takabatake et al., 2007; Choi et al., 2009; Cheval et al., 2013).

Binding of calmodulin to its targets could be either Ca2+ -dependent or - independent (Rhoads and Friedberg, 1997). Calmodulin has four Ca2+-binding EF-hand motifs. Upon binding Ca2+, it undergoes a conformational change, which allows it to bind and modulate the activity of numerous proteins involved in diverse cellular processes including plant defense (reviewed in Kudla et al., 2010; Reddy et al., 2011).

In tobacco, NtCaM3 activates NAD kinase that acts as a cofactor of NADPH oxidase to

produce reactive oxygen species (ROS) that is an important event during program cell death to defend plants against pathogens (Wen et al., 2013). Numerous other studies have demonstrated involvement of calmodulin-binding proteins including transcription factors, kinases, phosphatases, channels and pumps, and many uncharacterized proteins in plant defense (Kim et al., 2002; Truman et al., 2013; Reddy et al., 2000; Choi et al., 2009) but a holistic view of calcium signaling during plant-pathogen interplay is like a puzzle with lots of missing pieces.

Finding the host proteins targeted by pathogen effectors is one of the best strategies to unravel the structure of pathways involved in host-pathogen interactions.

But use of this strategy to find the calcium signaling related effector targets has been largely obscured due to the dependence of interaction of calcium signaling related proteins on Ca2+ homeostasis. Significant effort has been made to devise appropriate methods to find such interactions. The interaction of human CaMs with proteins of some human viruses was reported during the last decade (Taylor et al., 2012). Following the techniques used in these studies, recently, the interaction of three effectors of diverse pathogens including P. infestans with host CaMs was reported.

A Pseudomonas syringae effector HopE1 was found to interact with CaM to utilize it as a cofactor to target MAP65, which is an important component of cell microtubule network, and reduces the secretion of an immunity related protein PR-1 leading to inhibiting cell wall associated extracellular PTI responses (Guo et al., 2016).

CoDN3 a Colletotrichum orbiculare effector interacts with plant CaM to prevent cell death induced by one of its own effectors, NLP1 (Isozumi et al., 2019). A P. infestans effector SFI5 (Zheng et al., 2018a) is shown to interact with CaM for suppressing PTI.

Since CaMs are involved in shaping diverse physiological functions by modulating multiple pathways, these recent discoveries of pathogen effectors targeting host CaMs may be a major breakthrough towards untangling the complexity of host-pathogen interactions.

Phytophthora-Plant Interactions

Because of their high economic impact, significant research has been done to unravel the molecular basis of Phytophthora-plant Interactions. Like all other plant pathogens, interaction of Phytophthora spp. pathogens with their hosts follows the often cited zig-zag model (Jones and Dangl, 2006). They carry a variety of elicitors/PAMPS that might get recognized by plants and activate PAMP-triggered immunity (PTI). If encountered with PTI, Phytophthora spp. carry hundreds of effector proteins that can modulate host defense responses in a variety of ways to establish ETS conducive for infection. If a pathogen is not capable of overcoming this elicitin-induced PTI then the plant is a nonhost for that nonadapted pathogen.

Phytophthora effectors can be broadly categorized in to two classes with respect to their localization; apoplastic effectors that are secreted in the apoplast of the host cells and cytoplasmic effectors that are translocated to the inside of host cells.

Apoplastic effectors include diverse enzymes that can degrade host cell components to facilitate pathogen penetration into the host such as cutinases, glycoside hydrolases, pectinases and proteases. To protect themselves host produce proteases to degrade these hydrolytic effectors. To counter host proteases Phytophthora spp. have evolved protease inhibitors (McGowan and Fitzpatrick, 2017).

Most of the cytoplasmic effectors contain an N-terminal RXLR motif (Arg, any amino acid, Leu, Arg), which functions in translocating effectors into host cells, and a C-

terminal domain with effector activity. RXLR effectors enable pathogenic oomycetes to suppress basal immunity (Birch et al., 2008; Brasier, 2009; Bozkurt et al., 2012) as well as function as avirulence (Avr) factors to be recognized by their cognate R genes in the hosts to trigger ETI (Anderson et al., 2015). Since the discovery of these RXLR effectors, significant effort has been focused on understanding the molecular mechanisms behind their functions to either suppress or trigger plant immunity

(Anderson et al., 2015).

Recognition of Phytophthora spp. by Plants and Induction of PTI

Plants recognize Phytophthora spp. by sensing a variety of elicitors. The most well studied family of oomycetes elicitors are elicitins. Elicitins are extracellular cysteine-rich proteins secreted by oomycetes pathogens of the genus Pythium and

Phytophthora. These are considered to be oomycetes PAMPs because they are structurally conserved proteins with no sequence similarity to plant proteins, thus, they are recognized as non self molecules by the host plants leading to the induction of PTI

(Derevnina et al., 2016). A number of elicitins from different Phytophthora spp. have been identified and are being used to identify host factors that can offer broad spectrum non race specific resistance to diverse phytopathogens (Du et al., 2015b).

Elicitins are reported to bind sterols and lipids. Sterols are certain types of lipids that are essential structural and functional component of eukaryotic physiology.

Oomycetes of the genus Pythium and Phytophthora are devoid of sterol synthesis ability. They rely on their host’s sterols and thus have adapted efficient mechanisms of sterol scavenging from host cell membranes through elicitins (Mikes et al., 1998).

For example, cryptogein isolated from P. cryptogea is a one of the earliest identified sterol binding elicitins. It has been extensively used to explore defense

mechanisms underlying non-host resistance. The sterol-cryptogein complex is recognized by the plant PRRs and triggers a strong defense response (Lochman et al.,

2005) thus, making P. cryptogea , an avirulent pathogen of tobacco (Wendehenne et al., 2002). Cryptogein was reported to bind with a162-kDa glycoprotein having a N- linked carbohydrate part on the plasma membrane of tobacco cells. The authors of this study speculated that this glycoprotein is a potential tobacco PRR for cryptogein but there is no experimental evidence yet (Bourque et al., 1999). Multiple signaling pathways were suggested to be involved in cryptogein mediated early and late defense responses resulting in the development of local and systemic resistance in tobacco.

Early defense responses (PTI) encompasses changes in pH, activation of phospholipase C and plasma membrane localized protein kinases that in turn triggers calcium ions (Ca2+) influx leading to the calcium dependent transcriptional reprogramming of many defense components (Amelot et al., 2012). Induction of nitrate

- 2+ (NO3 ) efflux through anion channels regulated by Ca influx and phosphorylation events upstream of cryptogein leading to the induction of an oxidative burst is also reported during cryptogein-tobacco interactions (Wendehenne et al., 2002). Late responses like activation of PR proteins (after 24 hours) leading to the induction of HR and development of systemic acquired resistance (SAR) have also been reported

(Keller et al., 1996a; Keller et al., 1996b). PR proteins including plasma membrane and nuclear localized mitogen-activated protein kinases (MAPK) respond to cryptogein either in calcium dependent or independent manners (Lebrun-Garcia et al., 1998;

Lochman et al., 2005; Dahan et al., 2009). Calcium and protein kinase mediated signaling has been reported to be involved in depolarization of the plasma membrane,

cell wall modifications, protein phosphorylation, potassium (K+) and chloride (Cl-) efflux, and phenylpropanoid metabolism in tobacco cells upon cryptogein treatment (Lochman et al., 2005) (Amelot et al., 2011). Hoeberichts et al. investigated the effect of light during cryptogein-tobacco interaction and reported that cryptogein induced transcriptional changes are immensely affected by photoperiod and HR response is significantly delayed under light exposure (Hoeberichts et al., 2013).

The INF1 elicitin secreted by P. infestans shares 79% amino acid sequence identity with cryptogein. INF1 is shown to bind phytosterols but its sterol carrier capability for pathogen is not yet confirmed because the INF1 deficient P. infestans doesn’t lose pathogenicity (Kamoun et al., 1998; Fawke et al., 2015). Similar to cryptogein INF1 is shown to induce HR and SAR through calcium and protein kinase signaling mechanisms. Induction of INF1 induced HR has been shown to be independently regulated by calcium and protein kinase signaling pathways. Oxidative burst is incited but not necessarily required for HR however, it is reported to be involved in both calcium and protein kinase signaling pathways (Sasabe et al., 2000). Moreover,

MAPK interacting cytosolic proteins HSP70 and HSP90 are reported to be the essential components of INF1 induced HR (Kanzaki et al., 2003).

Induction of HR by INF1 in tobacco is mediated by cell surface receptor

BAK1/SERK3, a LRR repeat receptor-like kinase (LRR-RLK) that was first identified as an interactor of a plant hormone brassinosteroid receptor (BRI1). Later it was identified as an essential component of PTI against diverse PAMPs. Recognition of INF1 by RLP or RLK type PRR and location of INF1-PRR interaction was undefined for a priod but quite recently it was shown that multiple INFs are perceived by the extracellular domain

of a wild potato RLP elicitin response receptor (ELR) that associates with BAK1 to modulated localized cell death in potato (reviewed in (Fawke et al., 2015)). Recently a suppressor of BIR1 (SOBIR1) was identified as another required component in INF

(ParA1) induced ELR-BAK1 mediated non host resistance against P. parasitica in tomato and N. benthamiana (Peng et al., 2015). Wild potato ELR is also reported to recognize cryptogein (Du et al., 2015b) (Derevnina et al., 2016) but association of BAK1 and SOBIR1 with cryptogein induced HR has not been experimentally demonstrated.

Elicitins have been shown to induce systemic resistance by triggering multiple defense responses including secondary metabolism and plant hormone signaling pathways that enable the plant to fend off subsequent attackers. Elicitation of salicylic acid (SA)-mediated SAR after cryptogein application in tobacco was found effective against the black shank pathogen P. parasitica var nicotianae (Wendehenne et al.,

2002). In tomato INF1 doesn’t induce HR but it can activate systemic resistance through jasmonic acid (JA) and ethylene (ET) signaling (but not SA) pathways that was found effective against bacterial wilt disease pathogen Ralstonia solanacearum(Kawamura et al., 2009).

Other than elicitins, few other elicitors have also been identified from oomycetes.

For example, the extracellular glycoprotein Pep-13 elicitor was first isolated from P. sojae and later identified in other Phytophthora spp. is a well known PAMP that is required for their recognition by the hosts and induction of defense responses. The cognate plant PRR for Pep-13 is not yet identified. Calcium signaling has also been reported to be involved in inducing Pep-13-mediated plant defense responses (Blume et al., 2000).

Another unique oomycete elicitor OPEL, was identified from P. parasitica and has homologs in several other Phytophthora spp. OPEL is reported to induce strong PTI in tobacco leaves, making it resistant to P. parasitica as well as diverse pathogens like

Tobacco Mosaic Virus and the bacterium R. solanacearum. OPEL is large modular protein comprised of 556 amino acid with a signal peptide and three conserved domains; a glycine-rich protein domain, a thaumatin-like domain, and a glycosyl hydrolase domain harboring a laminarinase active site. This laminarinase active site is associated with the elicitor activity of OPEL either by direct recognition of this site by a particular PRR or by perceiving damage associated molecular patterns (DAMPs) generated through degradation of a specific polysaccharide substrate by enzymatic activity of OPEL laminarinase (Chang et al., 2015).

In addition to oomycete-specific elicitors, there are some PAMPs that are common among all filamentous (both fungi and oomycetes) pathogens for example, cellulose binding elicitor lectin (CBEL) and β-glucans. CBEL is a highly conserved protein, comprise of two cellulose binding domains. It was first identified in the cell wall of P. parasitica and found to be involved in pathogen adhesion to the host cell surface.

This interaction is detected by plant PRRs to trigger defense responses. β-glucans are the cell wall fractions having conserved characteristics patterns that are recognized by specific hosts. For example, β (1-6)-glucans or heptaglucans from P. sojae can be perceived as PAMPs by Soybean but not by tobacco plants. Moreover, different β- glucans from different pathogens initiate varied calcium oscillations during PTI implementation after their recognition (Fawke et al., 2015).

Conclusively, upon recognition by cognate plant PRR, Phytophthora spp.

PAMPs/elicitors can trigger PTI through crosstalk among several interlinked and independent defense signaling pathways leading to the development of non race specific resistance that can protect the plant against diverse pathogens.

Supression of PTI by Phytophthora Effectors

In order to establish infections, successful Phytophthora pathogens have either devised ways to overcome immune responses triggered by their elicitors or they modify their elicitors in such a way that they can’t be recognized by host PRRs. For example,

P. parasitica carriers INF1 like elicitin called ParA1 that is recognized by tobacco to induce defense (Kamoun et al., 1993) but it has been found that virulent strains of P. parasitica can bypass this recognition by suppressing ParA1 expression during their interaction with compatible hosts and this adaptation enables it to initiate black shank disease in tobacco (Colas et al., 2001).

A P. infestans RXLR effector Avr3a (PiAvr3a) is also reported to suppress INF1 elicitor induced PTI by interacting and stabilizing the U-box E3 ligase CMPG1 that functions in protein degradation and is essential component of INF1-induced cell death

(Bos et al., 2006; Gilroy et al., 2011). Another effector, Pi02860 that also interact with a different potato ubiquitin E3 ligase (NRL1) is reported to suppress INF1 mediated cell death (Yang et al., 2016). As described above MAPK and calcium signaling pathways contribute to shape PTI. A RXLR effector PexRD2 binds to MAPKKKε to interrupt associated signaling pathways involved in defense response (King et al., 2014).

Recently, SFI5 another RXLR effector from P. infestans is reported to interact with calmodulin, the calcium sensor for suppressing PTI (Zheng et al., 2018b).

RXLR effectors are also reported to interacts with host vesicle-trafﬁcking and secretion mechanisms to suppress or enhance host defense through depending upon the absence or presence of cognate R gene. For example, PiAvrblb2 (PITG_04090), a

P. infestans RXLR effector, interacts with a plant immune protease C14 and prevents its secretion in the apoplast, apparently to prevent degradation of the Phytophthora virulence proteins (Bozkurt et al., 2011). Rpi-blb2, a potato R gene has been shown to recognize PiAvrblb2 and trigger hypersensitive response (HR) through SGT1 mediated pathways in N. benthamiana (Oh et al., 2014). AVR1 that is recognized by R1 to induce

HR in the host and is reported to interact and stabilize an exocyst component Sec5 resulting in enhanced defense against P. infestans (Du et al., 2015c).

Phytophthora spp. also carry effectors to target the host’s RNA silencing machinery to promote disease progression. Two P. sojae RXLR effectors referred to as

Phytophthora Suppressor of RNA Silencing 1 and 2 (PSR1 and PSR2) have been shown to suppress RNA silencing and enhance plant susceptibility to promote pathogen infection (Qiao et al., 2013). PSR1 is reported to target a host RNA helicase PINP1

(PSR1-Interacting Protein 1) that regulates the accumulation of endogenous small interfering RNAs and microRNAs and in Arabidopsis (Qiao et al., 2015). A P. infestans effector, Pi14054 has been experimentally demonstrated to act as a suppressor of RNA silencing in N. benthamiana (Vetukuri et al., 2017).

RXLR effectors have also been found to interfere with plant growth hormone signaling. PiAvr2, another RXLR effector interacts with potato phosphatase BSL1 that is a component of brassinosteroid (BR) hormone signaling crucial for plant growth. BSL1 is also reported to interact with a disease resistance protein R2. Interestingly, although

both virulent and avirulent alleles of PiAvr2 interact with BSL1, only the avirulent variants mediate the interaction of R2 with BSL1 and result in defense (Saunders et al.,

2012). Recently, PiAvr2 was shown to significantly induce BR signaling and enhance susceptibility to P. infestans. Moreover, it also caused the constitutive overexpression of bHLH transcription factor that is known to regulate antagonism between growth and immunity. It was concluded that PiAvr2 acts to reduce plant immunity by switching plant hormonal machinery towards enhanced growth (Turnbull et al., 2017).

RXLR effectors have also been found to suppress endoplasmic reticulum (ER) mediated immunity. A P. capsici RXLR effector PcAvr3a12, was experimentally demonstrated to interact with an endoplasmic reticulum (ER) stress-sensing protein

FKBP15-2 encoding a peptidyl-prolyl cis-trans isomerase. This PcAvr3a12- FKBP15-2 interaction inhibits FKBP15-2 stress sensing activity thus, attenuating ER-mediated defense response against P. capsici (Zheng et al., 2018b).

Phytophthora Effectors Targeting Host Susceptibility Factors

Apart from the traditional effector-R gene model, the concept of dominant susceptibility (S) emerged after the discovery of susceptibility genes (S genes) that if present in a particular host favor disease development by a specific pathogen. One example of this concept is the MLO gene. Plants harboring this gene were found to be highly susceptible to powdery mildew disease (Bai et al., 2008). Recently, some

Phytophthora effectors targeting host susceptibility factors to promote virulence were discovered (Boevink et al., 2016a). For example, Pi04089 interacts with a RNA-binding protein KRBP1 to enhance P. infestans colonization. KRBP1 is reported to be a S factor for P. infestans (Wang et al., 2015). Pi02860 interacts with a potato S factor NRL1 that encodes a CULLIN3-associated ubiquitin E3 ligase to enhance pathogenicity (Yang et

al., 2016). Another RXLR effector, Pi04314 interacts with three isoforms of host PP1c

(protein phosphatase 1 catalytic) and translocate them from the nucleolus to the nulcleoplasm. This interaction has been shown to increase host susceptibility to P. infestans by suppressing JA and SA responsive gene expression (Boevink et al.,

2016bi). RXLR effector Pi03192 targets two N. benthamiana NAC transcription factors at the endoplasmic reticulum (ER) membrane and prevents their localization to the nucleus thus, resulting in increased host susceptibility to P. infestans by preventing transcriptional regulation of defense components (McLellan et al., 2013). Another P. infestans RXLR effector, PexRD54 binds to host autophagy protein ATG8CL to stimulate the formation of autophagosomes. This PexRD54-ATG8CL binding outcompetes ATG8CL binding with an autophagy cargo receptor Joka2 that plays positive role in defense (Dagdas et al., 2016). All these host proteins targeted by RXLR effectors are referred to as S factors for P. infestans because they were found to promote infection. Interaction of another RXLR effector PiAvr2, is known to suppress plant immunity by interacting with BSL1. Recently, BSL phosphatses were shown to act as susceptibility factors and interaction of PiAVR2 with all members of the BSL family has been experimentally demonstrated to be required for as an event of EST (Turnbull et al., 2019).

These studies suggest that RXLR effectors target diverse host proteins involved in different cellular processes to intervene in host defense in multiple ways. However, compared to the hundreds of Phytophthora effectors, these few examples represent only the tip of an iceberg. Our understanding of effector biology remains incomplete

and significantly more remains to be discovered and investigated to fully understand how effectors manipulate host cellular processes to modulate host defense responses.

Perspective on Developing Phytophthora-Resistant Plants

Management of Phytophthora diseases by harnessing host genetic resistance is the most effective, eco-friendly and in the long run cost effective strategy. In the past several P. infestans resistance genes (Rpi) have been identified from tomato, potato and other plants and are being used to develop Phytophthora-resistant plants

(Rodewald and Trognitz, 2013; Zhu et al., 2012). But most Rpi genes are very specific in recognizing their Phytophthora counter parts (avr genes) that are diverse not only among different Phytophthora species but also among different races and strains of the same species (Rodewald and Trognitz, 2013). As reviewed above, plant defense against Phytophthora spp. is a complex multilayered, interlinked phenomenon where each defense layer is intervened by either elicitors or effectors that have been rapidly coevolving during host-Phytophthora interactions. Thus, the R gene-mediated resistance is not sufficient alone to provide broad spectrum and durable resistance.

In order to develop sustainable resistance in crop plants against these noxious pathogens a thorough understanding of the molecular basis of Plant-Phytophthora interactions and discovery of multiple sources of genetic resistance are very important.

In this era of high throughput technology, Phytophthora-plant interactions are being widely studied, using different “omics” (genomics, transcriptomics, proteomics, metabolomics and effectromics) approaches to identify novel components of plant defense. For example, comparative transcriptome studies of P. infestans-resistant wild tomato enabled the identification of P. infestans resistant transcription factors

SpWRKY3 (Cui et al., 2018), and a long non coding RNA (lncRNA16397) along with glutaredoxin (SpGRX) gene (Cui et al., 2017).

Elicitor initiated nonhost resistance have also been explored widely in an effort to identify nonrace-specific sources of resistance. Recently, an elicitin response receptor

(ELR) was identified from a wild potato species that can recognize elicitins from diverse

Phytophthora spp (Derevnina et al., 2016).

Identification of such nonrace-specific sources of resistance and their pyramiding with multiple R genes and other defense genes can offer stable broad spectrum resistance against these rapidly evolving Phytophthora spp. Conventional breeding (if sources of resistance are intraspecies) and genetic engineering both should be utilized for developing resistance cultivars. However, since GMOs remian highly controversial these days, genome editing through CRISPER/Cas technology provides an excellent tool to knock out susceptibility factors for Phytophthora spp.

Research Objectives

As described above, plant-Phytophtora interactions comprise of complex cross talks among hundereds of host and pathogen genes. In the modern era of “omics”, affordable high-throughput data generation technologies like genomics and transcriptomics have enabled us to model a holistic concept of plant-pathogen interactions.

Transcriptome analysis of host responses to pathogens has become a widely deployed technique to narrow down the list of potential genes involved in plant defense.

Identification of genetic sources of resistance against P. parasitica can serve as a valuable resource for developing Phytophthora resistant crops in the future. In our laboratory, several accessions of a wild tomato species, Solanum pimpinellifolium were

tested against P. parasitica infection and some highly resistant and susceptible accessions were identified. Carrizo citrange, a common citrus rootstock is considered resistant against P. parasitica. My major goal is to analyze the transcriptome of Tomato-

Phytophthora and Citrus-Phytophthora interactions with an intent to identify genes that potentially contribute to defense against P. parasitica.

At the protein level, large scale protein-protein interactions studies among host and pathogen proteins and exploring their biological relevance enables us to track the functions of pathogen effectors and their targeted pathways to manipulate host defense.

In our laboratory, extensive screening of a tomato cDNA library against RXLR effectors known for their involvement in infection revealed interaction of a core effector with calmodulins (CAMs). CAMs are calcium binding proteins that serve as a primary calcium sensor in living cells. Upon pathogen invasion, fluctuations in the Ca2+ ion concentration of host cells is essentially the key event involved in the perception of threat (Aldon et al., 2018). These Ca2+ fluctuations are sensed by calcium sensors that can directly bind to Ca2+ ions and undergo conformational changes, enabling them to affect the function of their down-stream targets that ultimately lead to the development of defense responses (Ranty et al., 2016). Physical interaction of Phytophthora effectors with CAMs points towards the manipulation of host’s calcium signaling pathway by the pathogen to modulate defense responses. The second major goal of my research is to explore the implications of theses effector-calmodulin interactions on calcium signaling and its role in plant defense.

CHAPTER 2 COMPARATIVE TRANSCRIPTOME ANALYSIS BETWEEN A RESISTANT AND A SUSCEPTIBLE WILD TOMATO ACCESSION IN RESPONSE TO PHYTOPHTHORA PARASITICA

Introduction

Tomato is among the most widely cultivated crop plants that is not only consumed as raw fruit and culinary vegetable in meals but also constitute a major agricultural industry worldwide. The United States of America is among the top tomato producing countries of the world with total annual production of 246 million cwt, valued about $1.67 billion in 2017 (USDA, 2018). Apart from being an important commercial food crop, tomato is being widely used as a model plant to study different aspects of plant biology including host-pathogen interactions. Tomato production is threatened by numerous diseases caused by a variety of pathogens that belong to all major groups of pathogens like viruses, viroids, bacteria, nematodes, fungi and oomycetes. Because of its high economic value and the fact that it could be attacked by diverse pathogens, the tomato pathosystem is an excellent model to study plant-pathogen interactions.

Oomycetes in the genus Phytophthora are responsible for a number of devastating diseases in tomatoes for example, late blight, Phytopthora root, crown and buckeye rot caused by different Phytophthora species. These diseases not only damage tomato crop production but also cause major postharvest losses, threatening the tomato processing industry. Phytophthora parasitica is mainly known as a root and fruit pathogen of tomato associated with Phytophthora root rot and buckeye rot diseases

 This chapter has been republished with permissions from: Naveed, Z., and Ali, G. (2018). Comparative transcriptome analysis between a resistant and a susceptible wild tomato accession in response to Phytophthora parasitica. International journal of molecular sciences 19, 3735.

but leaf infection, stem canker, stem girdling, collar rot, blossom blight and damping off of seedlings have also been reported in tomatoes in different parts of the world.

Buckeye rot is a major tomato disease in almost all tomato growing areas of the world like India, China, many parts of Europe and the USA (Wani, 2011). The disease was first reported in Florida in 1907 followed by an epidemic causing 40% loss in tomato production in an experimental field in Indiana in 1921. It is still a major problem in many tomato growing states such as Florida, Maryland and Pennsylvania. Being unique among other phytopathogens, control strategies against Phytophthora diseases are very limited. Management of Phytophthora diseases by manipulating host genetic resistance is being considered the most effective, eco-friendly and in the long run cost effective resistance development strategy. Understanding and implementation of molecular basis of Tomato-Phytophthora interactions is very important. Identification of genetic sources of resistance against Phytophthora species can serve as a valuable resource for developing Phytophthora-resistant crops in the future.

Several P. infestans resistance genes (Rpi) have been identified from different

Solanum spp. mainly wild potato species. There are five major Rpi genes (Ph-1 – Ph-5) derived from tomatoes. Ph-1, Ph-2, Ph-3 and Ph-5 were identified from different accessions of Solanum pimpinellifolium, a wild tomato species, whereas the source of

Ph-4 was another wild tomato species S. habrochaites. Efforts to develop Phytophthora resistant potatoes and tomatoes by using individual Rpi genes were not successful.

Because most of these genes are very specific in recognizing their Phytophthora counter parts (avr genes) that are diverse not only among different Phytophthora spp.but also among different races and strains of the same species (Rodewald and

Trognitz, 2013; Zhang et al., 2014). Pyramiding of multiple Rpi genes in potato offered somewhat stable and broad-spectrum resistance against P infestans. Other than R genes, comparative transcriptome studies of P. infestans-resistant S. pimpinellifolium enabled the identification of P. infestans-resistant transcription factors SpWRKY3 (Cui et al., 2018), and a long non coding RNA (lncRNA16397) along with the glutaredoxin

(SpGRX) gene (Cui et al., 2017). Transgenic tobaccos containing another transcription factor, SpWRKY1 were found to be resistant to P. nicotianae (Li et al., 2015).

Identification of non-race specific sources of resistance and their pyramiding with multiple R genes can offer stable broad spectrum resistance against these rapidly evolving Phytophthora spp. In our laboratory, several accessions of a wild tomato species, S. pimpinellifolium were screened against P. parasitica infection and some highly resistant and susceptible accessions were identified. In this study, we are comparing the transcriptome of P. parasitica resistant (hereafter Sp-R) and susceptible

(hereafter Sp-S) accessions of S. pimpinellifolium to understand the molecular basis of resistance and susceptibility. Our main goal was to identity genes putatively involved in resistance against P. parasitica.

Materials and Methods

P. parasitica Inoculation, RNA Extraction and Sequencing

A highly infectious P. parasitica strain (12-1) was grown on solid V8 medium plates for 10 days. Zoospores were induced by applying dehydration stress on cultures from day 5 to day 7. On 10th day cultured plates were flooded with sterilized water to induce release of zoospores and diluted up to 105 cells/ml. Leaves of thirty days old seedlings of tomatoes grown under controlled conditions in sterilized soil were inoculated with zoospores. Two spots 10 µl of zoospore suspension were placed on the

underside of the leaf (each side of midrib) with the help of pipettor. Mock inoculation was done by using sterilized water without zoospores. Around two hundred spots were inoculated in each accession and experiment was repeated three times. After three days post inoculation (72 hpi), a spot was counted as HR if lesion was confined to the inoculation spot whereas a spot was considered a diseased spot if a water soaked lesion was observed on the area at least five times bigger than the initial inoculation site. For RNA extraction, leaf samples from treated and untreated plants were collected at 24hpi and 48hpi. Leaves were thoroughly washed, immediately frozen in liquid nitrogen and stored in -80 0C until further use. Total RNA by using QIAGEN RNeasy kit, was extracted from frozen leaf tissues. Twelve cDNA libraries were constructed and sequenced using Illumina Hiseq 2500 platform.

RNA-Seq Analysis

Most of the sequencing data analysis was done by using CLC Genomics

Workbench program (Version 9.5.3, QIAGEN Denmark). Raw data files were imported as paired end data for each individual sample. Quality of raw data was checked by generating quality check (QC) reports for each sample. Adapters were removed, and sequences were further trimmed and filtered using quality trim cutoff of 0.005

(equivalent to Phred score 27) and minimum read length of 20bp. All sequence libraries were rechecked by generating QC reports to make sure that trimming and quality filters were stringent enough. Clean reads from all samples were mapped to S. lycopersicum genome (SL2.50 version) with mismatch cost of 2 and insertion cost of 3. Genes with

RPKM value greater than zero were considered expressed. Differential expression analysis was done using RNA-Seq tool of CLC genomics workbench that is based on generalized linear model assuming that read counts follow a negative binomial

distribution. This method is similar to what is used in DESeq and edgeR packages and by far the best reported analysis for differential expression (Boevink et al., 2016). A gene was considered differentially expressed if its FDR p-value  0.001 and Log2 fold change  |2|. Heat map was based on normalized log CPM values (CLC Genomics

Manual) using Euclidean distance and complete linkage of clusters, 50 fixed number of features and minimum counts of 10 in at least one sample. All Venn Diagram analysis were done by using the webtool (http://bioinformatics.psb.ugent.be/webtools/Venn/).

ITAG annotations were extracted for all differentially expressed genes (hereafter

DEGs). Gene Ontology (GO) term enrichment analysis and Parametric analysis of gene set enrichment (PAGE) was done by feeding ITAG IDs to the Agriculture gene ontology

(AgriGO) tool (http://bioinfo.cau.edu.cn/agriGO). AgriGO is an integrated online tool designed for GO term enrichment analysis of agricultural species (Du et al., 2010).

Cross comparisons of individual SEA results of Sp-R and Sp-S DEGs was done by using SEACOMPARE (Cross comparison of SEA) tool of AgriGO.

In addition to ITAG annotations, KEGG (Kyoto Encyclopedia of Genes and

Genomes) terms for DEGs were obtained by feeding DEGs sequences to KAAS

(KEGG automatic annotation server) to assign KO terms (Moriya et al., 2007) using the

Nightshade family (sly, nta, ath, aly) database options. Pathview was used to generate

KEGG pathways (Luo et al., 2017). Mapman software was utilized to view DETs in biological plant specific processes (Thimm et al., 2004).

Validation of RNA-seq Results

To validate differential expression results obtained by using RNA-seq data, top

DETs that were potentially involved in the development of resistance response in R

against P. parasitica were subjected to quantitative real time PCR by using previously published method (El-Sayed et al., 2015;Patel et al., 2015). Gene specific primer pairs

(Table 2-1) were designed using NCBI primer BLAST tool

(https://www.ncbi.nlm.nih.gov/tools/primer-blast/). Actin gene was used as a reference to calculate the relative expression values. Melt curve analysis was done to verify specific amplification of single gene product by each primer pair.

Results

P. parasitica Infection on S. pimpinellifolium Accessions

Initial screening of thirty wild tomato accessions was done against P. parasitica that resulted in the identification of some resistant and susceptible accessions

(unpublished data). We chose a highly resistant (Sp-R) and a highly susceptible (Sp-S) accession for RNA-seq analysis. Leaves from thirty-day old plants were inoculated with the P. parasitica zoospore suspension and symptoms were observed overtime.

Localized HR like lesion was observed on the leaves of Sp-R leading to pathogen’s growth arrest and thus no disease whereas water soaked lesions were observed on the leaves of Sp-S that resulted in severe disease development and visible pathogen growth and sporulation on the leaf surface (Figure 2-1). Like other Phytophthora spp., P. parasitica is a hemi-biotroph, infection studies in tomato and A. thaliana have revealed a short biotrophic phase (between 0 to 24 hpi) followed by quick switch to a necrotrophic phase at about 30 hpi (Le Berre et al., 2008); (Attard et al., 2010). To capture transcriptional changes during biotrophic and transition phases with an objective to understand molecular basis of resistance, we collected two replicates of each P. parasitica treated and control Sp-R and Sp-S samples at 24 hpi and 48 hpi for RNA-seq analyses.

Analysis of RNA-seq Data

In total 329 million, 126 base pair long raw reads were generated from all twelve libraries with an average of 13 million paired end reads per library. The raw data was deposited in the NCBI SRA. After adaptor removal and quality filtering, 274 million clean reads were obtained. A very stringent criteria was used in cleaning data, all the sequences with Phred score of 27 were filtered out leaving clean reads ranging from 16 to 26 million for all samples (Table 2-2). Clean reads from all twelve libraries were mapped to the tomato genome SL2.50. All the samples showed overall mapping coverage of 96-97 percent. While the percentage of uniquely mapped reads of all the samples ranged between 76 to 85 %. Further analysis was carried out using only the uniquely mapped reads. Detailed sequencing and mapping statistics is given in Table 2-

Genes with RPKM value greater than zero in both replications for each group were considered expressed. More than 20,000 genes were expressed in all the treated and control groups of both genotypes at both time points with varied number of specifically expressed genes in each group, highest was 1,048 in P. parasitica infected

Sp-S samples at 48hpi (Figure 2-2). Overall, 22,715 expressed genes were found across all 6 groups. In total, 21,789 expressed genes were present in R samples and

21,924 in Sp-S samples, among them 791 and 926 were specifically expressed across all Sp-R and Sp-S samples respectively (Figure 2-2).

To check the quality of our data and to have an overview of variations among all our samples, principal components analysis (PCA) was conducted using RPKM values.

Principal component (PC) 1 explained 50% of the variation, PC2 comprised of 32% variance, together it explained 82% variance among 12 groups (24 samples). Samples

from resistant (Sp-R) susceptible (Sp-S) accessions were separated clearly in two groups, depicting transcriptome variation between the two genotypes. P. parasitica treated and mock inoculated samples showed considerable variations among them, revealing the transcriptome change in response to pathogen in both Sp-R and Sp-S genotypes. Relatively less variations were observed between both treatment time points of both Sp-R and Sp-S samples, indicating some transcription reprogramming overtime in response to pathogen. PCA plot revealed very little to no variance among all sample replicates, indicating very high reproducibility of biological replicates and high quality sequencing data (Figure 2-3).

Differential Expression Analysis

Differential expression was calculated for all treatments versus mock samples of both Sp-R and Sp-S genotypes over both time points. In addition to that all mock and treated samples of both genotypes were compared to each other. Expressed genes with

FDR p-value  0.001 and Log2 fold change  |2| were designated as Differentially expressed genes (DEGs). Overall, more DEGs were identified in Sp-S treated vs. Sp-S- control compared to Sp-R-treated vs. Sp-R-control and the number of DEGs found at

24hpi was greater than 48hpi in all treatment vs. control as well as Sp-R vs. Sp-S comparisons. Moreover, the number of down regulated genes were greater than upregulated ones in all comparison groups (Table 2-3). The highest number of DEGs,

3079 was found in Sp-S-24hpi treated vs Sp-S-control comparison with 1132 upregulated and 1947 downregulated DEGs. Whereas, a relatively lower number of

DEGs (2657) was observed in Sp-R-24hpi treated vs. Sp-R-control comparison. In Sp-R vs. Sp-S differential expression analysis, 1158 and 889 DEGs were observed at 24 and

48hpi respectively. We also compared controls of Sp-R and Sp-S genotypes and found

322 DEGs, the lowest count among all the comparisons. Detail DEGs counts over all comparisons is given in Table S2. Venn diagram analysis revealed 1173 genes were differentially expressed only in Sp-R accession upon P. parasitica inoculation, of them

519 and 293 DEGs were unique to 24hpi and 48hpi respectively (Figure 2-4). Among

DEGs exclusively found in Sp-R accession, many important plant defense related genes were highly induced for example, Serine protease inhibitors (Solyc09g084480.2,

Solyc09g084490.2 and Solyc09g089530.2) (Kim et al., 2009), Kunitz family trypsin and protease inhibitor protein (Solyc03g098790.1), Chitinases (Solyc10g074440.1,

Solyc02g082930.2, Solyc05g050130.2 and Solyc02g082960.2), Arabidopsis defensin- like protein (Solyc07g009260.2) (Thomma et al., 2002) and VQ motif-containing protein

(Solyc02g064570.1)(Le Berre et al., 2008) etc. On the other hand, many important defense gene like, many R genes, calcium signaling genes, BAK1-interacting receptor etc. were found downregulated. Mlo that is a well known powdery mildew susceptibility

(Bai et al., 2008) gene (Solyc03g095650.2) was found highly suppressed in Sp-R.

Heat map based on hierarchical cluster analysis of expressed genes across all samples, revealed specific expression trends of some core defense genes across Sp-R vs. Sp-S accessions as well as control vs. infected samples of both accessions. Among the top 50 highly correlated genes across samples, 22 clustered together to reveal higher expression levels in all four mock samples compared to very low expression in all eight P. parasitica treated samples regardless of genotype and timepoints (Figure 2-5).

Among rest of the clusters, several core plant defense related genes were seen highly induced in both resistant and susceptible accessions upon P. parasitica treatment.

In P. parasitica treated samples, 18 genes were found highly upregulated in Sp-S and comparatively less induced in Sp-R at both 24 and 48hpi. 14 out of these 18 were very important plant defense regulators including 5 pathogenesis related proteins

(Solyc09g007010.1, Solyc04g064880.2, Solyc01g106620.2, Solyc01g097270.2 and

Solyc01g097240.2 ), two chitinases CHI3 (Solyc02g082920.2), CHI-B

(Solyc10g055800.1), LOX1 (Solyc08g029000.2), MLP423 (Solyc09g090980.2) and

BG1 (Solyc10g079860.1) (Veronesi et al., 1996);(Van Loon and Van Strien, 1999);

(George et al., 2016); (Schoina et al., 2017). Another important plant defense gene, rubisco activase (Solyc10g086580.1) was expressed in both Sp-R and Sp-S P. parasitica treated plants only at 48hpi (Baebler et al., 2011). Two genes, 2-oxoglutarate

(2OG) and Fe (II)-dependent oxygenase superfamily protein (Solyc10g086580.1.1) and

GTP binding elongation factor (Solyc11g069700.1) were found highly induced only in P. parasitica infected Sp-R samples at both time points. Few genes were induced in all treatments, but induction was much higher in Sp-R plants compared to Sp-S. For example, a heat shock protein HSP70 (Solyc09g010630.2) and PHE ammonia lyase 1 or PAL1 (Solyc09g007920.2) were most highly upregulated in Sp-R at 24hpi. Highest induction of Peroxidase 2 (Solyc01g006300.2) was found in Sp-R at 48hpi (Figure 2-5).

All these genes are well known to be involved in plant defense against oomycete pathogens particularly Phytophthora spp. (Kanzaki et al., 2003; Zhang et al., 2017a);

(Zhang et al., 2017b).

Analysis GO Enrichment Analysis of DEGs

GO term enrichment analysis of 3208 Sp-R and 3426 Sp-S DEGs by Singular

Enrichment Analysis (SEA) tool using S. lycopersicum ITAG2.4 as background reference annotated 2301 and 2485 DEGs respectively. Significant enrichment of 22

and 30 GO terms was revealed among Sp-R and Sp-S DEGs respectively. The most significantly enriched GO term, common among DEGs of both Sp-R and Sp-S genotype was catalytic activity. Other significantly abundant molecular function (MF) terms were oxidoreductase activity, transferase activity, oxygen binding and hydrolase activity

(Table 2). Biological processes (BP) terms were response to stimulus, cellular and metabolic processes, biological regulation and regulation of biological processes, and cellular component (CC) terms like cell, cell part organelle and macromolecular complex were present among DEGs of both Sp-R and Sp-S genotypes, but they were not significantly enriched (Figure 2-6).

Comparisons of Sp-R, Sp-S and Sp-R vs. Sp-S SEA results revealed some interesting trends. For example, peptidase and endopeptidase inhibitor activity terms were only present in Sp-R and glutathione transferase activity was specific to Sp-S.

Significant enrichment of ten different transport activity related GO terms specifically assigned to S DEGs. Whereas, enrichment of carboxylesterase, cellulose synthase activity, peptidase and endopeptidase inhibitor activity were unique to be differentially expressed in Sp-R genotype only (Table 2-4).

Parametric Gene Set Enrichment Analysis Based on GO

To gain more insight about functional annotations of DEGs with relevance to their expression levels across all the groups at each time point separately, Parametric analysis of gene set enrichment (PAGE) was done for all six (P. parasitica treated vs. mock samples of both genotypes and P. parasitica treated Sp-R vs Sp-S samples at both time points) comparisons. PAGE is an efficient and sensitive method that ranks annotated gene clusters according to their expression levels.

Out of total 6,101 wild tomato DEGs found across all six comparisons of P. parasitica treated samples, 339 unique GO terms were assigned to 3654 DEGs. PAGE analysis revealed, 62 and 51 significant GO terms among Sp-S DEGs, 23 and 19 in R

DEGs and 14 and 26 in Sp-R vs Sp-S DEGs at 24hpi and 48hpi respectively. Greater number of significant GO terms were found among Sp-S pathogen treated vs. control groups that indicates more intense transcriptional reprogramming in susceptible genotypes upon infection with P. parasitica.

Most of the terms related to regular physiological activity, like cellular processes, cellular component organization, localization, transporter activity, cell and cell part were found to be downregulated among all groups with few exceptions in Sp-R vs. Sp-S groups. Binding, cell wall organization and or biogenesis catalytic activity enzyme regulator activity and transcription regulator activity were found to be upregulated in most of the groups. Metabolic processes were downregulated in Sp-R whereas, upregulated in Sp-S genotypes (Figure 2-7A).

DEGs of all six groups were found to be enriched in some important plant defense related GO categories with varied expression trends. Antioxidant activity was found highly upregulated in both Sp-R and Sp-S samples in response to P.parasitica infection (Figure 2-7A). Cell death was found to be progressively upregulated from 24 to 48hpi in Sp-S and was downregulated in Sp-R at 24hpi but highly up at 48hpi.

Response to stimulus was found to be upregulated across all comparisons in the bar graph (Figure 2-7A).

Cross comparison of GO biological processes hierarchy between Sp-R and Sp-S revealed induction of genes related to response to stimulus in both Sp-R and Sp-S

accessions upon P. parasitica infection but there were significant differences in the downstream hierarchy. Response to biotic stimulus was upregulated in both resistant and susceptible accessions but response to wounding and response to external stimulus were specifically only in resistant accession (Figure 2-7B and 2-7C). Response to wounding comprised of 11 wound induced serine-type endopeptidase inhibitors

(SIPs) whereas, other plant defense genes like defensin, hydrolyses, kinases and R genes were enriching the response to stress and external stimulus categories.

Differential expression of many different SIP genes was seen in control vs. infected samples of both Sp-R and Sp-S and in Sp-R vs. Sp-S comparisons but the number of induced SIPs were more in Sp-R as compared to Sp-S. Out of total ten highly induced

SIPs in response to P. parasitica treatment, six were significantly upregulated in only

Sp-R, two were specific to Sp-S and two were induced both in Sp-R and Sp-S (Figure 2-

8). Defensin (Solyc07g009260.2), another important defense related gene was only induced in Sp-R at 24hpi. An NB-ARC domain containing R gene (Solyc06g008790.2) was constitutively expressed at low level (RPKM around 2) in only Sp-R in both control and treated samples.

Overall, oxidoreductase activity was found significantly enriched among DEGs of both Sp-R and Sp-S in response to P. parasitica (Table 2-4). Further dissection revealed significant induction of genes categorized in daughter GO term oxidoreductase activity, acting on peroxide as acceptor and peroxidase activity were induced in both

Sp-R and Sp-S whereas oxidoreductase activity, acting on the CH-CH group of donors,

NAD or NADP as acceptor was found downregulated only in Sp-S. Significant differential induction of these core defense components in Sp-R compared to Sp-S

make them potential candidates responsible for resistance against P. parasitica in Sp-R accession.

KEGG Annotations

To visualize the functional involvement of DEGs in biological pathways and to look for the differences between successful plant defense and disease development responses of Sp-R and Sp-S wild tomato accessions in response to P. parasitica infection at 24 and 48hpi, KEEG ontology of DEGs was done by using KAAS and assigned KO terms for each individual group, were mapped on KEEG pathways using

Pathview. Most of the DEGs showed common trend of up or down regulation in both

Sp-R and Sp-S at each time point in general physiological pathways as well as plant defense related pathways with few exceptions.

Overall, 30 spots were mapped on MAPK signaling pathway (Figure 2-9). All mapped genes in ethylene responsive defense response were found upregulated in both Sp-R and Sp-S, except MAPK 3/6 that was downregulated in R and was not found among S DEGs. Some core plant defense related components, like PR1, ERF1,

EBF1/2, PP2C, ETR/ERS were found upregulated, whereas, SnRK2 and CAM4 were commonly down regulated in both Sp-R and Sp-S accessions at both time points.

Contrasting trends of up and down regulation between Sp-R and Sp-S were shown by only three genes. FLS2 that was found upregulated in Sp-S at both time points was downregulated in Sp-R at only 24hpi, CAT1 that is a component of stress tolerant response was down regulated both in Sp-R and Sp-S at 24hpi but was found upregulated only in Sp-S at 48hpi. Upstream to CAT1, abscisic acid responsive

PYR/PYL was up regulated in Sp-S at both time points but in Sp-R it showed contrasting trend of up regulation at 24hpi downregulation at 48hpi. WRKY2229 and

WRKY33 were found downregulated only in Sp-R and wasn’t differentially expressed in

S. MPK1/2/7/14, ChiB, RTE1 and ACS6 were found up regulated in both Sp-R and Sp-

S at 48hpi but were not found in Sp-S at 24hpi. Cell death related OX1 was found upregulated only in Sp-R at 24hpi and RbohD was only found upregulated at both time points in Sp-S (Figure 2-9).

Most of the plant hormone signal transduction pathway was represented by

DEGs in both genotype with only few differences. Auxin responsive AUX1 was found upregulated in Sp-R and down regulated in Sp-S at both time points. Contrastingly, GH3 and JAZ were down in Sp-R and up in Sp-S. NPR1 was downregulated in both Sp-R and Sp-S whereas its regulator TGA transcription factor was found up only in Sp-S at

48hpi. Downstream TGA there are PR1 genes possibly responsible for disease resistance were up in both Sp-R and Sp-S at both time points (Figure 2-10).

Visualization of DETs in Plant Pathways

Pathview. Most of the DEGs showed common trend of up or down regulation in both

Sp-R and Sp-S at each time point in general physiological pathways as well as plant defense related pathways with few exceptions.

among S DEGs. Some core plant defense related components, like PR1, ERF1,

EBF1/2, PP2C, ETR/ERS were found upregulated, whereas, SnRK2 and CAM4 were commonly down regulated in both Sp-R and Sp-S accessions at both time points.

PYR/PYL was up regulated in Sp-S at both time points but in Sp-R it showed contrasting trend of up regulation at 24hpi downregulation at 48hpi. WRKY2229 and

WRKY33 were found downregulated only in Sp-R and wasn’t differentially expressed in

S. MPK1/2/7/14, ChiB, RTE1 and ACS6 were found up regulated in both Sp-R and Sp-

S at 48hpi but were not found in Sp-S at 24hpi. Cell death related OX1 was found upregulated only in Sp-R at 24hpi and RbohD was only found upregulated at both time points in Sp-S (Figure 2-9).

Most of the plant hormone signal transduction pathway was represented by

48hpi. Downstream TGA there are PR1 genes possibly responsible for disease resistance were up in both Sp-R and Sp-S at both time points (Figure 2-10).

Validation of RNA-seq Results with qRT-PCR

To verify differential expression analysis based on RNA-seq data, transcriptional levels of selected DEGs that were potentially involved in defense against P. parasitica and represent different expression profiles across P. parasitica treated and control samples of both Sp-R and Sp-S, were determined by qRT-PCR analysis. For example,

2-oxoglutarate (2OG) and Fe (II)-dependent oxygenase (Solyc07g043420.2), lipoxygenase 1 (Solyc08g029000.2), heat shock protein 70 (Solyc09g010630.2), elongation factor (Solyc11g069700.1) and papain family cysteine protease

(Solyc04g080960.2). Relative expression values of all the tested genes were calculated by using constitutively expressed Actin gene. Expression profiles of all tested genes resulted from qRT-PCR agreed with our RNA-seq data analysis results (Figure 2-13).

Discussion

Solanum pimpinellifolium is a wild tomato species that is considered among the closest relatives of the cultivated tomato. Different accessions of S. pimpinellifolium have been reported to carry resistance against several diseases that threaten the production of cultivated tomato (Cui et al., 2017). In this study we have investigated transcriptional reprogramming triggered by P. parasitica in resistant (Sp-R) and susceptible (Sp-S) S. pimpinellifolium accessions with an aim to identify genes putatively involved in resistance development in Sp-R against P. parasitica.

Differential expression analysis was done on treatments vs. mock samples of both Sp-R and Sp-S genotypes over both time points, to identify the specific expression changes brought out by P. parasitica in both accessions during biotic and switch to necrotrophic phases of infection. Major transcriptional reprogramming of plant defense related mechanisms as well as diverse cellular and metabolic processes was observed

in both Sp-S and Sp-R in response to P. parasitica infection (Figure 2-7, Table 2-4).

Number of identified DEGs in treatment vs. control comparisons were higher in Sp-S so does the number of assigned GO terms (Table 2-3), depicting more intense transcriptional reprogramming during susceptible interaction compared to resistance development against P. parasitica. The mechanism of plant’s response to pathogens is a complex phenomenon that involves interconnected network of changes in several regular physiological processes (Jones and Dangl, 2006) that is clearly reflected from the transcriptome based studies of several diverse plant-pathogen systems (Adhikari et al., 2012) ; (Judelson et al., 2008); (Kunjeti et al., 2012); (Gao et al., 2013); (Chen et al.,

2014). Regarding plant defense, we found differential expression of several genes involved in multiple plant immunity related mechanisms in both Sp-R and Sp-S.

However, our focus was to identify induced genes specifically related to development of resistance response in Sp-R against Phytophthora parasitica.

Antimicrobial Genes in Relevance to Resistance Response Against P. parasitica

The class one PR genes are well known for their involvement in plant defense against multiple pathogens. Overexpression of PR-1a gene in transgenic tobacco plants resulted in enhanced resistance against P. parasitica (Alexander et al., 1993) whereas, transgenic tobacco plants with silenced PR-1 genes were found to be highly susceptible to P. parasitica as compared to the controls plants (Riviere et al., 2008). Thus, it could be speculated that PR-1 act as a positive regulator of plant resistance against P. parasitica. PR-1 is also considered as SA signature and its enhanced expression is linked to systemic acquired resistance (SAR) (Ali et al., 2018). In the present study, we found induction of a PR-1 gene only during the resistance response against P. parasitica in Sp-R accession (Figure 2-8).

Other PR genes encoding antimicrobial compounds for example, proteases, chitinases and proteinase inhibitors constitute the basal defense in plants. Antimicrobial activities essentially serve as a first line of host’s defense against the invading pathogens. Many of them are highly induced upon perception of pathogen’s elicitors or

MAMPs thus, could be considered as a component of PTI (Jashni et al., 2015).

Chitinases are considered as antifungal compounds because they have been reported to degrade chitin in the fungal cell walls, but they have also been found effective against oomycetes (Liu et al., 2017); (Osorio-Hern et al., 2016). Four different chitinases were found highly induced in Sp-R upon P. parasitica treatment (Figure 2-8).

A chitinase gene (ChiIV3) in Capsicum annuum was found to be induced by

Phytophthora capsici was shown to have an inhibitory effect on the growth of P. capsici.

Moreover, ChiIV3 was also found to act as a chitin (unknown) receptor that trigger cell death and defense signaling upon perceiving P. capsici attack (Liu et al., 2017). In addition to plants, chitinases from a biocontrol organism from Trichoderma spp. have also been found effective against P. parasitica (Osorio-Hern et al., 2016). Enhanced induction of chitinases specifically during resistance outcome against P. parasitica make them potential anti-Phytophthora candidates.

Protease inhibitors have been proposed to protect the plant against pathogen’s proteins and proteases. Antimicrobial activity of plant proteinase inhibitors was first observed in tomato-Phytophtora interaction where induced trypsin and chymotrypsin proteinase inhibitors were found to correlate with resistance response of plant against

P. infestans (Kim et al., 2009). PAGE analysis revealed significant induction of biotic stress related genes in both Sp-R and Sp-S but response to wounding, stress and

external stimulus were significantly induced only in Sp-R during resistance outcome against P. parasitica (Figure 2-7). Further dissection of these exclusively enriched GO terms revealed significant upregulation of antimicrobial activity related genes like wound induced SIPs, Kunitz family protease inhibitor protein and defensin (Figure 2-8). Four and six different SIPs homologues were specifically induced in Sp-S and Sp-R tomatoes during susceptible and resistant interaction with P. parasitica. Significant induction of

SIPs was observed in potato in response to P. infestans and wounding.

Plant defensins is another important subfamily of antimicrobial proteins.

Defensins are a family of small diverse antimicrobial peptides ubiquitous to throughout the plant kingdom. They are well known for their role in plant defense against a wide variety of pathogens including different Phytophthora species. Transgenic tomato expressing a chili defensin gene showed increased resistance against P. infestans

(Zainal et al., 2009). A wild tobacco defensin (NmDef02) presented very strong antimicrobial activity against P. parasitica. Transgenic tobacco and potato plants expressing this NmDef02 gene showed enhanced resistance to P. parasitica. and P. infestans (Portieles et al., 2010). In our study six SIPs and one defensin were specifically induced in Sp-R accession during resistant response against P. parasitica.

Significant induction of SIPs and defensin genes specifically in Sp-R make them potential candidates that might have some essential contribution in resistance development by serving as antimicrobial agents against P. parasitica.

Contrasting Expression of Downy Mildew Susceptibility Factor in Response to P. parasitica

Another gene (Solyc10g086580.1) annotated as 2-oxoglutarate (2OG) and Fe

(II)-dependent oxygenase protein presented a contrasting trend; significant induction in

Sp-R and repression in Sp-S during resistant and susceptible responses to P. parasitica infection respectively (Figure 2-4 and 2-8). In Arabidopsis, a downy mildew resistance gene (DMR6) encoding a 2-oxoglutarate (2OG) and Fe (II)-dependent oxygenase and other DMR6 like oxygenases (DLO1 and DLO2) are reported to be associated with plant defense but these reported to be required for susceptibility of Arabidopsis to downy mildew pathogen Hyaloperonospora parasitica. Arabidopsis mutants lacking DMR6 showed enhanced resistance to Phytophthora capsici. Enhanced expression of DMR6 was observed both in case of compatible and incompatible Arabidopsis-H. parasitica interactions with more strong activation during early stages of incompatible interaction compared to compatible interaction (Van Damme et al., 2008). Induced expression of an immune suppressor of related oomycete pathogens leading to resistance response to P. parasitca could be attributed to enhanced defense leading to pathogen arrest in

Sp-R whereas decreased expression in case of susceptible response could be either a counter defense strategy executed by P. parasitca or repression of a susceptibility factor by the plant to execute defense. Whether the tomato DMR6 identified here is a susceptibility factor for P. parasitca or not, cannot be concluded solely on the basis expression studies.

Induction of Plant Heat Stress Tolerance Related Genes

A gene encoding GTP binding elongation factor Tu (Solyc11g069700.1) was found highly induced only in Sp-R during resistance development against P. parasitica.

These kind of elongation factors serve as translation factors in protein biosynthesis and have been reported to play an essential role in plant heat stress tolerance ( (Li et al.,

2018) but their role in plant defense response is unknown. Another component of plant heat stress tolerance, HSP70 a heat shock protein (Solyc09g010630.2) was found

significantly upregulated in both Sp-S and Sp-R but highest induction was seen in Sp-R at 24hpi. HSP70 is reported to play an essential role in plant defense by interacting with

MAPKs and is required for the induction of INF1 (a P. infestans elicitor) mediated cell death (Kanzaki et al., 2003). Recently, in planta association between HSP70 and a P. infestans effector Pi23226 was shown to induced cell death in Nicotiana benthamiana

(Lee et al., 2018).

R Genes

Sp-R accession presented HR like response against P. parasitica infection that is generally considered an event of ETI activated by recognition of specific pathogen effector by particular R gene. Interestingly most of the R genes were either found to be downregulated or very sparsely expressed in Sp-R upon treatment with P. parasitica.

On the contrary many R genes were found highly upregulated in the susceptible accession Sp-S upon P. parasitica infection. A non traditional R gene called Mlo, was found highly suppressed only in Sp-R at 24hpi and slightly induced at 48hpi as well as during susceptible interaction of Sp-S to P. parasitica (Figure 2-8). Mlo is widely known as susceptibility factor for powdery mildew disease. Loss of function of a tomato Mlo resulted in enhanced resistance against powdery mildew (Bai et al., 2008). From our data, it could be speculated that resistance outcome in Sp-R could be the result of recessive resistance instead of dominant R gene mediated resistance (Sun et al.,

2016).

Transcriptional reprogramming of many cellular and physiological processes and all components of biotic stress responses especially high induction of several defense related genes, suppression of susceptibility factors, modulation of phytohormone signaling, proteolysis and signaling in Sp-R accession indicates significant activation of

multiple defense mechanisms leading to resistant outcome against P. parasitica.

Functional significance of defense related genes identified here needs further investigation.

Table 2-1. qRT-PCR primers for selected genes. Transcript ID Description Primers Solyc07g043420.2 2- GGAGGCAGGGTTCTTTGACA oxoglutarate ATGCATTGCGAATAGGCTGC and Fe dependent oxygenase Solyc08g029000.2 lipoxygenase AGCACACCCGATGAGTTTGA 1 TGAAGAACTTGAGGTGTTGGGA

Solyc09g010630.2 heat shock ATGGCTGGAAAGGGTGAAGG protein 70 ACGCTCAGAGTCGGTGAATC

Solyc11g069700.1 Elongation TGTGCCGATTTCTGGTTTCG factor Tu ACGACCAACAGGAACAGTCC family

Solyc04g080960.2 Papain GACTGGCGTGAAAAAGGAGC family GGGTCACACTCATGGTCACA cysteine protease

Table 2-2. Summary statistics of RNA-seq data and mapping results. Group 1, 2 and 3 comprise of mock, 24 hpi and 48 hpi P. parasitica treated samples of resistant accession (Sp-R) whereas 4, 5 and 6 represent mock, 24 hpi and 48 hpi P. parasitica treated samples of susceptible accession (Sp-S) respectively.

Grou Sample name Raw reads Clean Reads Mapped Uniquely No: of p reads (%) mapped mapped (%) genes

1 R-mock (Rep 1) 27,886,208 23,047,994 97.93 81.06 21,436 R-mock (Rep 2) 29,938,504 24,872,121 97.93 77.35 21,374 2 R-24hpi (Rep 1) 25,731,844 21,627,411 97.97 85.44 21,222 R-24hpi (Rep 2) 27,210,026 22,621,023 97.96 79.42 21,175 3 R-48hpi (Rep 1) 27,175,390 22,690,068 97.67 76.36 21,356 R-48hpi (Rep 2) 29,441,742 24,775,792 97.82 80.97 21,533 4 S-mock (Rep 1) 19,181,092 16,133,709 97.59 76.95 20,966 S-mock (Rep 2) 27,938,220 23,227,317 97.07 84.29 21,381 5 S-24hpi (Rep 1) 30,978,162 26,197,946 97.31 79.66 21,429 S-24hpi (Rep 2) 28,387,338 23,713,599 97.53 79.46 21,280 6 S-48hpi (Rep 1) 28,563,832 23,854,528 96.27 76.36 21,631 S-48hpi (Rep 2) 26,601,354 21,881,482 97.36 77.45 21,565 Total 329033712 274,642,990

Table 2-3. Number of DEGs found among all comparisons. Total Up Down DEGs unique DEGs regulated regulated to the DEGs DEGs comparison S-24hpi treated vs S-control 3079 1132 1947 775 S-48hpi treated vs S-control 1919 811 1108 152 R-24hpi treated vs R-control 2657 868 1789 519 R-48hpi treated vs R-control 1836 606 1230 293 4598* R vs. S R-control vs S-control 322 155 167 178 R-24hpi treated vs S-24hpi treated 1158 287 871 589 R-48hpi treated vs S-48hpi treated 889 207 682 302 1681* * Indicates total number of unique DEGs among all comparisons.

Table 2-4. Comparison of GO term enrichment analysis (by SEACOMPARE) of DEGs found in P. parasiticat infected vs. control samples of both resistant (R) and susceptible (S) wild tomato genotypes. Significance levels are based on enrichment and lowest FDR values with a cutoff of < 0.05 that is indicated by color model mapping significance on a red to yellow scale. Highly significant terms indicated by red. GO Term Description Color No: of Model DEGs R S R S GO:0003824 catalytic activity 899 974 GO:0004866 endopeptidase inhibitor activity 19 --- GO:0030414 peptidase inhibitor activity 20 --- GO:0019825 oxygen binding 52 57 GO:0016740 transferase activity 337 344 GO:0016798 hydrolase activity, acting on glycosyl bonds 47 53 GO:0016491 oxidoreductase activity 188 214 GO:0004364 glutathione transferase activity --- 18 GO:0004553 hydrolase activity, hydrolyzing O-glycosyl compounds 35 42 GO:0016758 transferase activity, transferring hexosyl groups 75 64 GO:0046527 glucosyltransferase activity 49 41 GO:0016757 transferase activity, transferring glycosyl groups 91 82 GO:0016168 chlorophyll binding 15 16 GO:0008194 UDP-glycosyltransferase activity 45 40 GO:0046906 tetrapyrrole binding 21 24 GO:0035251 UDP-glucosyltransferase activity 34 32 GO:0016705 oxidoreductase activity, acting on paired donors, with 46 45 incorporation or reduction of molecular oxygen GO:0004091 carboxylesterase activity 62 --- GO:0016759 cellulose synthase activity 15 --- GO:0016614 oxidoreductase activity, acting on CH-OH group of 37 41 donors GO:0015103 inorganic anion transmembrane transporter activity 22 25 GO:0005372 water transmembrane transporter activity 14 15 GO:0015250 water channel activity 14 15 GO:0022891 substrate-specific transmembrane transporter activity --- 127 GO:0022892 substrate-specific transporter activity --- 149 GO:0005215 transporter activity --- 178 GO:0022857 transmembrane transporter activity --- 147 GO:0022804 active transmembrane transporter activity --- 86 GO:0015291 secondary active transmembrane transporter activity --- 51 GO:0019203 carbohydrate phosphatase activity --- 11 GO:0008509 anion transmembrane transporter activity --- 34 GO:0015171 amino acid transmembrane transporter activity --- 19 GO:0016616 oxidoreductase activity, acting on the CH-OH group of --- 34 donors, NAD or NADP as acceptor GO:0005275 amine transmembrane transporter activity --- 20

A B

C 160 **

140

120 ** 100

80 60

40 spots Numberof inoculated 20

0 HR Disease Nothing

R S

Figure 2-1. Disease Symptoms and HR response. A) HR on the leaves of S. pimpinellifolium accession (Sp-R) resistant to P. parasitica and B) Disease symptoms on S. pimpinellifolium susceptible (Sp-S) accession in response to P. parasitica. C) Disease and HR spots on Sp-S and Sp-R leaves after 72 hours of P. parasitica inoculation. * * indicates p < 0.05. Photo courtesy of Zunaira Afzal.

A B

C D

Figure 2-2. Venn diagrams presenting the overlap of expressed genes between resistant (Sp-R) and susceptible (Sp-S) wild tomato accessions. Pink circle represents R and blue is for S samples. A) Comparison of P. parasitica infected R and S samples at 24hpi. B) Comparison of P. parasitica infected R and S samples at 48hpi. C) Comparison of mock inoculated R and S samples. D) Combined comparison between all (mock and infected) R and S libraries.

Figure 2-3. Principle component analysis (PCA) plot. R and S genotype factor showed variation among all their treated mock samples. Green oval is encompassing all R samples and red oval within the green oval is showing infected R samples at each time point. Yellow oval has enclosed all S samples within that red is encircling the infected ones. Color key presenting the sample groups is given at the left. Each colored dot represents a single sample. Replications are indicated by same color dots.

Figure 2-4. Venn diagram analysis of DEGs. DEGs from treatment vs control comparisons of both Sp-R and Sp-S at 24 and 48hpi. Pink ovals represent Sp-S treatment vs control and blue represents Sp-R treatment vs. control comparisons.

Solyc07g066310.2

Solyc02g085950.2

Solyc07g049530.2 Solyc02g082920.2 Solyc03g034220.2

(unknown)

(CAT2)

(unknown)

(ARA12)

(ASN1)

) (unknown

(PR)

(BGAL1)

(PA2)

(OSM34) (CHI3)

(MAN1)

(BG1)

(CDSP32)

(ACO1)

(PR)

(CAP/PR) (GLDP1)

(LHCA3)

(D/ARF)

(CHI

(LHCA1)

(MEE14)

(unknown)

(PR)

(PSBR)

(LOX1)

(CuAO)

(GCN5)

(HSP70)

(RD19)

(ALFAVPE)

(RBSC

) (MLP423

(PRK)

(FNR2)

(CAX)

(FTSH5) (CP12)

(GSTU25)

(STM)

(THI1)

(RBSC

(EF

(PAL1)

(2OG)

) (RCA

(CAB)

(TIM)

Tu)

2A)

3B)

Sp-S- Control-1 Sp-S- Control-2 Sp-R- Control-1 Sp-R- Control-2 Sp-S-48hpi- 1 Sp-S-48hpi- 2 Sp-R-48hpi- 1 Sp-R-48hpi- 2 Sp-S-24hpi- 1 Sp-S-24hpi- 2 Sp-R-24hpi- 1 Sp-R-24hpi- 2

Figure 2-5. Heat map showing hierarchical cluster analysis of top 50 highly expressed genes across all samples. It shows that many core plant defense genes have differential expression profiles both among treatment vs. control as well as Sp-R vs. Sp-S conditions. and Gradient scale is representing expression levels with red showing highest expression to blue with lowest expression.

A Sp-R

B Sp-S

Sp-R vs. Sp-S

GO Annotations

Figure 2-6. Overview of GO annotations assignment of DEGs by SEA tool of AgriGO. Percent distribution of DEGs across enriched GO terms for cellular component, molecular function and biological processes for all DEGs found in P. parasitica treated vs. control samples of Sp-R genotype at both 24 and 48hpi (A) for Sp-S genotype (B) DEGs identified by Sp-R treated vs. Sp-S treated samples at both time points (C). Green bars represent reference background (S. lycopersicum ITAG 2.4) used to calculate enrichment and blue bar represent input DEGs.

Figure 2-7. Gene set enrichment analysis by PAGE tool of AgriGO. A). Bar chart representing GO annotations across all groups. Z-scores plotted on Y axis were calculated using fold change values of DEGs. Six bars are representing six differential expression comparisons among the samples. Starting from the left, first and second bars represent Sp-R-treated vs. Sp-R-control comparison at 24 and 48hpi respectively, third and fourth bars are depicting Sp-S-treated vs. Sp-S-control at 24 and 48hpi respectively and fifth and sixth bars are for Sp-R- treated vs. Sp-S-treated at 24 and 48hpi respectively. Negative Z-score depicts down regulation, represented by blue bars whereas positive Z-score means upregulation, shown as red bars. Star indicates plant defense related annotations. (B & C) Comparative (Resistant vs. susceptible) GO biological process hierarchal analysis based on gene set enrichment analysis at 24hpi (B) and 48hpi (C) Significantly enriched upregulated terms are shown in color gradient red to yellow, red means highest level of upregulation and yellow means slightly upregulated. Downregulation is showed in blue color gradient. Boxes that have double borders means that gene set is significantly enriched in both R and S samples.

Key Low High Up-regulation

Down-regulation Low High

Significant in both columns C

Figure 2-7. Continued

Solyc09g089510.2 Solyc09g084490.2 Solyc03g093790.1 Solyc09g084480.2 PAL1 Solyc09g007920.2 Solyc07g009260.2defensin Solyc09g089530.2 PR-1 Solyc09g006010.2 CH Solyc10g074440.1 Solyc09g084460.2 Solyc09g084440.2 EF Solyc11g069700.1 Solyc09g084450.2 Solyc02g082930.2CH

Solyc02g064570.1VQ Solyc07g043420.22OG Solyc09g083440.2 Solyc08g080610.2OSM34 Solyc02g082960.2CH Solyc09g010630.2HSP70 HY Solyc01g057000.2 Solyc03g095650.2MLO R gene Solyc06g008790.2 Solyc08g080630.2 Solyc04g071600.2ASR Solyc10g086100.1 Solyc03g093860.1HY CH

Solyc05g050130.2

- - -

24hpi

24hpi 24hpi

48hpi

24hpi

48hpi

Control

- -

Sp Sp Sp Figure 2-8. Heatmap showing expression profiles of core defense genes that could be putatively involved in resistance development against P. parasitica in Sp-R. Gradient scale is showing z scores of DEGs where red represents most induced expression and blue depicts highest repression. Gene identifiers with red stars on the right side are protease inhibitors. Other abbreviations used are: CH- chitinases, HY- hydrolases, VQ-VQ motif containing gene, 2OG- 2OG-Fe oxygenase, EF- GTP binding Elongation factor Tu, OSM34- osmotin34, HSP70-heatshock protein 70, ASR- Abscisic acid stress ripening 5, PAL1- PHE ammonia lyase 1, and R gene- NB-ARC domain-containing disease resistance gene.

Figure 2-9. Visualization of DEGs involved in MAPK signaling pathway. Each box is depicting DEGs from four treatments vs. control comparisons thus divided into four segments, left half is representing Sp-R treatment vs. control with first segment (left) 24hpi and second 48hpi. Right half is depicting Sp-S treatment vs. control with first segment (left) 24hpi and second (right) 48hpi. White box or no color fill means no DEG was assigned to that KO term. Color gradient represents log2 fold ratios with red representing upregulation and blue representing downregulation in treatments vs. mock samples. White segment in between colors means that term was not differentially expressed in the respective comparison.

Figure 2-10. DEGs mapped on plant hormone signal transduction pathway. Each box is depicting DEGs from four treatments vs. control comparisons thus divided into four segments, left half is representing Sp-R treatment vs. control with first segment (left) 24hpi and second 48hpi. Right half is depicting Sp-S treatment vs. control with first segment (left) 24hpi and second (right) 48hpi.

White box or no color fill means no DEG was assigned to that KO term. Color gradient represents log2 fold ratios with red representing upregulation and blue representing downregulation in treatments vs. mock samples. White segment in between colors means that term was not differentially expressed in the respective comparison.

Figure 2-11. Visualization of differentially expressed genes in wild tomatoes in response to P. parasitica on biotic stress pathway, A) Resistant accession, B) Susceptible accession. Color gradient represents log2 fold ratios with red representing upregulation and blue representing downregulation in treatments over controls.

Solyc07g063710.1 Solyc09g007510.1 Solyc040796690.2 Solyc07g063750.2 Solyc07g066550.2 Solyc02g072070.2 Solyc09g007730.2 Solyc02g079590.2 Solyc01g007960.2 Solyc07g055690.1 Solyc07g005110.2 Solyc02g080080.2 Solyc07g063730.1 Solyc09g011330.1 Solyc02g078530.2 Solyc12g006840.1 Solyc05g005050.2 Solyc02g079530.2 Solyc03g111540.1 Solyc05g005060.2 Solyc09g075910.1 Solyc12g036320.1 Solyc05g008310.2 Solyc12g087920.1 Solyc05g005070.2 Solyc03g083480.2 Solyc03g083470.2 Solyc02g080040.2 Solyc02g080050.1 Solyc04g007380.1 Solyc02g068300.2 Solyc02g086590.2 Solyc02g071820.2 Solyc12g014350.1 Solyc04g077310.1 Solyc02g086210.1 Solyc04g015460.2 Solyc01g008500.2 Solyc02g071870.2 Solyc07g055630.2 Solyc01g006520.2 Solyc02g079990.2 Solyc02g079550.1 Solyc07g053120.2 Solyc07g055640.1 Solyc02g080010.2 Solyc02g079560.1 Solyc01g094830.2 Solyc05g009090.1 Solyc04g077390.2 Solyc04g077340.2 Solyc08g079460.2 Solyc10g076760.1 Solyc02g071810.2 Solyc02g078750.2 Solyc08g076050.2 Solyc07g063770.2 Solyc03g005960.2 Solyc04g077270.2 Solyc04g079690.2 Solyc01g007980.2

Solyc06g060690.2

1 2 2

- - -

48hpi

24hpi

48hpi

24hpi

48hpi 48hpi

24hpi

- -

Control

R Control

Control

R S

- -

Sp Sp

Sp Sp Sp

Figure 2-12. Heatmap showing expression profiles of DEGs encoding DUF26 type receptor like kinases (RLKs). Gradient scale is showing z scores of DEGs where red represents most induced expression and blue depicts highest repression.

Solyc08g029000.2 Solyc07g043420.2 b LOX 1 2OG-Fe oxygenase a 1200 800 2.5 3000

1100 700

1000 2500

2 900 b 600

800 2000

500 a

700 1.5 a 600 400 1500

a RPKM RPKM 500 1 c 300 400 1000 300 c 200 0.5

200 500 Relativevalues expression Relative expression values expression Relative a 100 a 100 a a 0 0 0 0 Sp-S-Control Sp-S-24hpi Sp-S-48hpi Sp-R-Control Sp-R-24hpi Sp-R-48hpi Sp-S-Control Sp-S-24hpi Sp-S-48hpi Sp-R-Control Sp-R-24hpi Sp-R-48hpi Relative Normalized Expression RPKM Relative Normalized Expression RPKM

Solyc09g010630.2 Solyc11g069700.1 HSP70 b b Elongation Factor 3 3500 8 1500 7.5 7 1300 3000 2.5

a 6.5

6 1100

2500

5.5 2 a 5 900

2000 4.5

RPKM RPKM 1.5 a a 4 700 a 1500 3.5 3 a 500 values expression 1 2.5 1000 2 a 300

Relativevalues expression 1.5 a 0.5 500 a Relative 1 100 0.5 a 0 0 0 -100 Sp-S-Control Sp-S-24hpi Sp-S-48hpi Sp-R-Control Sp-R-24hpi Sp-R-48hpi Sp-S-Control Sp-S-24hpi Sp-S-48hpi Sp-R-Control Sp-R-24hpi Sp-R-48hpi Relative Normalized Expression RPKM Relative Normalized Expression RPKM Solyc04g080960.2 Cysteine Protease 2 2500

2000 1.5

1500 a a

1 RPKM

1000

0.5 b Relativevalues expression b 500 b b

0 0 Sp-S-Control Sp-S-24hpi Sp-S-48hpi Sp-R-Control Sp-R-24hpi Sp-R-48hpi Relative Normalized Expression RPKM

Figure 2-13. qRT-PCR based validation of plant defense related DEGs of wild tomato in response to P. parasitica. Expression levels of tested genes were normalized based on transcript levels of Actin gene. RPKM values calculated from RNA- seq are compared to relative expression values determined by qRT-PCR analysis. Relative expression values of infected samples were determined by using the average expression value of all replicates of a particular group.

CHAPTER 3 TRANSCRIPTOME PROFILE OF CARRIZO CITRANGE ROOTS IN RESPONSE TO PHYTOPHTHORA PARASITICA INFECTION

Introduction

Citrus is one of the most important commercial fruit crops, grown in more than hundred countries in tropical and subtropical areas (Talon and Gmitter, 2008).

Altogether, citrus production is adversely impacted by numerous diseases, but major yield-affecting citrus diseases are distinct among different geographical regions of the world depending upon the climate, management practices and type of scion and root stocks (Davis, 1982). The Phytophtora disease complex on citrus is globally known to be the most damaging soil-borne disease of citrus. Phytophthora parasitica, P. palmivora and P. citrophthora are the three major Phytophthora species that can cause root rot, foot rot and sometimes in severe cases gummosis and brown rot of fruit (Savita and Nagpal, 2012). P. parasitica Dastur, also known as P. nicotianae, is the most widespread citrus infecting species throughout the tropical and subtropical areas of the world and is present in most citrus growing groves of Florida (Gade and Lad, 2018);

(Zitko and Timmer, 1994).

Phytophthora root rot is a serious problem of citrus particularly during nursery propagation of susceptible rootstocks, causing root decay and death of seedlings. Root rot in bearing orchards can cause slow tree decline and yield losses (Savita and

Nagpal, 2012). Swingle citrumelo and Carrizo citrange are considered Phytophthora resistant rootstocks (Graham, 1995); (Hutchison, 1974).

 This chapter has been republished with permissions from: Afzal Naveed, Z., Huguet-Tapia, J.C., and Ali, G.S. (2019). Transcriptome profile of Carrizo citrange roots in response to Phytophthora parasitica infection. Journal of Plant Interactions 14, 187-20

In general, these rootstocks offer considerable resistance to Phytophthora diseases, but this resistance is usually broken down under disease complex situations when insect and multiple pathogen infestations acts in synergism causing severe damage to citrus trees. For example, the Phytophthora-Diaprepes complex, in which damage of citrus roots by larvae of Diaprepes root weevil tends to aggravate

Phytophthora root rot, renders even highly resistant rootstocks susceptible to

Phytophthora diseases (Graham et al., 2003). Interaction between Phytophthora root rot and citrus greening (also known as Huang Long Bing, HLB), which is the most devastating citrus disease in Florida, has also been observed (Graham et al., 2007).

Phytophthora spp. prefers to infect new roots and HLB favors new growth of fibrous roots thus making this disease complex devastating for citrus trees. Understanding molecular bases of citrus defense against Phytophthora and its relevance to other citrus pathogens is required for designing long-term resistance development strategies to combat losses caused by these noxious pathogens.

The first citrus disease caused by P. parasitica was reported back in 1832, well before the advent of the discipline of plant pathology. Being so ancient and a widespread problem, a lot of research had been conducted to understand the biology, etiology and control of this disease. However, little is known about the molecular basis of citrus-Phytophthora interactions. A foliar microarray based differential gene expression analysis among resistant and susceptible hybrids and their resistant

(Poncirus trifoliata cv. Rubidoux) and susceptible (Citrus sunki) parents was conducted in response to stem inoculation of P.parasitica (Boava et al., 2011). Several genes that were differentially expressed during these comparisons were suggested to play a role in

the resistance or susceptibility of these hosts to citrus gummosis. Recently, Dalio et al. reported the molecular basis of compatible and incompatible citrus-P. parasitica interaction by using susceptible (Citrus sunki) and resistant (Poncirus trifoliata) citrus rootstocks, respectively, by comparing the gene expression of defense related host genes through quantitative PCR. They have shown significant induction of key defense genes in susceptible rootstock whereas little or no changes in the expression level of tested defense genes in the resistant rootstocks were observed by them in response to pathogen (Dalio et al., 2017). Similarly, on the pathogen side, transcriptional profiling of

P. parasitica during citrus gummosis was conducted to identify pathogenicity factors during infection (Rosa et al., 2007). These reports of citrus-Phytophthora interactions are either confined to the above ground infections where mostly leaf tissues were used for transcriptome analysis or restricted to few defense related genes in roots, but so far, no study has been done to determine whole transcriptional profiling of citrus roots during Phytophthora root rot.

In this study, using the RNA-seq approach, we have analyzed whole transcriptome of Carrizo citrange rootstock in response to P. parasitica at 24 and 48 hours post inoculation (hpi) to determine the transcriptional changes during this tolerant citrus-Phytophthora interaction. Our major focus was to identify citrus genes potentially involved in defense particularly R genes. Findings of our analyses provide insights into how citrus roots provide resistance to Phytophthora, and these findings provide a strong basis to further explore the molecular basis of tolerance offered by Carrizo rootstock.

Potential uses of the outcomes of our analyses for developing rootstocks that are tolerant to Phytophthora, especially in the context of disease complexes are discussed.

Materials and Methods

P. parasitica Inoculation, Sample Collection and Sequencing

P. parasitica strain 13-723A was isolated from citrus plants grown in local nursery in Florida. It was grown on 10% solid V8 medium plates at 12:12 hours dark:light cycle at 25 oC for 10 days. Zoospores were induced by applying dehydration stress on cultures from day 5 to day 7. On 10th day of culture, plates were flooded with sterilized water to induce release of zoospores. Zoospores were filtered through two to four layers of cheesecloth and adjusted to 105 cells/ml. Roots of sixty days old seedlings of Carrizo citrange rootstock grown under controlled conditions in sterilized soil were inoculated with 50 ml zoospore suspension. Mock inoculation was done using 50 ml sterilized water without zoospores.

Root samples from pathogen and mock treatments were collected at 24 hours post inoculation (hpi) and 48 hpi. Roots were rinsed in sterile water, immediately frozen in liquid nitrogen and stored at -80 0C until further use. Total RNA was extracted from frozen root tissues and six cDNA libraries were constructed and sequenced using the paired-end (PE) 125 x 2 sequencing reactions using the Illumina HiSeq 2500 platform.

RNA-Seq Quality Assessment

All raw reads were assessed for quality using FastQC (v0.11.5) and the CLC

Genomics Workbench 9.0 (Qiagen). All PE raw reads were processed for trimming the

Illumina sequencing adaptors and low-quality reads with the Trim Galore software

(version 0.4.3).

Reference Genome Mapping

Clean reads, after removing adaptors and low-quality reads, from all samples were mapped to two Citrus sinensis genomes (Ridge Pineapple and Valencia) and to

one Swingle citrumelo (Citrus x paradisi x Citrus trifoliata) genome separately. Samples containing Phytophthora parasitica reads were mapped to the P. parasitica INRA-310 genome (BROAD institute). Reference genome mapping was done using the bowtie2 algorithm of tophat aligner (Trapnell et al., 2009;Langmead et al., 2009).

De novo Assembly and Differential Expression Analysis

Pathogen contamination in the P. parasitica treated samples libraries were removed by filtering out the reads mapped to P. parasitica reference genome. Clean reads from all seven samples were de novo assembled using Trinity (Grabherr et al.,

2011) to get a pooled de novo reference assembly for citrus. Although pathogen reads were filtered before generating pooled host assembly, pathogen contamination were still there. In order to remove that, a local BLAST search was done against two customized databases based on all available Citrus and Phytophthora spp. genomes (Table 3-1) and all putative pathogen transcripts were removed.

Assembly optimization was done by following the previously described approach with few amendments (McCann et al., 2017). Briefly, all Trinity assembled transcripts were fed to TransDecoder package (Haas et al., 2013) for the prediction of open reading frames (ORFs). Longest ORF peptide sequences were then subjected to

BLASTp against uniport protein database as well as scanned for protein domains using

Pfam. Results from BLASTp and Pfam were integrated to retain the coding transcripts and rest were all filtered out. Isoforms having more than 98% similarities were removed to keep only the unique coding transcripts.

Optimized assembly was used as a reference for further analysis of citrus transcriptome and cleaned paired end reads from all samples were subjected to transcript quantification by using RSEM (RNA-seq by Expectation Maximization)

program (Li and Dewey, 2011), which is incorporated in the Trinity pipeline. Abundance estimation steps of RSEM involve mapping of clean PE reads of each sample back to the de novo pooled assembly followed by calculation of gene expression by counting reads mapped to trinity genes and transcripts. Read counts from all samples were combined in a matrix and normalization was done using the TMM method of RSEM (Li and Dewey, 2011). Quality of pooled assembly was checked by calculating N50, ExN50 statistics (Figure 3-1) and other summary statistics (Table 3-2). Differential expression analysis among the mock and P. parasitica-infected treatments at each time point was done using the EdgeR package (Robinson et al., 2010); (Seyednasrollah et al., 2013).

Since we had two replicates of the pathogen-treated samples and one replicate for the mock-inoculated samples at 24 hpi and 48 hpi, and one replicate at the 0 hr control, the dispersion parameter in EdgeR was set to 0.1. Differential expression was analysis was done based on both coding isoforms count and unigene count. Whole analysis pipeline is summarized in Figure 3-2.

Functional Annotation of Transcripts

Differentially expressed transcripts were annotated using Blast2GO with the default parameters (Conesa et al., 2005). CloudBlast search option was used to BLAST

DETs against viridiplantae database with an expect value (e-value) of 1.0e-3. To identify

R-genes, a local BLASTx search of optimized assembly transcript sequences against all available plant R gene protein sequences in the Plant Resistance Gene database

(PRGdb) was performed with an e-value 1.0E-10 and R genes were selected if percent identity was greater than 80 and bit score was greater than 100 (Sanseverino et al.,

2012).

For GO term enrichment analysis, GO terms for DETs obtained from annotation results were fed to the Singular Enrichment Analysis (SEA) tool of AgriGO using customized parameters (Du et al., 2010). Reduce and visualize gene ontology

(REViGO) was used to summarize and visualize most significantly enriched GO terms

(Supek et al., 2011).

KEGG (Kyoto Encyclopedia of Genes and Genomes) enzyme codes for DETs were obtained through Blast2GO plugin. In addition to that all DETs were fed to KAAS

(KEGG automatic annotation server) to assign KO terms (Moriya et al., 2007) and

Pathview was used to generate KEGG pathways (Luo et al., 2017). Mapman software was also utilized to view DETs in biological processes (Thimm et al., 2004;Usadel et al.,

2009). Mercator web application (Lohse et al., 2013) and homology search for citrus orthologs were used to assign MapMan bins.

Validation of DETs by Quantitative Real Time PCR

To validate differential expression results obtained by following Trinity-RSEM-

EdgeR pipeline, ten DETs were subjected to reverse transcriptase PCR (RT-PCR) for cDNA synthesis followed by quantitative real time PCR (qRT-PCR) by following the previously published protocol (El-Sayed et al., 2015); (Patel et al., 2015). GADPH gene was used as a reference to calculate the relative expression values. Gene specific primer pairs were designed using NCBI primer BLAST tool

(https://www.ncbi.nlm.nih.gov/tools/primer-blast/). Primers used for this experiment are given in Table 3-3. Melting curve analysis was done to ensure amplification of single gene product by each primer pair.

Results

Phytophthora parasitica Infects and Colonize Carrizo Citrange Roots

Carrizo citrange has been reported to be tolerant to P. parasitica infection. To investigate whether P. parasitica infects Carrizo citrange roots or not, we inoculated roots of 60 days old plants with zoospores and followed infection dynamics studies under confocal microscope. Zoospores attachment, encystement and cyst germination on roots surface were observed at approximately 3 hpi (Figure 3-3A). Inter and intracellular hyphal growth was clearly observed at 24 hpi followed by heavy colonization at 48 hpi. Haustoria like structures were also clearly observed at 24 and 48 hpi (Figure 3-3B and C). Contrastingly, no pathogen structures were observed during P. parasitica interaction with P. trifoliata that is another resistant citrus rootstock (Dalio et al., 2017), suggesting that different rootstocks respond differently to different strains of

P. parasitica. Like all other Phytophthora species, P. parasitica is a hemi-biotrophic pathogen. Infection studies of P.parasitica in A. thaliana roots have revealed a short biotrophic phase (between 0 to 24 hpi) followed by quick switch to a necrotrophic phase at about 30 hpi (Attard et al., 2010). Although, P. parasitica infection and colonization was observed on Carrizo citrange roots, disease symptoms in aerial parts of the plants were very mild. To capture transcriptional changes in the infected roots and understand molecular basis of disease resistance, we collected infected root samples at 24 hpi and

48 hpi for RNA-seq analyses. mRNA-seq Data Analyses and Quality Assessment

Total RNA was isolated from infected and mock-treated roots, processed and sequenced using paired end (PE) 125 x 2 reaction on the Illumina HiSeq 2500 platform according to the manufacturer’s protocol. In total, 165 million PE reads (20 billion base

pairs) with an average length of 120 bp were obtained. Average Phred scores of all reads was above 30 and more than 70% reads had Phred score > 35, indicating very high-quality sequence data. After removing Illumina adaptors and low-quality reads, approximately 165 million clean reads comprised of 19.8 billion base pairs remained, providing us with sufficient depth for differential expression analyses.

Reference Genome Mapping

All four P. parasitica containing samples were mapped to the P. parasitica INRA-

310 genome. At 24hpi both replications of pathogen treated citrus roots showed 0.8% mapping coverage on P. parasitica genome. Whereas, a higher mapping coverage of

1.8% was observed at 48 hpi in both samples (Table 3-4), indicating increased growth of P. parasitica on citrus roots from 24 to 48 hpi.

Carrizo citrange is a hybrid of Citrus sinensis and Poncirus trifoliata. Currently, two C. sinensis genomes are available but none for P. trifoliate (Adhikari et al., 2012);

(Wang et al., 2014). However, draft genome of Swingle citrumelo, which is also a hybrid of C. paradisi Macf and P. trifoliata, is available (Zhang et al., 2016) and it could be used as a reference to map the Carrizo citrange reads that correspond to the P. trifoliata genome. To determine the proportions of these two, one parent (C. sinensis) and one step sibling (Swingle) genomes in Carrizo transcriptome, all read samples were separately mapped to these three reference genomes. Results of mapping to all three genomes showed only a 55-73 % coverage of the Carrizo transcriptome suggesting that a significant proportion of the Carrizo genome might be missing from these three genomes (Table 3-4). Therefore, to improve coverage, we constructed a de novo genome assembly and used it as a reference for differential gene expression analyses as follows.

De novo Transcriptome Assembly and Gene Expression Analysis

After reference genome mapping, reads mapped to P.parasitca genomes were filtered out and pathogen free clean reads from all seven samples were used to generate a pooled de novo transcriptome assembly using Trinity (Grabherr et al., 2011), one of the most widely used de novo transcriptome assembler for short reads in both animal (Hsu et al., 2017) ; (McCann et al., 2017) and plant transcriptome studies

(Evangelisti et al., 2017); (Guo et al., 2016); (Xiong et al., 2017); (Yang et al., 2017).

Altogether, 60874 trinity genes and 87299 trinity assembled transcripts were obtained with an N50 values of 1428 bp and 1796 bp, respectively (Table 3-2). In additional checks for contamination, 1043 Phytophthora genes (comprised of 1082 trinity transcripts) were removed.

Trinity transcripts are artificially assembled so chances of redundant transcripts are high. Thus, before moving further an optimization of Trinity constructed de novo assembly is necessary to get rid of false positives and to save statistical power in the downstream analysis (McCann et al., 2017); (Ono et al., 2015). Assembly optimization was done by running homology-based search using BLAST and Pfam on Transdecoder predicted ORFs (see methods for detail) and 21,278 coding unigenes comprised of

40,943 coding transcripts were obtained (Table 3-2).

To estimate transcript abundance, reads were mapped back to the optimized reference assembly and gene expression in terms of normalized FPKM (fragment per kilo base of exon per million fragments mapped) was calculated using the TMM method in RSEM. RSEM is a Trinity supported program for accurate gene expression quantification using de novo assembled reference transcriptomes (Haas et al., 2013);

(Li and Dewey, 2011). To check the quality of our reference assembly, ExN50 statistics

that refers to N50 of the top most highly expressed transcripts was computed from normalized expression data, resulting in E90N50 value of 2056 bp, which means that

90% of the top highly expressed transcripts have an N50 value of 2056 bp and indicates a very high-quality assembly (Figure 3-1).

Estimation of Variation Among Samples

To determine the reproducibility of biological replicates and variations among P. parasitica-infected and mock-inoculated samples over both time points, we performed correlation and principal components analysis (PCA) among all samples. Sample correlation matrix based on log2 transformed counts revealed that pathogen-treated replicates and mock-treated controls at both time points displayed high correlation with each other. As is shown in the dendogram in Figure 3-4A, biological replicates at each time point clustered together indicating high reproducibility of the data. Consistent with the expectations of the experimental conditions, P. parasitica-infected samples at both time-points also clustered together. Similarly, mock-treated controls at both time points also clustered together but separate from the cluster of P. parasitica-infected samples.

These results were validated with PCA analysis (Figure 3-4B). PCA plots of PC1 vs

PC2 and PC2 vs PC3 revealed very little variation among replicates at both time points

(Figure 3-4B). Similarly, mock-inoculated samples both at 24 hpi and 48 hpi clustered together. As expected, PC1 (42.9%) explained most of the variations among mock- inoculated and pathogen-infected root samples, whereas, PC2 (18.7%) explained variation among the mock-treated and the 24 hpi and 48 hpi samples (Figure 3-4).

Differential Expression Analysis of Citrus Genes in Response to P. parasitica

Differential expression analysis was performed using EdgeR package, which has been shown to work better for data with fewer replicates (Robinson et al., 2010);

(Seyednasrollah et al., 2013). Volcano plots of all pairwise comparisons presenting log fold change distribution of differentially expressed transcripts along the x-axis and False

Discovery Rate (FDR) values on the y-axis showed differential expression of numerous transcripts among all comparisons (Figure 3-5).

Differentially expressed transcripts (DETs) with log2 fold change >2 threshold and FDR < 0.005 were selected for further analysis. Overall hierarchical cluster heatmap is shown in Figure 3-6A. Hierarchical cluster analysis of DETs across all samples revealed four sub-clusters representing distinct differential expression patterns among infected and mock inoculated roots (Figure 3-6B). Sub-cluster 1 consisted of

2290 down-regulated DETs that are further down-regulated from 24 hpi to 48 hpi, sub- cluster 2 contained of 1924 up-regulated DETs that are slightly up-regulated from 24 hpi to 48 hpi and sub-cluster 3, comprised of 419 transcripts that are very highly upregulated compared to controls and continue to increase expression from 24 to 48 hpi.

Whereas, sub-cluster4 represents down-regulated DETs which have same expression at both time points (Figure 3-6B).

In the 24 hpi treatment vs. 24 hpi mock comparison, a total 3960 DETs were identified with 2193 upregulated and 1767 downregulated. In the 48 hpi treatment vs. 48 hpi mock comparison, we identified 5521 DETs with 2526 upregulated and 2995 downregulated. Whereas, only 690 DETs were observed in the 24 hpi and 48 hpi pathogen-infected comparison. Venn diagram analysis of DETs among all pairwise comparisons revealed 844 and 2175 DETs uniquely expressed in citrus roots in response to P. parasitica at 24 hpi and 48 hpi respectively (Figure 3-7).

Functional Annotation of Citrus DETs and Mapping on Plant Defense Pathways

Comprehensive functional annotation of DETs was done by using Blast2GO software (Conesa et al., 2005). Annotation statistics of DETs from all the databases is shown in Figure 3-8. Out of the total 6692 citrus DETs, BLAST hits were found for 6544

DETs. The top hits species distribution showed that majority of sequences were aligned to Citrus sinensis followed by C. clementina whereas few hits were found against other six Citrus spp. with yet available genome information (Figure 3-9).

Gene ontology (GO) terms were assigned to 5234 DETs, in which around 41% of

GO terms were associated with molecular function (MF), 36% with biological processes

(BP), and 22% with cellular components (CC). Within the MF class, most of the DETs were place in the categories of metabolic and cellular processes, response to stimulus and localization. Most represented BP categories were catalytic activity, binding and transport activity (Figure 3-10). Detailed GO term visualization indicated presence of many cytoskeleton, microtubule, cell wall related terms in all three; BP, MF and CC GO classes (Figure 3-11). Interactive graphs of most significantly enriched GO terms revealed interlinked cell wall macromolecule and polysaccharide metabolism among biological processes, and oxidoreductase activity and diverse binding related molecular functions (Figure 3-12). Most of the DETs assigned to oxidoreductase were highly upregulated at both time points.

To visualize DETs in biological pathways, transcript sequences were subjected to

KEGG annotations and KEGG orthology (KO) terms were assigned to 2510 DETs

(Figure 4). The top KEGG pathways for DETs included metabolic pathway, biosynthesis of secondary metabolites, phenylpropanoid biosynthesis and biosynthesis of antibiotics.

Plant defense related pathways like Plant hormone signal transduction, Plant-

pathogen interaction and MAPK signaling pathway also showed enough mapping of

DETs (Table 3-5). Most of the transcripts mapped on Plant-Pathogen interaction (Figure

3-13) and MAPK signaling pathway were upregulated at both time points (Figure 3-14).

In plant hormone signal transduction, auxin, cytokinine and gibberellin pathways were all downregulated whereas, ethylene (ET), abscisic acid (ABA), jasmonic acid (JA) and salicylic acid (SA) were mostly upregulated (Figure 3-15). KEGG enzyme enrichment analysis showed highest mapping on Flavanoid biosynthesis pathway (Figure 3-16), pathview mapping of DETs on this pathway showed abundance of downregulated genes (Figure 3-17).

Citrus DETs were also visualized in plant functional pathways using the MapMan pathway analysis tool, which accurately assign hierarchal ontologies and provide visual representation of genes involved in different plant processes (Thimm et al., 2004);

(Usadel et al., 2009).

MapMan bins were assigned to 2417 DETs out of which 468 and 963 were mapped on biotic stress pathway revealing significant enrichment of DETs among all plant defense response categorize at 24 hpi and 48 hpi respectively (Figure 3-18). In general, most of the genes in cell wall metabolism, abiotic stress, peroxidation and secondary metabolism were upregulated at both 24 hpi and 48 hpi. PR-genes, and

MAPK signaling were mostly upregulated. Genes in redox state and general signaling were almost equally distributed between the up- and down-regulated categories.

Proteolysis was seen more upregulated at 48 hpi compared to 24 hpi. Most of the transcription factors (TFs) like, ERF, bZIP and DOF showed mixed trend of up- and down-regulation whereas most of the WRKY TFs were up-regulated. A notable

exception was genes in the MYB transcription family, a majority of which were downregulated. Consistent to KEEG pathway mappings, plant hormones involved in growth and general physiological processes like auxin and brassinosteroid showed downregulation, and ET and ABA that are generally considered as stress responsive hormones were mostly upregulated (Figure 3-15 and Figure 3-18). Overall, these analyses suggested that many different biological, molecular and cellular processes in host are affected by P. parasitica infection.

Putative R-genes Differentially Expressed in Citrus Roots in Response to P. parasitica

Functional annotations indicated presence of several R-genes among citrus

DETs. In addition to that we ran a local BLAST of all citrus transcripts against protein sequences of all R genes in Plant resistance genes database (PRGdb). PRGdp is a comprehensive database comprised about 104459 R genes from 233 plant species

(Sanseverino et al., 2012). In total, 645 transcripts corresponding to 454 genes showed

80-100% (e-value threshold of 1e-10) homologies to 490 unique R genes. Out of those

490 R genes, 284 had proper annotations and 277 of them matched to either C. sinensis or C. clementia R genes. These putative R genes represented all R gene classes with 33 containing NB-ARC and LRR domains, which are grouped broadly in NL class in PRGdb, 56 in the CNL class, 12 in the CN, 6 in the TNL, 135 in the RLP, 9 in the RLK (8 RLK-GNK2), 4 in the MLO-like, 27 in the N class and the rest of them were categorized in others class.

Out of total 454 R genes found in the root transcriptome of citrus, 186 were found differentially expressed in response to P. parasitica infection. Of those, 100 were upregulated in the range of 4 to 32,768 times and 80 were downregulated (4 to 256 fold

change). The top differentially expressed R gene was

(PRGDB00149304/TRINITY_DN23214_c3_g3) that is a G-type lectin S-receptor-like serine threonine- kinase (SD2-5) was around 8000 times more expressed in response to P. parasitica at 24 hpi and reached up to 32000 times upregulated at 48 hpi. Five other R genes of G-type lectin S-receptor-like serine threonine- kinases were also found to be upregulated at 1 or both time points. A receptor 12 like uncharacterized R gene was more than 2000 times more expressed upon P. parasitica infection at both time points. Out of 4 MLO-like genes, one (PRGDB00150044) was upregulated at both time points. Two RGA3 like R genes were also found with conflicted differential expression,

PRGDB00149184 was downregulated whereas PRGDB00148450 was highly upregulated. Out of 6 TNLs, 3 TMV resistant N-like R genes (PRGDB00152395,

PRGDB00158343 and PRGDB00149509) were found highly upregulated. Three RPP13 like disease resistance genes were found upregulated; PRGDB00148968 and

PRGDB00147909 only at 24hpi whereas PRGDB00149876 was highly expressed at both 24 and 48 hpi. In the CNL class, three different sub types containing of three, four and eight R genes that were predictive homologues of At1g12280, At4g27220 and

At5g63020 respectively were among the highly upregulated R genes (Table 3-6).

Upregulation of several different classes of R genes could have a significant role in the preventing disease progression in citrus by P. parasitica.

Validation of RNA-seq Results for Selected DETs by qRT-PCR

To verify differential expression analysis results from RNA-seq data, transcriptional levels of selected DETs including WRKY transcription factor

31(TRINITY_DN19772_c0_g1_i1), disease resistance At4g27220 isoform X1

(TRINITY_DN23376_c2_g4_i1) and myb family transcription factor APL isoform

X2(TRINITY_DN16689_c0_g1_i1), expansin (TRINITY_DN28627_c0_g1_i1), G-type lectin S-receptor-like serine threonine- kinase SD2-5 (TRINITY_DN23214_c3_g3_i1) etc. that were potentially involved in the development of tolerance response in citrus roots against P. parasitica and represent different expression profiles across P. parasitica treated and control citrus roots samples, were determined by qRT-PCR analysis. Relative expression values of all the tested genes were calculated by using constitutively expressed citrus GADPH gene. Expression profiles of all tested genes resulted from qRT-PCR were in agreement with our RNA-seq data analysis results

(Figure 3-19).

Discussion

Genetic resistance among wild and cultivated plants against wide variety of phytopathogens are the best resources that can be utilized to develop the most economical and environmentally friendly strategies to fight against Phytophthora diseases. In citrus, resistance rootstocks are excellent breeding material for developing

Phytophthora resistant citrus trees with high quality fruits (Lima et al., 2018). Few rootstocks like Carrizo citrange, Swingle citrumelo and P. trifoliata have been reported to be tolerant to P.parasitica infection except in disease complexes (Dalio et al., 2017);

(Graham, 1995); (Graham et al., 2003); (Hutchison, 1974). Molecular mechanism of resistance of most of these citrus rootstocks to Phytophthora pathogens is not well understood. This study was designed to identify R genes and overall transcriptional changes in Carrizo citrange rootstock in response to Phytophthora parasitica through whole transcriptome sequencing of infected and control samples during early stages of infection.

Microscopic studies of P. parasitica treated citrus roots clearly showed the attachment and germination of zoospores on the root surface followed by mycelium colonization (Figure 3-3), indicating successful initial infection. Mapping coverage of P. parasitica genome in the pathogen-inoculated samples increased three times from 24 to

48 hpi (Table 3-4), which closely corresponded to increased root colonization of pathogen from 24 to 48 hpi observed under the microscope (Figure 3-3) and is a clear evidence of successful infection and colonization of Carrizo citrange roots by P. parasitica. However, disease symptoms were not evident in the above ground parts of the plant suggesting that the understudy rootstock is a host of P. parasitica and exhibits a tolerant response unlike P. trifoliata that showed non-host resistance (Dalio et al.,

2017). To understand the molecular basis of this tolerant response on transcriptome level, we did differential isoform expression analysis in citrus roots during initial stages of P. parasitica infection.

A de novo transcriptome assembly was generated because read mapping coverage on reference genomes of related Citrus spp. (Adhikari et al., 2012); Wang et al., 2014 (Wang et al., 2014) were quite low (Table 3-4) which is understood because

Carrizo citrange is a hybrid. Assembly optimization was done to keep only biologically meaningful transcripts and to save statistical power in the differential expression analysis (McCann et al., 2017); (Ono et al., 2015). Initial differential expression was done by using both coding isoforms counts and unigene counts. Considering the fact that Carrizo citrange is a hybrid and shared exons between two parental genomes could cause ambiguities and significant loss of information in case of unigenes, most of the downstream analysis presented here are based on protein coding transcripts. Although

statistically exhaustive and complicated, transcript level differential expression analysis has been proved to provide higher resolution and several new algorithms enable individual transcript abundance estimation (Soneson et al., 2015). In our studies results from both methods were almost similar except the fact that gene-level analysis missed some of the differential expressed transcripts, contrastingly transcript-level analysis covered most of the genes predicted to be differentially expressed by gene-level analysis. Differential expression analysis using transcripts count was also validated by qRT-PCR and expression profiles of all tested genes were found similar to RNA-seq results (Figure 3-19).

Numerous plant defense related genes putatively involved in preventing diseases progression were identified to be differentially expressed in citrus roots in response to P. parasitica infection. In addition to plant defense, major transcriptional reprogramming of diverse cellular and metabolic processes in citrus root upon P. parasitica infection was observed. The mechanism of plant’s response to pathogens and development of resistance or tolerant response is a complex phenomenon that involves interconnected network of changes in several regular physiological processes (Jones and Dangl, 2006) that is clearly reflected in transcriptome-based studies of several different plant- pathogen systems (Adhikari et al., 2012); (Chen et al., 2014); (Gao et al., 2013);

(Judelson et al., 2008); (Kunjeti et al., 2012). Consistent with gene expression profiles of

Nicotiana benthamiana leaves and A. thaliana roots infected with P. parasitica during compatible interaction (Le Berre et al., 2017); (Shen et al., 2016), transmembrane transport, protein kinase activity, iron ion binding, carbohydrate metabolism, phosphotransferase activity and many other daughter terms of biological and molecular

processes were significantly enriched among both up and down regulated DETs in citrus roots. Unlike susceptible A. thaliana roots where number of DEGs didn’t increase considerably with time during P. parasitica infection (Le Berre et al., 2017), total number of DETs increased in citrus roots from 24 hpi to 48 hpi suggesting induction of stronger defense response towards P. parasitica overtime that can be related to tolerance response during later stages (Figure 3-7 and Figure 3-18).

Upon perception of pathogen generation of reactive oxygen species (ROS) has been reported as one of the earliest defense strategies in plants (Elmayan and Simon-

Plas, 2007). Oxidoreductase activity was found to be the most significantly differentially regulated MF in roots upon P. parasitica infection and several different oxidases and reductases were found to be highly upregulated. Upregulation in oxidation reduction processes has been observed in resistant response of many plants to different pathogens including Phytophthora species (Reeksting et al., 2016). Differential proteome analysis of a resistant soybean cultivar in response to P. sojae also revealed significant upregulation in oxidoreductase activity related proteins (QIU et al., 2009). A tobacco NADPH oxidase was found responsible for triggering HR in response to a

Phytophthora cryptogea elicitor leading to acquired resistance (Elmayan and Simon-

Plas, 2007). An NADPH oxidase (TRINITY_DN6244_c1/c0) was 512 times more expressed in citrus roots infected with P. parasitica compared to control so it might be involved in ROS signaling during early stages of infection but that was not enough to prevent pathogen colonization. P. parasitica is a hemi biotroph so is adapted to survive oxidative burst during initial stages of infection (Attard et al., 2010).

After perception and attachment to the host’s surface, the next step for successful infection is penetration into the host. Cytoskeleton reorganization of infected cells is considered an important component of plant defense against pathogen penetration (Hardham et al., 2007); (Underwood, 2012); (Higaki et al., 2011). Significant enrichment of membrane and cell wall macromolecule metabolism among DETs (Figure

3-12), and representation of microtubule and overall cytoskeleton organization among all three GO (MF, BP & CC) classes were observed (Figure 3-11). In addition to that enrichment of cell wall related genes in the biotic stress overview of MapMan is also evident (Figure 3-18). Expansins are reported to play an important role in plant growth and involved in cell wall loosening, disassembly and separation during cell growth. We found differential modulation of 29 different expansins at both extremes of up and downregulation. For instance, expansin B 1 (TRINITY_DN21518_c1_g2_i8) was 3565 time more expressed and unclassified expansin (TRINITY_DN28627_c0_g1_i1) was

304 times downregulated in P. parasitica infected roots as compared to uninfected roots. Unlike our results only downregulation of four expansins was seen during the compatible interaction of P. parasitica in A. thaliana roots (Le Berre et al., 2017). Most of the other cell wall and microtubule related genes were significantly down regulated in the infected cells, which could be either an attribute of plant defense response against pathogen infection or pathogen’s strategy to manipulate host cell for successful penetration.

Consistent with the findings of Le Berre et al. in A. thaliana roots, some VQ motif contacting genes were found to be upregulated in citrus roots during early stages of P. parasitica infection (Le Berre et al., 2017) but at 48 hpi mix trend of up and down

regulation was seen. Unlike their finding, VQ29 reported to be specifically involved in resistance against P. parasitica was not found differentially expressed in our study (Le

Berre et al., 2017).

Phytohormones, specifically SA, JA and ET are considered an integral part of immune response (Denance et al., 2013), which lead to broad spectrum systemic acquired resistance (SAR) and induced systemic resistance (ISR) (Shoresh et al.,

2010). SA is considered an important component of SAR, which involves induction of numerous PR genes not only in the infected cells but also in non-infected systemic tissues (Durrant and Dong, 2004). Another important component of SAR, downstream to SA is NPR1 (Vanacker et al., 2001), which interacts with TGA transcription factor of the bZIP family to induce expression of PR genes ultimately giving rise to SAR (Durrant and Dong, 2004). Some PR genes, NPR homologs, one NPR suppressor and three

TGA transcription factors were differentially modulated in citrus roots in response to P. parasitica infection. Both JA and ET are more closely associated with ISR, usually antagonistic to the SA pathway, and the balance between these three signaling pathways is determined by the lifestyle of the pathogen (Dong, 1998); (Heil and

Bostock, 2002). DETs involved in ET pathway were mostly upregulated in our studies, whereas, mixed trend of up and down regulation was observed in the putative JA DETs

(Figure 3-15). Foliar application of JA in potato and tomato have been shown to induce resistance against Phytophthora infestans Local and systemic protection against

Phytophthora infestans induced in potato and tomato plants by jasmonic acid and jasmonic methyl ester (Cohen et al., 1993). In case of N. benthamiana-P. parasitica interactions, induction of JA and ET pathway is reported but SA pathway was not seen

to be activated (Shen et al., 2016). Plants generally activates JA/ET pathway and repress the SA pathway infected by necrotrophic pathogens. In contrast, activation of

SA and repression of JA/ET pathways have been observed in response to biotrophic pathogens (Spoel et al., 2007). In contrast to our results, hormonal cross talk was shown not to be involved in resistant response of P. trifoliata to P. parasitica (Dalio et al., 2017). Flavonoid biosynthesis that is also linked to plant hormones regulation especially SA pathway, was also found highly enriched differentially modulated genes

(Figure 3-16 and 3-17). Differential induction or repression of JA, ET and SA pathways in citrus roots in response to P.parasitica infection represents a cross talk of these defense hormones in dealing with the hemi-biotrophic nature of P. parasitica infection, which might be responsible for the host’s tolerance response to disease despite heavy pathogen colonization of roots.

Functional annotations revealed DETs involved in reproduction (Figure 3-11) seven of them were found with pollen allergen signatures. These pollen-pistil interaction proteins are generally known as allergens as their homologs in many plants cause allergies in animals (Rogers et al., 1991;Rogers et al., 1992). They are characterized as pectate lyase enzymes that might have some role in cell death of unwanted reproductive structures during fruit development (Wu and Cheung, 2000). Role of pectate lyase in deterioration of host cells during infection has also been reported in host-pathogen interaction studies but most of the known pectate lyases involve in infection are of pathogen origin (Vorwerk et al., 2004). Recently, Mondragon et al (2017) have reported some transcriptional level resemblances in Arabidopsis flower in response to fungal infection and pollination. They have observed similar transcriptional

trends in some genes putatively involved in pollination and immune response to

Fusarium graminearum that penetrates through pistils of Arabidopsis flowers

(Mondragon et al., 2017). Differential expression of pectate lyase genes in roots may have some role in HR or pollen pistil interactions might have some association between plant response to intracellular hyphal penetration and pollen tube penetration phenomenon in flowers.

R genes are an integral component of host-pathogen interactions, involved in sensing pathogen patterns (RLPs and RLKs) and pathogen-induced manipulations in host cells followed by triggering an array of responses including local cell death or hypersensitive response (HR), phytohormone signaling, activation of PR genes and many other defense related pathways (Thomma et al., 2011);(Wanderley-Nogueira et al., 2012). More than 40% of the total R genes identified in citrus root transcriptome were found to be differentially modulated upon P. parasitica infection. Significant induction of several important R gene classes was observed that could be putatively involved in tolerant response to P. parasitica (Table 3-6).

Infection of P. parasitica in N. benthamiana leaves caused up-regulation of 56

LRR-RLKs (Shen et al., 2016). We found upregulation of around 68 RLPs and RLKs indicating significant induction of pathogen sensors in citrus roots in response to P. parasitica infection. Lectin receptor kinases are an important class of RLKs, comprising of three (G-type, L-type and C-type) further subclasses. Six G-type RLKs including the top induced R gene (32000 times upregulated) were induced in citrus roots upon P. parasitica infection compared to control (Table 3-6). L-type RLKs contribute in plant defense and reported to confer resistance against Phytophthora species (Wang and

Bouwmeester, 2017); (Wang et al., 2015) but G-type has never been reported as

Phytophthora resistant gene. In rice BpH3, a resistance gene comprises of three G-type

RLKs is known to provide broad spectrum stable resistance against rice brown plant hopper since last thirty years. Pi-d2 another G-type RLK in rice confer resistance to rice blast fungus Magnaporthe grisea (Wang et al., 2018). Receptor 12 was another uncommon uncharacterized R gene that was highly responsive to P. parasitica in citrus roots (Table 3-6). Whether, these receptors play a role in resistance to Phytophthora would be an interesting to investigate in future studies.

Boava et al. (2011) have reported upregulation of two R genes (one TNL and one RPS4 type) in P. parasitica resistant Poncirus trifoliata plants compared to Citrus sunki genotypes, which are reported to be susceptible to P. parasitica (Boava et al.,

2011). Consistent to their findings, one RPS4 homologue (TRINITY_DN22857_c1_g1) and four TNLs including three TMV resistant N-like genes were among the highly induced R genes in citrus against P. parasitica. In a late blight resistance potato cultivar

SD20, TMV and RGA3 type resistant genes were found upregulated in response to P. infestans (Yang et al., 2018). One RGA3 homologue was induced and one depressed in our studies. RPP13 locus of A. thaliana has been reported to offer broad spectrum resistance against different strains of a biotrophic oomycete pathogen Peronospora parasitica that cause downy mildew in different plants (Bittner-Eddy et al., 2000). Three

RPP13 homologs were found upregulated in citrus roots in response to P. parasitica.

Apart from these examples, several other genes belong to different classes of R genes were highly induced pathogen challenged citrus roots. Involvement of these R genes

100

along with other defense regulators have been suggested to play a significant role in preventing disease progression in Carrizo citrange in response to P. parasitica.

Transcriptional reprogramming of most of the cellular and physiological processes and all components of biotic stress responses especially high induction of several R genes, modulation of phytohormone signaling, proteolysis, TFs and signaling in Carrizo roots indicates significant activation of defense during infection that increase overtime and thus prevent diseases progression after colonization and confer a tolerant response against P. parasitica.

101

Table 3-1. List of genomes that were used to make custom BLAST database.

Sr.no: Available Genomes

Citrus

1 Citrus sinensis (Ridge Pineapple)

2 Citrus sinensis (Valencia)

3 Citrus maxima

4 Citrus medica

5 Citrus clementina

6 Citrus x paradisi x Citrus trifoliata

7 Citrus unshiu

8 Citrus cavaleriei

Phytophthora Spp.

1 Phytophthora parasitica

2 Phytophthora kernoviae

3 Phytophthora lateralis

4 Phytophthora infestans

5 Phytophthora capsici

6 Phytophthora ramorum

7 Phytophthora pluvialis

8 Phytophthora taxon totara

9 Phytophthora sojae

10 Phytophthora megakarya

11 Phytophthora cactorum

12 Phytophthora x alni

13 Phytophthora nicotianae

14 Phytophthora agathidicida

15 Phytophthora multivora

16 Phytophthora pisi

17 Phytophthora rubi

18 Phytophthora fragariae

19 Phytophthora pinifolia

20 Phytophthora cryptogea

21 Phytophthora cambivora

22 Phytophthora cinnamomi

102

Table 3-2. The statistics of Trinity generated de novo assembly. Feature Value General assembly statistics Total trinity assembled genes 60874 Total trinity assembled transcripts 87299 Percent GC content 40.51 Statistics based on longest isoform per trinity gene Contig N10 3797 Contig N20 2847 Contig N30 2260 Contig N40 1831 Contig N50 1482 Average contig length 766.97 Total assembled bases 46688332 Statistics based on all transcripts Contig N10 3996 Contig N20 3113 Contig N30 2560 Contig N40 2147 Contig N50 1796 Average contig length 1011.41 Total assembled bases 88294993 Statistics based on optimized assembly Total transcripts 40943 Total unigenes 21278 Transcript Contig N50 1774 Gene Contig N50 1981 Total assembled amino acids 51049025

103

Table 3-3. qRT-PCR primers for selected genes. Transcript ID Description Primers TRINITY_DN19772_c0_g1_i1 WRKY 31 GAGATTGATTGGCCCTGGCT ATTTGGCCTTTTGGGGAGGT TRINITY_DN23376_c2_g4_i1 disease GTCACAGCCCCACACTATCA resistance GCGTGCTTAGAGAGAGAGGTC At4g27220 TRINITY_DN16689_c0_g1_i1 Myb APL ATTGCCGCACACCCCTTTTA isoform2 GAGCGAGCGAAGAGAGAAGA TRINITY_DN28627_c0_g1_i1 expansin

TRINITY_DN23214_c3_g3_i1 G-type lectin S- RLK SD2-5 TRINITY_DN15632_c0_g1_i1 GADPH CATTTCTGGCTTGTGCCTGC AGGCCGTAGATCTGAGGAGA

Table 3-4. Mapping coverage of all seven RNA-Seq samples to Citrus and P. parasitica genomes. Sample name Genomes C.sinensis (V) C.sinensis Swingle P. parasitica (RP) 0 hpi-control 60.2% 68.5% 66.3% 5% 24 hpi-control 62.9% 70.4% 73.2% - 24 hpi-treatment 60.2% 72.3% 69.9% 0.8% Rep1 24 hpi-treatment 63.2% 71.3% 68.6% 0.8% Rep2 48 hpi-control 57.1% 73.3% 76% - 48 hpi-treatment 57.4% 70.0% 68.4% 1.8% Rep1 48 hpi-treatment 55.2% 70.0% 68.3% 1.8% Rep2

104

Table 3-5. Top KEGG Pathways for DETs. KO terms Description Number of hits KO01100 Metabolic pathway 323 KO01110 Biosynthesis of secondary metabolites 192 KO00940 Phenylpropanoid biosynthesis 87 KO01130 Biosynthesis of antibiotics 85 KO01230 Biosynthesis of amino acids 47 KO01200 Carbon metabolism 40 KO00500 Starch and sucrose metabolism 29 KO04075 Plant hormone signal transduction 27 KO04626 Plant-pathogen interaction 25 KO04016 MAPK signaling pathway - plant 22 KO00941 Flavonoid biosynthesis 18

105

Table 3-6. R gene DETs induced upon P. parasitica infection in citrus roots. Transcript ID R gene ID Symbol Description Log2FC 24hpi 48hpi DN23214_c3_g3_i1 PRGDB00149304 G-type G-type lectin S- 13.07 15.18 RLK receptor-like serine threonine- kinase SD2-5 DN22851_c0_g1_i1 PRGDB00148464 RLP Leucine-rich 12.36 10.11 repeat protein kinase family protein DN23591_c4_g3_i1 PRGDB00152395 TNL TMV resistance 11.59 11.58 N-like DN10162_c0_g1_i1 PRGDB00149639 receptor 12 11.27 10.34 DN23638_c1_g5_i2 PRGDB00147721 CNL disease 10.45 10.37 resistance At4g27220 isoform X1 DN23509_c6_g3_i17 PRGDB00154261 CNL disease 10.36 10.28 resistance At1g12280 DN22574_c0_g1_i7 PRGDB00149656 serine threonine- 10.34 10.64 kinase EDR1 isoform X1 DN23532_c2_g1_i5 PRGDB00155473 CNL disease 10.27 10.07 resistance At4g27220 isoform X2 DN10546_c2_g1_i1 PRGDB00150116 LRR receptor-like 9.96 4.27 serine threonine- kinase GSO2 DN11631_c0_g1_i1 PRGDB00156448 RLP LRR receptor- 9.88 7.16 like serine threonine- kinase At3g47570 isoform X3 DN21819_c0_g4_i1 PRGDB00155514 TNL NB-ARC domain- 9.7 9.46 containing disease resistance isoform 2

106

Table 3-6. Continued Transcript ID R gene ID Symbol Description Log2FC 24hpi 48hpi DN22789_c0_g1_i2 PRGDB00148452 NL disease resistance 9.64 8.64 At5g63020 DN23214_c3_g3_i2 PRGDB00149304 RLK G-type lectin S- 9.42 9.7 receptor-like serine threonine- kinase SD2-5 DN21140_c3_g1_i7 PRGDB00155884 RLP leucine-rich repeat 9.32 8.25 receptor kinase At2g19210 DN20583_c0_g1_i3 PRGDB00155495 NL disease resistance 9.3 9.12 At5g63020 DN23591_c4_g3_i13 PRGDB00152395 TNL TMV resistance N- 9.17 7.44 like DN23603_c0_g1_i15 PRGDB00148450 NL disease resistance 9.09 RGA3 isoform X1 DN20545_c3_g3_i5 PRGDB00149431 mitogen-activated 8.63 8.23 kinase kinase kinase YODA isoform X1 DN23140_c0_g1_i5 PRGDB00147871 receptor kinase 8.46 At2g23200 DN16425_c0_g1_i1 PRGDB00154920 RLK G-type lectin S- 8.3 receptor-like serine threonine- kinase CES101 isoform X1 DN22648_c0_g1_i3 PRGDB00148631 RLP leucine-rich repeat 8.26 9.03 receptor-like serine threonine- kinase At3g14840 DN11990_c1_g1_i1 PRGDB00152848 NL disease resistance 8.22 9.19 At5g63020 DN18693_c2_g1_i2 PRGDB00147575 CNL disease resistance 8.01 At5g63020 DN22857_c1_g1_i2 disease resistance 7.92 RPS4-like DN23509_c6_g3_i10 PRGDB00149626 CNL AF506028_20NBS- 7.49 7.72 LRR type disease resistance

107

Table 3-6. Continued Transcript ID R gene ID Symbol Description Log2FC 24hpi 48hpi DN22388_c0_g1_i2 PRGDB0014790 CNL disease resistance 7.36 9 RPP13-like DN23413_c0_g1_i2 PRGDB0015624 CNL disease resistance 7.36 7 At4g27190-like isoform X2 DN17517_c0_g2_i1 PRGDB0014874 CNL disease resistance 7.28 3.93 4 At4g27190-like DN18028_c0_g1_i1 PRGDB0014835 RLK G-type lectin S- 7.25 3 receptor-like serine threonine- kinase SD2-2 DN23314_c1_g1_i2 PRGDB0014856 NL disease resistance 6.59 7.05 5 At4g27220 DN21925_c0_g1_i1 PRGDB0014935 NL NB-ARC domain- 6.41 5.36 2 containing disease resistance DN20349_c1_g1_i1 PRGDB0014844 RLP receptor kinase 5 5.68 6.53 5 DN23142_c0_g1_i1 PRGDB0014836 RLP LRR receptor-like 5.62 4.35 9 serine threonine- kinase At4g08850 DN23600_c1_g2_i1 PRGDB0014902 CNL disease resistance 5.04 10.6 3 8 At1g12280 6 DN23217_c0_g1_i1 PRGDB0015420 RLP BRASSINOSTEROI 4.96 6.81 5 D INSENSITIVE 1- associated receptor kinase 1 DN23127_c2_g1_i1 PRGDB0015834 TNL TMV resistance N- 4.95 3 like DN21210_c1_g1_i1 PRGDB0015012 NL Disease resistance 4.48 7 CC-NBS-LRR class family DN23478_c0_g1_i6 PRGDB0014877 receptor-like serine 4.3 4.02 4 threonine- kinase At5g57670 DN16482_c0_g1_i1 PRGDB0015462 L-type lectin-domain 4.3 4.44 5 containing receptor kinase DN11381_c0_g1_i1 PRGDB0014926 wall-associated 4.28 2 receptor kinase 2-like

108

Table 3-6. Continued Transcript ID R gene ID Symbol Description Log2FC 24hpi 48hpi DN2209_c0_g1_i1 PRGDB00152958 CNL disease resistance 4.24 4.16 At5g63020 DN23600_c1_g2_i10 PRGDB00149028 CNL disease resistance 4.2 At5g63020 DN20545_c3_g3_i1 PRGDB00149431 mitogen-activated 4.07 5.24 kinase kinase kinase YODA isoform X1 DN22161_c0_g1_i2 PRGDB00148380 RLP LRR receptor-like 3.95 serine threonine- kinase At4g08850 DN14920_c0_g1_i3 PRGDB00158497 receptor kinase 3.88 At5g39030 isoform X1 DN15425_c0_g1_i2 PRGDB00153071 receptor kinase 3.86 5.86 At1g11050 DN17248_c3_g1_i1 PRGDB00148841 RLK- receptor kinase 3.76 GNK2 At4g00960 DN21119_c0_g1_i1 PRGDB00150044 Mlo- MLO 6 3.74 3.79 like DN18954_c0_g1_i1 PRGDB00148615 RLP leucine-rich repeat 3.71 4.55 receptor-like serine threonine- kinase At3g14840 DN19972_c0_g1_i1 PRGDB00148970 receptor kinase 3.68 4.33 At1g11050 DN15425_c0_g1_i1 PRGDB00153071 receptor kinase 3.67 3.26 At1g11050 DN19991_c0_g1_i2 PRGDB00153430 RLK chitin elicitor 3.53 receptor kinase 1- like DN10442_c0_g1_i1 PRGDB00152598 L-type lectin- 3.5 3.2 domain containing receptor kinase DN21551_c0_g1_i1 PRGDB00157847 RLP LRR receptor-like 3.41 3.78 serine threonine- kinase At1g74360

109

Table 3-6. Continued Transcript ID R gene ID Symbol Description Log2FC 24hpi 48hpi DN6015_c0_g1_i1 PRGDB00147533 leucine-rich repeat 3.41 4.31 receptor kinase At1g35710 DN23595_c0_g2_i6 PRGDB00148807 RLP LRR receptor-like 3.37 5.35 serine threonine- kinase RFK1 DN19347_c0_g1_i1 PRGDB00150195 nematode 3.35 4.14 resistance -like HSPRO2 DN12565_c0_g1_i1 PRGDB00155269 kinase and PP2C- 3.35 like domain- containing DN20962_c0_g1_i1 PRGDB00149924 LRR receptor-like 3.34 serine threonine- kinase At1g53420 isoform X1 DN16770_c0_g1_i1 PRGDB00149696 CNL disease resistance 3.31 3.43 At5g63020 DN20228_c0_g1_i1 PRGDB00153886 RLK G-type lectin S- 3.23 3.09 receptor-like serine threonine- kinase B120 isoform X1 DN23654_c5_g1_i5 PRGDB00149876 CNL disease resistance 3.12 2.73 RPP13 1 DN19159_c0_g1_i1 PRGDB00154214 RLP LRR receptor-like 2.98 2.62 serine threonine- kinase At1g05700 DN23254_c0_g2_i2 PRGDB00148451 RLP somatic 2.98 embryogenesis receptor kinase 1- like isoform X1 DN22569_c0_g1_i1 PRGDB00148366 LRR receptor-like 2.94 serine threonine- kinase GSO1 DN23209_c1_g1_i2 PRGDB00149784 RLP LRR receptor-like 2.94 4.65 serine threonine- kinase At3g47570 DN23291_c0_g1_i1 PRGDB00148137 Other disease resistance 2.87 2.63 At4g19050

110

Table 3-6. Continued Transcript ID R gene ID Symbol Description Log2FC 24hpi 48hpi DN22367_c0_g1_i2 PRGDB00154320 CN U-box domain- 2.86 4.32 containing 35 DN23209_c1_g1_i1 PRGDB00149784 RLP LRR receptor-like 2.86 3.06 serine threonine- kinase At3g47570 DN21140_c2_g1_i1 PRGDB00155885 RLP leucine-rich repeat 2.85 receptor kinase At2g19210 DN23053_c0_g1_i1 PRGDB00148968 NL disease resistance 2.82 RPP13 1 DN13924_c0_g1_i1 PRGDB00148931 RLK G-type lectin S- 2.81 receptor-like serine threonine- kinase At4g27290 DN23547_c3_g1_i4 PRGDB00149509 TNL TMV resistance N- 2.8 like DN23532_c2_g1_i1 PRGDB00155469 CNL disease resistance 2.8 2.87 At4g27220 isoform X1 DN23573_c0_g1_i15 PRGDB00150359 RLP LRR receptor-like 2.71 serine threonine- kinase At1g56130 isoform X4 DN15500_c0_g1_i1 PRGDB00148227 L-type lectin-domain 2.65 2.55 containing receptor kinase DN23424_c3_g1_i4 PRGDB00148915 RLK G-type lectin S- 2.64 3.26 receptor-like serine threonine- kinase At4g27290 isoform X1 DN23509_c6_g3_i12 PRGDB00149626 CNL AF506028_20NBS- 2.59 LRR type disease resistance DN23217_c0_g1_i2 PRGDB00154205 RLP BRASSINOSTEROID 2.59 2.31 INSENSITIVE 1- associated receptor kinase 1 DN22596_c5_g4_i1 PRGDB00148242 RLP receptor kinase 2.59 3.35 At3g47110 DN21140_c3_g1_i1 PRGDB00147853 leucine-rich repeat 2.57 receptor kinase At2g19210

111

Table 3-6. Continued Transcript ID R gene ID Symbol Description Log2FC 24hpi 48hpi DN23603_c0_g1_i10 PRGDB00148450 NL disease 2.56 resistance RGA3 isoform X1 DN23600_c1_g2_i11 PRGDB00149028 CNL disease 2.55 resistance At1g12280 DN22329_c0_g1_i1 PRGDB00152417 RLP receptor kinase 2.44 HSL1 DN21632_c0_g1_i1 PRGDB00148028 lysM domain 2.43 2.55 receptor-like kinase 4 DN17400_c0_g1_i1 PRGDB00148064 RLP inactive LRR 2.41 3.12 receptor-like serine threonine- kinase BIR2 DN20819_c0_g1_i1 PRGDB00153879 L-type lectin- 2.41 2.95 domain containing receptor kinase DN23424_c3_g1_i1 PRGDB00148915 RLK G-type lectin S- 2.38 2.96 receptor-like serine threonine- kinase At4g27290 isoform X1 DN21112_c0_g1_i2 PRGDB00153906 RLK- cysteine-rich 2.33 GNK2 receptor kinase 15 isoform X1 DN23254_c0_g2_i1 PRGDB00148451 RLP somatic 2.27 embryogenesis receptor kinase 2 DN23529_c1_g1_i1 PRGDB00148969 RLP LRR receptor-like 2.25 3.64 serine threonine- kinase FLS2 DN22416_c0_g1_i1 PRGDB00157892 RLP LRR receptor-like 2.24 2.53 serine threonine- kinase RPK2 DN17840_c2_g3_i1 PRGDB00152378 CNL NB-ARC domain- 9.63 containing disease resistance protein

112

Table 3-6. Continued Transcript ID R gene ID Symbol Description Log2FC 24hpi 48hpi DN20363_c0_g1_i3 PRGDB00149410 8.29 DN19645_c0_g1_i1 PRGDB00152830 serine threonine- 8.15 kinase At4g35230 DN21734_c0_g1_i3 PRGDB00149685 CNL NB-ARC domain- 7.14 containing disease resistance protein DN21420_c0_g1_i5 PRGDB00153161 RLP Leucine-rich 7.13 repeat protein kinase family protein DN18752_c0_g1_i2 PRGDB00152719 RLK- cysteine-rich RLK 7.09 GNK2 (RECEPTOR-like protein kinase) 3 DN23603_c0_g1_i1 PRGDB00148450 NL NB-ARC domain- 7 containing disease resistance protein DN20385_c0_g4_i3 PRGDB00155039 6.96 DN23167_c2_g2_i1 PRGDB00149184 NL NB-ARC domain- 4.59 containing disease resistance protein DN19159_c0_g1_i4 PRGDB00154222 RLP Leucine-rich 3.46 repeat protein kinase family protein DN23456_c0_g1_i4 PRGDB00147698 3.4 DN23654_c5_g1_i2 PRGDB00149876 CNL disease 3.28 resistance RPP13 DN22115_c1_g1_i1 PRGDB00150180 3.18 DN23595_c0_g2_i10 PRGDB00148807 RLP receptor-like 3.15 kinase in flowers 1 DN22596_c5_g4_i3 PRGDB00153002 RLP Leucine-rich 3.12 repeat protein kinase family protein

113

Table 3-6. Continued Transcript ID R gene ID Symbol Description Log2FC 24hpi 48hpi DN18066_c0_g1_i1 PRGDB00157415 receptor kinase 3.06 At5g39000 DN18294_c0_g1_i2 PRGDB00148514 3.04 DN18888_c0_g1_i1 PRGDB00147497 2.89 DN17102_c0_g1_i1 PRGDB00153015 2.76 DN23184_c1_g1_i1 PRGDB00148370 receptor kinase 2.69 At1g49730 DN21934_c0_g1_i1 PRGDB00201112 2.68 DN20208_c0_g1_i1 PRGDB00153654 2.62 DN17781_c0_g1_i1 PRGDB00152944 NL NB-ARC domain- 2.59 containing disease resistance protein DN22307_c0_g1_i1 PRGDB00150346 RLP phytosulfokin 2.52 receptor 1 DN22574_c0_g1_i5 PRGDB00149656 2.51 DN20258_c0_g1_i1 PRGDB00148176 2.5 DN23461_c1_g1_i3 PRGDB00153902 2.44 DN23013_c0_g3_i1 PRGDB00148022 NL LRR and NB- 2.42 ARC domains- containing disease resistance protein DN20832_c0_g1_i1 PRGDB00156273 serine threonine- 2.41 kinase EDR1 isoform X1 DN16699_c0_g1_i1 PRGDB00147960 CN P-loop containing 2.28 nucleoside triphosphate hydrolases superfamily protein DN18981_c0_g1_i1 PRGDB00152279 2.26 DN19159_c0_g1_i2 PRGDB00154222 RLP Leucine-rich 2.24 repeat protein kinase family protein

114

Table 3-6. Continued Transcript ID R gene ID Symbol Description Log2FC 24hpi 48hpi DN18613_c0_g1_i1 PRGDB00155437 G-type lectin S- 2.17 receptor-like serine threonine- kinase At5g35370 DN23595_c0_g2_i3 PRGDB00148807 RLP receptor-like 2.17 kinase in flowers 1 DN22851_c0_g1_i1 PRGDB00148464 RLP Leucine-rich 12.36 10.11 repeat protein kinase family protein

E90N50=2056

Figure 3-1. Plot showing ExN50 statistics. E90N50 value indicates 90% of expressed transcripts have N50 value of 2056bp that indicates a very good quality assembly.

115

Quality check FASTQC

Adaptor removal and cleaning TrimGalore

Reference genome mapping

Tophat Filter Trinity Pathogen

Assembly Optimization TransDecoder BLAST

Abundance estimation RSEM

Differential expression analysis EdgeR

Functional Annotation Blast2go

Figure 3-2. Pipeline of RNA-Seq data analysis.

116

10 µm 10 µm

10 µm

10 µm 10 µm

Figure 3-3. Infection dynamics of Phytophthora parasitica in Carrizo roots from 3hpi to 48hpi. A) P. parasitica zoospore attachment and germination on the root surface indicated by arrows, observed at 3hpi in confocal microscope under Differential interference contrast white field (DICWF) filter. B) Inter and intracellular hyphal growth (indicated by arrow) was observed at 24hpi. C) A heavy colonization of P. parasitica mycelium was seen at 48hpi. After trypan blue staining hook like invaginations that are probably haustoria are clearly visible, indicated by arrows. Photo courtesy of Zunaira Afzal.

117

Pearson correlation A coefficient 24hr-Control 48hr-Control 24hr-treatment 48hr-treatment

48hr-treatment-Rep2

48hr-treatment-Rep1

24hr-treatment-Rep2

24hr-treatment-Rep1

48hr-Control

24hr-Control

Rep1

Rep2

- -

Control Control

- - 48hr

24hr

treatment

24hr 48hr

48hr 24hr B

Figure 3-4. Sample correlation and Principle component analysis (PCA). A) Dendogram and sample correlation matrix heatmap shows high correlation between biological replicates than among treatments. Green to red coloring refers to low to high correlations. B) PCA analysis shows that biological replicates of P. parasitica-inoculated samples at both time points (2hr_Treatment and 48hr_Treatment) cluster together reassuring the heatmap analyses shown in A. Similarly, mock-treated samples at both time points also clustered together.

118

B A 48hpi treatment Vs. Control 24hpi treatment Vs. Control

C 48hpi Vs. 24hpi 48hpi Vs. 24hpi treatments

Figure 3-5. Volcano plots showing pairwise comparisons between P. parasitica infected and mock-treated control roots samples. Dispersion of total differentially expressed transcripts found in: A) 48 hpi infected roots Vs. 24 hpi infected roots comparison, B) 24 hpi infected roots Vs 24 hpi mock treated roots transcriptome comparison and C) 48 hpi infected roots Vs 48 hpi mock treated roots comparison. Log fold change values are plotted at X-axis Vs. False Discovery Rate (FDR) at the Y-axis. Red color is indicating transcripts with statistically valid FDR values.

119

A Centered log2

Sub-cluster 2 B

Sub-cluster 3

Sub-cluster 1

2290 transcripts

Sub-cluster 2

1924 transcripts Centeredlog2(fpkm+1)

Sub-cluster 3

419 transcripts Centeredlog2(fpkm+1)

Figure 3-6. Hierarchical cluster analysis of differential expressed transcripts (DETs) across all samples. A) Heatmap showing overall hierarchical clustering of all DETs in citrus roots in response to P. parasitica infection. Yellow color is indicating induced expression and purple is pointing towards repression. B) Based on similarities in their expression profiles, total DETs are divided in three different sub-clusters. Sub-cluster 1 is showing highest number of DETs with increased expression from mock vs. P. parasitica treated samples at both time points. Similarly, second sub-cluster is representing down-regulated DETs upon infection at both time points. Third cluster is not only representing highly up-regulated DETs in mock vs infection but also slight expression variations from 24hpi to 48hpi.

120

A 24hpi 48hpi

48hpi Vs. 24hpi

B 48hpi C 48hpi 24hpi 24hpi

48hpi Vs. 24hpi 48hpi Vs. 24hpi

Figure 3-7. Venn diagrams shown number of citrus DETs. Total DETs (P-value < 0.001 and log2 fold change > 2) across all comparisons are shown in (A) Upregulated DETs are shown in (B) and Downregulated DETs are shown in (C). Pink circle represents 24hpi treatment vs 24hpi mock comparison, green circle is 48hpi treatment vs. 48hpi mock and yellow is 48hpi treatment vs 24hpi treatment.

121

MapMan 2417

C R KEGG

U 2510

I GO 5234

N BLAST 6544

InterPro 6484

Total 6692

0 1000 2000 3000 4000 5000 6000 7000 NUMBER OF SEQUENCES

Total InterPro BLAST GO KEGG MapMan

Figure 3-8. Functional annotation statistics of citrus DETs in response to P. parasitica infection, from different databases.

122

* * *

* *

Figure 3-9. Top hit species distribution graph. * are indicating citrus species.

123

Number of DETs

Figure 3-10. Distribution of GO categories assigned to differentially expressed citrus transcripts in response Phytophthora parasitica infection. Number of sequences assigned to top 20 categories in all three classes (molecular function (MF), biological processes (BP) and cellular component (CC)) are shown.

124

B C

Figure 3-11. Detailed GO term annotations of DETs in: A) cellular component, B) molecular function and C) biological process classes.

125

Figure 3-12. Interactive graphs for summary visualization of most abundant GOs among all significantly enriched GO terms. Most enriched GO terms molecular functions were related to diverse oxidoreductase and binding activities (A). Among biological processes, cell wall macromolecule metabolism, dephosphorylation and ion transmembrane transport were dominant (B). Significance levels are based on enrichment and FDR values with highly significant terms indicated by red nodes. Edges are biologically meaningful links between significant terms.

126

Figure 3-13. Visualization of differentially expressed transcripts involved in MAPK signaling pathway. Color gradient represents log2 fold ratios with red representing upregulation and green representing downregulation in treatments over mock roots. Left and right-hand side of each KO box show DETs mapping at 24hpi and 48hpi respectively. White box or no color fill means no DET was assigned to that KO term. Half white box on either side means that term was not differentially expressed at the respective time point. Significant dominance of up regulated transcripts was seen in whole pathway.

127

Figure 3-14. Mapping of differentially expressed transcripts on Plant-pathogen interaction pathway. Color gradient represents log2 fold ratios with red representing upregulation and green representing downregulation in treatments over mock roots. Left and right-hand side of each KO box show DETs mapping at 24hpi and 48hpi respectively. White box or no color fill means no DET was assigned to that KO term. Half white box on either side means that term was not differentially expressed at the respective time point.

128

Figure 3-15. Differentially expressed transcripts mapped on Plant hormone signaling pathway. Color gradient represents log2 fold ratios with red representing upregulation and green representing downregulation in treatments over mock roots. Left and right-hand side of each KO box show DETs mapping at 24hpi and 48hpi respectively. White box or no color fill means no DET was assigned to that KO term. Half white box on either side means that term was not differentially expressed at the respective time point.

129

Figure 3-16. KEGG enzyme enrichment analysis showed maximum mapping on Flavanoid biosynthesis pathway. Color key represents names of different enzymes involved in the pathway.

130

Figure 3-17. DETs mapped on Flavanoid biosynthesis pathway. Color gradient represents log2 fold ratios with red representing upregulation and green representing downregulation in treatments over mock roots. Left and right- hand side of each KO box show DETs mapping at 24hpi and 48hpi respectively. White box or no color fill means no DET was assigned to that KO term. Half white box on either side means that term was not differentially expressed at the respective time point.

131

Figure 3-18. Visualization of differentially expressed genes involved in biotic stress pathway in response to P. parasitica. Color gradient represents log2 fold ratios with red representing upregulation and green representing downregulation in treatments over mock roots. Each box represents one transcript. Significant dominance of downregulated transcripts in cell wall, peroxidases, MAPK, heat shock proteins and secondary metabolites are evident. In hormone signaling Auxins, Brassinosteroid and ABA were more downregulated whereas Ethylene and JA categorize are more encompassed by upregulated DETs. In transcription factor categorize WARKY and EFR have more upregulated DETs whereas MYB has more downregulated DETs. Most DETs designated as pathogenesis- related genes (PR-genes) can be seen upregulated. Mixed trends of up and down regulation were seen other biotic stress related processes.

132

myb family transcription factor APL WRKY31 ( TRINITY_DN23376_c2_g4_i1 ) ( TRINITY_DN19772_c0_g1_i1 ) 1.2 35 14 250

n 12

n 30

o i

o 1 i

s 200

r r

p 10 p

25 x

E 0.8

d e

150 e

e e

z l

8 l

l a

20 a

0.6 M

N M

15 e 100

v e

t l

a 0.4 e

l 4 R

e 10 R 50 0.2 2 5

0 0 0 0 24hpi-control 24hpi-infected 48hpi-control 48hpi-infected 24hpi-control 24hpi-infected 48hpi-control 48hpi-infected

Relative Normalized Expression TMM Relative Normalized Expression TMM

G-type lectin receptor Putative R-Gene ( TRINITY_DN23214_c3_g3_i1 ) ( TRINITY_DN23376_c2_g4_i1 ) 10000 4000 800 35 9000 700 3500

30 n

i s

o 8000

i s

s 600 e

s 3000

25 r

p r

x 7000

p E

x 500

E 20 d

2500

e e

d 6000 z

e l

400 l

l l

15 v

m 5000 2000

m M

300 o

10 T

N e

M 4000

i 1500

e 200

l a

5 e 3000

l R

e 100 1000 R 0 2000 0 500 24hpi-control 24hpi-infected 48hpi-control 48hpi-infected 1000 -100 -5 0 0 -200 -10 24hpi-control 24hpi-infected 48hpi-control 48hpi-infected

Relative Normalized Expression TMM Relative Normalized Expression TMM

Expansin arogenate dehydrogenase chloroplastic ( TRINITY_DN28627_c0_g1_i1) (TRINITY_DN20364_c0_g1_i2) 14 4000 1.2 250

3500

n 12

n o

o 1 200 s

i s

s e

s 3000 r

p 10 x

p E

E 0.8 150

d 2500 e

e z

e i

l l z 8

u a

l a

l v

v 2000 r

0.6 100 M

r o

N M

N T

M e

T v

i 1500

v t

i a

t l

a 0.4 50 e l 4

e R R 1000 0.2 0 2 500

0 0 0 -50 24hpi-control 24hpi-infected 48hpi-control 48hpi-infected 24hpi-control 24hpi-infected 48hpi-control 48hpi-infected Relative Normalized Expression TMM Relative Normalized Expression TMM germacrene D synthase-like elongation factor 2-like (TRINITY_DN23786_c0_g1_i1 ) (TRINITY_DN12559_c0_g1_i1) 1.2 400 100 800 90

1 350 n

n o

o 700

i s

s 80

e e

0.8 300 r

r p p 600

x 70

d 0.6 250

e e

e 60

z i

500 u

m m 50

0.4 200 r

M M

o 400

N N

T e

e 40

v i

i 0.2 150 t

t 300

l e

e 30 R R 0 100 200 24hpi-control 24hpi-infected 48hpi-control 48hpi-infected 20 -0.2 50 100 10 -0.4 0 0 0 24hpi-control 24hpi-infected 48hpi-control 48hpi-infected Relative Normalized Expression TMM Relative Normalized Expression TMM Figure 3-19. qRT-PCR based validation of plant defense related and top DETs in citrus roots in response to P. parasitica. Expression levels of tested genes were normalized based on transcript levels of GADPH gene. TMM values calculated from RNA-seq are compared to relative expression values determined by qRT-PCR analysis. Relative expression values of infected samples were determined against the average values of mock treated samples at each time point. Expression profiles DETs in both analysis correlate well for all the tested genes.

133

CHAPTER 4 THE PHYTOPHTHORA RXLR EFFECTOR AVRblb2 MODULATES PLANT IMMUNITY BY INTERFERING WITH Ca2+ SIGNALING PATHWAY

Introduction

Phytophthora spp. carry hundreds of effector proteins that are secreted either to the apoplast or inside of host cells (Haas et al., 2009; Raffaele et al., 2010). Most of the translocated Phytophthora effectors contain an N-terminal RXLR motif (Arg, any amino acid, Leu, Arg), which is reported to be involved in translocating effectors into host cells and a C-terminal domain with effector activity. RXLR effectors enable oomycetes to suppress basal immunity (Birch et al., 2008; Brasier, 2009; Fan et al., 2011; Bozkurt et al., 2012) as well as function as avirulence (Avr) factors in activating their cognate resistance genes (R genes) in the hosts (Anderson et al., 2015). Since the discovery of these RXLR effectors, great efforts are devoted to understanding the molecular mechanisms of how these effectors suppress or trigger plant immunity (Anderson et al.,

2015). The RXLR effectors share little sequence homology to characterized proteins thus precluding homology-based functional inferences. Identifying host targets of these effectors is a widely used method for providing clues about their function.

Current findings suggest that RXLR effectors primarily suppress host defense mechanisms (Brasier, 2009; Fan et al., 2011) but also alter other cellular processes

(Birch et al., 2008; Bozkurt et al., 2012).

Functional annotation as well as in planta functional analyses have shown that these effectors affect autophagy (Dagdas et al., 2016), protein degradation and stability

 Part of this chapter has been republished with permissions from: Naveed, Z.A., Bibi, S., and Ali, G.S. (2019). The Phytophthora RXLR Effector Avrblb2 Modulates Plant Immunity by Interfering With Ca2+ Signaling Pathway. Frontiers in Plant Science 10.

134

(CMPG1) (Bos et al., 2010; Gilroy et al., 2011), kinase and phosphatase signaling

(MAPKKK, PP1c) (King et al., 2014; Boevink et al., 2016a; Boevink et al., 2016b), transcription (NAC) (McLellan et al., 2013), RNA binding and small RNA biogenesis

(Qiao et al., 2013; Du et al., 2015; Jing et al., 2016), protein secretion (Du et al., 2015), endoplasm reticulum stress-mediated immunity (Jing et al., 2016) and brassinosteroid hormone signaling (Saunders et al., 2012).

Some RXLR effectors interact with susceptibility (S) factors of the host to promote virulence (Boevink et al., 2016b). For example, Pi04089 interacts with a RNA- binding protein KRBP1 to enhance P. infestans colonization (Du et al., 2015). Host vesicle-trafﬁcking and secretion mechanisms have also found to be modulated by RXLR effectors. AVR1 that is recognized by R1 to induce HR in the host, is reported to interact and stabilize an exocyst component Sec5 resulting in enhanced defense against P. infestans (Du et al., 2015).

AVRblb2 is another core RXLR effector, activated during infection. PiAVRblb2 interacts with a plant immune protease C14 and prevents its secretion in the apoplast, apparently to prevent degradation of the Phytophthora virulence proteins (Bozkurt et al.,

2011). Rpi-blb2, a coiled-coil-nucleotide-binding leucine-rich repeat (CC-NBS-LRR) receptor type R gene has been shown to recognize PiAvrblb2 and trigger hypersensitive response (HR) (Oh et al., 2009). Transformation of an Rpi-blb2 genomic clone from S. bulbocastanum, spanning the entire Rpi-blb2 gene cassette consisting of its native promoter, coding regions and 3` untranslated regulatory regions, into potato conferred broad spectrum late blight resistance (Song et al., 2003).

135

In our lab, Y2H screening of some core Phytophthora spp. effectors against a tomato (Solanum lycopersicum) cDNA library revealed interaction of AVRblb2 effectors with tomato calmodulins (CaMs) (Figure 4-1A and B). Further verification of Y2H results was done by performing BiFC (Figure 4-1C and D) and CaM overlay assays (Figure 4-2 and 4-3) using AVRblb2 homologs from P. infestans and P. parasitica. All these experiments confirmed a Ca2+ dependent interaction of AVRblb2 homologs with diverse

CaMs of S. lycopersicum, Nicotiana benthamiana and Arabidopsis thaliana (Figure 4-1,

4-2 and 4-3).

It is well established that spatio-temporal oscillations in plant cellular Ca2+ levels are early events in response to microbes including Phytophthora (Tavernier et al., 1995;

Zimmermann et al., 1997; Xu and Heath, 1998; Blume et al., 2000; Grant et al., 2000;

Lecourieux et al., 2002; Gust et al., 2007; Ranf et al., 2008). Pathogen-induced Ca2+- signatures are decoded by various Ca2+-binding proteins, among which calmodulin is relatively well studied in numerous plant microbe interactions (Harding et al., 1997; Do

Heo et al., 1999; Chiasson et al., 2005; Takabatake et al., 2007; Reddy et al., 2011;

Truman et al., 2013; Du et al., 2015).

Calmodulin has four Ca2+-binding EF-hand motifs. Upon binding Ca2+, it undergoes a conformational change, which allows it to bind and modulate the activity of numerous proteins involved in diverse cellular processes including plant defense

(reviewd in Kudla et al., 2010; Reddy et al., 2011). Numerous studies have demonstrated involvement of calmodulins and calmodulin-binding proteins including transcription factors, kinases, phosphatases, channels and pumps, and many

136

uncharacterized proteins in plant defense (Kim et al., 2002);(Choi et al., 2009);(Reddy et al., 2000).

Physical interaction among Phytophthora effectors and calmodulins points towards an interplay between pathogen and host calcium signaling pathways that could either aid pathogen in disease development by subverting plant defense or it could be a strategy used by plant to prevent pathogen invasion. Thus far, direct physical interaction of two other effector with CaM is shown. A Pseudomonas syringae effector HopE1 utilizes CaM as a cofactor to target MAP65, which is an important component of cell microtubule network, and reduces the secretion of an immunity related protein PR-1 leading to inhibiting cell wall associated extracellular PTI responses (Guo et al., 2016).

Recently, physical association of another RXLR effector SFI5 wih CaM is shown to regulate MTI response (Zheng et al., 2018).

In our laborartory, it was discovered that CaM binding domain (CBD) lies in the effector domain of Avrblb2. It was found by making a series of C- and N-terminal deletion mutants of the PiAvrblb2 protein (Figure 4-4C) followed by their in vitro testing for interaction with calmodulin using calmodulin binding overlay assays. These analyses revealed that deletion mutants lacking C-terminal aa 77 – 100 (77 – 100) or

N-terminal aa 1 – 87 (1 – 87) lost calmodulin binding activity, whereas constructs longer than these two these two mutants retained calmodulin-binding activity (Figure 4-

4A and C). These results were verified in in vivo BiFC analyses (Figure 4-4B). To make sure that only this region is sufficient for calmodulin binding, all five amino acids (aa 78

– 82) in this region were replaced to alanine (mutant designated as NCB), which completely abolished its interaction with calmodulin (Fig. 4A and B). All together, these

137

analyses suggest that amino acids 78 – 82 in the C-terminal region (within the effector domain) of Avrblb2 are important for binding to CaM (Fig. 4-4C).

CBD domains usually contain hydrophobic anchor amino acids, which occur in different configurations, usually 1-8-14, 1-5-10 in different CaM binding proteins (Tidow and Nissen, 2013). Based on this information and results from deletion analyses, amino acids at position 64, 65, 74, and 79 were identified as potential anchor sites. To determine if any of these amino acids are responsible for interaction with CaM, each of these predicted anchor sites was changed to alanine (A) one by one and was tested for their interaction to CaM. Calmodulin-binding overlay assays showed, that none of these amino acids affected calmodulin binding suggesting that these amino acids are not solely responsible for interaction with calmodulin (Fig. 4-5A, right panel). Absence of interaction was not due to lack of protein expression as all these mutants were expressed very well (Fig. 4-5A, upper panels). Since deletion analyses delineated CBD to amino acids 78 – 82, using alanine scanning, we mutated single amino acids in this region to alanine and tested their interaction with calmodulin. All mutants bound CaM

(Fig. 4-5A) but when all five amino acids from position 78 to 82 were changed to alanine the interaction was broken (Fig. 4-4A, (NCB) marked by *), suggesting that different amino combinations between amino acids 78 – 82 contribute to calmodulin binding.

The amino acid position 69 in Avrblb2 is highly polymorphic and affects its recognition by Rpi-blb2. To investigate whether this position affects CaM binding with

Avrblb2, different variants were tested and it was found that regardless of amino acid type at position 69, all Avrblb2 homologs (PITG20300 with A69, PITG18683 with I69,

PITG04085 with I69, and PITG20303 with F69) bound equally well to calmodulin (Fig. 4-

138

5A). To further verify aa A69 was changed to F69 in PITG04090, which is recognized by

Rpi-blb2 and which was used as template for constructing deletion and single amino acid mutants. Calmodulin-binding assays showed that the PITG04090-A69F mutant also retained calmodulin binding (Fig. 4-5A arrow) suggesting that aa 69-dependent

Rpi-blb2 activation by Avrblb2 is independent of calmodulin binding.

My goal was to determine the functional significance of AVRblb2-CaM interaction by connecting the dots among AVRblb2, CaM, C14 and Rpiblb2. In this respect, my first objective was to find out whether CaM binding to Avrblb2 is required for its recognition by Rpi-blb2 or not and the second objective was to find out whether the Avrblb2-CaM interaction affects Avrblb2-C14 interaction or not. My results indicate that the interaction of Avrblb2 to calmodulin is essential for the Avrblb2/Rpi-blb2 effector/R gene-mediated

HR response. Whereas, it was found out that Avrblb2-C14 interaction is not affected by

CaM binding to Avrblb2 but is highly dependent on the presence or absence of Ca2+.

Materials and Methods

Plasmid Construction and Transient Expression

Gateway cloning of sequence-verified wild type AVRblb2 (WT), mutant Avrblb2

(NCB) and C14 constructs cloned in pENTR/D-TOPO entry vector was done in pGWB series binary vectors. Both WT and NCB AVRblb2 were clone in pGWB442 with N- terminal eYFP fusion tag and C14 was cloned in pGWB414 with C-terminal 3xHA fusion tag (Figure 4-6A). These constructs were verified through restriction digestion analysis followed by sequencing. Confirmed constructs were transformed into Agrobacterium tumefaciens GV3101 and their presence was confirmed through gene specific PCR. For protein expression confirmation each of these constructs were transiently expressed in

N. benthamiana using the Agrobacterium-mediated transient transformation system

139

using the previously described protocol (Bos et al., 2006) followed by immunoblotting

(Figure 4-6B and C). Briefly, 4 to 6 weeks old N. benthamiana plants grown under controlled conditions (25 °C temperature and 16:8 hours light:dark period) were infiltrated (spot or whole leaf) with A. tumefaciens carrying the constructs. Culture

OD600 of 0.8–0.9 was maintained in agroinfiltration medium (pH 5.6) constituted of 10 mM MES, 10 mM MgCl2 and 200 μM acetosyringone. Confocal microscopy was done one day post inoculation (dpi) using the Olympus Spinning Disk (Ix81) confocal microscope.

Protein Extraction and Western Blotting

For protein isolation inoculated leaf tissues were collected 2 dpi, frozen immediately in liquid nitrogen and processed immediately or stored at -80 °C until further use. Frozen tissue was grounded in liquid nitrogen. For initial expression confirmation protein extraction was performed as follows: 100mg of grounded tissues was mixed with 200µl of ice cold extraction buffer (10 % Glycerol , 50mM Tris (pH 7.5),

1mM EDTA, 150mM NaCl, 10mM DTT , 1X cpi (protease inhibitor cocktail) and 0.1%

Tween 20) in 1.5 ml microcentrifuge tube. Tube was kept on ice until the powder was thawed completely in the extraction buffer. After that the mixture was thoroughly mixed by vortexing (low speed) for almost 10 seconds and kept on ice for 30 minutes with vortexing after every 10 minutes interval. Then centrifuged at 3,000×g for 10 min at 4°C and the supernatant was transferred to a new 1.5-mL microcentrifuge tube that was again centrifuged at full speed for 10 min at 4°C. Protein extract (supernatant) was transferred to a new tube.

For immune blot assay, samples were prepared as follows: 40 µl protein extract was mixed with 20 µl of 3X SDS loading dye, samples were then heated at 95oC for 5

140

minutes and incubated on ice for 2mins followed by centrifugation at 3,000 x g for 2 mins. Around 20 µl of this sample was loaded per well in 12% SDS/PAGE gell and run for 1 hour at 150 V in 1X electrode buffer. For transblotting PVDF membrane was used.

Membrane was cut, labelled, washed with methanol (2 times), dipped in ice cold transblot buffer (800mL of 1.25X trans blot buffer + freshly added 200mL methanol) and kept in freezer until used. The whole transfer assembly cassette was dipped in transblot buffer and kept in freezer. After setting transfer assembly, ran at 100V for 1 hour under cold conditions. Membrane was removed, rinsed with distilled water followed by rinsing with TBST. Membrane was blocked in TBST + 5% milk for 2 hours by gently agitating on moving plate. Membrane was again rinsed with TBST and 5ml of TBST + 5% milk + antibody (Anti-GFP-HRP [abcam # ab6663] 1:5000 or Anti-HA-HRP [abcam # ab1190]

1:2500 or Anti-HIS-HRP [clontech # 631210] 1:10,000 dilution in TBST ) was added to the container and kept on moving plate for 2 hours. Antibody solution was removed and

3-5 washings (each for 5 mins on moving plate) with TBST were done followed by staining with 10 ml developing buffer for 30 mins to 1 hour. After developing, membrane was washed with distilled water and air dried.

In planta Co-immunoprecipitation

For co-immunoprecipitation, agroinfiltration was done using A. tumefaciens cultures (OD600 of 0.8) carrying the respective plant expression constructs either individually (controls) or mixed in a 1:1 ratio (for Co-IP). Protein extraction was done as stated above with few amendments in the buffers (EDTA free). Extraction buffer was comprised of: 10mM Tris (pH 7.5), 150mM NaCl, 1X cpi (EDTA free protease inhibitor cocktail) and 0.1% Tween 20). GFP-Trap®_A beads(chromotek) beads were used for pull down according to manufacturer’s protocol with few amendments. All buffers used

141

in the process were EDTA free. Protein extract was spitted into two (100 µl in each) microcentrifuge tube, 1 for adding calcium and other for EGTA to make it calcium free.

Protein samples were then diluted to make the sample volume 550 µl. Save 50 µl of diluted sample at this stage for use as inputs in the blots.

Dilution buffers were prepared separately for each treatment by adding either

2mM CaCl2 or 2mM EGTA to extraction buffer. Diluted lysates were added to equilibrated GFP-Trap®_A beads (after washing) and were tumble end-over- end for 1 hour at 4°C followed by centrifugation at 2.500x g for 2 min at 4°C. Supernatant was discarded and beads were resuspend in the 500 μl of respective ice-cold dilution buffer for washing. They were again tumble end-over- end for 5 minutes at 4°C followed by centrifugation at 2.500x g for 2 min at 4°C. Washing was repeated 3 times. Finally, beads were resuspended in 100 μl 2x SDS-sample buffer and boiled resuspended GFP for 10 min at 95°C to dissociate immunocomplexes from GFP-Trap®_A beads. Beads were separated by centrifugation at 2.500x g for 2 min at 4°C and SDS-PAGE was performed with the supernatant as described above.

Hypersensitive Response Analysis in N. benthamiana Plants

HR was analyzed in wild type N. bethamiana (Nb/WT) plants or N. bethamiana that were stably transformed with the Rpi-Blb2 gene (Nb/Rpi-blb2) using the

Agrobacterium-mediated transient expression as described above. A. tumefaciens

GV3101 carrying appropriate WT Avrblb2 or non-calmodulin-binding Avrblb2 constructs were syringe-infiltrated on at least 20 spots into fully expanded leaves of Nb/WT or

Nb/Rpi-blb2, and visual HR (cell death) was recorded 2 days, 5 days and 15 days after inoculation. Each experiment was repeated three times.

142

Results

Calmodulin-binding to Avrblb2 is not Required for Localization of Avrblb2

Avrblb2 localizes to plasma membrane of the plant cells (Bozkurt et al., 2011).

Our BiFC assays also showed interaction of CaM with WT Avrblb2 as well as deletion its mutants, takes place at plasma membrane of the N. bethamiana cells (Figure 4-1, 4-

2 and 4-4). To determine whether the CaM binding to Avrblb2 is required for its localization to plant cell plasma membrane, both WT and mutant (NCB) Avrblb2 were tagged with YFP in plant expression vector (Figure 4-6A). Transient expression of these constructs in N. bethamiana plants followed by confocal microscopy at about 1-1 ½ dpi showed that both WT and NCB were localized at plasma membrane suggesting that

CaM binding to Avrblb2 is not contributing to its localization in the plant cells (Figure 4-

7).

Calmodulin-binding to AVRblb2 is not Required for AVRblb2 binding with C14

To determine whether the CaM binding to Avrblb2 affects its binding with C14 or not, C14 gene was fused with 3xHA at C-terminal in plant expression vector (Figure 4-

6A). Both YFP tagged WT-Avrblb2 and NCB-Avrblb2 were transiently coexpressed with

C14 in N. benthamiana followed by Co-IP either in the presence of Ca2+ or in Ca2+ free environment (Figure 4-9). Our results indicate that C14 binds with both WT-Avrblb2

(Figure 4-9A) and NCB-Avrblb2 (Figure 4-9B) but only in Ca2+ free environment that is only in the presence of Ca2+ chelator EGTA. Presence of Ca2+ in the reaction seems to break down this Avrblb2-C14 interaction (Figure 4-9). Our findings suggest that calmodulin binding to Avrblb2 is not required for it binding to C14 but Avrblb2-C14 binding is strongly affected by Ca2+ homeostasis in plant cells.

143

Calmodulin-binding to AVRblb2 is Required for Rpi-blb2-Dependent Cell Death

To determine whether CaM binding to Avrblb2 is essential for Rpi-blb2 mediated

HR or not, we used Rpi-blb2 transgenic N. benthamiana to test our non-calmodulin binding mutant of AVRblb2 (NCB-Avrblb2) with mutated CBD (78-82aa).

Agrobacterium-mediated infiltration of YFP tagged NCB-Avrblb2, WT-Avrblb2 (Figure 4-

6A) and A69F-Avrblb2 (BiFC construct with nYFP) was done in both wild type and Rpi- blb2 transgenic N. benthamiana plants and HR symptoms were followed from 1 to 15 days post inoculation (dpi). WT-Avrblb2 was used as positive control and Avrblb2-A69F, which is not recognized by Rpiblb2 (Oh et al., 2009) was used as a negative control.

Protein expression of these constructs was confirmed by looking for YFP under confocal microscope at 1dpi (Figure 4-7) and through western blot by using anti-GFP-HRP antibody (Figure 4-9C). As expected, none of the Avrblb2 variants caused any HR symptoms on WT N. benthamiana (Fig. 4-8A). Consistent with previous results, WT

Avrblb2 (PITG04090) did cause HR on transgenic N. benthamiana expressing Rpi-blb2 gene. HR symptoms were obvious at around 45% of spots infiltrated with WT-Avrblb2 only in Rpi-blb2 containing transgenic N. benthamiana (NB/Rpiblb2) leaves at 2dpi and almost 100% HR spots were observed at 15dpi (Fig. 4-8B). Mild HR symptoms started to appear as early as 1dpi and obvious cell degradation and disrupted YFP signals were observed under confocal microscope (Figure. 4-7). In contrast, NCB-Avrblb2 did not cause any HR just like A69F-Avrblb2 either 2 to 5 (Figure 4-8A and B) or even15 dpi.

Moreover, NCB-Avrblb2 expression and plasma membrane localization in both WT and transgenic plants were exactly like WT-Avrblb2 in WT plants (Figure. 4-7). These findings suggest that calmodulin binding to Avrblb2 does play a role in its recognition by

Rpi-blb2 receptor to induce cell death.

144

Discussion

Calmodulin is a ubiquitous Ca2+ sensor, reported to be involved in host defense responses. Several studies based on differential expression, overexpression or silencing of CaMs consistently showed their involvement in plant’s response to pathogens (Do Heo et al., 1999; Takabatake et al., 2007; Choi et al., 2009; Ali et al.,

2013; Cheval et al., 2013). However, direct physical interaction between pathogen effectors and host calmodulins is reported very recently in case of a Pseudomonas syringae effector HopE1 and a P. infestans effector SFI5 (Guo et al., 2016), (Zheng et al., 2018). Here we report the interaction of Avrblb2, a core RXLR effector of

Phytophthora species, with calmodulin. The expression of Avrblb2 is specifically induced during the early biotrophic stages of infection (Brasier, 2009; Haas et al., 2009) and the Avrblb2-CaM interaction could be speculated as an early event of host- pathogen cross talk to modulate Ca2+ dependent-defense signaling during the early stages of infection.

Avrblb2s belong to a polymorphic gene family reported to have undergone diversifying selection. Multiple Avrblb2 paralogs have been found in the genomes of different Phytophthora spp. (Oliva et al., 2015). We found that divergent Avrblb2 homologs of P. infestans, P. parasitica and P. sojae interact with calmodulins suggesting that the Avrblb2-CaM interactions are general, and they could be playing a conserved role in plant-Phytopthora interactions. CaMs are the calcium sensors that have multiple downstream protein targets to operate diverse pathways, combined outcomes of which results in specific response to particular stimuli. Thus, Avrblb2-CaM interaction could be a key for the pathogen to intervene multiple pathways. Moreover, reports on an effector targeting multiple diverse host proteins are scarce.

145

Bozkurt et al. reported that Avrblb2 promotes virulence of P.infestans in N. benthamiana by preventing the secretion of an important plant immune protease (C14) in the apoplast of infected cells hence, guarding the pathogen proteins from degradation by C14 (Bozkurt et al., 2011). Localization of Avrblb2 is important for its virulence function. Consistent with the previous findings we observed plasma membrane localization of Avrblb2 in planta (Bozkurt et al., 2011). Moreover, we found that Avrblb2 interacts with CaM primarily at the plasma membrane, which is also consistent with the localization of Avrblb2-C14 interaction (Bozkurt et al., 2011). We hypothesized that CaM binding to Avrblb2 might have some role in its localization to plasma membrane and interaction with C14. But our results show that calmodulin binding to Avrblb2 is neither required for its localization nor for its interaction with C14.

Surprisingly, we observed that Avrblb2-C14 interaction is only seen in the absence of Ca2+ (in the presence of Ca2+ chelator EGTA) and is broken upon adding

Ca2+ to the reaction environment. On the other hand, Avrblb2-CaM interaction takes place in the presence of Ca2+. On the basis of these findings, it can be speculated that the switch in the interacting partners of Avrblb2 pertains to Ca2+ homeostasis in the host cells. Spatio-temporal osscillations in free Ca2+ concentration (Ca2+ signatures) in response to biotic stress stimuli has been reported as a key early event and CaMs are primary Ca2+ sensors (Ranty et al., 2016). Different Ca2+ signatures are generated even by slightly different stimuli leading to either compatible/incompatible interactions (Reddy et al., 2011). It is well known that CaMs get activated upon binding Ca2+ and undergo conformational changes to bind its downstream targets (Kilhoffer et al., 1988).

Considering these facts, we propose that Avrblb2-CaM interaction takes place if there is

146

enough stimulus generate Ca2+ to activate CaM otherwise Avrblb2 interacts with C14 to promote infection. Outcome of Avrblb2-CaM interaction needs further investigation.

Rpi-blb2, the cognate R gene of Avrblb2 is well known to confer durable broad spectrum resistance against P. infestans and is being widely used for developing late blight resistant commercial potato cultivars (Oliva et al., 2015; Orbegozo et al., 2016).

The Avrblb2 CBD identified in our studies is located within the HR-eliciting domain of

Avrblb2 prompting us to hypothesize that calmodulin binding could have a role in the

Rpi-blb2 mediated HR. Our results indicate that the non-calmodulin-binding Avrblb2-

NCB failed to induce any HR in Rpi-blb2 expressing N. benthamiana plants, suggesting that Rpi-blb2 induced HR depends upon CaM binding with Avrblb2. Based on these observations, we speculate that CaM serve as a guard or decoy, which is monitored by

Rpi-blb2 for binding to Avrblb2 to induce HR. This is in contrast to the Avrblb2/C14 interaction, which was reported to not be involved in the recognition of Avrblb2 by Rpi- blb2 (Bozkurt et al., 2011). Currently, there is no evidence for direct binding of calmodulin to any canonical NB-LRR type R proteins. However, direct interaction of

CaM with the barley MLO, a non-NB-LRR type recessive resistance gene, or more accurately a dominant susceptibility S gene, was shown to play an essential role in regulating powdery mildew resistance (Kim et al., 2002).

Rpi-blb2 mediated cell death in response to PiAvrblb2 has been shown to require

SGT1, a eukaryotic co-chaperone widely reported to be involved in both PTI and ETI associated cell death responses (Abascal et al., 2014; Liu et al., 2016). The P. infestans INF1 elicitor-induced HR also require SGT1 (Abascal et al., 2014). Since

SGT1 is involved in Ca2+ signaling-dependent HR induction in pepper plants in

147

response to P. capsici INF1 elicitor (Liu et al., 2016), HR induction by Rpi-blb2 in response to Avrblb2-CaM interaction might have some direct or indirect link with SGT1 pathway. In the light of all these findings we suggest that calmodulin-dependent Ca2+ signaling plays a critical role in determining virulence or avirulence activity of Avrblb2 in host cells.

148

A B

M kDa pACT

pACT- CaM

YPD -L-W-H

C 2

M YFPn Avrblb2 A

n YFPc CaM3 c

Y YFPc GUS Y

D YFP RFP Merged

Y Y

Figure 4-1. The Phytophthora Avrblb2 effectors interact with calmodulin. A) The GAL4- DBD-Avrblb2 fusion protein was expressed in the yeast Mav203 cells. Control, Mav203 cells without any vector [left lane] and Mav203 carrying pDEST32-DBD-Avrblb2 (Avrblb2 fused to the Gal4-DNA binding domain) [right lane] were subjected to Western blot using anti-GAL4-DBD-HRP antibody. (B) DBD-Avrblb2 interacted with calmodulin fused to the Gal4- Activation Domain (pACT-CaM) in yeast, as is shown by growth on –L-W- H+3AT plates. pAS and pACT are empty vector, that express the GAL4-DBD and GAL4-ACT domain, respectively, and were used as negative controls. (C) Left panel: Schematics of constructs used for BiFC analyses. Avrblb2 was fused to YFPn (YFPn-PiAvrblb2) and CaM3 was fused to YFPc (YFPc- CaM3). GUS fused to YFPc was used as negative control. Right panel: BiFC analyses of the interaction of YFPn-PiAvrblb2 co-expressed with calmodulin YFPc-CaM3 in Arabidopsis protoplasts. Scale bar represents 10 µm. (D) In vivo BiFC analyses showing interaction of CaM3 with the Phytophthora infestans and the P. parasitica Avrblb2 homologs at the plasma membrane of leaf cells of N. benthamiana. Scale bar represents 100 µm. Photo courtesy of Gul Shad Ali.

149

Figure 4-2. The P. infestans and P. parasitica Avrblb2 interact with calmodulin in a Ca2+–dependent manner. (A) Schematics of 6x His-tagged CaM and GST- tagged Avrblb2 homologs: P. parasitica (PpAvrblb2), P. infestans (PiAvrblb2), and the P. infestans truncated Avrblb2 (Δ56-100). (B & C) Immunoblots of in vitro co-IP assays showing: the interaction of GST-PiAvrblb2 with HIS- CaM1(B) and the interaction of both GST-PiAvrblb2 and GST-PpAvrblb2 with HIS-CaM3 (C) in the presence of Ca2+ but not EGTA, a Ca2+-chelator. Both negative controls Δ56-100 and GUS didn’t show any CaM pull down indicating specificity of Avrblb2-CaM interactions. All fusion proteins were expressed in E. coli, inputs and IP were run on duplicate 12% SDS polyacrylamide gels and transferred to PVDF membranes. Expression of GST-Avrblb2and His-CaM fusion proteins was detected using an Anti-GST- HRP and Anti-His-HRP antibodies respectively.

150 4

1 A 7

a:GST-HRP

CaM-HRP

PITG20300 PITG18683 PITG04085 PITG20303

PPTG18954 PSTG159181 PSTG159257 GUS

1 A 7

5 5

P e C a:GST-HRP

CaM-HRP

PITG20300 PITG18683 PITG04085 PITG20303

PPTG18954 PSTG159181 PSTG159257 GUS

2+ Figure 4-3C. Most of the divergent Avrblb2 homologs interact with calmodulin in a Ca - dependent manner. (A) In vitro CaM-HRP overlay assay show that GST- tagged divergent Avrblb2 homologs from three Phytophthora species, P. infestans (PITG20300, PITG18683, PITG04085 and PITG20303), P. parasitica (PPTG18954) and P. sojae (PSG159181 and PSG159257) interact with calmodulin (CaM:HRP). GST-Avrblb2 fusion proteins, expressed in E. coli, were run on duplicate 12% SDS polyacrylamide gels and transferred to PVDF membranes. Expression of GST-Avrblb2 fusion proteins was detected using an Anti-GST-HRP antibody (Upper panel) and CaM-binding of GST- Avrblb2 proteins was assessed using CaM-binding overlay assays using AtCaM2::HRP antibody. (B) Transient BiFC analyses show that divergent Avrblb2 homologs interact with calmodulin in vivo. The indicated Avrblb2 homologs fused to YFPn were co-expressed with CaM3-YFPc in N. benthamiana leaves stably expressing a membrane-localized red fluorescent protein marker (RFP::PMRK). Green/Yellow fluorescence indicates BiFC of Avrblb2 with calmodulin. Scale bar represents 50 µm. (C) Phylogeny of Avrblb2 homologs. Amino acids 34 (48-81) were aligned using ClustalW and phylogenetic tree was constructed using the Maximum Likelihood (ML) method with 1000 bootstrap replication (Tamura and Nei, 1993; Tamura et al., 2011). Photo courtesy of Gul Shad Ali.

151

Figure 4-4. Calmodulin-binding site mapping of the P. infestans Avrblb2. A) CaM- overlay assays of PiAvrblb2 mutants performed to find out CaM binding region show that CaM binding activity was lost when a five amino acids C- terminal region from position 78 to 82 was deleted. Mutant NCB indicated by * in which all five amino acids (78-82) were replaced with Alanine also didn’t show any CaM binding. Expression of GST-Avrblb2 fusion proteins was detected using an Anti-GST-HRP antibody (Upper panel) and CaM-binding of PiAvrblb2 mutant proteins was assessed using CaM-binding overlay assays using AtCaM2::HRP antibody. (B) In planta BiFC assays showing interaction of PiAvrblb2 mutants fused to YFPn with CaM fused to YFPc. Just like CaM- overlay assay, mutants having intact CBD (77-81): Δ86-100, Δ82-100, Δ1-76, Δ1-45 and full length Avrbl2 show interaction with CaM and give green/yellow fluorescence. All others including NCB didn’t show interaction. BiFC analyses were carried out using transgenic N. benthamiana leaves expressing a membrane-localized red fluorescent protein marker (RFP::PMRK). Scale bar represents 50 µm. (C) Schematics of N-terminal and C-terminal deletion constructs of PiAvrblb2 and their binding with CaM. Boundaries of CBD are shown by black dotted lines. Photo courtesy of Gul Shad Ali.

152

A A A

a:GST-HRP

CaM-HRP

Figure 4-5. Identification of amino acids in Avrblb2, which are critical for interaction to calmodulin. (A) Immunoblot showing CaM-overlay assays of PiAvrblb2 variants mutated at potential anchor sites (left panels) and variants having polymorphism at position 69 (right panel). All these Avrblb2 variants showed CaM binding activity depicting that none of these amino acids play any role in CaM binding. A69F indicated by arrow is Avrblb2 variant that is not recognized by cognate R gene Rpi-blb2. All Avrblb2 variants were GST tagged and their expression was detected using an Anti-GST-HRP antibody (Upper panel) and CaM-binding of GST-Avrblb2 proteins was assessed using CaM-binding overlay assays using AtCaM2::HRP antibody. (B) Amino acid (64-90) alignment of Avrblb2 homologs of P. infestans, P. parasitica and P. sojae to visualize sequence variations and to identify potential anchor sites. and at position 69, in CBD. Positions 64, 65, 74 and 79 were identified as potential anchor sites. Position 69 is highly polymorphic. CBD show substantial variations in different Avrblb2 homologs. (C) Wheel projection diagram of Avrblb2 to show CBD.

153

A YFP AVRblb2

YFP AVRblb2(NCB)

C14-FL HA

AVRblb2 +ve -ve WT NCB C14 -ve +ve B C

55kda

33kda

Western blot: GFP-HRP Antibody Western blot: HA-HRP

Figure 4-6. Protein expression confirmation of Avrblb2 mutants and C14 cloned in plant expression vectors. A) N-terminal YFP tagged WT and mutant (NCB) Avrblb2 and C-terminal HA tagged C14 constructs. Western blots developed using Anti-GFP-HRP and Anti-HA-HRP exhibiting expression of: B) YFP tagged Avrblb2 constructs and C) HA tagged C14 in N. benthamiana plants 2 dpi respectively.

154

Nb/WT Nb/Rpi

Avrblb2 (Wt)

Avrblb2 (NCB)

Non YFP control

Figure 4-7. Visualization of YFP tagged WT and NCB in Nb/WT (left) and Nb/Rpi (right) leaves. No difference in plasma membrane localization of Wt and NCB was observed in both Nb/WT and Nb/Rpi leaves. Significant degradation of Nb/Rpi leaf spots infiltrated with WT Avrblb2 was observed that indicates initiation of HR symptoms at 1dpi. Scale bar represents 10 µm. Photo courtesy of Zunaira Afzal.

155

A WT Rpi-Blb2 WT Rpi-Blb2 F69 Wt Wt Wt F69 F69 Wt F69

I NCB

I NCB NCB NCB P NCB P NCB NCB D NCB

5 2 Wt Wt F69 F69 Wt F69 Wt F69

Nb/Rpi C Nb/WT Avrblb2 Avrblb2 Wt 69F NCB 69F Wt NCB _ 30 _ α: GFP-HRP antibody 25

Figure 4-8. The Avrblb2/Rpi-blb2-induced cell death is dependent on calmodulin binding to Avrblb2. A). Representative N. benthamiana leaves, 2 and15 days post Agro-infiltration of YFP tagged WT-Avrblb2 (Wt), Avrblb2-NCB (NCB) and Avrblb2-A69F (F69). HR symptoms are clearly visible at spots infiltrated with WT-Avrblb2 only in Rpi- blb2 containing transgenic N. benthamiana (NB/Rpiblb2) leaves both at 2 and 15 dpi whereas Avrblb2-NCB just like the negative control (Avrblb2-A69F) didn’t show any signs of HR even at 15 dpi. As expected, no HR spots were seen on Nb/WT leaves both at 2 and 15 dpi. B) Bar graph of infiltrated sites showing HR after 2 days, and 15 days of inoculation. Our non-calmodulin binding Avrblb2 mutant (NCB) didn’t show any HR spot compared to positive control, WT-Avrblb2 that showed around 45% visible HR spots at 2dpi and around 100% strong HR at 15dpi. C) Western blots of proteins extracted from Nb/WT and Nb/Rpi leaves at 1dpi. In Nb/WT (left three lanes) all three proteins Wt, F69 and NCB are expressed whereas in Nb/Rpi expression of F69 and NCB was comparable to their expression in Nb/WT but very faint band was observed for Wt Avrblb2 that indicates protein degradation even before visible HR. Anti-GFP-HRP antibody was used to detect expression of all fusion proteins. Note: BiFC construct of Avrblb2-A69F fused to YFPn was used as negative control, so smaller bands in F69 lanes are due to partial YFP. Photo courtesy of Zunaira Afzal. .

156

B A AvrRblb2-YFP (NCB) AvrRblb2-YFP (WT) C14-HA + – + – C14-HA + – + –

t α:GFP u

u α:GFP p

p n

n I I

α:HA α:HA

α:GFP α:GFP

P I

P -

- o

o C

C α:HA α:HA

EGTA Ca2+ EGTA Ca2+

Figure 4-9. In planta co-immunoprecipitation of Avrblb2 mutants with C14. A) YFP- Avrblb2 wild type and B) YFP-Avrblb2 non calmodulin binding mutant (NCB) were transiently expressed alone or coexpressed with C14-HA in N. benthamiana. Immuno-precipitates either in the presence of Ca2+ or EGTA obtained with GFP-Trap®_A beads were immuno-blotted with appropriate antisera. C14 binding can be seen with both WT-Avrblb2 and NCB-Avrblb2 only in in the presence of EGTA. Photo courtesy of Zunaira Afzal.

157

CHAPTER 5 SUMMARY AND CONCLUSIONS

Phytopathogenic oomycetes are associated with destructive diseases in many economically important crops worldwide. They are mainly distributed in the genus

Phytophthora. Currently, all known species within the genus Phytophthora are plant pathogens with varied life styles, infection modes and host ranges. Management of

Phytophthora diseases by manipulating host genetic resistance is being considered the most effective, eco-friendly and, in the long run cost-effective resistance development strategy. In order to develop sustainable resistance in crop plants against these noxious pathogens a thorough understanding of the molecular basis of plant-Phytophthora interactions and discovery of multiple sources of genetic resistance is very important.

Phytophthora parasitica is the most prevalent soilborne oomycete that can infect roots and stems of up to sixty different plant families including the model plant

Arabidopsis thaliana. Due to its widespread, broader host range and ability to infect model plants, P. parasitica has now emerged as a model of soilborne oomycete pathogens. Identification of genetic sources of resistance against P. parasitica can serve as a valuable resource for developing Phytophthora resistant crops in the future.

In our laboratory, several accessions of wild tomato species were tested against

P. parasitica infection and some highly resistant and susceptible accessions were identified. Current understanding of tomato-Phytophthora parasitica interaction is very limited and nothing is known at the whole genome or transcriptome level. In order to identify genes that potentially contribute to defense against P. parasitica in tomato, transcriptomes of a resistant and a susceptible Solanum pimpinellifolium accession in response to P. parasitica infection were analyzed and compared using RNA-seq

158

technology. In treatment vs control comparison, 2657 and 3079 differentially expressed genes (DEGs) were identified in resistant (Sp-R) and susceptible (Sp-S) samples respectively. Our results revealed 1173 genes that were differentially expressed only in

Sp-R accession upon P. parasitica inoculation. These exclusively found DEGs in Sp-R accession included some core plant defense genes for example, several protease inhibitors, chitinases, defensin, PR-1, a downy mildew susceptibility factor etc. were all highly induced. Whereas, several R genes, WRKY transcription factors and a powdery mildew susceptibility gene (Mlo) were highly repressed during the resistance outcome.

P. parasitica is known to cause root rot, foot rot/gummosis and brown rot of fruits in citrus. Carrizo citrange, a famous citrus rootstock is considered resistant against P. parasitica. Transcriptome of this rootstock in response to P. parasitica infection was analyzed with an aim to identify R genes involved in resistance development. In total,

6692 differentially expressed transcripts among P. parasitica-inoculated and mock- treated roots were identified. Of these, 3960 were differentially expressed at 24 hours post inoculation and 5521 were differentially expressed at 48 hours post inoculation.

Out of a total of 454 R genes found in the root transcriptome of citrus, 186 were found differentially expressed in response to P. parasitica infection. Of those, 100 were upregulated in the range of 4–32,768 times and 80 were downregulated (4 – 256 fold change). The top differentially expressed R gene was a G-type lectin S-receptor-like serine threonine- kinase (G-type RLK) was around 8000 times more expressed in response to P. parasitica at 24 hpi and reached up to 32000 times upregulated at 48 hpi. Five other G-type RLKs were also found to be upregulated at 1 or both time points.

Involvement of G-type RLKs is being reported first time in plant-Phytophthora

159

interaction. In addition to this, several other genes putatively involved in plant immunity were found differentially modulated in citrus roots in response to P. parasitica infection.

Transcriptional reprogramming of most of the cellular and physiological processes and all components of biotic stress responses like modulation of phytohormone signaling, proteolysis, R genes, transcription factors and calcium signaling in both tomato and citrus indicated significant activation of defense responses against P. parasitica. The results of these transcriptome analysis lay out a strong foundation for future studies aimed at improving genetic resistance of crop against

Phytopphthora species. Functional significance of defense related genes identified here needs further investigation.

As observed in our tomato-Phytophthora and citrus-Phytophthora transcriptome studies as well as tons of reports from literature, it is evident that calcium signaling plays an essential role in plant-Phytophthora interactions. In our laboratory, extensive screening of a tomato cDNA library against RXLR effectors (involved in infection), revealed interaction of Avrblb2, a core effector involved in infection with calmodulins

(CAMs), the primary calcium sensors. Upon pathogen invasion, fluctuations in the Ca2+ ion concentration of the host’s cells is essentially the key event involved in the perception of threat. These Ca2+ fluctuations are sensed by CaMs that can directly bind to Ca2+ ions and undergo conformational changes, enabling them to affect the function of their down-stream targets that ultimately lead to the development of defense responses. Physical interaction of Phytophthora effectors with CAMs points towards the manipulation of host’s calcium signaling pathway by the pathogen to modulate defense

160

responses. The second major goal of my research was to explore the implications of

Avrblb2-CaM interaction on plant defense responses.

Virulence function of Avrblb2 was associated with preventing the secretion of

C14, a plant defense protease to the apoplast apparently to protect the intercellular growth of Phytophthora infestans. Localization of Avrblb2 is important for its virulence function. Avrblb2 interacts with CaM at the plasma membrane, which is also consistent with the localization of Avrblb2-C14 interaction. We hypothesized that CaM binding to

Avrblb2 might have some role in its localization to plasma membrane and interaction with C14 and its secretion. But our results show that calmodulin binding to Avrblb2 is neither required for its localization nor for its interaction with C14. Avrblb2-CaM interaction takes place in the presence of Ca2+ whereas, Avrblb2-C14 interaction was only seen in the absence of Ca2+ and is broken upon adding Ca2+ to the reaction environment. Spatio-temporal oscillations in free Ca2+ concentration (Ca2+ signatures) in response to biotic stress stimuli has been reported as a key early event and different

Ca2+ signatures are generated even by slightly different stimuli leading to either compatible or incompatible interactions. Considering these facts, it is proposed that the switch in the interacting partners of Avrblb2 pertains to Ca2+ homeostasis in the host cells and Avrblb2-CaM interaction takes place if there is enough stimulus generate Ca2+ to activate CaM otherwise Avrblb2 interacts with C14 to promote infection (compatible interaction).

161

eliciting domain of Avrblb2 prompting us to hypothesize that Avrblb2-CaM interaction could have a role in the Rpi-blb2 mediated HR. The non-calmodulin-binding Avrblb2-

NCB failed to induce any HR in Rpi-blb2 expressing N. benthamiana plants, suggesting that Rpi-blb2 induced HR depends upon CaM binding with Avrblb2. Based on these findings, it is suggested that Avrblb2 by interacting with calmodulin interfere with plant defense associated Ca2+ signaling in plants. However, the outcome (in disease promotion or suppression) of Avrblb2-CaM interaction in the absence of Rpi-blb2 gene needs further investigation.

162

LIST OF REFERENCES

Aldon, D., Mbengue, M., Mazars, C., and Galaud, J.-P. (2018). Calcium signalling in plant biotic interactions. International journal of molecular sciences 19, 665. Amelot, N., Dorlhac de Borne, F., San Clemente, H., Mazars, C., Grima-Pettenati, J., and Brière, C. (2012). Transcriptome analysis of tobacco BY-2 cells elicited by cryptogein reveals new potential actors of calcium-dependent and calcium- independent plant defense pathways. Cell Calcium 51, 117-130. Amelot, N., Carrouche, A., Danoun, S., Bourque, S., Haiech, J., Pugin, A., Ranjeva, R., Grima-Pettenati, J., Mazars, C., and Briere, C. (2011). Cryptogein, a fungal elicitor, remodels the phenylpropanoid metabolism of tobacco cell suspension cultures in a calcium-dependent manner. Plant Cell Environ 34, 149-161. Anderson, J.P., Gleason, C.A., Foley, R.C., Thrall, P.H., Burdon, J.B., and Singh, K.B. (2010). Plants versus pathogens: an evolutionary arms race. Functional plant biology 37, 499-512. Anderson, R.G., Deb, D., Fedkenheuer, K., and McDowell, J.M. (2015). Recent progress in RXLR effector research. Molecular Plant-Microbe Interactions 28, 1063-1072. Atkinson, M.M., Keppler, L.D., Orlandi, E.W., Baker, C.J., and Mischke, C.F. (1990). Involvement of plasma membrane calcium influx in bacterial induction of the K+/H+ and hypersensitive responses in tobacco. Plant physiology 92, 215-221. Attard, A.s., Evangelisti, E., Kebdani-Minet, N.m., PanabiÃ¨res, F., Deleury, E., Maggio, C., Ponchet, M., and Gourgues, M. (2014). Transcriptome dynamics of Arabidopsis thaliana root penetration by the oomycete pathogen Phytophthora parasitica. BMC genomics 15, 538. Bai, Y., Pavan, S., Zheng, Z., Zappel, N.F., ReinstÃ¤dler, A., Lotti, C., De Giovanni, C., Ricciardi, L., Lindhout, P., and Visser, R. (2008). Naturally occurring broad- spectrum powdery mildew resistance in a Central American tomato accession is caused by loss of Mlo function. Molecular Plant-Microbe Interactions 21, 30-39. Becktell, M., Smart, C., Haney, C., and Fry, W. (2006). Host Pathogen Interactions Between Phytophthora infestans and the Solanaceous Hosts Calibrachoa x hybridus, Petunia x hybrida, and Nicotiana benthamiana. Plant Disease 90, 24-32. Bergot, M., Cloppet, E., Perarnaud, V., Deque, M., Marcais, B., and Desprez-Loustau, M.-L. (2004). Simulation of potential range expansion of oak disease caused by Phytophthora cinnamomi under climate change. Global Change Biology 10, 1539- 1552. Blume, B., Nurnberger, T., Nass, N., and Scheel, D. (2000). Receptor-mediated increase in cytoplasmic free calcium required for activation of pathogen defense in parsley. The Plant Cell 12, 1425-1440. Boava, L.P., Cristofani-Yaly, M.n., Mafra, V.r.S., Kubo, K., Kishi, L.T., Takita, M.A., Ribeiro-Alves, M., and Machado, M.A. (2011). Global gene expression of Poncirus

163

trifoliata, Citrus sunki and their hybrids under infection of Phytophthora parasitica. BMC genomics 12, 39. Boevink, P.C., McLellan, H., Gilroy, E.M., Naqvi, S., He, Q., Yang, L., Wang, X., Turnbull, D., Armstrong, M.R., and Tian, Z. (2016a). Oomycetes seek help from the plant: Phytophthora infestans effectors target host susceptibility factors. Molecular Plant 9, 636-638. Boevink, P.C., Wang, X., McLellan, H., He, Q., Naqvi, S., Armstrong, M.R., Zhang, W., Hein, I., Gilroy, E.M., and Tian, Z. (2016b). A Phytophthora infestans RXLR effector targets plant PP1c isoforms that promote late blight disease. Nature communications 7. Bos, J.I., Kanneganti, T.D., Young, C., Cakir, C., Huitema, E., Win, J., Armstrong, M.R., Birch, P.R., and Kamoun, S. (2006). The C terminal half of Phytophthora infestans RXLR effector AVR3a is sufficient to trigger R3a-mediated hypersensitivity and suppress INF1-induced cell death in Nicotiana benthamiana. The Plant Journal 48, 165-176. Bourque, S., Binet, M.N., Ponchet, M., Pugin, A., and Lebrun-Garcia, A. (1999). Characterization of the cryptogein binding sites on plant plasma membranes. J Biol Chem 274, 34699-34705. Bozkurt, T.O., Schornack, S., Win, J., Shindo, T., Ilyas, M., Oliva, R., Cano, L.M., Jones, A.M., Huitema, E., and van der Hoorn, R.A. (2011). Phytophthora infestans effector AVRblb2 prevents secretion of a plant immune protease at the haustorial interface. Proceedings of the National Academy of Sciences 108, 20832-20837. Bruni, G.N., Weekley, R.A., Dodd, B.J., and Kralj, J.M. (2017). Voltage-gated calcium flux mediates Escherichia coli mechanosensation. Proceedings of the National Academy of Sciences 114, 9445-9450. Chang, Y.H., Yan, H.Z., and Liou, R.F. (2015). A novel elicitor protein from Phytophthora parasitica induces plant basal immunity and systemic acquired resistance. Mol Plant Pathol 16, 123-136. Cheval, C.c., Aldon, D., Galaud, J.-P., and Ranty, B.t. (2013). Calcium/calmodulin- mediated regulation of plant immunity. Biochimica et Biophysica Acta (BBA)- Molecular Cell Research 1833, 1766-1771. Choi, H.W., Lee, D.H., and Hwang, B.K. (2009). The pepper calmodulin gene CaCaM1 is involved in reactive oxygen species and nitric oxide generation required for cell death and the defense response. Molecular Plant-Microbe Interactions 22, 1389- 1400. Colas, V., Conrod, S., Venard, P., Keller, H., Ricci, P., and Panabières, F. (2001). Elicitin genes expressed in vitro by certain tobacco isolates of Phytophthora parasitica are down regulated during compatible interactions. Mol Plant Microbe Interact 14, 326- 335.

164

Cui, J., Luan, Y., Jiang, N., Bao, H., and Meng, J. (2017). Comparative transcriptome analysis between resistant and susceptible tomato allows the identification of lnc RNA 16397 conferring resistance to Phytophthora infestans by co‐expressing glutaredoxin. The Plant Journal 89, 577-589. Cui, J., Xu, P., Meng, J., Li, J., Jiang, N., and Luan, Y. (2018). Transcriptome signatures of tomato leaf induced by Phytophthora infestans and functional identification of transcription factor SpWRKY3. Theoretical and Applied Genetics 131, 787-800. Dagdas, Y.F., Belhaj, K., Maqbool, A., Chaparro-Garcia, A., Pandey, P., Petre, B., Tabassum, N., Cruz-Mireles, N., Hughes, R.K., and Sklenar, J. (2016). An effector of the Irish potato famine pathogen antagonizes a host autophagy cargo receptor. Elife 5, e10856. Dahan, J., Pichereaux, C., Rossignol, M., Blanc, S., Wendehenne, D., Pugin, A., and Bourque, S. (2009). Activation of a nuclear-localized SIPK in tobacco cells challenged by cryptogein, an elicitor of plant defence reactions. Biochem J 418, 191-200. Derevnina, L., Dagdas, Y.F., De la Concepcion, J.C., Bialas, A., Kellner, R., Petre, B., Domazakis, E., Du, J., Wu, C.H., Lin, X., Aguilera-Galvez, C., Cruz-Mireles, N., Vleeshouwers, V.G., and Kamoun, S. (2016). Nine things to know about elicitins. New Phytol 212, 888-895. Do Heo, W., Lee, S.H., Kim, M.C., Kim, J.C., Chung, W.S., Chun, H.J., Lee, K.J., Park, C.Y., Park, H.C., and Choi, J.Y. (1999). Involvement of specific calmodulin isoforms in salicylic acid-independent activation of plant disease resistance responses. Proceedings of the National Academy of Sciences 96, 766-771. Doughari, J. (2015). An Overview of Plant Immunity. Plant Pathology & Microbiology. Du, H., Wang, Y., Yang, J., and Yang, W. (2015a). Comparative transcriptome analysis of resistant and susceptible tomato lines in response to infection by Xanthomonas perforans race T3. Frontiers in Plant Science 6, 1173. Du, J., Verzaux, E., Chaparro-Garcia, A., Bijsterbosch, G., Keizer, L.C., Zhou, J., Liebrand, T.W., Xie, C., Govers, F., Robatzek, S., van der Vossen, E.A., Jacobsen, E., Visser, R.G., Kamoun, S., and Vleeshouwers, V.G. (2015b). Elicitin recognition confers enhanced resistance to Phytophthora infestans in potato. Nat Plants 1, 15034. Du, Y., Mpina, M.H., Birch, P.R., Bouwmeester, K., and Govers, F. (2015c). Phytophthora infestans RXLR effector AVR1 interacts with exocyst component Sec5 to manipulate plant immunity. Plant Physiology, pp. 01169.02015. Fawke, S., Doumane, M., and Schornack, S. (2015). Oomycete interactions with plants: infection strategies and resistance principles. Microbiol Mol Biol Rev 79, 263-280. Gilroy, E.M., Taylor, R.M., Hein, I., Boevink, P., Sadanandom, A., and Birch, P.R. (2011). CMPG1-dependent cell death follows perception of diverse pathogen elicitors at

165

the host plasma membrane and is suppressed by Phytophthora infestans RXLR effector AVR3a. New Phytologist 190, 653-666. Goodwin, P., Kirkpatrick, B., and Duniway, J. (1989). Cloned DNA probes for identification of Phytophthora parasitica. Phytopathology 79, 716-721. Grant, J.J., Yun, B.-W., and Loake, G.J. (2000). Oxidative burst and cognate redox signalling reported by luciferase imaging: identification of a signal network that functions independently of ethylene, SA and Me-JA but is dependent on MAPKK activity. The Plant Journal 24, 569-582. Guo, M., Kim, P., Li, G., Elowsky, C.G., and Alfano, J.R. (2016). A Bacterial Effector Co- opts Calmodulin to Target the Plant Microtubule Network. Cell host & microbe 19, 67-78. Gust, A.A., Biswas, R., Lenz, H.D., Rauhut, T., Ranf, S., Kemmerling, B., Gotz, F., Glawischnig, E., Lee, J., and Felix, G. (2007). Bacteria-derived peptidoglycans constitute pathogen-associated molecular patterns triggering innate immunity in Arabidopsis. Journal of Biological Chemistry 282, 32338-32348. Haas, B.J., Kamoun, S., Zody, M.C., Jiang, R.H., Handsaker, R.E., Cano, L.M., Grabherr, M., Kodira, C.D., Raffaele, S., and Torto-Alalibo, T. (2009). Genome sequence and analysis of the Irish potato famine pathogen Phytophthora infestans. Nature 461, 393-398. Hansen, E.M. (2015). Phytophthora species emerging as pathogens of forest trees. Current Forestry Reports 1, 16-24. Harding, S.A., Oh, S.H., and Roberts, D.M. (1997). Transgenic tobacco expressing a foreign calmodulin gene shows an enhanced production of active oxygen species. The EMBO Journal 16, 1137-1144. Hoeberichts, F.A., Davoine, C., Vandorpe, M., Morsa, S., Ksas, B., Stassen, C., Triantaphylidès, C., and Van Breusegem, F. (2013). Cryptogein-induced transcriptional reprogramming in tobacco is light dependent. Plant Physiol 163, 263-275. Isozumi, N., Inoue, Y., Imamura, T., Mori, M., Takano, Y., and Ohki, S. (2019). Ca2+- dependent interaction between calmodulin and CoDN3, an effector of Colletotrichum orbiculare. Biochemical and biophysical research communications. Jones, J.D., and Dangl, J.L. (2006). The plant immune system. Nature 444, 323. Kamoun, S., Klucher, K.M., Coffey, M.D., and Tyler, B.M. (1993). A gene encoding a host- specific elicitor protein of Phytophthora parasitica. Mol Plant Microbe Interact 6, 573-581. Kamoun, S., van West P, Vleeshouwers, V.G., de Groot KE, and Govers, F. (1998). Resistance of Nicotiana benthamiana to Phytophthora infestans is mediated by the recognition of the elicitor protein INF1. Plant Cell 10, 1413-1426.

166

Kamoun, S., Furzer, O., Jones, J.D., Judelson, H.S., Ali, G.S., Dalio, R.J., Roy, S.G., Schena, L., Zambounis, A., and PanabiÃ¨res, F. (2015). The Top 10 oomycete pathogens in molecular plant pathology. Molecular plant pathology 16, 413-434. Kanzaki, H., Saitoh, H., Ito, A., Fujisawa, S., Kamoun, S., Katou, S., Yoshioka, H., and Terauchi, R. (2003). Cytosolic HSP90 and HSP70 are essential components of INF1-mediated hypersensitive response and non-host resistance to Pseudomonas cichorii in Nicotiana benthamiana. Molecular plant pathology 4, 383-391. Kawamura, Y., Hase, S., Takenaka, S., Kanayama, Y., Yoshioka, H., Kamoun, S., and Takahashi, H. (2009). INF1 Elicitin Activates Jasmonic Acid- and Ethylene- mediated Signalling Pathways and Induces Resistance to Bacterial Wilt Disease in Tomato (Journal of Phytopathology), pp. 287-297. Keller, H., Blein, J.-P., Bonnet, P., and Ricci, P. (1996a). Physiological and molecular characteristics of elicitin-induced systemic acquired resistance in tobacco. Plant Physiology 110, 365-376. Keller, H., Bonnet, P., Galiana, E., Pruvot, L., Friedrich, L., Ryals, J., and Ricci, P. (1996b). Salicylic acid mediates elicitin-induced systemic acquired resistance, but not necrosis in tobacco. Molecular plant-microbe interactions: MPMI (USA). Kim, M.C., Panstruga, R., Elliott, C., Muller, J., Devoto, A., Yoon, H.W., Park, H.C., Cho, M.J., and Schulze-Lefert, P. (2002). Calmodulin interacts with MLO protein to regulate defence against mildew in barley. Nature 416, 447-451. King, S.R., McLellan, H., Boevink, P.C., Armstrong, M.R., Bukharova, T., Sukarta, O., Win, J., Kamoun, S., Birch, P.R., and Banfield, M.J. (2014). Phytophthora infestans RXLR effector PexRD2 interacts with host MAPKKKε to suppress plant immune signaling. The Plant Cell 26, 1345-1359. Kliejunas, J.T. (2010). Sudden oak death and Phytophthora ramorum: a summary of the literature. Koornneef, M., and Meinke, D. (2010). The development of Arabidopsis as a model plant. The Plant Journal 61, 909-921. Kudla, J., Batistic, O., and Hashimoto, K. (2010). Calcium signals: the lead currency of plant information processing. The Plant Cell 22, 541-563. Latijnhouwers, M., de Wit, P.J., and Govers, F. (2003). Oomycetes and fungi: similar weaponry to attack plants. Trends in microbiology 11, 462-469. Le Berre, J.Y., Engler, G., and Panabieres, F. (2008). Exploration of the late stages of the tomato Phytophthora parasitica interactions through histological analysis and generation of expressed sequence tags. New Phytologist 177, 480-492. Lebrun-Garcia, A., Ouaked, F., Chiltz, A., and Pugin, A. (1998). Activation of MAPK homologues by elicitors in tobacco cells. Plant J 15, 773-781. Lecourieux, D., Ranjeva, R., and Pugin, A. (2006). Calcium in plant defence-signalling pathways. New Phytol 171, 249-269.

167

Lecourieux, D., Mazars, C., Pauly, N., Ranjeva, R., and Pugin, A. (2002). Analysis and effects of cytosolic free calcium increases in response to elicitors in Nicotiana plumbaginifolia cells. The Plant Cell 14, 2627-2641. Lochman, J., Kasparovsky, T., Damborsky, J., Osman, H., Marais, A., Chaloupkova, R., Ponchet, M., Blein, J.P., and Mikes, V. (2005). Construction of cryptogein mutants, a proteinaceous elicitor from Phytophthora, with altered abilities to induce a defense reaction in tobacco cells. Biochemistry 44, 6565-6572. Marchadier, E., Oates, M.E., Fang, H., Donoghue, P.C., Hetherington, A.M., and Gough, J. (2016). Evolution of the calcium-based intracellular signaling system. Genome biology and evolution 8, 2118-2132. McGowan, J., and Fitzpatrick, D.A. (2017). Genomic, Network, and Phylogenetic Analysis of the Oomycete Effector Arsenal. mSphere 2. McLellan, H., Boevink, P.C., Armstrong, M.R., Pritchard, L., Gomez, S., Morales, J., Whisson, S.C., Beynon, J.L., and Birch, P.R. (2013). An RxLR effector from Phytophthora infestans prevents re-localisation of two plant NAC transcription factors from the endoplasmic reticulum to the nucleus. PLoS pathogens 9, e1003670. Meng, Y., Zhang, Q., Ding, W., and Shan, W. (2014). Phytophthora parasitica: a model oomycete plant pathogen. Mycology 5, 43-51. Messenger, B., Menge, J., and Pond, E. (2000). Effects of gypsum soil amendments on avocado growth, soil drainage, and resistance to Phytophthora cinnamomi. Plant Disease 84, 612-616. Mikes, V., Milat, M.L., Ponchet, M., Panabières, F., Ricci, P., and Blein, J.P. (1998). Elicitins, proteinaceous elicitors of plant defense, are a new class of sterol carrier proteins. Biochem Biophys Res Commun 245, 133-139. Mitra, M. (1929). Phytophthora parasitica dast. Causing damping off disease of cotton seedlings and fruit-rot of Guava in India. Transactions of the British Mycological Society 14, 249-254. Mounika, K., Panja, B., and Saha, J. (2017). Diseases of Peace lily [Spathiphyllum sp.] caused by fungi, bacteria and viruses: A review. The Pharma Innovation 6, 103. Mondragón-Palomino, M., John-Arputharaj, A., Pallmann, M., Dresselhaus, T. (2017). Similarities between reproductive and immune pistil transcriptomes of Arabidopsis species. Plant Physiol. 174, 1559–1575. Neher, D., and Duniway, J. (1992). Dispersal of Phytophthora parasitica in tomato fields by furrow irrigation. Plant Disease 76, 582-586. Nelson, S.C. (2008). Late blight of tomato (Phytophthora infestans). Nowicki, M., Foolad, M.R., Nowakowska, M., and Kozik, E.U. (2012). Potato and tomato late blight caused by Phytophthora infestans: an overview of pathology and resistance breeding. Plant Disease 96, 4-17.

168

Oh, S.-K., Kwon, S.-Y., and Choi, D. (2014). Rpi-blb2-mediated hypersensitive cell death caused by Phytophthora infestans AVRblb2 requires SGT1, but not EDS1, NDR1, salicylic acid-, jasmonic acid-, or ethylene-mediated signaling. The plant pathology journal 30, 254. Orlikowski, L.B., Ptaszek, M., Orlikowska, T., and Trzewik, A. (2010). Phytophthora species, new pathogens in landscape and gardens. Vytauto Didziojo universiteto Botanikos sodo rastai, 2010, t. 14, p. 133-138. Park, S.-W., Kaimoyo, E., Kumar, D., Mosher, S., and Klessig, D.F. (2007). Methyl salicylate is a critical mobile signal for plant systemic acquired resistance. Science 318, 113-116. Peng, K.C., Wang, C.W., Wu, C.H., Huang, C.T., and Liou, R.F. (2015). Tomato SOBIR1/EVR Homologs Are Involved in Elicitin Perception and Plant Defense Against the Oomycete Pathogen Phytophthora parasitica. Mol Plant Microbe Interact 28, 913-926. Pinckard, J., and Guidroz, G.F. (1973). A boll rot of cotton caused by Phytophthora parasitica. Phytopathology 63, 896-899. Podger, F. (1978). Phytophthora cinnamomi a cause of lethal disease of indigenous plant communities. Phytopathology 62, 972-981. Qiao, Y., Shi, J., Zhai, Y., Hou, Y., and Ma, W. (2015). Phytophthora effector targets a novel component of small RNA pathway in plants to promote infection. Proceedings of the National Academy of Sciences 112, 5850-5855. Qiao, Y., Liu, L., Xiong, Q., Flores, C., Wong, J., Shi, J., Wang, X., Liu, X., Xiang, Q., and Jiang, S. (2013). Oomycete pathogens encode RNA silencing suppressors. Nature genetics 45, 330-333. Ranf, S., Wunnenberg, P., Lee, J., Becker, D., Dunkel, M., Hedrich, R., Scheel, D., and Dietrich, P. (2008). Loss of the vacuolar cation channel, AtTPC1, does not impair Ca2+ signals induced by abiotic and biotic stresses. The Plant Journal 53, 287- 299. Ranty, B.t., Aldon, D., Cotelle, V.r., Galaud, J.-P., Thuleau, P., and Mazars, C. (2016). Calcium sensors as key hubs in plant responses to biotic and abiotic stresses. Frontiers in Plant Science 7. Reddy, A., Reddy, V.S., and Golovkin, M. (2000). A calmodulin binding protein from Arabidopsis is induced by ethylene and contains a DNA-binding motif. Biochemical and biophysical research communications 279, 762-769. Reddy, A.S., Ali, G.S., Celesnik, H., and Day, I.S. (2011). Coping with stresses: roles of calcium-and calcium/calmodulin-regulated gene expression. The Plant Cell 23, 2010-2032. Rhoads, A.R., and Friedberg, F. (1997). Sequence motifs for calmodulin recognition. The FASEB Journal 11, 331-340.

169

Rodewald, J., and Trognitz, B. (2013). Solanum resistance genes against Phytophthora infestans and their corresponding avirulence genes. Molecular plant pathology 14, 740-757. Sanseverino, W., Roma, G., De Simone, M., Faino, L., Melito, S., Stupka, E., Frusciante, L., and Ercolano, M.R. (2009). PRGdb: a bioinformatics platform for plant resistance gene analysis. Nucleic acids research 38, D814-D821. Sasabe, M., Takeuchi, K., Kamoun, S., Ichinose, Y., Govers, F., Toyoda, K., Shiraishi, T., and Yamada, T. (2000). Independent pathways leading to apoptotic cell death, oxidative burst and defense gene expression in response to elicitin in tobacco cell suspension culture. Eur J Biochem 267, 5005-5013. Saunders, D.G., Breen, S., Win, J., Schornack, S., Hein, I., Bozkurt, T.O., Champouret, N., Vleeshouwers, V.G., Birch, P.R., and Gilroy, E.M. (2012). Host protein BSL1 associates with Phytophthora infestans RXLR effector AVR2 and the Solanum demissum immune receptor R2 to mediate disease resistance. The Plant Cell 24, 3420-3434. Shearer, B., Crane, C., and Cochrane, A. (2004). Quantification of the susceptibility of the native flora of the South-West Botanical Province, Western Australia, to Phytophthora cinnamomi. Australian Journal of Botany 52, 435-443. Takabatake, R., Karita, E., Seo, S., Mitsuhara, I., Kuchitsu, K., and Ohashi, Y. (2007). Pathogen-induced calmodulin isoforms in basal resistance against bacterial and fungal pathogens in tobacco. Plant and Cell Physiology 48, 414-423. Tavernier, E., Wendehenne, D., Blein, J.-P., and Pugin, A. (1995). Involvement of free calcium in action of cryptogein, a proteinaceous elicitor of hypersensitive reaction in tobacco cells. Plant Physiology 109, 1025-1031. Taylor, J.E., Chow, J.Y., Jeffries, C.M., Kwan, A.H., Duff, A.P., Hamilton, W.A., and Trewhella, J. (2012). Calmodulin binds a highly extended HIV-1 MA protein that refolds upon its release. Biophysical journal 103, 541-549. Truman, W., Sreekanta, S., Lu, Y., Bethke, G., Tsuda, K., Katagiri, F., and Glazebrook, J. (2013). The calmodulin-binding protein60 family includes both negative and positive regulators of plant immunity. Plant Physiology 163, 1741-1751. Turnbull, D., Yang, L., Naqvi, S., Breen, S., Welsh, L., Stephens, J., Morris, J., Boevink, P.C., Hedley, P.E., Zhan, J., Birch, P.R.J., and Gilroy, E.M. (2017). RXLR Effector AVR2 Up-Regulates a Brassinosteroid-Responsive bHLH Transcription Factor to Suppress Immunity. Plant Physiol 174, 356-369. Tyler, B.M. (2007). Phytophthora sojae: root rot pathogen of soybean and model oomycete. Molecular plant pathology 8, 1-8. Vernooij, B., Friedrich, L., Morse, A., Reist, R., Kolditz-Jawhar, R., Ward, E., Uknes, S., Kessmann, H., and Ryals, J. (1994). Salicylic acid is not the translocated signal responsible for inducing systemic acquired resistance but is required in signal transduction. The Plant Cell 6, 959-965.

170

Vetukuri, R.R., Whisson, S.C., and Grenville-Briggs, L.J. (2017). Phytophthora infestans effector Pi14054 is a novel candidate suppressor of host silencing mechanisms. European Journal of Plant Pathology, 1-7. Wang, X., Boevink, P., McLellan, H., Armstrong, M., Bukharova, T., Qin, Z., and Birch, P.R. (2015). A host KH RNA-binding protein is a susceptibility factor targeted by an RXLR effector to promote late blight disease. Molecular Plant 8, 1385-1395. Wen, Y., Lim, G.X.-Y., and Wong, S.-M. (2013). Profiling of genes related to cross protection and competition for NbTOM1 by HLSV and TMV. PLoS ONE 8, e73725. Wendehenne, D., Lamotte, O., Frachisse, J.M., Barbier-Brygoo, H., and Pugin, A. (2002). Nitrate efflux is an essential component of the cryptogein signaling pathway leading to defense responses and hypersensitive cell death in tobacco. Plant Cell 14, 1937-1951. Xu, H., and Heath, M.l.C. (1998). Role of calcium in signal transduction during the hypersensitive response caused by basidiospore-derived infection of the cowpea rust fungus. The Plant Cell 10, 585-597. Yang, L., McLellan, H., Naqvi, S., He, Q., Boevink, P.C., Armstrong, M., Giuliani, L.M., Zhang, W., Tian, Z., and Zhan, J. (2016). Potato NPH3/RPT2-Like Protein StNRL1, targeted by a Phytophthora infestans RXLR effector, is a susceptibility factor. Plant Physiology 171, 645-657. Zheng, X., Wagener, N., McLellan, H., Boevink, P.C., Hua, C., Birch, P.R., and Brunner, F.J.N.P. (2018a). Phytophthora infestans RXLR effector SFI 5 requires association with calmodulin for PTI/MTI suppressing activity. Zheng, X., Wagener, N., McLellan, H., Boevink, P.C., Hua, C., Birch, P.R.J., and Brunner, F. (2018b). Phytophthora infestans RXLR effector SFI5 requires association with calmodulin for PTI/MTI suppressing activity. New Phytol 219, 1433-1446. Zhu, S., Li, Y., Vossen, J.H., Visser, R.G., and Jacobsen, E. (2012). Functional stacking of three resistance genes against Phytophthora infestans in potato. Transgenic research 21, 89-99. Zimmermann, S., Nürnberger, T., Frachisse, J.-M., Wirtz, W., Guern, J., Hedrich, R., and Scheel, D. (1997). Receptor-mediated activation of a plant Ca2+-permeable ion channel involved in pathogen defense. Proceedings of the National Academy of Sciences 94, 2751-2755. Zipfel, C., Kunze, G., Chinchilla, D., Caniard, A., Jones, J.D., Boller, T., and Felix, G. (2006). Perception of the bacterial PAMP EF-Tu by the receptor EFR restricts Agrobacterium-mediated transformation. Cell 125, 749-760.

171

BIOGRAPHICAL SKETCH

Zunaira Afzal was born and raised in Faisalabad, Punjab, Pakistan. She did her bachelor’s degree in agriculture with the major plant breeding and genetics from

University of Agriculture Faisalabad, Punjab, Pakistan in 2009. She got her master’s degree in biotechnology from National Institute of Biotechnology and Genetic

Engineering (NIBGE) that is affiliated with Pakistan Institute of Engineering and Applied

Sciences (PIEAS), Islamabad, Pakistan in 2012. Her master’s project was focused on diversity of Begomoviruses from different weed plants and development novel method to generate transgenic elite cotton cultivars. After her master’s, she started working as a

Research Associate in the same laboratory with an aim to develop cotton leaf curl resistant cotton plants. In 2015, she received a Fulbright scholarship to pursue her

Doctoral studies in the United States. She joined the Department of Plant Pathology at the University of Florida in 2016, under the guidance of Dr. Gul Shad Ali. She conducted her research on investigating the Phytophthora-plant interactions at the Mid Florida

Research and Education Center, Apopka, FL. She received her Ph.D. from the

University of Florida in the Fall of 2019.

172