Functional Characterization of Polyketide-Derived

Home , Alternaria carthami, Ascochyta fabae, Ascochyta pisi, Cochliobolus victoriae, Didymella pinodes, Didymella rabiei

FUNCTIONAL CHARACTERIZATION OF POLYKETIDE-DERIVED

SECONDARY METABOLITES SOLANAPYRONES PRODUCED BY

THE CHICKPEA BLIGHT PATHOGEN, ASCOCHYTA RABIEI:

GENETICS AND CHEMICAL ECOLOGY

WONYONG KIM

A dissertation submitted in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

WASHINGTON STATE UNIVERSITY Department of Plant Pathology

AUGUST 2015

To the Faculty of Washington State University:

The members of the Committee appointed to examine the dissertation of

WONYONG KIM find it satisfactory and recommend that it be accepted

______Weidong Chen, Ph.D., Chair

______Tobin L. Peever, Ph.D.

______George J. Vandemark, Ph.D.

______Lee A. Hadwiger, Ph.D.

______Ming Xian, Ph.D.

ACKNOWLEDGEMENTS

I take this opportunity to thank my major advisor, Dr. Weidong Chen. I have learned a tremendous amount from him in framing hypothesis and critical thinking in science. He gave me every possible opportunity to attend conferences to present my research and interact with scientific communities. I would also like to thank my committee members Drs. Tobin L. Peever,

George J. Va ndemark, Lee A. Hadwiger and Ming Xian for their open-door policy when questions arose and for giving me ideas and suggestions that helped develop this dissertation research. I am very fortunate to have such a nice group of committee members who are experts each in their own fields such as Systematics, Genetics, Molecular Biology and Chemistry.

Without their expertise and helps the research presented in this dissertation could not have been carried out.

I thank to Drs. Jeong-Jin Park and Chung-Min Park for long term collaboration during my doctoral study and being as good friends. Drs. Frank M. Dugan and Akamatsu O. Hajime were also invaluable sources of advice and research materials. I thank to Dr. Lori M. Carris as her

Mycology class introduced me to the wondrous world of fungi. I am also appreciative of the the support that I received from the Department of Plant Pathology in Washington State University, the USDA Cool Season Food Legume Research Program, and the USA Dry Pea and Lentil

Council. Special thanks to my labmates, Liangsheng Xu, Renuka Attanayake, and Tony Chen, for their technical as well as emotional supports.

iii

Lastly, but most importantly, I thank my parents, Hyung-wook Kim and Seong-joo Yeo, for their loving support and continuous sacrifices throughout my life. They from an early age gave me inspiration to study agricultural science and be a person who contributes to the society in which we live. This encouraged me to continue during rough times. Special acknowledgement goes to my wife, Eunju Nam, for her understanding, encouragement and love, and lots of love to our son, Jung-wee.

FUNCTIONAL CHARACTERIZATION OF POLYKETIDE-DERIVED

SECONDARY METABOLITES SOLANAPYRONES PRODUCED BY

THE CHICKPEA BLIGHT PATHOGEN, ASCOCHYTA RABIEI:

GENETICS AND CHEMICAL ECOLOGY

Abstract by Wonyong Kim, Ph.D. Washington State University August 2015

Chair: Weidong Chen

Ascochyta rabiei causes Ascochyta blight of chickpea. The pathogen is known to produce polyketide-derived secondary metabolites, solanapyrones A, B and C, of which solanapyrone A has long been considered a key virulence factor in the chickpea-A. rabiei interaction, due to its phytotoxicity to chickpea. In order to determine the role of solanapyrones during infection process, solanapyrone-minus mutants were generated from A. rabiei strains of different pathotypes by targeting sol5 and sol4 genes, which encode the last step enzyme and a pathway-specific regulator for solanapyrone biosynthesis, respectively. In addition, phytotoxicity of solanapyrones was examined with various legumes including chickpea plants with varying degree of resistance to the disease. As a result, purified solanapyrone A showed a broad spectrum of phytotoxicity, causing necrotic lesions on all tested plants. The resulting solanapyrone-minus mutants were equally virulent as their corresponding wild-type progenitors, indicating that

solanapyrone A is neither a host-selective toxin nor a virulence factor of A. rabiei. Despite the dispensability in parasitic growth of the pathogen, solanapyrones are produced by all strains investigated in this and several previous studies. The universal production of solanapyrones prompted us to examine possible ecological roles other than parasitism.

Synteny analyses of A. rabiei genome with genomes of related species revealed that the solanapyrone biosynthesis gene cluster was placed in a ‘genomic island’, where genes implicated in niche adaptation are often found. The expressions levels of the solanapyrone cluster genes were high during saprobic growth, but negligible during infection processes. Also, the gene expressions culminated at the formation of pycnidia, indicating that solanapyrones are produced in a growth and tissue-specific manner. To investigate ecological roles of solanapyrones, wild-type strains or solanapyrone-minus mutants were co-cultured with saprobic fungi that have been isolated from chickpea debris left in a field. Wild-type strains effectively suppressed the growth of the saprobic fungi, whereas solanapyrone-minus mutants did not. The half maximal inhibitory concentration of solanapyrone A against the saprobic fungi ranged from 50 to 80 μM.

These results suggest that solanapyrone A plays an important role in competition and presumably in survival of the pathogen in nature.

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ...... iii

ABSTRACT ...... v

LIST OF TABLES ...... x

LIST OF FIGURES ...... xi

PREFACE ...... xvii

CHAPTER 1: Introduction

General Introduction ...... 1

References ...... 26

CHAPTER 2: Functional Analyses of the Diels-Alderase Gene sol5 of Ascochyta rabiei

and Alternaria solani Indicate that the Solanapyrone Phytotoxins Are Not

Required for pathogenicity

Abstract ...... 38

Introduction ...... 39

Materials and Methods ...... 43

Results ...... 55

Discussion ...... 62

Acknowledgements ...... 69

References ...... 85

vii

CHAPTER 3: Genomic Analysis of Ascochyta rabiei Identifies Dynamic Genome

Environments of Solanapyrone Biosynthesis Gene Cluster and a Novel

Type of Pathway-specific Regulator

Abstract ...... 92

Introduction ...... 93

Materials and Methods ...... 97

Results ...... 103

Discussion ...... 110

Acknowledgements ...... 118

References ...... 136

CHAPTER 4: Growth and Tissue-specific Production of a Secondary Metabolite,

Solanapyrone A, and Its Ecological Role during Saprobic Growth

of the Fungal Pathogen, Ascochyta rabiei

Abstract ...... 145

Introduction ...... 146

Materials and Methods ...... 149

Results ...... 156

Discussion ...... 162

Acknowledgements ...... 167

References ...... 184

viii

CHAPTER 5: Use of Non-targeted Metabolic Profiling as a Tool for the Chemotaxonomy

of Legume-associated Ascochyta and Allied Anamorphic Genera

Abstract ...... 191

Introduction ...... 192

Materials and Methods ...... 197

Results ...... 205

Discussion ...... 212

Acknowledgements ...... 218

References ...... 228

CHAPTER 6: Summary ...... 235

APPENDIX A: Supplementary Table ...... 238

APPENDIX B: Supplementary Figures ...... 244

LIST OF TABLES

Page

Table 2-1. Isolates of Ascochyta rabiei and Alternaria solani used in this study ...... 71

Table 2-2. Mean lesion length and lesion number produced in chickpea cv. CDC Frontier

stems inoculated with mixed inoculum of Ascochyta rabiei isolate AR628 and

its Δsol5 mutant ...... 72

Table 2-3. Lesion size caused by Alternaria solani isolates and Δsol5 mutants on leaves

of potato and tomato cultivars ...... 73

Table 2-4. List of isolates examined for the presence of sol5 gene ...... 74

Table 3-1. The solanapyrone biosynthesis gene cluster and flanking open reading frames

and pseudogenes in Ascochyta rabiei ...... 119

Table 3-2. The locations of putative binding motifs for a Zn(II)2Cys6 zinc cluster

transcription factor in the promoter regions of biosynthetic genes for

solanapyrone production ...... 120

Table 4-1. Peak assignments of compounds detected in extracts of chickpea stems and

autoclaved chickpea straws inoculated with A. rabiei strain AR628 ...... 168

Table 4-2. Correlation of the expression levels between GFP and solanapyrone cluster

genes in the sol5prom::GFP reporter strain ...... 169

Table 5-1. List of Ascochyta and Phoma strains used in this study ...... 219

SupplementaryTable. Primers used in this study ...... 239

LIST OF FIGURES

Page

Fig. 1-1. The disease cycle of A. rabiei (teleomorph, Didymella rabiei), the causal

agent of Ascochyta blight of chickpea ...... 22

Fig. 1-2. Chemical structure of solanapyrones produced by A. rabiei ...... 23

Fig. 1-3. Chemical structures of secondary metabolites (SMs) produced by

legume-associated Ascochyta and Phoma ...... 24

Fig. 1-4. Chemical structures of biologically active solanapyrone analogues ...... 25

Fig. 2-1. Gene replacement via split-marker strategy and PCR analysis of sol5-deletion .... 75

Fig. 2-2. Detection of metabolites in cultures of A. rabiei and Al. solani and

their Δsol5 mutants ...... 76

Fig. 2-3. Expression pattern of solanapyrone biosynthesis genes ...... 77

Fig. 2-4. Comparison of growth phenotypes of A. rabiei and Al. solani wild-type strains

and their respective Δsol5 mutants ...... 78

Fig. 2-5. Production of pseudothecia and ascospores ...... 79

Fig. 2-6. Purification and verification of solanapyrones and the immediate

precursor compound ...... 80

Fig. 2-7. Comparative analysis of phytotoxic effects of solanapyrones on legumes ...... 81

Fig. 2-8. Effects of solanapyrone A on root growth of Arabidopsis thaliana ...... 82

Fig. 2-9. Pathogenicity tests of A. rabiei and its Δsol5 mutants ...... 83

Fig. 2-10. Detached leaf assay for virulence of Al. solani isolates and the corresponding

Δsol5 mutants on potato cultivars ...... 84

Fig. 3-1. Gene replacement via split-marker strategy and PCR analysis of sol4-deletion .... 121

Fig. 3-2. Solanapyrone biosynthesis gene cluster and its genomic environment ...... 122

Fig. 3-3. Degeneracy of transposable elements and ORF10 located near the

solanapyrone gene cluster ...... 123

Fig. 3-4. Structural organization of typical C6 zinc cluster transcription factor (Gal4)

and Sol4 protein ...... 125

Fig. 3-5. Lack of solanapyrone production in Δsol4 mutant culture ...... 126

Fig. 3-6. Indeterminate in vitro growth of Δsol4 mutants ...... 127

Fig. 3-7. Conidial morphology of wild-type strains and their corresponding

Δsol4 mutants ...... 129

Fig. 3-8. Pathway-specific control of sol4 gene for solanapyrone cluster genes ...... 130

Fig. 3-9. Synteny blocks shared between the A. rabiei contig harboring solanapyrone

biosynthesis gene cluster and scaffolds of related Dothideomycetes ...... 132

Fig. 3-10. Gene content of the flanking region of sol1 gene in Alternaria solani ...... 133

Fig. 3-11. Comparison of expression levels of solanapyrone cluster genes between

AR628 wild-type and its Δsol4 mutant via RT-PCR analysis ...... 134

Fig. 3-12. Comparison of virulence between Δsol4 mutants and their wild-type

Progenitors ...... 135

xii

Fig. 4-1. Visualization of A. rabiei during parasitic and saprobic growth ...... 170

Fig. 4-2. Growth phase-dependent solanapyrone production ...... 171

Fig. 4-3. Time-course expression of solanapyrone biosysnthesis cluster genes during

the parasitic and saprobic growth of A. rabiei ...... 173

Fig. 4-4. pH shifts and solanapyrone A production of A. rabiei growing on different

carbon sources ...... 174

Fig. 4-5. Effects of carbon sources on production of solanapyrone A ...... 175

Fig. 4-6. Tissue-specific production of solanapyrone A ...... 176

Fig. 4-7. No GFP induction during plant infection ...... 177

Fig. 4-8. Secretion of solanapyrones outside of fungal cells ...... 178

Fig. 4-9. Antagonistic potentials of A. rabiei (AR628 strain) against fungal competitors .... 179

Fig. 4-10. Antagonistic potentials of A. rabiei (AR21 strain) against fungal competitors .... 181

Fig. 4-11. Antifungal activities of solanapyrone A ...... 182

Fig. 4-12. Front sides of confrontation plates with A. rabiei AR628 strain or

solanapyrone-minus mutants on the left side of the plates ...... 183

Fig. 5-1. Colony morphology of Ascochyta and Phoma species ...... 221

Fig. 5-2. Consistent metabolic profiles between strains within the same species ...... 222

Fig. 5-3. Chromatograms relating to the detection of ascochitine and pinolidoxin ...... 224

Fig. 5-4. Chemical diversity and grouping observed in Ascochyta/Phoma species ...... 225

Fig. 5-5. Bayesian phylogeny of Ascochyta/Phoma species ...... 226

xiii

LIST OF SUPPLEMENTARY FIGURES

Page

Fig. S2-1. Pairwise alignment of DNA sequences, and deduced amino acid sequences

of sol5 genes between Al. solani and A. rabiei ...... 245

Fig. S2-2. The correlation details among protons and between protons and carbons of pro-

solanapyrone II-diol by 300 MHz 1H NMR and 75 MHz 13C NMR analyses .... 246

Fig. S2-3. 300 MHz COSY NMR of prosolanapyrone II-diol ...... 247

Fig. S2-4. 300 MHz HMBC NMR of prosolanapyrone II-diol ...... 248

Fig. S2-5. Verification of Arabidopsis polymerase lamda mutant genotypes ...... 249

Fig. S3-1. Verification of integration of pHNU3PelA plasmid carrying sol4 open reading

frame in sol4-overexpression strains ...... 250

Fig. S3-2. Multiple DNA sequence alignment of Molly transposon from Stagonospora

nodorum and transposon-like sequences located proximal to the solanapyrone

gene cluster in A. rabiei ...... 251

Fig. S3-3. Pairwise sequence alignment of ORF10 and a P450 homologous gene found

in Coniosporium apollinis...... 253

Fig. S3-4. Multiple amino acid sequence alignment of A. rabiei Sol4 and two hypothetical

proteins from Aspergillus niger and Podospora anserina ...... 255

Fig. S4-1. Strategy to construct the pSol5::GFP plasmid ...... 257

Fig. S4-2. Verification of integration of pSol5::GFP and pII99dsRed plasmids into

xiv

the sol5 genomic locus ...... 259

Fig. S4-3. Dose response curves of solanapyrone A ...... 261

Fig. S4-4. Growth phase-dependent solanapyrone production ...... 262

Fig. S4-5. pH shifts and colony morphology of A. rabiei growing on different carbon

sources ...... 263

Fig. S5-1. Mass spectra of peaks corresponding to ascochitine and pinolidoxin ...... 264

Fig. S5-2. Structural identification of ascochitine and pinolidoxin ...... 265

Dedication

This work is dedicated to my grandmother, 김점순

xvi

PREFACE

This dissertation research has been published or will be submitted in peer-reviewed scientific journals. A modified version of Chapter 2 was published in Molecular Plant-Microbe

Interactions 2015, 28:486-496. Wonyong Kim was responsible for the study design, experiment execution, data analyses and manuscript writing. Dr. Chung-Min Park performed NMR analyses and contributed to the wiriting. Dr. Jeong-Jin Park assisted in LC-MS analyses. Dr. Hajime O.

Akamatsu generated a GFP-expressing Ascochyta rabiei strain and contributed to the writing.

Drs. Tobin L. Peever and George Vandemark edited the manuscript. Drs. Ming Xian and David

R. Gang provided instruments for the study. Dr. Weidong Chen provided advice and assistance in designing the experiments and analyses, and contributed to the writing.

The Chapter 3 was written in manuscript format and will be submitted to BMC Genomics or similar journal. The co-authors will be Drs. Jeong-Jin Park, David R. Gang, Tobin L. Peever and Weidong Chen. Wonyong Kim was responsible for the study design, experiment execution, data analyses and manuscript writing. Dr. Park assisted in LC-MS and data analyses. Dr. Gang contributed to the writing. Dr. Peever provided a draft genome of Alternaria solani. Dr. Chen contributed to the data analyses and writing.

The Chapter 4 was written in manuscript format and will be submitted to Environmental

Microbiology or similar journal. The co-authors will be Drs. Jeong-Jin Park, Frank M. Dugan,

Hajime O. Akamatsu, Tobin L. Peever, David R. Gang and Weidong Chen. Wonyong Kim was

xvii

responsible for the study design, experiment execution, data analyses and manuscript writing. Dr.

Park assisted in LC-MS analyses and data analyses. Dr. Dugan provided chickpea debris-inhabiting fungal strains and edited the manuscript. Dr. Akamatsu provided a

GFP-expressing Ascochyta rabiei strain. Drs. Peever and Gang edited the manuscript. Dr. Chen directed the study, helped design experiment and contributed to the data analyses and the writing.

The Chapter 5 was written in manuscript format and will be submitted to Fungal

Diversity or similar journal. The co-authors will be Drs. Jeong-Jin Park, Chung-Min Park, Tobin

L. Peever, David R. Gang, Ming Xian, Walter J. Kaiser and Weidong Chen. Wonyong Kim was responsible for the study design, experiment execution, data analyses and manuscript writing. Dr.

Park, J-.J assisted in LC-MS analyses and multivariate data analyses. Dr. Park, C-.M performed

NMR analyses. Drs. Peever and Kaiser contributed to the collection of fungal strains. Drs. Peever,

Gang, Xian, and Chen edited the manuscript.

xviii

CHAPTER ONE

GENERAL INTRODUCTION

Ascochyta Blight of Chickpea

Ascochyta rabiei (Pass.) Labrousse [teleomorph: Didymella rabiei (Kovachevski) v.

Arx] causes Ascochyta blight, a foliar disease of chickpea (Cicer arietinum L.). The disease affects all aboveground tissues of the host plant (Nene and Reddy, 1987). Infection with A. rabiei leads to reduction of yield and quality of the harvested grain, and often to total crop loss when environmental conditions are conducive to disease development, i.e., when periods of cool, wet weather persist during the growing season (Nene, 1982; Singh et al., 1984).

Domesticated chickpea is a cool-season grain legume crop and the second most widely grown legume in semi-arid regions of the world, next to the common pea (Pisum sativum L.). It is also important as a rotation crop in the cereal-based production systems of the U.S. Pacific

Northwest or a specialty crop in other areas. There are two main types of chickpea, desi and kabuli. The desi-type is characterized as having angular and smaller seeds with darker pigmentation, while the kabuli-type has characteristics with beaked and larger seeds with white to creamy coloration. Production and consumption of desi-type chickpeas are largely restricted to the Middle East and South Asia, whereas kabuli-type chickpeas is a popular and valuable global commodity being produced in most of other regions and especially in the western hemisphere

(Chen, 2008).

Ascochyta blight of chickpea is one of the most destructive diseases of domesticated chickpea (Chen et al., 2011). The disease is manifested through necrotic lesions on green tissues including leaves, stems, and pods. The causal agent, A. rabiei, is a haploid heterothallic (i.e. outcrossing) ascomycete. Ascospores produced by sexual reproduction between strains of opposite mating types have been considered to be important for long-distance dissemination and initiation of disease outbreaks (Trapero-Casas et al., 1996; Kaiser, 1997; Milgroom and Peever,

2003). Pseudothecia, the sexual fruiting bodies, develop on infested chickpea plant debris and give rise to ascospores, which are forcibly discharged from the fruiting bodies early in the growing season. Although either ascospores or conidia can serve as initial inoculum for the disease, ascospores may be better able to germinate and penetrate host tissues under harsh environmental conditions compared to conidia (Trapero-Casas and Kaiser, 2007). From necrotic lesions on the host tissues, A. rabiei produces the asexual fruiting bodies, pycnidia, containing numerous conidia that can serve as secondary inoculum. Conidia are disseminated to adjacent tissues and neighboring plants by rain splash. Each infection cycle only takes 10–14 days. Thus, several secondary infections can occur during the growing season and lead to severe disease epidemics (Trapero-Casas and Kaiser, 1992). Seed infection or contamination may have a role in spread of the disease to new production areas that were previously free of the disease. The A. rabiei disease cycle is summarized in Figure 1-1.

Solanapyrone Phytotoxins

The mode of infection of A. rabiei is thought to be necrotrophic, i.e., the pathogen kills plant cells in advance of or during mycelia invasion. Phytotoxins and cell-wall degrading enzymes being secreted are often presumed to be determinants of pathogenicity in necrotrophic plant pathogens. A .rabiei is known to produce polyketide-derived secondary metabolites (SMs), solanapyrones A, B, and C (Fig. 1-2) (Alam et al., 1989; Chen and Strange, 1991). Phytotoxicity of solanapyrones has been evaluated in vitro with intact chickpea plants and chickpea cell culture suspensions (Höhl et al., 1991; Latif et al., 1993; Kaur, 1995; Hamid and Strange, 2000).

Exogenous application of A. rabiei culture filtrates containing a mixture of solanapyrones caused disintegration of the intact leaf tissue of chickpea (Höhl et al., 1991). In addition, when detached chickpea shoots were incubated with purified solanapyrone A, the stem internode below the uppermost leaf lost its turgor within 3 days (Hamid and Strange, 2000). Kaur (1995) determined relative phytotoxicity of solanapyrones by measuring inhibition rates of chickpea root growth; solanapyrone A was the most toxic followed by solanapyrones C and B (Kaur, 1995). Positive correlations were observed between virulence and solanapyrone production of A. rabiei isolates

(Hamid and Strange, 2000), and between the Ascochyta blight resistance of chickpea cultivars and their levels of tolerance to solanapyrone phytotoxins (Kaur, 1995).

Solanapyrone production appears to be associated with the production of conidia in A. rabiei. Strains that sporulate poorly in growth media tend to produce less solanapyrones (Höhl et al., 1991). However, when the strains were grown in growth media which supported better

conidiation, they produced more solanapyrones (Höhl et al., 1991). The authors also argued that solanapyrones were produced from germinating spores in liquid culture. In the study, however, the solanapyrone production was measured once at 24 h after incubation. The result cannot rule out the possibility that the spore suspension prepared for initiation of the liquid culture already contained solanapyrones as “contaminants”, since solanapyrones are produced in large quantity during conidiation. To date it is not known if solanapyrones are produced in the host tissues.

Attempts to detect solanapyrones from infected chickpea tissues were unsuccessful (Höhl et al.,

1991). To explain this phenomenon, it has been proposed that the absence of solanapyrones in the infected chickpea is ascribed to rapid metabolism of solanapyrones by the host plant, and that the ability to metabolize solanapyrones can be used as an indicator of host resistance (Hamid and

Strange, 2000; Bahti and Strange, 2004).

Many studies have demonstrated the phytotoxicity of solanapyrones to chickpea, but there is no direct evidence supporting the involvement of solanapyrones in virulence or pathogenicity. It would be useful to have solanapyrone-minus mutants for unequivocal demonstration that the toxins play a role in pathogenicity or virulence by comparing virulence of the mutants to that of the wild-type progenitors. To obtain such mutants, T-DNA insertion mutants that produce a reduced amount of solanapyrones were generated via

Agrobacterium-mediated random mutagenesis (Mogensen et al., 2006). However, unfortunately, these mutants were not characterized with respect to the T-DNA insertion sites or virulence.

Infection Strategy of A. rabiei

The infection process of A. rabiei in host tissues has been described (Höhl et al., 1990;

Köhler et al., 1995; Ilarslan and Dolar, 2002; Rea et al., 2002; Nizam et al., 2010). Although slight variation in the infection process has been observed likely due to the use of different A. rabiei strains and chickpea cultivars, it can be generalized as follows.

The fungus penetrates directly through the cuticle of leaf and stem tissues 1–2 days post inoculation (dpi), forming an appressorium-like structure (Nizam et al., 2010). Entry through stomata, hydathodes, and wounds is also common (Höhl et al., 1990). Around 3 dpi, mycelia have spread through the intercellular spaces of the subepidermal layer of leaves and stems of a susceptible cultivar (Höhl et al., 1990). The distortion of 2–3 layers of cortical parenchyma cells is evident in the susceptible cultivar at this early stage of infection, whereas in a resistant cultivar cell disruption is delayed. In the resistant cultivar, visible symptoms such as chlorotic lesions can be observed only after 5 dpi (Höhl et al., 1990). Small necrotic spots are frequently found on leaves and stems of resistant cultivars, which are considered as the results of resistant response.

Within 5–7 dpi, the cells near intercellular hyphae in the susceptible cultivar undergo plasmolysis, in which most of cellular organelles have disintegrated without direct hyphal intrusion (Höhl et al., 1990; Rea et al., 2002). The lesions formed on host tissues expand further, and the growth of mycelia along leaf petioles towards the stem is observed (Höhl et al., 1990). The phloem tissues of petioles are extensively colonized by the pathogen, but colonization of xylem tissues with lignified sclerenchyma cells is rare (Rea et al., 2002). After 5 dpi, a large number of pycnidia

start to form on susceptible cultivars, whereas only a few pycnidia are produced in resistant cultivars even after prolonged incubation. Pycnidia are often developing on the vascular tissues which likely provide a physical matrix for their formation, especially when the host tissues are macerated by severe infection (Höhl et al., 1990).

Quantitative Nature of Virulence of A. rabiei and the Host Resistance

To clarify terminology used throughout this study, the term ‘pathogenicity’ is defined as a qualitative term (i.e., the pathogen’s ability to cause disease), whereas the term ‘virulence’ is defined as a quantitative term (i.e., the level of disease that a pathogen causes in compatible interaction with its host, the degree of pathogenicity). A high degree of genetic and pathogenic variation exists in field populations of A. rabiei (Chongo et al., 2004; Peever et al., 2004).

Classification of this variation as pathotypes (or races) is crucial for the identification of genetic sources of host resistance and for developing resistant cultivars. However, there is little agreement regarding the classification of pathogenic variation in A. rabiei due to lack of information on virulence factors which may facilitate the classification of strains into pathotypes.

A recent trend is the broad classification of strains into two pathotypes: pathotype I (less virulence) and pathotype II (more virulence) (Chen et al., 2004b; Jayakumar et al., 2005; Chen et al., 2011). However, a genetic study showed that progeny from a cross between a pathotype I and a pathotype II strain displayed intermediate virulence, suggesting that the pathotype designation represented the extremes of a continuous distribution of virulence, rather than representing two discrete virulence types (Peever et al., 2012).

Despite the economic and agricultural importance of the A. rabiei–chickpea interaction, little attention has been paid to this interaction at a molecular level, and only a few studies have described the molecular aspects of this interaction. In earlier studies, Cho and Muehlbauer (2004) examined the effect of the plant signaling molecules, salicylic acid (SA) and jasmonic acid (JA), known to mediate plant defense responses, using recombinant inbred lines (RILs) generated from a cross of Ascochyta blight-resistant and Ascochyta blight-susceptible germplasm. However, the study showed that blight resistance in the RILs did not cosegregate with the expression of several defense-related genes that were induced by exogenously applied SA or JA, suggesting that the systemic regulation of the gene expression by the plant hormones are not required for the resistance to A. rabiei, unlike other plant model systems such as Arabidopsis (Cho and

Muehlbauer, 2004). Another study showed that establishment of thickened cells in response to A. rabiei infection can inhibit the spread of the pathogen through chickpea stem tissues (Rea et al.,

2002). When plants are damaged by infection or wounding, JA acts as a potent inducer of copper amine oxidase (CuAO) that plays a role in H2O2 production in the intercellular space (Rea et al.,

2002). In a resistant cultivar, the accumulated H2O2 was utilized by plant peroxidases to establish the lignin and suberin barriers against the invasion of A. rabiei. Inhibition of CuAO activity and the associated H2O2 production caused the resistant cultivar to be more susceptible to the disease, indicating that in the resistant cultivar the induced physical barriers are critical for limiting the growth of the necrotrophic pathogen (Rea et al., 2002).

Complete resistance to the blight disease has not been reported yet in any commercial chickpea cultivar or germplasm line. Notably, the recently published genome sequence of the chickpea cultivar, CDC Frontier (kabuli-type) revealed that, compared to other related legume species, the genome contains a markedly reduced number of the nucleotide-binding site leucine-rich repeat (NBS-LRR) disease resistance genes that often confer a full-resistance to plant pathogens; the number of NBS-LRR disease resistance genes is nearly one-fifth that found in Medicago truncatula and one-third that in Glycine max (Varshney et al., 2013). Given the quantitative nature of the virulence of A. rabiei strains, preformed and induced chickpea defense systems (reviewed in Jayakumar, 2005) are likely more important in the A. rabiei–chickpea interaction than the deployment of NBS-LRR disease resistance genes that have been shown to play a key role in incompatible interactions in other pathosystems. Legumes are capable of producing antimicrobial metabolites such as alkaloids, coumarins, stilbenes and isoflavonoids

(Kamphuis et al., 2012). Chickpea is also known to produce a variety of isoflavonoids including biochanin A and formononetin (constitutive isoflavonoids) as well as chickpea phytoalexins, medicarpin and maackiain (induced isoflavonoids) (Lv et al., 2009; Wu et al., 2012). To counteract to these chemical defense systems of chickpea, most A. rabiei strains are able to convert the constitutively expressed isoflavonoids and the induced phytoalexins into less- or non-toxic metabolites (Kraft and Barz, 1985; Kraft et al., 1987). Also, it was reported that a pronounced accumulation of phytoalexins was observed in resistant cultivars infected with A. rabiei (Daniel et al., 1990). It is currently unclear whether the extent of isoflavonoids production

and their detoxification are key determinants in the A. rabiei–chickpea interaction. Nevertheless, this arms race-type interaction may imply that A. rabiei have evolved to overcome chemical and physical barriers deployed by hosts, and at the same time chickpea may have forged more rigid, repulsive barriers, establishing the quantitative nature of the host resistance.

Host Specificity and Taxonomy of Ascochyta spp.

It has been noted that the anamorphic genera Ascochyta and Phoma both are polyphyletic, being found in other fungal taxa in the order Pleosporales, although a majority of the members (≈

70%) are found in the family Didymellaceae (Boerema and Bollen, 1975; Aveskamp et al., 2010).

Recent phylogenetic studies demonstrated that Phoma should be restricted to the Didymellaceae

(Aveskamp et al., 2010). Species that belong to Ascochyta and Phoma are highly similar in morphology, physiology, and pathogenicity (Aveskamp et al., 2010). In the Saccardoan system,

Ascochyta and Phoma were only distinguished by conidial morphology; two-celled conidia in the former and one celled conidia in the latter (Boerema and Bollen, 1975; Boerema et al., 2004).

However, two-celled conidia appear to have evolved independently multiple times during evolution of several lineages in the Pleosporales.

In addition to A. rabiei, several Ascochyta spp. infect economically important cool season food legumes. These include A. pisi Lib., A. fabae Speg., A. lentis Vassiljevsky, and A. viciae-villosae Ondrej, which are pathogens of pea (Pisum sativum L.), faba bean (Vicia faba L.), lentil (Lens culinaris Medik.), and hairy vetch (Vicia villosa Roth), respectively (Nene et al.,

1988; Kaiser, 1997; Chen et al., 2011). These fungi are morphologically similar and, the biological species concept has been successfully applied to circumscribe the legume-associated

Ascochyta species (Kaiser et al., 1997; Hernandez-Bello et al., 2006; Peever et al., 2007). In vitro genetic crosses were made among strains of A. rabiei, A. fabae and A. lentis (Kaiser et al., 1997).

A. rabiei was able to mate neither with A. fabae nor A. lentis. In contrast, the genetic crosses between A. fabae and A. lentis strains produced sexual fruiting bodies with viable ascospores, although intrinsic postzygotic mating barriers were observed, such as nonstandard numbers of ascospores in each ascus and variable size of ascospores (Kaiser et al., 1997). These results suggested that A. rabiei is more distantly related to the other two species, and that the Ascochyta spp. infecting chickpea, faba bean, and lentil are each distinct biological species. In a separate study, genetic crosses were made between A. pisi and A. fabae and between A. lentis and A. viciae-villosae (Hernandez-Bello et al., 2006). In both combinations, normal ascus and viable ascospores were produced, i.e., no obvious intrinsic mating barriers between A. pisi and A. fabae and between A. lentis and A. viciae-villosae. However, the pathogenic ability of the progeny derived from the cross between A. pisi and A. fabae were severely impaired, indicating that extrinsic postzygotic mating barriers likely operate to prevent hybrid formation and thereby maintain species integrity. These results revealed that the species used in each cross are closely related, but still biologically distinct (Hernandez-Bello et al., 2006). The biological species of the legume-infecting Ascochyta were well supported by phylogenetic analyses with DNA sequences

of multiple protein-coding genes, each being grouped into separate well-supported clades

(Peever et al., 2007; Chilvers et al., 2009).

Chemical Ecology of Legume-associated Ascochyta and Allied Anamorphic Genera

Filamentous fungi are rich sources of bioactive natural compounds produced as the result of diverse secondary metabolic pathways. Ascochyta species parasitic to different legumes are also known to produce a variety of SMs (Fig. 1-3). Ascochitine (syn. ascochytine) is a quinone methide with yellow fluorescence, produced via a polyketide biosynthetic pathway. It was originally found in culture of A. pisi (Bertini, 1956), and later in A. fabae (Oku and Nakanishi,

1963). Ascochitine is structurally similar to the well-known mycotoxin citrinin produced by

Monascus, Aspergillus, and Penicillium species in the class Eurotiomycetes (Shimizu et al.,

2005). Although no information on genes involved in ascochitine biosynthesis is available, the identification of a polyketide synthase gene (pksCT) for citrinin biosynthesis in M. purpureus

(Shimizu et al., 2005) may provide an opportunity to search for a homologous gene responsible for ascochitine biosynthesis in the Ascochyta species. In this dissertation research, ascochitine was also found in cultures of other legume-infecting Ascochyta and Phoma, such as A. viciae-villosae and Phoma koolunga (see Chapter 5). Ascochitine production is not restricted to the legume-infecting species but it is also found in the marine endophyte Stagonosporopsis salicorniae [cf. formerly A. salicorniae, the genus Stagonosporopsis is related to Phoma section

Heterospora (Aveskamp et al. 2010)] and non-legume plant pathogens A. hyalospora and P. clematidina (Venkatasubbaiah and Chilton, 1992; Smith et al., 1994; Seibert et al., 2006).

Ascochitine showed tissue-specific phytotoxicity, inhibiting root growth but not shoot growth of blue pea seedlings (Clitoria ternatea L.) (Lakshmanan and Padmanabhan, 1968). Also, ascochitine caused electrolyte leakage of susceptible but not resistant Clematis cultivars, and the toxin was retrieved from the plant tissues infected by P. clematidina, suggesting its involvement in pathogenicity or virulence (Smith et al., 1994). The authors also argued that aggressive isolates tend to produce higher amounts of ascochitine. However, no such association was observed in A. fabae isolates (Beed et al., 1994). Ascochitine was originally reported to be a selective antifungal agent, affecting only a certain member of fungi (Oku and Nakanishi, 1964;

Nakanishi and Oku, 1969). The compound was metabolized by some insensitive fungal species, and the amount of ascochitine absorption by the fungi was proportional to the degree of tolerance

(Nakanishi and Oku, 1969). Also, given the fact that ascochitine inhibits the enzymatic activity of MPtpB (Mycobacterium tuberculosis protein tyrosine phosphatase 1B) (Seibert et al., 2006), ascochitine is likely toxic to bacteria as well. However, its antibacterial activity remains to be investigated. Ascosalitoxin and ascochital are a naturally-occurring precursor and another end product derived from the same biosynthetic pathway, respectively (Seibert et al. 2006).

Ascosalitoxin exhibited phytotoxicity on pea and bean (Evidente et al., 1993a). Ascochital found in some marine-derived ascomycetous fungi showed an antibacterial activity (Kusnick et al.,

2002; Seibert et al., 2006).

A. lentis causing disease of lentil is closely related to A. viciae-villosae, A. pisi and A. fabae (Peever et al., 2007). However, ascochitine has not been isolated from A. lentis so far,

instead the fungus is known to produce anthraquinones and some previously known SMs that were absent in other legume-associated Ascochyta species (Andolfi et al., 2013).

Anthraquinone-class SMs are widely found in different fungal species and thought to play diverse roles in fungal biology, such as pigmentation, antibiosis, phytotoxin and cellular signaling (Bick and Rhee, 1966; Davies and Hodge, 1974; Jalal et al., 1992; Kachlicki and

Wakuliński, 2002; Bouras and Strelkov, 2008; Miethbauer et al., 2009; Lin et al., 2012). A. lentis produces two anthraquinones, pachybasin and lentisone (Andolfi et al., 2013). Pachybasin was originally identified from a culture of Boeremia foveata [cf. formerly, P. foveata, the genus

Boeremia is a newly described genus from Phoma sections (Aveskamp et al. 2010)] (Bick and

Rhee, 1966). It was recently demonstrated that pachybasin produced by Trichoderma harzianum is implicated in the mycoparasitic activity of the fungus (Lin et al., 2012). Lentisone, the newly identified anthraquinone from A. lentis, exhibits a strong phytotoxicity, causing necrotic lesions when externally applied to leaf disks of different legumes (Andolfi et al., 2013). The mode of action, cellular targets and biosynthetic genes for the anthraquinone compounds remain to be identified. A. lentis also produces tyrosol and pseurotin A (Andolfi et al., 2013). Tyrosol was identified as a quorum-sensing compound of Candida albicans, inducing filamentous growth and biofilm formation of this dimorphic fungus (Chen et al., 2004a). Pseurotin A is a potent inhibitor for chitin synthase and thus likely plays a role in antagonistic interactions with other fungi

(Wenke et al., 1993). Recently, a hybrid type polyketide synthase responsible for pseurotin A biosynthesis was identified from Aspergillus fumigatus (Maiya et al., 2007).

Pinolidoxin is a 10-membered macrolide with a medium-sized lactone ring, produced by

Peyronellaea pinodes [cf. formerly A. pinodes, the genus Peyronellaea was recently elevated from Phoma sections (Aveskamp et al. 2010)] (Evidente et al., 1993b). Pey. pinodes also produces several structurally related 10-membered macrolides such as herbarumin II, pinolide, and pinolidoxin derivatives, among which only pinolidoxin showed significant phytotoxicity to legumes (Cimmino et al., 2012). Putaminoxin is another macrolide produced by P. putaminum

(Evidente et al., 1995). P. putaminum is primarily a soil saprobe, but also regarded as an opportunistic pathogen infecting roots of various plants (Farr and Rossman, 2015). Interestingly, both pinolidoxin and putaminoxin are potent inhibitors of induced phenylalanine ammonia lyase activity. This enzyme mediates the phenylpropanoid pathway of higher plants, without any effect on the growth and viability of the plant cells, suggesting their role in suppressing plant defense responses (Vurro and Ellis, 1997). The structure-activity relationship study with the two macrolides and their synthetic derivatives showed that the two hydroxyl groups and the unmodified propyl side chain at the lactone ring are important for the observed phytotoxicity

(Evidente et al., 1998). The study also reported a mild zootoxicity and zero antifungal activity for pinolidoxin and putaminoxin. However, the lack of antifungal activities was based on tests with only a single mitosporic fungus, Geotrichum candidum. Thus, to assure lack of a possible selectivity, the compounds should be tested with more diverse fungal species.

P. medicaginis, an opportunistic or weak pathogen of Medicago spp., produces a

13-membered macrolide brefeldin A that was originally identified in distantly-related species,

Eupenicillium brefeldianum (Härri et al., 1963) and later in Alternaria carthami, the causal agent of Alternaria leaf spot of safflower (Carthamus tinctorius L.) (Tietjen et al., 1983; Driouich et al.,

1997). Brefeldin A mimicked disease symptoms, causing necrosis and wilting when exogenously applied to the leaves and roots of safflower (Tietjen et al., 1983). However, another study with P. medicaginis showed that brefeldin A was only produced during saprobic growth of the fungus, but not in living plant tissues (Weber et al., 2004). Also, brefeldin A suppressed spore germination and growth of epiphytic fungi, suggesting its protective role from other organisms that might compete for limited source of nutrients after the host decays (Weber et al., 2004).

Brefeldin A has been used for studies of intracellular membrane trafficking system, since it blocks protein transport from the endoplasmic reticulum to the Golgi apparatus (Colanzi et al.,

2013; Silletta et al., 1999). However, its pathogenic and ecological roles in the fungal biology remain investigated. Some other macrolides are produced in culture by Ascochyta and Phoma species that cause diseases of non-legume plants. P. herbarum is known to produce herbarumin I,

II, and III, all of which showed significant phytotoxicity (Fausto Rivero-Cruz et al., 2000;

Boruwa et al., 2006). Recently, herbarumin I was reported to be antibacterial (Jangili et al., 2014).

A. hyalospora is known to produce an antifungal macrolide, pyrenolide A (Venkatasubbaiah and

Chilton, 1992), which was first identified in culture of the barley pathogen Pyrenophora teres

(Nukina et al., 1980a; Nukina et al., 1980b).

A. viciae is pathogenic mainly to Vicia spp., and known to produce two structurally related terpene-derived SMs, ascofuranone and ascochlorin (Tamura et al., 1968; Sasaki et al.,

1973). Several natural analogues with antiviral, antifungal, or antitumoric activities were also found across various fungal taxa, such as Fusarium, Acremonium, Colletotrichum, Verticillium, and Cylindrocladium (Ellestad et al., 1969; Kato et al., 1970; Kosuge et al., 1973;

Cagnoli-Bellavita et al., 1975; Takamatsu et al., 1994). Ascofuranone and ascochlorin produced by A. viciae have received great attention owing to their potential as chemotherapeutic agents

(Takatsuki et al., 1969; Minagawa et al., 1996). The two terpenoids are structural analogues of ubiquinol, an essential component of the respiratory chain for ATP synthesis, and thus inhibit the enzymatic activities of protozoan alternate oxidase by acting at the ubiquinol binding domain

(Nihei et al., 2003; Mogi et al., 2009). In addition, ascochlorin was also reported to inhibit the respiratory chain of the ascomycetous yeast Pichia anomala by targeting mitochondrial cytochrome bc1 complex (a.k.a. coenzyme Q) (Berry et al., 2009). Intuitively, given the mode of action, the compounds likely play an antagonistic role in interactions with different classes of microorganisms like saprobic protozoa and fungi, though their ecological roles in the fungal biology have not been studied yet.

The production of solanapyrones is unique to A. rabiei and has not been reported from related Ascochyta and Phoma species. Previous studies have shown that all tested A. rabiei strains produce at least one of the solanapyrones A, B, and C (Höhl et al., 1991; Kaur, 1995).

Cytochalasin D, an antimicrobial polyketide was also isolated from one strain among nine A. rabiei strains tested (Latif et al., 1993). The cytochalasin-producing strain did not produce solanapyrones, and was later discovered not to be A. rabiei (personal communication with S.S.

Alam). In fact, the cytochalasin family-SMs are produced by some related species like A. lathyri and P. exigua (Capasso et al., 1987; Vurro et al., 1992). The biosynthetic pathway of solanapyrones has been extensively studied as it involves the Diels-Alder reaction, a [4+2] cycloaddition with a high degree of regio- and stereo-selectivity, that catalyzes the formation of the decalin ring of solanapyrones (Oikawa et al., 1995; Oikawa et al., 1998; Kasahara et al.,

2010). The Diels-Alder reaction is the key step for the biosynthesis of industrially important SMs such as lovastatin (a cholesterol-lowering drug) and spinosyn A (a potent insecticide) (Auclair et al., 2000; Kim et al., 2011). Recently, the solanapyrone biosynthesis gene cluster was identified in Alternaria solani (Kasahara et al., 2010). The gene cluster comprises six genes, among which sol5 gene was implicated in the biotransformation of prosolanapyrone II into solanapyrone A with good exo-selectivity that is consistent with a Diels-Alder reaction (Katayama et al., 1998;

Kasahara et al., 2010).

Solanapyrone A is known to specifically bind to X-family DNA polymerases in vitro

(Mizushina et al., 2002). This particular DNA polymerase family exerts its function exclusively on DNA repair process and cell cycle control in the DNA replication processes during mitosis and meiosis (García-Díaz et al., 2000; Yamtich and Sweasy, 2010). Recently, there has been increasing evidence that DNA damage and subsequent repair processes are linked to the induction of plant immune responses (Durrant et al., 2007; Wang et al., 2010; Song et al., 2011;

Yan et al., 2013). Therefore, it is plausible that solanapyrone A inhibits DNA repair processes, induces cell cycle arrest, and possibly causes apoptosis. Alternatively, solanapyrone A could

affect a defense signaling pathway induced by DNA damage and subsequent repair processes in plants.

Solanapyrone Analogues

Solanapyrones A, B, and C were originally isolated from cultures of Al. solani, the causal agent of early blight of potato/tomato (Ichihara et al., 1983). Since this first discovery, many natural solanapyrone analogues have been found in several fungal species occupying different ecological niches (Jenkins et al., 1998; Schmidt et al., 2007; Trisuwan et al., 2009; Wu et al.,

2009; Wang et al., 2014). Three previously unidentified solanapyrones (E–G), in addition to solanapyrone C, were identified from culture of a filamentous marine fungus associated with the green alga Halimeda monile (Jenkins et al., 1998). Unfortunately, the fungus was not taxonomically identified due to its rapid loss of viability in growth media. The three new solanapyrones produced by the marine-derived fungus did not exhibit any significant phytotoxicity to the host alga. Only solanapyrone C showed an algistatic effect to a unicellular algal species (Jenkins et al., 1998).

Four additional solanapyrones (J–M) were isolated from an unidentified fungicolous fungus (Schmidt et al., 2007). Solanapyrone J and K both exhibited strong antifungal and antibacterial activities to Candida albicians and Staphylococcus aureus. The observed toxicities were comparable to those of widely used antimicrobial agents, filipin and gentamycin sulfate.

However, solanapyrone J and K did not show any toxicity to Escherichia coli, a gram negative

bacterium. Solanapyrones L and M that are structurally similar to solanapyrones K and J, respectively, did not show any antimicrobial activities, indicating that a slight difference in functional groups on the pyrone ring has a considerable effect on bioactivity of the compounds.

In two separate studies, several solanapyrones were found in cultures of two fungal species in the genus Nigrospora: one is marine-derived and the other is isolated as an endophyte from a medicinal plant (Trisuwan et al., 2009; Wu et al., 2009). The endophytic Nigrospora sp. produced two previously unidentified solanapyrones N and O, and the known solanapyrone C, of which only solanapyrone N was effective in suppressing the growth of Botrytis cinerea and

Penicillium islandicum among the 7 tested fungal species, the other fungi were tolerant or insensitive to the compound (Wu et al., 2009). The marine-derived Nigrospora sp. produced three solanapyrone analogues, and the known solanapyrone A. The authors coined new names for the newly identified analogues as nigrosporapyrones A, B, and C. However, these compounds are of essentially the same chemical structure with solanapyrone A and C, except for having one hydroxyl group on the decalin ring (Trisuwan et al., 2009). The authors examined antibacterial activities of nigrosporapyrone A, solanapyrone A and some other metabolites produced by the fungus. As a result, only nigrosporapyrone A showed a moderate toxicity to clinical bacteria strains, Staphylococcus aureus and methicillin-resistant S. aureus (Trisuwan et al., 2009).

Most recently, a species phylogenetically related to Al. tenuissima produced three solanapyrone analogues (P–R), as well as the known solanapyrones (A–C) (Wang et al., 2014).

The authors conducted a comprehensive bioassay to test antibacterial activities of the six

solanapyrones against several bacterial strains of ecological and clinical importance. Among the six solanapyrones, solanapyrones A and C exhibited antibacterial activities as strong as ampicillin. Solanapyrone A showed the best activity to Bacillus subtilis and Micrococcus tetragenus, whereas solanapyrone C was only toxic to B. megaterium. However, it is noteworthy that the toxicity of solanapyrone C to the particular bacterial species was approximately 4 times stronger than solanapyrone A (Wang et al., 2014).

In summary, solanapyrones found in nature have been shown to be toxic to a variety of organisms including bacteria (gram positives only), fungi, unicellular algae and plants. Although a dozen solanapyrone analogues have been found in several fungal species, only a few of them were shown to be biologically active and may represent authentic natural analogues (Fig. 1-4).

The other analogues showing marginal bioactivities may be biosynthetic derivatives or artifacts during the sample preparation and/or extraction procedure. For example, solanapyrone A was considered to be the parent compound for solanapyrones P, Q and R found in an Alternaria sp., and the derivatives show significantly reduced antibacterial activities compared to solanapyrone

A (Wang et al., 2014). The mutually exclusive selectivity of solanapyrones A and C against

Bacillus spp. is particularly of interest. One may argue that it is the structural diversity of SMs that renders the producing strain competitive against a broad range of organisms. A comprehensive structure-activity relationship study of solanapyrones analogues would reveal novel aspects regarding their specific activities against organisms sharing the same ecological niches with the producing fungi.

Aims of This Study

The amount of documented phytotoxicity of solanapyrones calls into question whether the phytotoxins are involved in pathogenicity or virulence of A. rabiei. This research initially focused on the development of an efficient screening technique for superior breeding lines with

Ascochyta blight resistance. It is laborious and time-consuming to inoculate the pathogen and assess the levels of resistance for individual breeding lines. The use of solanapyrone phytotoxin would expedite the screening processes, provided that the phytotoxin is a key determinant for disease. To establish the role of solanapyrones in pathogenicity or virulence, it would be required to generate solanapyrone-minus mutants and compare their disease-causing abilities with that of wild-type progenitors. Therefore, the aims of this study were to search for the solanapyrone biosynthetic genes in A. rabiei, to disrupt one of the biosynthetic genes, and to elucidate the possible pathogenic and ecological roles of solanapyrones.

Figure 1-1. The disease cycle of A. rabiei (teleomorph, Didymella rabiei), the causal agent of

Ascochyta blight of chickpea (Kaiser, 1997).

Figure 1-2. Chemical structure of solanapyrones produced by A. rabiei.

Figure 1-3. Chemical structures of secondary metabolites (SMs) produced by legume-associated

Ascochyta and Phoma. A. SMs produced by A. pisi and A. fabae, the structure of citrinin produced by Monascus purpureus is presented as o-quinone form for the structural comparison with ascochitine. B. SMs produced by A. lentis. C. Macrolides produced by Ascochyta and

Phoma species. D. Terpenoids produced by A. viciae.

Figure 1-4. Chemical structures of biologically active solanapyrone analogues.

REFERENCES

Alam, S.S., Bilton, J.N., Slawin, A.M.Z., Williams, D.J., Sheppard, R.N., and Strange, R.N. 1989. Chickpea blight: Production of the phytotoxins solanapyrones A and C by Ascochyta rabiei. Phytochemistry 28:2627-2630.

Andolfi, A., Cimmino, A., Villegas-Fernández, A.M., Tuzi, A., Santini, A., Melck, D., Rubiales, D., and Evidente, A. 2013. Lentisone, a new phytotoxic anthraquinone produced by Ascochyta lentis, the causal agent of Ascochyta Blight in Lens culinaris. Journal of Agricultural and Food Chemistry 61:7301-7308.

Auclair, K., Sutherland, A., Kennedy, J., Witter, D.J., Van den Heever, J.P., Hutchinson, C.R., and Vederas, J.C. 2000. Lovastatin nonaketide synthase catalyzes an intramolecular Diels-Alder reaction of a substrate analogue. Journal of the American C hemical Society 122:11519-11520.

Aveskamp, M.M., de Gruyter, J., Woudenberg, J.H.C., Verkley, G.J.M., and Crous, P.W. 2010. Highlights of the Didymellaceae: A polyphasic approach to characterise Phoma and related pleosporalean genera. Studies in Mycology 65:1-60.

Bahti, P., and Strange, R.N. 2004. Chemical and biochemical reactions of solanapyrone A, a toxin from the chickpea pathogen, Ascochyta rabiei (Pass.) Labr. Physiological and Molecular Plant Pathology 64:9-15.

Beed, F.D., Strange, R.N., Onfroy, C., and Tivoli, B. 1994. Virulence for faba bean and production of ascochitine by Ascochyta fabae. Plant Pathology 43:987-997.

Berry, E.A., Huang, L.-S., Lee, D.-W., Daldal, F., Nagai, K., and Minagawa, N. 2009. Ascochlorin is a novel, specific inhibitor of the mitochondrial cytochrome bc1 complex. Biochimica et Biophysica Acta 1797:360-370.

Bertini, S. 1956. Su di un composto ad antibiotica prodotto da Ascochyta pisi Lib. Annali Sperimentaz Agraria (Roma) 11:545-556.

Bick, I.R.C., and Rhee, C. 1966. Anthraquinone pigments from Phoma foveata Foister. Biochemical Journal 98:112-116.

Boerema, G.H., and Bollen, G.J. 1975. Conidiogenesis and conidial septation as differentiating

criteria between Phoma and Ascochyta. Persoonia 8:111-144.

Boerema, G.H., de Gruyter, J., Noordeloos, M.E., and Hamers, M.E.C. 2004. Phoma Identification Manual: Differentiation of specific and infra-specific taxa in culture, Wallingford, UK: CAB International.

Boruwa, J., Gogoi, N., and Barua, N.C. 2006. A short enantioselective total synthesis of the phytotoxic lactone herbarumin III. Organic & Biomolecular Chemistry 4:3521-3525.

Bouras, N., and Strelkov, S.E. 2008. The anthraquinone catenarin is phytotoxic and produced in leaves and kernels of wheat infected by Pyrenophora tritici-repentis. Physiological and Molecular Plant Pathology 72:87-95.

Cagnoli-Bellavita, N., Ceccherelli, P., Fringuelli, R., and Ribaldi, M. 1975. Ascochlorin: a terpenoid metabolite from Acremonium luzulae. Phytochemistry 14:807.

Capasso, R., Evidente, A., Randazzo, G., Ritieni, A., Bottalico, A., Vurro, M., and Logrieco, A. 1987. Isolation of cytochalasins A and B from Ascochyta heteromorpha. Journal of Natural Products 50:989-990.

Chen, H., Fujita, M., Feng, Q., Clardy, J., and Fink, G.R. 2004a. Tyrosol is a quorum-sensing molecule in Candida albicans. Proceedings of the National Academy of Sciences of the United States of America 101:5048-5052.

Chen, W. 2008. Chickpea diseases: Ecology and control. Encyclopedia of Pest Management 1:1-5.

Chen, W., Sharma, H.C., and Muehlbauer, F.J. 2011. Compendium of Chickpea and Lentil Diseases and Pests, St. Paul, MN, APS Press.

Chen, W., Coyne, C.J., Peever, T.L., and Muehlbauer, F.J. 2004b. Characterization of chickpea differentials for pathogenicity assay of ascochyta blight and identification of chickpea accessions resistant to Didymella rabiei. Plant Pathology 53:759-769.

Chen, Y.M., and Strange, R.N. 1991. Synthesis of the solanapyrone phytotoxins by Ascochyta rabiei in response to metal cations and development of a defined medium for toxin production. Plant Pathology 40:401-407.

Chilvers, M.I., Rogers, J.D., Dugan, F.M., Stewart, J.E., Chen, W., and Peever, T.L. 2009. Didymella pisi sp. nov., the teleomorph of Ascochyta pisi. Mycological Research

113:391-400.

Cho, S., and Muehlbauer, F.J. 2004. Genetic effect of differentially regulated fungal response genes on resistance to necrotrophic fungal pathogens in chickpea (Cicer arietinum L.). Physiological and Molecular Plant Pathology 64:57-66.

Chongo, G., Gossen, B.D., Buchwaldt, L., Adhikari, T., and Rimmer, S.R. 2004. Genetic diversity of Ascochyta rabiei in Canada. Plant Disease 88:4-10.

Cimmino, A., Andolfi, A., Fondevilla, S., Abouzeid, M.A., Rubiales, D., and Evidente, A. 2012. Pinolide, a new nonenolide produced by Didymella pinodes, the causal agent of Ascochyta blight on Pisum sativum. Journal of Agricultural and Food Chemistry 60:5273-5278.

Colanzi, A., Grimaldi, G., Catara, G., Valente, C., Cericola, C., Liberali, P., Ronci, M., Lalioti, V.S., Bruno, A., Beccari, A.R., Urbani, A., De Flora, A., Nardini, M., Bolognesi, M., Luini, A., and Corda, D. 2013. Molecular mechanism and functional role of brefeldin A-mediated ADP-ribosylation of CtBP1/BARS. Proceedings of the National Academy of Sciences of the United States of America 110:9794-9799.

Daniel, S., Tiemann, K., Wittkampf, U., Bless, W., Hinderer, W., and Barz, W. 1990. Elicitor-induced metabolic changes in cell cultures of chickpea (Cicer arietinum L.) cultivars resistant and susceptible to Ascochyta rabiei. Planta 182:270-278.

Davies, D.G., and Hodge, P. 1974. Chemistry of quinones. Part V. Structure of cladofulvin, a bianthraquinone from Cladosporium fulvum cooke. Journal of the Chemical Society, Perkin Transactions 1:2403-2405.

Driouich, A., Jauneau, A., and Staehelin, L.A. 1997. 7-Dehydrobrefeldin A, a naturally occurring brefeldin A derivative, inhibits secretion and causes a cis-to-trans breakdown of Golgi stacks in plant cells. Plant Physiology 113:487-492.

Durrant, W.E., Wang, S., and Dong, X. 2007. Arabidopsis SNI1 and RAD51D regulate both gene transcription and DNA recombination during the defense response. Proceedings of the National Academy of Sciences of the United States of America 104:4223-4227.

Ellestad, G.A., Evans, R.H., and Kunstmann, M.P. 1969. Some new terpenoid metabolites from an unidentified fusarium species. Tetrahedron 25:1323-1334.

Evidente, A., Capasso, R., Vurro, M., and Bottalico, A. 1993a. Ascosalitoxin, a phytotoxic trisubstituted salicylic aldehyde from Ascochyta pisi. Phytochemistry 34:995-998.

Evidente, A., Lanzetta, R., Capasso, R., Vurro, M., and Botralico, A. 1993b. Pinolidoxin, a phytotoxic nonenolide from Ascochyta pinodes. Phytochemistry 34:999-1003.

Evidente, A., Capasso, R., Andolfi, A., Vurro, M., and Zonno, M.C. 1998. Structure–activity relationship studies of putaminoxins and pinolidoxins: phytotoxic nonenolides produced by phytopathogenic Phoma and Ascochyta species. Natural Toxins 6:183-188.

Evidente, A., Lanzetta, R., Capasso, R., Andolfi, A., Bottalico, A., Vurro, M., and Zonno, M.C. 1995. Putaminoxin, a phytotoxic nonenolide from Phoma putaminum. Phytochemistry 40:1637-1641.

Farr, D.F., and Rossman, A.Y. (2015). Fungal Databases, Systematic Mycology and Microbiology Laboratory, Agricultural Research Service, The United States Department of Agriculture.

Fausto Rivero-Cruz, J., García-Aguirre, G., Cerda-García-Rojas, C.M., and Mata, R. 2000. Conformational behavior and absolute stereostructure of two phytotoxic nonenolides from the fungus Phoma herbarum. Tetrahedron 56:5337-5344.

García-Díaz, M., Domínguez, O., López-Fernández, L.A., de Lera, L.T., Saníger, M.L., Ruiz, J.F., Párraga, M., García-Ortiz, M.J., Kirchhoff, T., del Mazo, J., Bernad, A., and Blanco, L. 2000. DNA polymerase lambda, a novel eukaryotic DNA polymerase with a potential role in meiosis. Journal of Molecular Biology 301:851-867.

Höhl, B., Pfautsch, M., and Barz, W. 1990. Histology of disease development in resistant and susceptible cultivars of chickpea (Cicer arietinum L.) inoculated with spores of Ascochyta rabiei. Journal of Phytopathology 129:31-45.

Höhl, B., Weidemann, C., Höhl, U., and Barz, W. 1991. Isolation of solanapyrones A, B and C from culture filtrates and spore germination fluids of Ascochyta rabiei and aspects of phytotoxin action. Journal of Phytopathology 132:193-206.

Härri, E., Loeffler, W., Sigg, H.P., Stähelin, H., and Tamm, C. 1963. Über die Isolierung neuer Stoffwechselprodukte aus Penicillium brefeldianum Dodge. Helvetica Chimica Acta 46:1235-1243.

Hamid, K., and Strange, R.N. 2000. Phytotoxicity of solanapyrones A and B produced by the chickpea pathogen Ascochyta rabiei (Pass.) Labr. and the apparent metabolism of solanapyrone A by chickpea tissues. Physiological and Molecular Plant Pathology 56:235-244.

Hernandez-Bello, M.A., Chilvers, M.I., Akamatsu, H., and Peever, T.L. 2006. Host specificity of Ascochyta spp. infecting legumes of the Viciae and Cicerae tribes and pathogenicity of an interspecific hybrid. Phytopathology 96:1148-1156.

Ichihara, A., Tazaki, H., and Sakamura, S. 1983. Solanapyrones A, B and C, phytotoxic metabolites from the fungus Alternaria solani. Tetrahedron Letters 24:5373-5376.

Ilarslan, H., and Dolar, F.S. 2002. Histological and ultrastructural changes in leaves and stems of resistant and susceptible chickpea cultivars to Ascochyta rabiei. Journal of Phytopathology 150:340-348.

Jalal, M.A.F., Hossain, M.B., Robeson, D.J., and Van der Helm, D. 1992. Cercospora beticola phytotoxins: cebetins that are photoactive, magnesium ion-binding, chlorinated anthraquinone-xanthone conjugates. Journal of the American Chemical Society 114:5967-5971.

Jangili, P., Kashanna, J., Kumar, C.G., Poornachandra, Y., and Das, B. 2014. Stereoselective total synthesis of a novel regiomer of herbarumin I and its cytotoxic and antimicrobial activities. Bioorganic & Medicinal Chemistry Letters 24:325-327.

Jayakumar, P., Gan, Y.T., Gossen, B.D., Warkentin, T.D., and Banniza, S. 2005. Ascochyta blight of chickpea: Infection and host resistance mechanisms. Canadian Journal of Plant Pathology 27:499-509.

Jenkins, K.M., Toske, S.G., Jensen, P.R., and Fenical, W. 1998. Solanapyrones E-G, antialgal metabolites produced by a marine fungus. Phytochemistry 49:2299-2304.

Köhler, G., Linkert, C., and Barz, W. 1995. Infection studies of Cicer arietinum (L.) with GUS-(E. coli β-glucuronidase) transformed Ascochyta rabiei strains. Journal of Phytopathology 143:589-595.

Kachlicki, P., and Wakuliński, W. 2002. Analysis of anthraquinone pigments of Pyrenophora tritici-repentis. Acta Agrobotanica 53:97-105.

Kaiser, W.J. 1997. Inter- and intranational spread of ascochyta pathogens of chickpea, faba bean, and lentil. Canadian Journal of Plant Pathology 19:215-224.

Kaiser, W.J., Wang, B.C., and Rogers, J.D. 1997. Ascochyta fabae and A. lentis: Host specificity, teleomorphs (Didymella), hybrid analysis, and taxonomic status. Plant Disease 81:809-816.

Kamphuis, L.G., Williams, A.H., Küster, H., Trengove, R.D., Singh, K.B., Oliver, R.P., and Ellwood, S.R. 2012. Phoma medicaginis stimulates the induction of the octadecanoid and phenylpropanoid pathways in Medicago truncatula. Molecular Plant Pathology 13:593-603.

Kasahara, K., Miyamoto, T., Fujimoto, T., Oguri, H., Tokiwano , T., Oikawa, H., Ebizuka, Y., and Fujii, I. 2010. Solanapyrone synthase, a possible Diels–Alderase and iterative type I polyketide synthase encoded in a biosynthetic gene cluster from Alternaria solani. ChemBioChem 11:1245-1252.

Katayama, K., Kobayashi, T., Oikawa, H., Honma, M., and Ichihara, A. 1998. Enzymatic activity and partial purification of solanapyrone synthase: first enzyme catalyzing Diels-Alder reaction. Biochimica et Biophysica Acta - Protein Structure and Molecular Enzymology 1384:387-395.

Kato, A., Ando, K., Tamura, G., and Arima, K. 1970. Cylindrochlorin, a new antibiotic produced by Cylindrocladium. Journal of Antibiotics 23:168-169.

Kaur, S. 1995. Phytotoxicity of solanapyrones produced by the fungus Ascochyta rabiei and their possible role in blight of chickpea (Cicer arietinum). Plant Science 109:23-29.

Kim, H.J., Ruszczycky, M.W., Choi, S.-H., Liu, Y.-N., and Liu, H.-W. 2011. Enzyme-catalysed [4+2] cycloaddition is a key step in the biosynthesis of spinosyn A. Nature 473:109-112.

Kosuge, Y., Suzuki, A., Hirata, S., and Tamura, S. 1973. Structure of Colletochlorin from Colletotrichum nicotianae. Agricultural and Biological Chemistry 37:455-456.

Kraft, B., and Barz, W. 1985. Degradation of the isoflavone biochanin A and its glucoside conjugates by Ascochyta rabiei. Applied and Environmental Microbiology 50:45-48.

Kraft, B., Schwenen, L., Stöckl, D., and Barz, W. 1987. Degradation of the pterocarpan phytoalexin medicarpin by Ascochyta rabiei. Archives of Microbiology 147:201-206.

Kusnick, C., Jansen, R., Liberra, K., and Lindequist, U. 2002. Ascochital, a new metabolite from the marine ascomycete Kirschsteiniothelia maritima. Pharmazie 57:510-512.

Lakshmanan, M., and Padmanabhan, D. 1968. Effect of ascochitine on the in vitro growth of embryos of Clitoria ternatea L. Current Science 37:321-322.

Latif, Z., Strange, R.N., Bilton, J., and Riazuddin, S. 1993. Production of the phytotoxins, solanapyrones A and C and cytochalasin D among nine isolates of Ascochyta rabiei. Plant

Pathology 42:172-180.

Lin, Y.-R., Lo, C.-T., Liu, S.-Y., and Peng, K.-C. 2012. Involvement of pachybasin and emodin in self-regulation of Trichoderma harzianum mycoparasitic coiling. Journal of Agricultural and Food Chemistry 60:2123-2128.

Lv, Q., Yang, Y., Zhao, Y., Gu, D., He, D., Yili, A., Ma, Q., Cheng, Z., Gao, Y., Aisa, H.A., and Ito, Y. 2009. Comparative study on separation and purification of isoflavones from the seeds and sprouts of chickpea by high-speed countercurrent chromatography. Journal of Liquid Chromatography & Related Technologies 32:2879-2892.

Maiya, S., Grundmann, A., Li, X., Li, S.-M., and Turner, G. 2007. Identification of a hybrid PKS/NRPS required for pseurotin A biosynthesis in the human pathogen Aspergillus fumigatus. ChemBioChem 8:1736-1743.

Miethbauer, S., Gaube, F., Möllmann, U., Dahse, H.-M., Schmidtke, M., Gareis, M., Pickhardt, M., and Liebermann, B. 2009. Antimicrobial, antiproliferative, cytotoxic, and tau inhibitory activity of rubellins and caeruleoramularin produced by the phytopathogenic fungus Ramularia collo-cygni. Planta Medica 75:1523-1525.

Milgroom, M.G., and Peever, T.L. 2003. Population biology of plant pathogens: The synthesis of plant disease epidemiology and population genetics. Plant Disease 87:608-617.

Minagawa, N., Yabu, Y., Kita, K., Nagai, K., Ohta, N., Meguro, K., Sakajo, S., and Yoshimoto, A. 1996. An antibiotic, ascofuranone, specifically inhibits respiration and in vitro growth of long slender bloodstream forms of Trypanosoma brucei brucei. Molecular and Biochemical Parasitology 81:127-136.

Mizushina, Y., Kamisuki, S., Kasai, N., Shimazaki, N., Takemura, M., Asahara, H., Linn, S., Yoshida, S., Matsukage, A., Koiwai, O., Sugawara, F., Yoshida, H., and Sakaguchi, K. 2002. A plant phytotoxin, solanapyrone A, is an inhibitor of DNA polymerase β and λ. Journal of Biological Chemistry 277:630-638.

Mogensen, E.G., Challen, M.P., and Strange, R.N. 2006. Reduction in solanapyrone phytotoxin production by Ascochyta rabiei transformed with Agrobacterium tumefaciens. FEMS Microbiology Letters 255:255-261.

Mogi, T., Ui, H., Shiomi, K., Ōmura, S., Miyoshi, H., and Kita, K. 2009. Antibiotics LL-Z1272 identified as novel inhibitors discriminating bacterial and mitochondrial quinol oxidases. Biochimica et Biophysica Acta - Bioenergetics 1787:129-133.

Nakanishi, T., and Oku, H. 1969. Mechanism of selective toxicity: Absorption and detoxication of an antibiotic, ascochitine, by sensitive and insensitive fungi. Phytopathology 59:1563-1565.

Nene, Y.L. 1982. A review of Ascochyta blight of chickpea. Tropical Pest Management 28:61-70.

Nene, Y.L., and Reddy, M.V. 1987. Chickpea diseases, their control. Pages 233-270 in The chickpea (Saxena, M.C., and Singh, K.B., eds.). Wallingford, UK: CAB International.

Nene, Y.L., Hanounik, S.B., Qureshi, S.H., and Sen, B. 1988. Fungal and bacterial foliar diseases of pea, lentil, faba bean and chickpea. Pages 577-589 in World Crops: Cool Season Food Legumes (Summerfield, R.J., ed.). Dordrecht, The Netherlands: Kluwer Academic Publisher.

Nihei, C., Fukai, Y., Kawai, K., Osanai, A., Yabu, Y., Suzuki, T., Ohta, N., Minagawa, N., Nagai, K., and Kita, K. 2003. Purification of active recombinant trypanosome alternative oxidase. FEBS Letters 538:35-40.

Nizam, S., Singh, K., and Verma, P. 2010. Expression of the fluorescent proteins DsRed and EGFP to visualize early events of colonization of the chickpea blight fungus Ascochyta rabiei. Current Genetics 56:391-399.

Nukina, M., Ikeda, M., and Sassa, T. 1980a. Two new pyrenolides, fungal morphogenic substances produced by Pyrenophora teres (Diedicke) Drechsler. Agricultural and Biological Chemistry 44:2761-2762.

Nukina, M., Sassa, T., and Ikeda, M. 1980b. A new fungal morphogenic substance, pyrenolide a from Pyrenophora teres. Tetrahedron Letters 21:301-302.

Oikawa, H., Katayama, K., Suzuki, Y., and Ichihara, A. 1995. Enzymatic activity catalysing exo-selective Diels-Alder reaction in solanapyrone biosynthesis. Journal of the Chemical Society, Chemical Communications:1321-1322.

Oikawa, H., Kobayashi, T., Katayama, K., Suzuki, Y., and Ichihara, A. 1998. Total synthesis of (-)-solanapyrone A via enzymatic Diels-Alder reaction of prosolanapyrone. Journal of Organic Chemistry 63:8748-8756.

Oku, H., and Nakanishi, T. 1963. A toxic metabolite from Ascochyta fabae having antibiotic activity. Phytopathology 53:1321-1325.

Oku, H., and Nakanishi, T. 1964. Reductive detoxification of an antibiotic, ascochitine, by an

insensitive fungus, Fusarium lycopersici Sacc. Naturwissenschaften 51:538-538.

Peever, T.L., Barve, M.P., Stone, L.J., and Kaiser, W.J. 2007. Evolutionary relationships among Ascochyta species infecting wild and cultivated hosts in the legume tribes Cicereae and Vicieae. Mycologia 99:59-77.

Peever, T.L., Chen, W., Abdo, Z., and Kaiser, W.J. 2012. Genetics of virulence in Ascochyta rabiei. Plant Pathology 61:754-760.

Peever, T.L., Salimath, S.S., Su, G., Kaiser, W.J., and Muehlbauer, F.J. 2004. Historical and contemporary multilocus population structure of Ascochyta rabiei (teleomorph: Didymella rabiei) in the Pacific Northwest of the United States. Molecular Ecology 13:291-309.

Rea, G., Metoui, O., Infantino, A., Federico, R., and Angelini, R. 2002. Copper amine oxidase expression in defense responses to wounding and Ascochyta rabiei invasion. Plant Physiology 128:865-875.

Sasaki, H., Hosokawa, T., Sawada, M., and Ando, K. 1973. Isolation and structure of ascofuranone and ascofranol, antibiotics with hypolipidemic activity. Journal of Antibiotics 26:676-680.

Schmidt, L.E., Gloer, J.B., and Wicklow, D.T. 2007. Solanapyrone analogues from a Hawaiian fungicolous fungus. Journal of Natural Products 70:1317-1320.

Seibert, S.F., Eguereva, E., Krick, A., Kehraus, S., Voloshina, E., Raabe, G., Fleischhauer, J., Leistner, E., Wiese, M., Prinz, H., Alexandrov, K., Janning, P., Waldmann, H., and Konig, G.M. 2006. Polyketides from the marine-derived fungus Ascochyta salicorniae and their potential to inhibit protein phosphatases. Organic & Biomolecular Chemistry 4:2233-2240.

Shimizu, T., Kinoshita, H., Ishihara, S., Sakai, K., Nagai, S., and Nihira, T. 2005. Polyketide synthase gene responsible for citrinin biosynthesis in Monascus purpureus. Applied and Environmental Microbiology 71:3453-3457.

Silletta, M., Colanzi, A., Weigert, R., Di Girolamo, M., Santone, I., Fiucci, G., Mironov, A., Matteis, M., Luini, A., and Corda, D. 1999. Role of brefeldin A-dependent ADP-ribosylation in the control of intracellular membrane transport. Pages 43-51 in ADP-Ribosylation Reactions: From Bacterial Pathogenesis to Cancer (Alvarez-Gonzalez, R., eds.) Dordrecht, The Netherlands: Kluwer Academic Publisher.

Singh, K.B., Reddy, M.V., and Nene, Y.L. 1984. International testing of chickpeas for resistance

to Ascochyta blight. Plant Disease 68:782-784.

Smith, G.R., Munro, M.H.G., Fineran, B.A., and Cole, A.L.J. 1994. Evidence for the involvement of ascochitine in phoma leafspot-wilt disease of Clematis. Physiological and Molecular Plant Pathology 45:333-348.

Song, J., Durrant, W.E., Wang, S., Yan, S., Tan, E.H., and Dong, X. 2011. DNA repair proteins are directly involved in regulation of gene expression during plant immune response. Cell Host & Microbe 9:115-124.

Takamatsu, S., Rho, M.-C., Masuma, R., Hayashi, M., Komiyama, K., Tanaka, H., and Omura, S. 1994. A novel testosterone 5α-reductase inhibitor, 8', 9'-dehydroascochlorin produced by Verticillium sp. FO-2787. Chemical & Pharmaceutical Bulletin 42:953-956.

Takatsuki, A., Tamura, G., and Arima, K. 1969. Antiviral and antitumor antibiotics. XIV. Effects of ascochlorin and other respiration inhibitors on multiplication of Newcastle disease virus in cultured cells. Applied Microbiology 17:825-829.

Tamura, G., Suzuki, S., Takatsuki, A., Ando, K., and Arima, K. 1968. Ascochlorin, a new antibiotic, found by paper-disc agar-diffusion method. Journal of Antibiotics 21:539-544.

Tietjen, K.G., Schaller, E., and Matern, U. 1983. Phytotoxins from Alternaria carthami Chowdhury: Structural identification and physiological significance. Physiological Plant Pathology 23:387-400.

Trapero-Casas, A., and Kaiser, W.J. 1992. Influence of temperature, wetness period, plant age, and inoculum concentration on infection and development of Ascochyta blight of chickpea. Phytopathology 82:589-596.

Trapero-Casas, A., and Kaiser, W.J. 2007. Differences between ascospores and conidia of Didymella rabiei in spore germination and infection of chickpea. Phytopathology 97:1600-1607.

Trapero-Casas, A., Navas-Cortés, J.A., and Jiménez-Díaz, R.M. 1996. Airborne ascospores of Didymella rabiei as a major primary inoculum for Ascochyta blight epidemics in chickpea crops in southern Spain. European Journal of Plant Pathology 102:237-245.

Trisuwan, K., Rukachaisirikul, V., Sukpondma, Y., Preedanon, S., Phongpaichit, S., and Sakayaroj, J. 2009. Pyrone derivatives from the marine-derived fungus Nigrospora sp. PSU-F18. Phytochemistry 70:554-557.

Varshney, R.K., Song, C., Saxena, R.K. et al. 2013. Draft genome sequence of chickpea (Cicer arietinum) provides a resource for trait improvement. Nature Biotechnology 31:240-246.

Venkatasubbaiah, P., and Chilton, W.S. 1992. Phytotoxins of Ascochyta hyalospora, causal agent of lambsquarters leaf spot. Journal of Natural Products 55:461-467.

Vurro, M., and Ellis, B.E. 1997. Effect of fungal toxins on induction of phenylalanine ammonia-lyase activity in elicited cultures of hybrid poplar. Plant Science 126:29-38.

Vurro, M., Zonno, M.C., Evidente, A., Capasso, R., and Bottaiico, A. 1992. Isolation of cytochalasins A and B from Ascochyta lathyri. Mycotoxin Research 8:17-20.

Wang, S., Durrant, W.E., Song, J., Spivey, N.W., and Dong, X. 2010. Arabidopsis BRCA2 and RAD51 proteins are specifically involved in defense gene transcription during plant immune responses. Proceedings of the National Academy of Sciences of the United States of America 107:22716-22721.

Wang, X.-Z., Luo, X.-H., Xiao, J., Zhai, M.-M., Yuan, Y., Zhu, Y., Crews, P., Yuan, C.-S., and Wu, Q.-X. 2014. Pyrone derivatives from the endophytic fungus Alternaria tenuissima SP-07 of Chinese herbal medicine Salvia przewalskii. Fitoterapia 99:184-190.

Weber, R.W.S., Stenger, E., Meffert, A., and Hahn, M. 2004. Brefeldin A production by Phoma medicaginis in dead pre-colonized plant tissue: A strategy for habitat conquest? Mycological Research 108:662-671.

Wenke, J., Anke, H., and Sterner, O. 1993. Pseurotin A and 8-O-demethylpseurotin A from Aspergillus fumigatus and their inhibitory activities on chitin synthase. Bioscience, Biotechnology, and Biochemistry 57:961-964.

Wu, S.H., Chen, Y.W., Shao, S.C., Wang, L.D., Yu, Y., Li, Z.Y., Yang, L.Y., Li, S.L., and Huang, R. 2009. Two new solanapyrone analogues from the endophytic fungus Nigrospora sp. YB-141 of Azadirachta indica. Chemistry & Biodiversity 6:79-85.

Wu, Z., Song, L., Feng, S., Liu, Y., He, G., Yioe, Y., Liu, S.Q., and Huang, D. 2012. Germination dramatically increases isoflavonoid content and diversity in chickpea (Cicer arietinum L.) seeds. Journal of Agricultural and Food Chemistry 60:8606-8615. Yamtich, J., and Sweasy, J.B. 2010. DNA polymerase family X: Function, structure, and cellular roles. Biochimica et Biophysica Acta - Proteins and Proteomics 1804:1136-1150.

Yan, S., Wang, W., Marqués, J., Mohan, R., Saleh, A., Durrant, W.E., Song, J., and Dong, X.

2013. Salicylic acid activates DNA damage responses to potentiate plant immunity. Molecular Cell 52:602-610.

CHAPTER TWO

FUNCTIONAL ANALYSES OF THE DIELS-ALDERASE GENE SOL5 OF

ASCOCHYTA RABIEI AND ALTERNARIA SOLANI INDICATE THAT

THE SOLANAPYRONE PHYTOTOXIN ARE NOT REQUIRED

FOR PATHOGENECITY

ABSTRACT

Ascochyta rabiei and Alternaria solani, the causal agents of Ascochyta blight of chickpea

(Cicer arietinum) and early blight of potato (Solanum tuberosum), respectively, produce a set of phytotoxic compounds including solanapyrones A, B, and C. Although both the phytotoxicity of solanapyrones and their universal production among field isolates have been documented, the role of solanapyrones in pathogenicity is not well understood. Here we report the functional characterization of the sol5 gene, which encodes a Diels-Alderase that catalyzes the final step of solanapyrone biosynthesis. Deletion of sol5 in both A. rabiei and Al. solani completely prevented production of solanapyrones and led to accumulation of the immediate precursor compound prosolanapyrone II-diol, which is not toxic to plants. Deletion of sol5 did not negatively affect growth rate or spore production in vitro, and led to overexpression of the other solanapyrone biosynthesis genes, suggesting a possible feedback regulation mechanism. Phytotoxicity tests showed that solanapyrone A is highly toxic to several legume species and Arabidopsis thaliana.

Despite the apparent phytotoxicity of solanapyrone A, pathogenicity tests showed that solanapyrone-minus mutants of A. rabiei and Al. solani were equally virulent as their corresponding wild-type progenitors, suggesting that solanapyrones are not required for pathogenicity.

INTRODUCTION

Ascochyta rabiei is a necrotrophic plant pathogenic fungus and causes Ascochyta blight of chickpea (Cicer arietinum L.). The disease is manifested through necrotic lesions on all above ground tissues of the host plant (Cicer spp.) and may cause complete crop loss under conducive environmental conditions. Severe disease epidemics have been reported in various regions of the world (Nene, 1982; Singh et al., 1984). The pathogen is known to produce polyketide-derived secondary metabolites including solanapyrone A (SolA), solanapyrone B (SolB) and solanapyrone C (SolC) (Alam et al., 1989; Chen and Strange, 1991). The same set of secondary metabolites were previously detected and identified in cultures of Alternaria solani, the causal agent of the early blight of potato and tomato (Ichihara et al., 1983).All reliably identified A. rabiei isolates produce solanapyrones and their phytotoxicity has been investigated using chickpea cell suspension culture (Höhl et al., 1991; Latif et al., 1993) and detached or intact plants (Kaur, 1995; Hamid and Strange, 2000). In order to demonstrate involvement of the toxins in pathogenesis, efforts were made to correlate sensitivity of chickpea cultivars to the solanapyrone toxins with their susceptibility to the pathogen. A positive, but non-significant

correlation was detected among nine cultivars (Hamid and Strange, 2000). Toxin production varies significantly by isolate and prolific solanapyrone-producers tend to be more virulent (Kaur,

1995). Although many studies have focused on the phytotoxicity of solanapyrones, the occurrence of solanapyrones during infection process has not been demonstrated and attempts to extract solanapyrones from infected plant tissues were unsuccessful (Höhl et al., 1991).

Therefore, it has been proposed that the absence of solanapyrones in infected plants may be attributed to its rapid metabolism by host plants and the ability to metabolize solanapyrones has been used as an indicator of host resistance to the disease (Hamid and Strange, 2000; Bahti and

Strange, 2004).

Recently, the solanapyrone biosynthesis gene cluster was identified in Al. solani

(Kasahara et al., 2010). The cluster comprises six genes (sol1 to sol6) covering approximately 20 kb of the genome. Among the six sol genes in the cluster, sol5 encodes a Diels-Alderase that catalyzes the final step of solanapyrone biosynthesis. Purified recombinant Sol5 protein was able to convert in vitro synthesized precursor substrate, prosolanapyrone II, to SolA, demonstrating its involvement in both oxidation and the subsequent cyclization of the precursor compound via

Diels-Alder reaction (Kasahara et al., 2010). The gene sol1 encodes a polyketide synthase (PKS) that initiates the solanapyrone biosynthesis pathway. Like many other secondary metabolite gene clusters, the solanapyrone gene cluster also includes a transcription factor-encoding gene (sol4).

Roles of other sol genes, sol2 (an O-methyltransferase), sol3 (a dehydrogenase) and sol6 (a cytochrome P450), were tentatively assigned to the proposed biosynthetic scheme for

solanapyrones (Kasahara et al., 2010).

Despite the strong phytotoxicity of solanapyrones, relatively little information is available on its mechanism of toxicity or the range of plants to which these compounds are phytotoxic. It was originally reported that SolA showed variable toxicity to a somaclonal population of potato

(cv. Russet Burbank) in conjunction with an unidentified toxic compound in culture of Al. solani

(Matern et al., 1978). Also, Kaur (1995) suggested solanapyrones as a host-selective toxin with an observation that non-host legumes of A. rabiei, such as pea (Pisum sativum), cowpea (Vigna unguiculata), and green bean (Phaseolus vulgaris), were less sensitive to solanapyrones than chickpea. Both A. rabiei and Al. solani belong to the class Dothideomycetes which includes all plant-pathogenic fungi that produce host-selective toxins (Turgeon and Lu, 2000).

Host-selective toxins often target specific cellular components and may induce apoptosis

(Walton, 1996; Tsuge et al., 2013). SolA was shown to specifically bind to X-family DNA polymerases in vitro (Mizushina et al., 2002). This particular DNA polymerase family exerts its function exclusively on DNA repair processes during mitosis and meiosis (García-Díaz et al.,

2000; Yamtich and Sweasy, 2010). The mammalian genome possesses several different genes belonging to X-family DNA polymerase, whereas in higher plants DNA polymerase λ is the sole member of X-family DNA polymerase (Roy et al., 2009). Recently, it has become evident that proteins involved in DNA repair processes play crucial roles in the regulation of plant defense response (Song et al., 2011). Through a genetic screening of Arabidopsis thaliana mutants, several DNA repair proteins were characterized to be key components of plant immune signaling

and defense gene expression (Durrant et al., 2007; Wang et al., 2010).

On the basis of its ability to cause disease on chickpea differentials, A. rabiei isolates can be broadly classified into two pathotypes named pathotype I (less virulent) or pathotype II

(virulent) (Chen et al., 2004). However, recent studies of virulence in A. rabiei employing crosses between pathotype I and pathotype II isolates revealed that most of the progeny exhibited intermediate virulence suggesting that virulence in A. rabiei is under polygenic control (Peever et al., 2012). No complete resistance to Asochyta blight has been reported in chickpea germplasm and resistance to the pathogen is known to be controlled by quantitative trait loci (Santra et al.,

2000; Iruela et al., 2006). The quantitative nature of both host resistance and pathogen virulence suggests that preformed or induced physical/chemical defense responses of the host plant and the subversion of these responses by the pathogen may be important determinants of the overall outcome (disease severity) of chickpea–A. rabiei interactions (reviewed in Jayakumar et al.,

2005).

Solanapyrones have long been considered to play a crucial role in pathogenicity or virulence, since solanapyrones are the sole chemical family of secondary metabolites with phytotoxic activities produced by A. rabiei. However, the role of the toxins in pathogenicity has not been critically examined. More recently, T-DNA insertion mutants that exhibited reduced solanapyrone production were generated via Agrobacterium-mediated random mutagenesis.

However, the mutants were not characterized with respect to the T-DNA insertion sites or pathogenicity (Mogensen et al., 2006). The recent identification of the solanapyrone biosynthesis

gene cluster in Al. solani allows us to examine the role of the toxins in both A. rabiei and Al. solani through the use of solanapyrone-minus mutants generated with a targeted gene deletion approach.

The objectives of this study were to isolate solanapyrone toxins from A. rabiei and test their toxicity to host and selected non-host plants, develop solanapyrone-minus mutants through targeted-gene replacement, and characterize the toxin-minus mutants for growth and pathogenicity.

MATERIALS AND METHODS

Fungal Strains and General Culture Conditions

A. rabiei isolates AR19 (MAT1-2, ATCC 24891), AR21 (MAT1-1, ATCC 76502) and

AR628 (MAT1-1, ATCC 201622) were obtained from the A. rabiei collection maintained at the

USDA Western Regional Plant Introduction Station. A GFP-expressing strain of AR628

(Akamatsu et al., 2006) was also used for this study. Two field isolates ALS1 and ALS2 of Al. solani, collected from separate fields in 2010 and 2011 were provided by Dr. Dennis Johnson

(Washington State University, USA). Details on the isolates used in this study are listed in Table

2-1. For conidial production, A. rabiei isolates were grown on V8 agar media (200 mL V8 juice,

3 g CaCO3, 20 g agar in 1 L distilled water) for 2 weeks. For Al. solani isolates, spores were harvested from 2-week-old colonies on a modified PDA (1 g potato dextrose agar [BD

Diagnostic Systems, USA], 6 mL ethanol, and 20 g agar in 1 L distilled water).

Generation of Gene Replacement Constructs and Solanapyrone-minus Mutants

The sol5 gene was specifically deleted from the genome of A. rabiei and Al. solani, using the split marker method (Catlett et al., 2003). DNA fragments of the gene replacement constructs overlapping within the hygromycin B phosphotransferase (hph) gene were amplified using a double-joint PCR (Yu et al., 2004), with a minor modification. In the first round PCR from genomic DNA of AR628 isolate, an 800 bp upstream region of the sol5 coding region was amplified using primer 5-1/primer 5-2 pair, and a 1,354 bp downstream of the sol5 coding sequence was amplified using primer 5-3/primer 5-4 pair. A 1,372 bp hph cassette was amplified from pDWJ5 using HYG-F/HYG-R primer pair. The primer 5-2 and primer 5-3 carried 27-bp sequence tails that overlapped with the 5′ and 3′ ends of the hph cassette, respectively. In the second round PCR, each sol5 flanking DNA fragment was fused to the hph cassette through PCR by overlap extension. The resulting PCR products were cloned to pGEM-T Easy vector

(Promega, USA). DNA sequences of the split-marker constructs were verified by sequencing with T7 or SP6 primer. The vectors cloned with the upstream flanking DNA construct and the downstream flanking DNA construct were used as templates for the third round PCR with nested primer pairs, primer 5-7/HY-R pair for the upstream flanking DNA construct and primer

5-8/YG-F pair for the downstream flanking DNA construct. The upstream and downstream split-marker constructs were used for transformation of both A. rabiei and Al. solani. Primers and their sequences used in this study are listed in Appendix A.

Fungal Tra nsformation and Molecular Confirmation of Transformants

Preparation of fungal protoplasts and transformation was conducted as previously described with a minor modification (Akamatsu et al., 2010). Instead of using a disruption plasmid vector, one to 5 μg of each upstream and downstream split-marker DNA construct were used for genetic transformation of A. rabiei and Al. solani. Putative transformants were subcultured on V8 agar containing hygromycin B (200 μg/mL), and two rounds of single spore isolation were conducted to obtain homokaryotic transformants. Selected transformants were screened by a diagnostic PCR with the primer 5-11/primer 5-12 pair to amplify the full length

DNA sequence of sol5 gene in wild-types and hph cassette in putative Δsol5 mutants (Fig. 2-1).

Homologous integration of the split-marker DNA constructs were verified by PCR with an upstream flanking primer, primer 5-1, and HYG-R, confirming correct integration of the split-marker constructs into the sol5 gene locus.

Growth Rate and in vitro Sexual Stage Induction

Agar plugs (3 mm diam.) containing actively growing mycelia on PDA were transferred to fresh PDA and the cultures were incubated under 12 h fluorescent light at 20 ± 2°C. Colony diameters in orthogonal directions in five replicate plates (13.5 cm diam.) for each strain were measured in 3-day intervals for up to 36 days. Isolates AR628 (MAT1-1) and AR19 (MAT1-2) or their corresponding Δsol5 mutants were crossed as previously described by Wilson and Kaiser

(1995). Production of pseudothecia and ascospores on chickpea straw was examined 8 weeks

after incubation at 10°C.

RNA Extraction, RT-PCR, and Quantitative Real-time RT-PCR Analyses

RNA was extracted from cultures (mycelia and spores) of isolate AR628 and its Δsol5 mutant using RNeasy Plant Mini Kit (Qiagen). On-column digestion of possible genomic DNA contamination was performed using RNase-Free DNase Set (Qiagen). For RT-PCR analysis of solanapyrone biosynthesis genes, 200 ng of RNAs were reverse transcribed and amplified using

OneStep RT-PCR kit (Qiagen). sol genes (sol1–sol6) were amplified for 30-32 cycles; Actin1 for

28 cycles (annealing temperature, 62°C). The primer sets for RT-PCR analysis were designed to amplify flanking exons, including one intron in order to distinguish mRNA from pre-mRNA

(and/or possible gDNA contamination). Real-time RT-PCR assays were used to monitor expression levels of sol1and sol5 genes. Isolate AR628 was grown on PDA and sampled for

RNA extraction at 3-day intervals up to 21 days. For comparison of sol gene expression on PDA between the wild-type and the Δsol5 mutant, RNA was extracted from PDA culture 9 days after growth. First-strand cDNA synthesis was prepared from total RNA using the iScript cDNA

Synthesis Kit (Bio-Rad) according to manufacturer’s instructions. Real-time RT-PCR analyses were performed using the Bio-Rad iQ5 Real-Time PCR Detection System. RT-PCR mixtures were composed of 12 pmol of each primer, 12.5 μL of SYBR Green Supermix (Bio-Rad), 2 μL of cDNA (a 1:2.5 dilution of the 20 μL cDNA product), and nuclease-free water to a final volume of 25 μL. The PCR conditions consisted of a denaturing step at 95°C for 2 min, and

DNA denaturation step at 95°C for 10 s and both annealing and extension steps at 62°C for 30 s

for 45 cycles, and 65–95°C with a 0.5°C increment, each temperature for 5 s to obtain the melting curve. The quantification of the relative transcript levels of the solanapyrone biosynthesis genes in A. rabiei was performed with the comparative CT method normalization

(Schmittgen and Livak, 2008). The transcript levels of target genes were normalized against the

Actin1 gene transcript levels. The averages of the three biological replicates and standard deviations of the relative expression values were presented.

Solanapyrone Extraction and Purification

For extraction of solanapyrones from liquid cultures of A. rabiei and Al. solani isolates, thirty milliliters of half-strength PDB in 250-mL Erlenmeyer flasks were inoculated with 100 μL of conidial suspension of A. rabiei isolates (1 × 107 spores mL-1), or Al. solani isolates (1 × 105 spores mL-1), and incubated at 20°C without shaking. After 16-18 days, cultures were filtered through four layers of Miracloth (Calbiochem, USA) in a vacuum to remove mycelium.

Solanapyrones were extracted using Sep-Pak® Vac 6 cc (1g) tC18 cartridges (Waters Corp.,

Milford, MA, USA). Individual cartridges were eluted with 2 mL of MeOH and the eluents were subjected to UPLC-MS analysis.

For purification of solanapyrones A and C, 200 grams of autoclaved oat kernels in a mason jar (about 1 L size) were inoculated with ~50 agar plugs (5 mm diam.) containing actively growing mycelia of isolate AR628 or the Δsol5 mutant on V8 agar. The culture was incubated at

20°C under a 12/12 light regime and shaken daily to disperse the inoculums. After 16 days,

solanapyrones were extracted from the A. rabiei-colonized oat kernels with 200 mL of ethyl acetate (EtOAc). Extracts were dried over anhydrous MgSO4 and filtered through four layers of cheesecloth to remove oat kernel debris and MgSO4 and concentrated on a rotary evaporator. The residue was further purified by preparative TLC on silica gel; solanapyrone A [Rf 0.58, eluent

CHCl3/EtOH (95:5, v/v), yielding 2.8 mg] and solanapyrone C [Rf 0.44, eluent CHCl3/EtOH

(95:5, v/v), 1.3 mg]. Analytical and preparative thin layer chromatography was performed on

® TLC LuxPlate Silica gel 60 F254 plates, 0.25 and 1.0 mm (Merck, Germany). The spots were visualized by exposure to UV radiation (254 nm). For purification, UV-active spots of individual compounds on silica gel were collected separately, and dissolved in EtOAc/MeOH (9:1, v/v) to elute target compounds out from the silica gel.

For purification of the major compound eluted at 2.26 minutes in the chromatograms of the Δsol5 mutants, culture filtrates of AR628Δsol5 grown in half-strength PDB were partitioned with the same volume of EtOAc, and the organic phases were combined and partitioned again with distilled water (pH 3.0). Finally, the organic phases were combined and dried over anhydrous MgSO4 and concentrated on a rotary evaporator. The residue was purified by preparative TLC on silica gel; prosolanapyrone II-diol [Rf 0.26, eluent EtOAc/MeOH/H2O

(80:5:5, v/v), 4.1 mg], and subjected to spectrometric analyses for structural identification.

UPLC-MS Analysis

Chromatographic separation was achieved using an ACQUITY UPLC system (Waters

Corp.) on a 2.1 mm × 50 mm i.d., 1.7 µm, ACQUITY UPLC BEH C18 column (Waters Corp.).

Mobile phase A was water with 0.1% formic acid, and mobile phase B was acetonitrile with

0.1% formic acid. The gradient started from 7% B, followed by an 8 min linear gradient from 7% to 99% B. The gradient held it for 2 min and was finally stepped to 7% B to equilibrate the column for 2 min. The total LC run time was 13 min, the column temperature was maintained at

30°C, and the flow rate was 0.45 mL min-1.

MS analysis was performed on an inline Synapt G2-S HDMS (Waters Corp.) time of flight mass spectrometer. Positive ion electrospray mode was used for data collection. The desolvation gas was nitrogen (800 L h-1), and the collision gas was argon (2.0 mL min-1). The data acquisition range was m/z 50–1,000. The source temperature was 120°C, and the desolvation temperature was 250°C. The cone voltage was 30 V. The lock mass compound used was leucine encephalin with a reference mass at m/z 556.2771. For the TOF experiments, data were acquired in the MSE mode in which two separate scan functions were programmed for the

MS acquisition method. One scan function was set at low collision energy (trap at 4 eV and transfer at 2 eV), and the other scan function was set at high collision energy (trap ramped from

15–50 eV and transfer at 2 eV). The mass spectrometer switched rapidly between the two functions during data acquisition. As a result, information on intact precursor ions and on product ions was obtained from a single LC run.

For identification of the peak eluted at 4.32 in the chromatograms of the wild-type culture extracts, the following methodology was used. First, the elemental composition of the detected

peak was determined by comparing the exact mass of the precursor ion detected in the low collision energy spectrum and the exact mass of the proposed candidate (only differences below

5 mDa were accepted). A good fit in the isotopic pattern of both compounds was also required for the selection of the best candidate. The suitability of possible oxidized forms of SolA was checked through MassFragment software (Waters Corp.). MassFragment made it possible to evaluate whether the product ions detected in the high collision energy spectrum are linked to the fragments generated from the chemical structures of the candidates proposed. A score between 1 and 14 was provided by the software for each product ion. The lower the score value, the higher the plausibility of the fragment proposed. An oxidized form of SolA was considered confirmed by MassFragment software because most of the product ions showed the score provided by the software was equal to or below 3.

Phytotoxicity Tests

The phytotoxicity of solanapyrone A, solanapyrone C, and prosolanapyrone II-diol was tested on chickpea genotypes W6 22589 (resistant), FLIP 84-92C (intermediate), chickpea cv.

Spanish White (susceptible), pea cv. Dark Skin Perfection (Pisum sativum), lentil cv. Pardina

(Lens culinaris), and sweet pea Spencer-type mix (Lathyrus odoratus). Three plants of each species were grown in 1-litre plastic pots filled with Sunshine mix #4 (Sun Gro Horticulture,

Canada) and maintained for 4 weeks in a growth chamber (Conviron Model PGR15, Winnepeg,

Manitoba, Canada) at 12 h day (20°C) and 12 h night (16°C). The youngest, fully expanded leaves (leaves of the second or the third nodes from the top) were detached from each plant, and

the adaxial sides of leaflets were gently punctured with a pin to facilitate absorption of test compounds. The leaflets were placed on two layers of plastic mesh in a glass tray (22 × 30 cm) with wet paper towels, and the trays were covered with a plastic bag to keep leaflets moist and kept at 20°C. Purified compounds were adjusted to 16.5 μg/μL by adding MeOH and then brought up to a final concentration of 1 μg/μL in distilled water. A drop (3 μL) of each solution or solvent control (6% MeOH) was placed onto the punctured wound on the epidermis (9–12 leaflets per treatment). Necrotic lesions were photographed 3 days after application. The areas of necrotic lesions were calculated using Assess 2.0 Image Analysis Software for Plant Disease

Quantification (APS Press, USA) and used as an indicator of phytotoxicity.

Arabidopsis thaliana seeds, POLL-overexpressing transgenic line (POLLOE), and T-DNA insertion line (SALK_075391, POLLKD) were obtained from Drs. Rino Cella and Giovanni Maga

(University of Pavia, Italy). The overexpression and knockdown of POLL gene were verified via real-time RT-PCR analysis using the method described by Amoroso et al. (2011) (Fig. S2-5).

Surface-sterilized Arabidopsis seeds were sown on half strength Murashige and Skoog (MS) growth media (pH 5.7) supplemented with 1.0% sucrose and 0.7% phytoagar in square plastic

Petri dishes with grids (BD Biosciences, USA). After 3 days vernalization, the plates were kept in a growth chamber (16/8 h, light/dark cycle at 22oC) for 7 days. The seedlings were gently transferred to MS growth media containing different concentrations of solanapyrone A (20, 40, or

60 μM). The plates were further incubated for 6 days and growth of primary roots were measured and photographed. For media without solanapyrone A (0 μM), MeOH was amended to the MS

growth media, as a solvent control (0.5%, v/v).

Pathogenicity Tests

Pathogenicity test, using chickpea cvs. Spanish White and Dwelley, and disease assessment were conducted as described in Chen et al. (2004). For co-inoculation experiment, three seeds of chickpea cv. CDC Frontier were sown in 1-litre plastic pot (12 pots for the experiment) containing Sunshine mix #4 (Sun Gro Horticulture) and maintained in a growth chamber (Conviron Model PGR15) under a 12 h photoperiod with 20°C day and 16°C night at >

95% relative humidity. Spores of wild-type strain (AR628WT) and its sol5 mutant

(AR628sol5) were harvested from V8 agar media and mixed at 1:1 ratio in sterilized water, making a final concentration of 2 × 105 spores mL-1. Inoculation method was as described in

Chen et al. (2004). In order to calculate relative infection rates of AR628WT and AR628sol5, total number of lesions produced after incubation for 2 weeks was counted and each lesion was numbered and cut off from the plant along with a stem segment (≈ 3 cm long). After measuring lesion length, the fungal genotype that caused the lesion was isolated and determined by assessing colony morphology on acidified PDA (39 g PDA, 1 mL lactic acid per liter added after autoclaving, pH 3.5); fungal colonies exhibiting dark and restricted growth pattern were assigned to AR628WT, while lighter colonies with expansive growth were assigned to AR628sol5. The identity of fungal colony was further confirmed by subculturing on V8 agar containing 200

μg/mL of hygromycin B; AR628sol5 can grow on medium containing hygromycin B, whereas

AR628WT cannot. Lesions from which both AR628WT and AR628sol5 were obtained (less

than 10% of total lesions) were not included in chi-square goodness of fit test.

A detached leaf assay was used for measuring virulence of Al. solani isolates, ALS1,

ALS2 and their respective Δsol5 mutants. Leaves were excised from 6-week-old plants of two potato cultivars (Russet Norkotah and Ranger Russet) and one tomato cultivar (Moneymaker) grown in a greenhouse under a 16-h photoperiod with 25°C day and 18°C of night. Detached leaves were inoculated with 10 µL of spore suspension (2 × 104 spores mL-1) contained in a lens paper (0.7 cm diameter). Leaves were placed on a glass tray with five layers of paper towels at room temperature, and covered with a transparent plastic bag to maintain near 100% relative humidity. The inoculated leaves were photographed at 6 dpi. The area of the necrotic lesion on each leaf was calculated by the analysis of images as described above. The lesion areas were calibrated based on area of a square (1 cm2) paper that was included in each photograph.

DATA Analyses

For pathogenicity tests with chickpea cvs. Spanish White and Dwelley, experiments were conducted in completely randomized designs. Disease severity ratings (1 to 9) were recorded for each plant and the mean of the two plants in a pot represented one experimental unit (n = 6).

Levene’s test for equal variance assumption was used to determine if variance was independent of the mean and Shapiro-Wilk test was used to check for normality assumption. Student’s t-test was used for mean comparisons between wild-type strain and its corresponding sol5 mutant.

For co-inoculation experiment, chi-square goodness of fit test was used to detect any deviation

from the expected 1:1 ratio in total number of stem lesions caused by wild-type (AR628WT) and

sol5 mutant (AR628sol5). The mean lesion length and lesion number caused by AR628sol5 were compared with those caused by AR628WT, using Student’s t-test. Mean lesion length and lesion number produced on three chickpea plants (cv. CDC Frontier) in a pot represented one experimental unit (n = 12).

For detached leaf assay, differences in average lesion size produced on potato or tomato leaves by each strain were analyzed using one-way ANOVA (main effect, strain). No significant difference in mean lesion size was observed at a significance level of 0.05. Lesion produced on each detached leaf represented one experimental unit (n = 8). For phytotoxicity test, square-root transformed data was used for one-way ANOVA (main effect, legume species) as the original data violated the normality and equal variance assumptions. After data transformation, the assumptions were checked by Shapiro-Wilk’s and maximum likelihood ratio tests. Multiple comparisons of mean lesion size incited by application of 3 μg of SolA between different legume species were performed by LSD test using Bonferroni correction. Null hypotheses of no difference were rejected when P ≤ 0.05. Lesion produced on excised leaflets represented one experimental unit (n = 10–12). The analyses were performed using the MIXED procedure in the

SAS v9.2 software (SAS Institute Inc., USA).

RESULTS

Solanapyrone Biosynthesis Gene Cluster in A. rabiei

Based on the DNA sequence of the solanapyrone biosynthesis gene cluster identified from Al. solani (GenBank accession AB514562) (Kasahara et al., 2010), we recovered ≈ 10 kb genomic DNA sequence through a homology-based PCR approach. This DNA sequence included the full-length DNA sequences of sol5 and sol4 genes (GenBank accessions KM244526 and KM244525), coding for a solanapyrone synthase (a Diels-Alderase) and a putative transcription factor, respectively. Partial DNA sequences of other clustered genes (sol1, sol2, sol3 and sol6) were also successfully amplified from A. rabiei genomic DNA (isolate AR628,

ATCC 201622). The homologous sol5 DNA sequence and its deduced amino acid sequence were

97% identical to Al. solani sol5 (Fig. S2-1). No obvious orthologs of sol5 were found in 61 complete or draft Dothideomycetes genome databases that include Didymella exigua, the type species of the genus Didymella (Grigoriev et al., 2014). All other genes in the cluster were approximately 97% similar between A. rabiei and Al. solani. This result showed that the solanapyrone gene clusters are highly conserved between the distantly related fungal species that belong to the families, Didymellaceae (A. rabiei) and Pleosporaceae (Al. solani) in the class

Dothideomycetes.

Generation of Solanapyrone-minus Mutants

To study the role of solanapyrone toxins in pathogenicity, we generated solanapyrone-minus mutants from both A. rabiei and Al. solani by targeting the sol5 gene that catalyzes the final step of solanapyrone biosynthesis (Kasahara et al., 2010). For targeted sol5 gene deletion, three A. rabiei and two Al. solani isolates were used as progenitors (Table 2-1), and sol5 was deleted from the genomes using the gene replacement strategy illustrated in Figure

2-1.

To check if sol5-deletion (Δsol5) mutants lost the ability to produce solanapyrones, we compared chemical profiles of culture extracts of Δsol5 mutants to those of the wild-type progenitors. LC-MS analysis of 18-d old cultures of wild-type AR628 (A. rabiei) and ALS1 (Al. solani) isolates confirmed the production of SolA (eluted at 4.21 min) and SolC (eluted at 4.03 min) (Fig. 2-2). While SolA was a major component and consistently found in cultures, the amount of SolC production was highly variable from culture to culture and the timing of extraction, similar to previous observations for Al. solani (Oikawa et al., 1998b). Regardless of type of media and extraction method, SolB was never detected by a high resolution

UPLC-MS/MS system (Synapt G2, Waters Corp.) in cultures of A. rabiei or Al. solani in this study. Instead, a new peak (eluted at 4.32 min) was detected in the chromatogram (Fig. 2-2). The elemental composition and tandem mass analysis of the detected peak suggested the compound was an oxidized form of SolA having the molecular formula C18H22O5 (see Materials and

Methods). Structural identification of the compound is currently in progress. In Δsol5 culture

extracts, there was no detectable solanapyrone production. Several new peaks were detected that were absent in the chromatograms of the wild-type culture extracts (Fig. 2-2).

Structural Identification of the Precursor Compound in Δsol5 Mutants

To elucidate the structure of the commonly detected compound eluted at 2.26 min in the

LC-MS analysis of Δsol5 mutant cultures of A. rabiei and Al. solani (Fig. 2-2), it was purified by preparative TLC, and subsequently analyzed using 1D and 2D NMR techniques. The correlation details among protons and between protons and carbons were illustrated in Figure S2-2. Based on1H-1H COSY spectrum and by interpretation of HMBC data (Figs. S2-3 and S2-4), the compound was unambiguously identified as prosolanapyrone II-diol having the molecular

+ formula C18H26NaO6 (m/z = 361.1 [M + Na] ). This intensive spectrometric identification further confirmed the deletion of sol5 gene, which led to the accumulation of an oxidized form of prosolanapyrone II being used previously in in vitro synthesis of solanapyrones (Oikawa et al.,

1998a).

Solanapyrone Biosynthesis Gene Expression and Its Feedback Regulation

A. rabiei is known to produce solanapyrones during the onset of conidial sporulation

(Höhl et al., 1991). To gain further insights into kinetics and the timing of solanapyrone production in culture, we investigated transcript levels of the key solanapyrone biosynthesis genes, sol1 and sol5 involved in the first and the last steps of the solanapyrone biosynthesis pathway, respectively (Kasahara et al., 2010). As with secondary metabolite biosynthesis gene

regulation in other fungi (Payne et al., 1993; Peplow et al., 2003), sol1 and sol5 appear to be co-regulated, showing a similar expression pattern (Fig. 2-3 A). Only a trace amount of transcripts were detected from young mycelia at 3 days post incubation (dpi). As the fungal colony become mature, the expression levels of sol1 and sol5 gradually increased, peaking at 12 or 15 dpi when the fungal colonies were profusely sporulating followed by a reduction in expression thereafter. The kinetics of sol gene expression was consistent with the association of solanapyrone production with the conidiation process.

To examine the effect of sol5 deletion on the expression of other solanapyrone biosynthesis genes, transcript levels of sol genes (sol1-sol6) were compared between AR628 and its Δsol5. The sol5 gene deletion was confirmed by a complete lack of sol5 gene transcript in

Δsol5 (Fig. 2-3 B). The sol5 gene seemed to not be efficiently spliced in the wild-type as the pre-mRNA band was of nearly equal intensity as the mRNA band, although no pre-mRNA band was detected in the reference gene, Actin1. More interestingly, the RT-PCR analysis revealed that sol4 was markedly overexpressed in Δsol5. Real-time RT-PCR analysis indicated that expression levels of sol1 and sol4 genes in Δsol5 were approximately 5-fold and 25-fold higher than in the wild-type, respectively (Fig. 2-3 C).

Effect of Solanapyrones on Growth of A. rabiei and Al. solani

Although growth rates of A. rabiei and Al. solani differ greatly, wild-type isolates of A. rabiei and Al. solani both showed a restricted growth phenotype on nutrient-rich medium, PDA

(Fig. 2-4 A). Growth rates of the wild-type strains gradually decreased as the colonies matured

(Fig. 2-4 B). In contrast, growth rates of Δsol5 mutants were constant until they reached the edge of the agar plates (13.5 cm diam.). Conidial morphology and timing of conidiation of the Δsol5 were similar to their respective progenitors of either A. rabiei or Al. solani (data not shown). A. rabiei is an outcrossing fungal species requiring mating between two strains of opposite mating types to develop the sexual stage (Didymella rabiei) (Wilson and Kaiser, 1995). The sexual stage

(teleomorph) of A. rabiei was also successfully induced in the laboratory by crossing

AR628Δsol5 (MAT1-1) and AR19Δsol5 (MAT1-2) (Fig. 2-5). Therefore, solanapyrone production does not appear to be required for normal growth and development nor for sexual or asexual reproduction, thus fulfilling the criterion as secondary metabolites.

Phytotoxicity of Solanapyrones on Host and Non-host Plants

The phytotoxicity of solanapyrones to chickpea, the host of A. rabiei, has been well documented. However, the spectrum of the toxicity on different plant species and the mode of action remain largely unknown. Therefore, toxicity of SolA, SolC and the precursor, prosolanapyrone II-diol (Pros-II-diol) to host and non-host legume species was evaluated. The compounds were purified from culture extracts of wild-type (AR628) or Δsol5 mutant, and the purity and identities were verified via LC-MS analysis (Fig. 2-6). Among the tested compounds, only SolA showed significant phytotoxicity to all the tested plant species. Multiple comparisons of mean lesion size produced by 3 μg application of SolA (1 μg/μL, 3 μL) showed no significant difference in sensitivity to the toxin among chickpea genotypes with varying disease resistance to

Ascochyta blight (Fig. 2-7). SolC exhibited a moderate phytotoxic effect at the tested concentrations, while the precursor showed no toxic effect (Fig. 2-7). The solvent control (6%

MeOH) did not produce any detectable necrosis. This result suggests that SolA is a major phytotoxic compound with non-host selectivity.

Previously, SolA was shown to specifically bind to X-family DNA polymerases in vitro

(Mizushina et al., 2002). To test if SolA indeed targets the specific DNA polymerase such as

DNA polymerase λ (POLL), the sole member of X-family DNA polymerase in plants (Roy et al.,

2009), we utilized the model Arabidopsis system in which the modes of action of many fungal- phytotoxins have been studied (Stone et al., 2000; Lorang et al., 2004; Nishiuchi et al., 2006).

The seedling growth of Arabidopsis wild-type (Col-0 ecotype), POLL knockdown mutant

(POLLKD), and POLL overexpression transgenic plants (POLLOE) were monitored on MS agar amended with different concentrations of SolA. As the concentration of SolA increased, primary root growth was progressively restricted until it was completely inhibited at 60 μM (Fig. 2-8 A).

Levels of inhibition were not significantly different among the tested Arabidopsis genotypes, suggesting that Arabidopsis POLL is unlikely the target of SolA. Intriguingly, while primary root growth was inhibited by SolA treatment, lateral root branching was rather induced by increasing dose of SolA (Fig. 2-8 B). The lateral root branching was not due to the inactivation of SolA over time because fresh seedlings exhibited the same phenotype on 10-d old media containing 60 μM

SolA, indicating that this phenotype resulted from specific, but as yet unknown, actions of SolA.

Solanapyrones Not Required for Pathogenicity

The role of solanapyrones in pathogenicity of A. rabiei to chickpea was evaluated by comparing the disease-causing abilities of Δsol5 mutants with their respective progenitors which include isolates of different pathotypes. Since the precursor (Pros-II-diol) that accumulated in the

Δsol5 mutant cultures did not show toxicity to plants, it is assumed that the precursor would not contribute to pathogenicity or virulence. Overall disease symptoms and severity caused by Δsol5 mutants in susceptible chickpea cvs. Spanish White and Dwelley were not different from those caused by the wild-type strains (Fig. 2-9 A). In addition, microscopic examination at the early stage of infection further confirmed that pathogenic behaviors of the Δsol5 mutant were similar to the wild-type strain in terms of conidial germination, germ tube elongation on leaf surface and hyphal branching in the subepidermal layer at 1–2 dpi (Fig. 2-9 B).

Disease incidence rates of wild-type and Δsol5 mutant were also evaluated by a co-inoculation method on a resistant cv. CDC Frontier (Vail and Banniza, 2008). In the first experiment the total number of lesions caused by the mutant was significantly higher than that caused by the wild-type (P = 0.030) (Table 2-2). However in the second experiment, there was no statistical difference in total number of lesions incited by the wild-type or the Δsol5 mutant (P

= 0.479). The average length and number of lesions produced by the mutant was not different from those produced by the wild-type strains, based on Student’s t-test (Table 2-2). The result demonstrated that the solanapyrone-minus mutant was equally virulent to chickpea irrespective of the host genotype differing in the Ascochyta blight resistance.

With the Al. solani wild-type isolates and their corresponding Δsol5 mutants, we also tested the involvement of solanapyrones in the Al. solani-potato/tomato pathosystems. The sizes of necrotic lesions incited by the Δsol5 mutants were not significantly different from those incited by the wild-type strains on both potato and tomato (Table 2-3; Fig. 2-10).

Taken together, this series of pathogenicity tests with solanapyrone-minus mutants provided evidence that solanapyrones are not required either for Ascochyta blight of chickpea or for early blight of potato/tomato.

DISCUSSION

Fungi produce a diverse array of secondary metabolites that show various biological activities. It is tempting to define the role of secondary metabolites produced by plant pathogenic fungi as pathogenicity or virulence factors, especially when the chemical compounds exhibit toxicity (chlorosis or necrosis) to plant tissues when applied to plants independent of the pathogen. However, more direct evidence is needed before claiming that a compound is a pathogenicity factor, since the phytotoxicity could arise from a non-target effect of the test compound on living organisms. Here, we report generation of solanapyrone-minus mutants to elucidate the role of solanapyrones in pathogenicity. The resulting mutants all lacked the ability to produce solanapyrone and instead accumulated the immediate precursor that showed no apparent toxicity to plants. It has been reported that different genotypes of a given pathogen species may deploy different arsenals controlling disease development (Siewers et al., 2005). To

rule out the possibility of strain-specific use of solanapyrones in pathogenicity, we generated solanapyrone-minus mutants from three A. rabiei isolates varying in virulence and geographic origin, as well as from two Al. solani isolates. A series of pathogenicity tests consistently showed that solanapyrone-minus mutants were all as virulent as their respective wild-type progenitors.

We assumed that the precursors of solanapyrones play no roles in virulence, which was confirmed by targeted deletion of sol4 gene, a transcription factor in the solanapyrone gene cluster. The sol4-deletion mutants produced neither solanapyrone nor the precursors found in

Δsol5 mutants, yet they were equally virulent on chickpea as their progenitors (see Chapter 3).

These results strongly suggest that solanapyrones are not a virulence factor in both A. rabiei and

Al. solani, but may play an important role in other aspects of pathogen biology. Possible roles of solanapyrones in saprobic growth stage will be discussed below.

Evolutionary Perspective on the Origin of Solanapyrone Biosynthesis Gene Cluster

Horizontal gene transfer (HGT) is defined as the movement of stable genetic material between different strains or species (Doolittle, 1999). Recently, several HGT events of secondary metabolite gene clusters between fungi have been reported (Akagi et al., 2009; Khaldi and Wolfe,

2011; Slot and Rokas, 2011). The presence of solanapyrone biosynthesis gene cluster in the two different species, A. rabiei (Didymellaceae) and Al. solani (Pleosporaceae), suggests its acquisition by a HGT event between the two species, or from a third organism independently.

The DNA sequence identity of sol5 is over 97% between the two fungi. The similarity of these two genes was higher than those of housekeeping genes such as Actin1 (GenBank accessions

KM244530, JQ671723) and β-tubulin (GenBank accessions KM244529, HQ413317), which showed 91% and 90% of sequence identity, respectively and supporting the HGT hypothesis.

Genes involved in secondary metabolite biosynthesis, such as PKS and non-ribosomal peptide synthetase, are known to evolve rapidly (Inderbitzin et al., 2010). For example, T-toxin and PM-toxin that have similar chemical structures and the same biological activity are produced by Cochliobolus heterostrophus (Pleosporaceae) and Didymella zeae-maydis (Didymellaceae)

(Yoder, 1973). The homologous PKS genes responsible for biosynthesis of T-toxin and PM-toxin showed only 60% DNA sequence identity, suggesting divergence of the gene from a common ancestor, rather than HGT between the two species (Inderbitzin et al., 2010). The genome sequence data of A. rabiei isolate AR628 showed that the orthologous sol1 genes of A. rabiei and

Al. solani shared ≈ 97% identity (see Chapter 3). This remarkably high degree of similarity may indicate that the solanapyrone biosynthesis gene cluster was horizontally transferred, rather than vertically transmitted from a common ancestor of A. rabiei and Al. solani.

Solanapyrone Biochemistry

The sol5 gene encodes a Diels-Alderase that catalyzes the last step of solanapyrone biosynthesis. Diels-Alderase is a unique enzyme rarely found in nature and believed to be engaged in biosynthesis of commercially important chemicals, such as lovastatin (a cholesterol-lowering drug) and spinosyn A (a potent insecticide) (Auclair et al., 2000; Kim et al.,

2011). Sol5 is the first example of a naturally occurring Diels-Alderase (Katayama et al., 1998).

The structural identification of the major metabolite accumulated in Δsol5 mutants suggested it is prosolanapyrone II with a vicinal diol on its hydrocarbon chain. The structural difference from prosolanapyrone II could have resulted from a high degree of substrate promiscuity of cytochrome P450 monooxygenase. Some P450 enzymes are prone to react with multiple substrates (Ekroos and Sjögren, 2006; O'Reilly et al., 2013). Therefore, it is likely that the prosolanapyrone II that accumulated in Δsol5 mutants is oxidized by the physiologically irrelevant enzymatic reaction of a P450 (e.g. Sol6), forming an epoxide, which subsequently opens to form a vicinal diol. Notably, sol6 gene was overexpressed in the Δsol5 mutant (Fig. 2-3

B). To the best of our knowledge, this is the first genetic evidence that Sol5 catalyzes both oxidation and Diels-Alder reaction in vivo.

Interestingly, sol1 and sol4 genes were overexpressed in the Δsol5 mutant, suggesting that solanapyrone production is tightly controlled in the fungal cell. The accumulation of the precursor compound may lead to activation of the clustered genes possibly via a positive feedback loop. However, identification of upstream genetic factors affecting solanapyrone production would be necessary for improved understanding of the regulation mechanism. The velvet protein complex is conserved across filamentous fungi and known to coordinate developmental processes and secondary metabolite production (Bok and Keller, 2004; Wu et al.,

2012). Given the association of solanapyrone production with the conidiation processes, the homologous velvet complex proteins in A. rabiei and Al. solani may be involved in the regulation

of solanapyrone production as shown in other polyketide-derived secondary metabolites such as

T-toxin, aflatoxin, and lovastatin.

Non-host-selectivity of Solanapyrone A

In our comparative toxicity assay of solanapyrones, we did not detect a difference in necrotic lesion size caused by SolA among chickpea accessions with different levels of resistance to the disease. This is in contrast to previous findings of positive correlations between susceptibility to Ascochyta blight and toxin sensitivity in chickpea accessions (Matern et al.,

1978; Kaur, 1995; Hamid and Strange, 2000). Although only three accessions were tested in our study, we included the most resistant (W6 22589, nearly immune to most of A. rabiei isolates) and the most susceptible (Spanish White susceptible to both pathotypes I and II) (Chen et al.,

2004).

In this study we did not observe any specificity of solanapyrone toxins to specific legumes. Larger necrotic lesions were produced on non-host legumes (pea, lentil and sweet pea) than on chickpea. This may indicate a varying degree of sensitivity of plant species to SolA, as is the case for a phytotoxic compound produced by a closely related species, A. lentis (Andolfi et al., 2013). However, in our analysis it was observed that the epidermis of the non-host legumes was waxier than that of chickpea, and that droplets containing the test compounds remained for a longer period of times on the wounds. Due to the different texture of leaf surfaces, we were unable to test phytotoxicity on potato and tomato (the host of Al. solani). The droplets placed

onto leaflets of potato/tomato spread out soon after application, and as a result, caused very small or no lesions. Therefore, the prolonged contact of test compounds on leave wounds of the non-host legumes may account for larger lesion size caused by SolA. Nevertheless, SolC and the precursor Pros-II-diol exhibited moderate or no effects on host and non-host plants even at the highest concentration. These results suggest that SolA is a major non-host-selective phytotoxin produced by A. rabiei and Al. solani.

Effects of Solanapyrone A on Arabidopsis thaliana.

Many phytotoxic compounds target one or more cellular components, thereby disrupting normal cell function (Möbius and Hertweck, 2009). SolA is also known to specifically bind to

X-family DNA polymerases in vitro (Mizushina et al., 2002). However, whether SolA exerts its toxic effect on plants by targeting the particular DNA polymerase is unknown. To shed light on the mode of action of SolA, we utilized the Arabidopsis model system. Upon SolA treatment, primary root growth of Arabidopsis was suppressed at a concentration of 60 μM. This substantial toxic effect of SolA on Arabidopsis is comparable to that of a well-known phytotoxin deoxynivalenol (Masuda et al., 2007). However, Arabidopsis transgenic plants either lacking or overexpressing POLL gene responded to SolA similarly, indicating that POLL may not be the molecular target of SolA in Arabidopsis.

It is notable that SolA enhanced lateral root branching at concentrations that the primary root was completely inhibited. This unique phenotype induced by SolA is reminiscent of primary

root-specific inhibition caused by a synthetic compound used in a chemical genetics screen, which inhibits early abscisic acid signaling (Kim et al., 2012). Further genetic studies revealed that a member of nucleotide binding site-leucine-rich repeat (NB-LRR) proteins is responsible for the phenotype induced by the compound. This phenotypic similarity may suggest SolA functions in the signal transduction of the plant immune response in Arabidopsis. More recently, it was reported that victorin toxin produced by the necrotrophic fungus, Cochliobolus victoriae, was known to activate the plant immune system through a NB-LRR protein (Lorang et al., 2012).

For further studies, it will be of interest to search for a genetic factor contributing to the primary root arrest phenotype upon SolA treatment.

Possible Roles of Solanapyrones in Pathogen Biology

Genomic loci implicated in pathogenicity and secondary metabolite production tend to be highly polymorphic and divergent even within the same species (Cuomo et al., 2007; Rouxel et al., 2011). Among the worldwide collection of A. rabiei isolates, intraspecific karyotype variation was also observed (Akamatsu et al., 2012). Despite the high degree of genome plasticity of A. rabiei, sol5 gene was amplified from all 30 isolates chosen from the collection

(Table 2-4). Additionally, previous studies have detected production of solanapyrones in all the A. rabiei isolates used in the studies (Höhl et al., 1991; Kaur, 1995). The universal production of solanapyrones and the highly conserved cluster genes between the two different fungal species, A. rabiei and Al. solani, suggests its indispensable role in the pathogen biology other than parasitism.

A. rabiei undergoes the saprobic phase in its life cycle between chickpea growing seasons.

The pathogen is able to survive in chickpea debris and remain viable for more than 2 years at a low relative humidity (< 30%) (Kaiser, 1973). During this stage of the life cycle, the pathogen encounters numerous microbial competitors including bacteria, fungi, and nematodes. In fact, many phytotoxins produced by phytopathogenic fungi display antibiotic activities (Waring et al.,

1988; Boudart, 1989; Ali-Vehmas et al., 1998). Also, it has been reported that solanapyrone analogues found in cultures of non-pathogenic fungi exhibited antibiotic activities (Schmidt et al.,

2007; Trisuwan et al., 2009; Wu et al., 2009). Solanapyrones appear to be toxic to the fungi that produce them because the growth of the wild-type strains was restrained, while solanapyrone-minus mutants were able to grow almost indeterminately (Fig. 2-4). The restricted growth of a maturing wild-type fungal colony may be a consequence of the antifungal activity of accumulated solanapyrones in culture. We are currently investigating possible ecological roles in niche specialization during saprobic growth of the pathogen life cycle such as competition withother microbes and survival on chickpea debris.

ACKNOWLEDGMENTS

I am grateful to Drs. Rino Cella and Giovanni Maga (University of Pavia, Italy) for kindly providing Arabidopsis transgenic lines (POLLOE and POLLKD), and to Dr. Amber Vanden

Wymelenberg (University of Wisconsin) and Dr. Daniel Cullen (USDA Forest Service) for providing pTEFEGFP. I also acknowledge Dr. Hideaki Oikawa (Hokkaido University) for

providing solanapyrone A standard and Dr. Hak Joong Kim (Korea University) for valuable suggestions on solanapyrone biochemistry.

Table 2-1. Isolates of Ascochyta rabiei and Alternaria solani used in this study

Isolate and strain Mating Origin of Geographic Year of typea isolation location isolation

Ascochyta rabiei AR628WT (ATCC 201622) MAT1-1 C. arietinumc Syria 1995 AR628Δsol5 - - - -

AR628GFP (TEFprom::GFP)b - - - - AR628GFPΔsol5 - - - - AR21WT (ATCC 76502) MAT1-1 C. arietinum ID, USA 1986 AR21Δsol5 - - - - AR19WT (ATCC 24891) MAT1-2 C. arietinum Iran 1990

AR19Δsol5 - - - - Alternaria solani ALS1WT NA S. tuberosumd WA, USA 2011 ALS1Δsol5 - - - - ALS2WT NA S. tuberosum ID, USA 2010 ALS2Δsol5 - - - - a Mating type designation according to Barve et al. (2003). NA = not available. b Green fluorescent protein (GFP) gene fused to translation elongation factor promoter. c Cicer arietinum L. d Solanum tuberosum L.

Table 2-2. Mean lesion length and lesion number produced in chickpea cv. CDC Frontier stems inoculated with mixed inoculum of Ascochyta rabiei isolate AR628 (WT) and its Δsol5 mutant

Lesion Lesion Total Treatment lengtha numbera lesionsb

Exp.1 WT 1.08 4.50 54 Δsol5 0.93 6.58 79 P valuec 0.896 0.550 0.030

Exp.2 WT 0.89 7.92 95 Δsol5 0.86 8.75 105 P valuec 0.629 0.939 0.479 a The length and number of lesions produced in one pot containing 3 chickpea plants were averaged out and considered experimental unit (n = 12), lesion length in cm. b Lesions produced in 12 pots were assigned either to WT or Δsol5, based on colony morphology and selection marker resistance during subculture. c P values for lesion length and lesion number were obtained by Student’s t-tests, and P values for total lesion number were obtained by χ2 goodness of fit test.

Table 2-3. Lesion size (cm2) caused by Alternaria solani isolates and Δsol5 mutants on leaves of potato and tomato cultivars.

Potato Tomato

Strain cv. Russet Norkotah cv. Ranger Russet cv. Moneymaker

ALS1WT 0.87 0.59 1.06

ALS1sol5 0.78 0.51 0.80

ALS2WT 0.77 0.47 1.15

ALS2sol5 0.65 0.41 0.82

P valuea 0.919 0.403 0.611 a Null hypothesis of no difference in mean lesion sizes produced by the wild type and mutant strains was not rejected by one-way ANOVA (α = 0.05)

Table 2-4. List of isolates examined for the presence of sol5 gene

Mating Karyo- Number of Geographic Year of Isolate typea typeb chromosomeb location isolation AR19 MAT1-2 B 14 Iran 1990 AR20 MAT1-2 B 14 USA, ID 1986 AR21 MAT1-1 B 15 USA, ID 1986 AR45 MAT1-1 B 13 Morocco 1991 AR51 MAT1-1 B 12 Algeria 1991 AR66 MAT1-1 B 12 Bulgaria 1991 AR74 MAT1-2 B 13 Tunisia 1991 AR97 MAT1-1 C 13 Itary 1991 AR116 MAT1-1 B 13 Spain 1991 AR160 MAT1-1 B 12 Bulgaria 1992 AR164 MAT1-1 C 14 Syria 1992 AR181 MAT1-2 C 14 France 1991 AR208 MAT1-2 B 13 India 1993 AR231 MAT1-2 A 13 USA, ND 1993 AR254 MAT1-1 B 13 Bulgaria 1994 AR398 MAT1-1 B 13 USA, ID 1987 AR415 MAT1-1 B 15 USA, ID 1987 AR607 MAT1-2 B 13 Canada 1995 AR628 MAT1-1 B 14 Syria 1995 AR649 MAT1-2 B 13 Morocco 1992 AR735 MAT1-2 A 13 Lebanon 1995 AR745 MAT1-2 A 14 Australia 1995 AR746 MAT1-2 A 16 Australia 1996 AR823 NA B 14 Bolivia 2000 AR827 NA B 14 Bolivia 2000 AR830 MAT1-2 B 14 Australia 1997 AR837 MAT1-2 B 14 Australia 1991 AR844 NA B 13 Ethiopia 1997 BOYD-99 MAT1-1 B 14 USA, ID 1999 MSR-98 MAT1-1 B 14 USA, ID 1998 a Mating type designation according to Barve et al. (2003). NA = not available b Karyotype and estimated chromosome numbers are as described in Akamatsu et al. (2012).

Figure 2-1. Gene replacement via split-marker strategy and PCR analysis of sol5-deletion. A.

Schematic diagram of targeted gene replacement strategy. B. PCR analysis of A. rabiei wild-type isolates and their Δsol5 mutants. Replacement of sol5 gene to hph cassette was verified in two independent mutants from each isolate using a primer pair, primer 5-11 and primer 5-12

(wild-types; 2,330 bp, Δsol5 mutants; 1,952 bp). Note that AR21Δsol5 2-1 and AR19Δsol5 5-2 had heterokaryotic nuclei; both wild-type- and mutant-specific bands were amplified. C. PCR verification for homologous integration of the replacement fragment to the correct genomic site with a primer pair, primer 5-1 and HYG-R (2,179 bp).

Figure 2-2. Detection of metabolites in cultures of A. rabiei and Al. solani and their Δsol5 mutants. Total ion chromatograms of equivalent quantities of culture extracts of isolates AR628

(A. rabiei), ALS1 (Al. solani) and the respective Δsol5 mutants, indicated by observed m/z values with their respective retention times (in parenthesis) for each compound at the peak.

Figure 2-3. Expression pattern of solanapyrone biosynthesis genes. A, Time-course gene expression levels of sol1 and sol5 in the wild-type strain AR628 of A. rabiei on PDA. B,

Comparison of expression levels of solanapyrone cluster genes (sol1–sol6) between AR628 wild-type (WT) and its Δsol5 mutant via RT-PCR analysis. Total RNA was extracted from 9-d old colony grown on PDA. M = 100 bp DNA ladder marker. Total genomic DNA (gDNA) of the wild-type was used to compare PCR products from DNA (The size differences were due to presence of an intron in genomic DNA). C, Relative expression levels of sol1, sol4, and sol5 genes in the AR628 wild-type (WT) and its Δsol5 mutant. Quantitative RT-PCR was performed and the gene expression of sol genes were normalized by Actin1 expression levels in respective samples. Error bars are standard deviation of three biological replicates.

Figure 2-4. Comparison of growth phenotypes of A. rabiei and Al. solani wild-type strains and their respective Δsol5 mutants. A, Representative pictures of colony morphology of wild-type isolates AR628 and AR19, and their corresponding Δsol5 mutants of A. rabiei 5 weeks after incubation on PDA. B, Growth curves of the wild-type isolates (WT) and their corresponding

Δsol5 mutants on PDA. Error bars are standard deviation (n = 5). C, Representative pictures of colony morphology of wild-type isolates ALS1 and ALS2, and their corresponding Δsol5 mutants of Al. solani 3 weeks after incubation on PDA. D, Growth curves of the wild-type isolates (WT) and their corresponding Δsol5 mutants on PDA.

Figure 2-5. Production of pseudothecia and ascospores. Genetic crosses were made either between wild-type strains AR628 (MAT1-1) and AR19 (MAT1-2) or between the solanapyrone-minus mutants AR628Δsol5 and AR19Δsol5

Figure 2-6. Purification and verification of solanapyrones and the immediate precursor compound. Base peak ion chromatograms of culture extract of A. rabiei isolate AR628 grown on half strength PDB (A), preparative TLC-purified solanapyrone A (B) and solanapyrone C (C) from the wild-type strain, and prosolanapyrone II-diol (D) from the Δsol5 mutant, with mass spectrum (insets left) and UV spectrum (insets right) for each purified compound at the peak.

Figure 2-7. Comparative analysis of phytotoxic effects of solanapyrones on legumes. Droplets (3

μL) of different concentrations (0.04, 0.2 or 1.0 μg/μL) of the purified compounds were applied to pin-punctured wounds on leaflets. Necrotic areas produced by solanapyrone A (solid lines), solanapyrone C (dashed lines) and prosolanapyrone II-diol (dotted lines) 3 days after application.

Different alphabetic letters indicate significant difference in mean lesion size produced on legumes by 1.0 μg/μL of solanapyrone A (LSD test using Bonferroni correction, P ≤ 0.05). Bars represent standard deviation (n = 9–12). Chickpea (R) = resistant germplasm line W6 22589,

Chickpea (I) = intermediate resistant germplasm line FLIP 84-92C, Chickpea (S) = susceptible chickpea cv. Spanish White, Pea = cv. Dark Skin Perfection (Pisum sativum), Lentil = cv.

Pardina (Lens culinaris), Sweet Pea = Spencer-type mix (Lathyrus odoratus).

Figure 2-8. Effects of solanapyrone A on root growth of Arabidopsis thaliana. A, Inhibition of primary root elongation by solanapyrone A. Primary root growth was measured 6 days after transfer of Arabidopsis seedlings (Col-0 ecotype, POLLOE and POLLKD) to the half strength of

MS medium containing the indicated concentrations of solanapyrone A. Bars represent standard deviation (n = 17–21). B, Representative pictures showing lateral root branching induced by solanapyrone A. Pictures were taken 7 days after treatments. Note enhanced root branching by increasing concentrations of solanapyrone A.

Figure 2-9. Pathogenicity tests of A. rabiei and its Δsol5 mutants. A, Mean disease severity of two chickpea cvs. Spanish White and Dwelley caused by three isolates AR628 (pathotype II),

AR19 (pathotype I), AR21 (pathotype I), and their corresponding Δsol5 mutants. Disease score was assessed 2 weeks after inoculation based on the 1–9 rating scale (1 = no symptom, 9 = dead plants). Student’s t-test showed that the disease severity caused by the Δsol5 mutants was not different from that caused by their respective wild-type progenitors. B, Visualization of germinating spores and subsequent hyphal branching of green fluorescent-labeled strain

(AR628GFP) and its Δsol5 mutant version (AR628GFPΔsol5) on leaflets of cv. Spanish White.

Bar = 20 μm.

Figure 2-10. Detached leaf assay for virulence of Al. solani isolates (ALS1 and ALS2) and the corresponding Δsol5 mutants on potato cvs. Russet Norkotah (A), Ranger Russet (B), and tomato cv. Moneymaker (C). Arrangement of strains is as in A.

REFERENCES

Akagi, Y., Akamatsu, H., Otani, H., and Kodama, M. 2009. Horizontal chromosome transfer, a mechanism for the evolution and differentiation of a plant-pathogenic fungus. Eukaryotic Cell 8:1732-1738.

Akamatsu, H.O., Chilvers, M.I., Stewart, J.E., and Peever, T.L. 2010. Identification and function of a polyketide synthase gene responsible for 1,8-dihydroxynaphthalene-melanin pigment biosynthesis in Ascochyta rabiei. Current Genetics 56:349-360.

Akamatsu, H.O., Chilvers, M.I., Kaiser, W.J., and Peever, T.L. 2012. Karyotype polymorphism and chromosomal rearrangement in populations of the phytopathogenic fungus, Ascochyta rabiei. Fungal Biology 116:1119-1133.

Akamatsu, H.O., Stone, L.J., Sigler, A.A., Chilvers, M.I., Arie, T., and Peever, T.L. 2006. Heterologous expression of fluorescence protein in the phytopathogenic fungus Ascochyta rabiei and visualization of the fungus in planta. Phytopathology 96:S4.

Ali-Vehmas, T., Rizzo, A., Westermarck, T., and Atroshi, F. 1998. Measurement of antibacterial activities of T-2 toxin, deoxynivalenol, ochratoxin A, aflatoxin B1 and fumonisin B1 using microtitration tray-based turbidimetric techniques. Journal of Veterinary Medicine Series A 45:453-458.

Amoroso, A., Concia, L., Maggio, C., Raynaud, C., Bergounioux, C., Crespan, E., Cella, R., and Maga, G. 2011. Oxidative DNA damage bypass in Arabidopsis thaliana requires DNA polymerase λ and proliferating cell nuclear antigen 2. Plant Cell 23:806-822.

Andolfi, A., Cimmino, A., Villegas-Fernández, A.M., Tuzi, A., Santini, A., Melck, D., Rubiales, D., and Evidente, A. 2013. Lentisone, a new phytotoxic anthraquinone produced by Ascochyta lentis, the causal agent of Ascochyta blight in Lens culinaris. Journal of Agricultural Food Chemesity 61:7301-7308.

Bahti, P., and Strange, R.N. 2004. Chemical and biochemical reactions of solanapyrone A, a toxin from the chickpea pathogen, Ascochyta rabiei (Pass.) Labr. Physiological and Molecular

Plant Pathology 64:9-15.

Barve, M.P., Arie, T., Salimath, S.S., Muehlbauer, F.J., and Peever, T.L. 2003. Cloning and characterization of the mating type (MAT) locus from Ascochyta rabiei (teleomorph: Didymella rabiei) and a MAT phylogeny of legume-associated Ascochyta spp. Fungal Genetics and Biology 39:151-167.

Bok, J.W., and Keller, N.P. 2004. LaeA, a regulator of secondary metabolism in Aspergillus spp. Eukaryotic Cell 3:527-535.

Boudart, G. 1989. Antibacterial activity of sirodesmin PL phytotoxin: Application to the selection of phytotoxin-deficient mutants. Applied and Environmental Microbiology 55:1555-1559.

Catlett, N.L., Lee, B.N., Yoder, O.C., and Turgeon, B.G. 2003. Split-marker recombination for efficient targeted deletion of fungal genes. Fungal Genetics Newsletter 50:9-11.

Chen, W., Coyne, C.J., Peever, T.L., and Muehlbauer, F.J. 2004. Characterization of chickpea differentials for pathogenicity assay of ascochyta blight and identification of chickpea accessions resistant to Didymella rabiei. Plant Pathology 53:759-769.

Cuomo, C.A., Güldener, U., Xu, J.-R. et al. 2007. The Fusarium graminearum genome reveals a link between localized polymorphism and pathogen specialization. Science 317:1400-1402.

Doolittle, W.F. 1999. Lateral genomics. Trends in Cell Biology 9:M5-M8.

Ekroos, M., and Sjögren, T. 2006. Structural basis for ligand promiscuity in cytochrome P450 3A4. Proceedings of the National Academy of Sciences of the United States of America 103:13682-13687.

Grigoriev, I.V., Nikitin, R., Haridas, S., Kuo, A., Ohm, R., Otillar, R., Riley, R., Salamov, A., Zhao, X., Korzeniewski, F., Smirnova, T., Nordberg, H., Dubchak, I., and Shabalov, I. 2014. MycoCosm portal: Gearing up for 1000 fungal genomes. Nucleic Acids Research

42:D699-D704.

Ichihara, A., Tazaki, H., and Sakamura, S. 1983. Solanapyrones A, B and C, phytotoxic metabolites from the fungus Alternaria solani. Tetrahedron Letters 24:5373-5376.

Inderbitzin, P., Asvarak, T., and Turgeon, B.G. 2010. Six new genes required for production of T-toxin, a polyketide determinant of high virulence of Cochliobolus heterostrophus to maize. Molecular Plant-Microbe Interactions 23:458-472.

Iruela, M., Rubio, J., Barro, F., Cubero, J.I., Millán, T., and Gil, J. 2006. Detection of two quantitative trait loci for resistance to ascochyta blight in an intra-specific cross of chickpea (Cicer arietinum L.): Development of SCAR markers associated with resistance. Theoretical and Applied Genetics 112:278-287.

Jayakumar, P., Gan, Y.T., Gossen, B.D., Warkentin, T.D., and Banniza, S. 2005. Ascochyta blight of chickpea: Infection and host resistance mechanisms. Canadian Journal of Plant Pathology 27:499-509.

Kaiser, W.J. 1973. Factors affecting growth, sporulation, pathogenicity, and survival of Ascochyta rabiei. Mycologia 65:444.

Kaur, S. 1995. Phytotoxicity of solanapyrones produced by the fungus Ascochyta rabiei and their possible role in blight of chickpea (Cicer arietinum). Plant Science 109:23-29.

Khaldi, N., and Wolfe, K.H. 2011. Evolutionary origins of the fumonisin secondary metabolite gene cluster in Fusarium verticillioides and Aspergillus niger. International Journal of Evolutionary Biology, Article ID 423821.

Kim, H.J., Ruszczycky, M.W., Choi, S.-H., Liu, Y.-N., and Liu, H.-W. 2011. Enzyme-catalysed [4+2] cycloaddition is a key step in the biosynthesis of spinosyn A. Nature 473:109-112.

Kim, T.-H., Kunz, H.-H., Bhattacharjee, S., Hauser, F., Park, J., Engineer, C., Liu, A., Ha, T., Parker, J.E., Gassmann, W., and Schroeder, J.I. 2012. Natural variation in small molecule-induced TIR-NB-LRR signaling induces root growth arrest via EDS1- and PAD4-complexed R protein VICTR in Arabidopsis. Plant Cell 24:5177-5192.

Latif, Z., Strange, R.N., Bilton, J., and Riazuddin, S. 1993. Production of the phytotoxins, solanapyrones A and C and cytochalasin D among nine isolates of Ascochyta rabiei. Plant Pathology 42:172-180.

Lorang, J., Kidarsa, T., Bradford, C.S., Gilbert, B., Curtis, M., Tzeng, S.-C., Maier, C.S., and Wolpert, T.J. 2012. Tricking the guard: Exploiting plant defense for disease susceptibility. Science 338:659-662.

Lorang, J.M., Carkaci-Salli, N., and Wolpert, T.J. 2004. Identification and characterization of victorin sensitivity in Arabidopsis thaliana. Molecular Plant-Microbe Interactions 17:577-582.

Möbius, N., and Hertweck, C. 2009. Fungal phytotoxins as mediators of virulence. Current Opinion in Plant Biology 12:390-398.

Masuda, D., Ishida, M., Yamaguchi, K., Yamaguchi, I., Kimura, M., and Nishiuchi, T. 2007. Phytotoxic effects of trichothecenes on the growth and morphology of Arabidopsis thaliana. Journal of Experimental Botany 58:1617-1626.

Matern, U., Strobel, G., and Shepard, J. 1978. Reaction to phytotoxins in a potato population derived from mesophyll protoplasts. Proceedings of the National Academy of Sciences of the United States of America 75:4935-4939.

Nene, Y.L. 1982. A review of Ascochyta blight of chickpea. Tropical Pest Management 28:61-70.

Nishiuchi, T., Masuda, D., Nakashita, H., Ichimura, K., Shinozaki, K., Yoshida, S., Kimura, M., Yamaguchi, I., and Yamaguchi, K. 2006. Fusarium phytotoxin trichothecenes have an elicitor-like activity in Arabidopsis thaliana, but the activity differed significantly among their

molecular species. Molecular Plant-Microbe Interactions 19:512-520.

O'Reilly, E., Corbett, M., Hussain, S., Kelly, P.P., Richardson, D., Flitsch, S.L., and Turner, N.J. 2013. Substrate promiscuity of cytochrome P450 RhF. Catalysis Science & Technology 3:1490-1492.

Oikawa, H., Kobayashi, T., Katayama, K., Suzuki, Y., and Ichihara, A. 1998a. Total synthesis of (-)-solanapyrone A via enzymatic Diels-Alder reaction of prosolanapyrone. Journal of Organic Chemistry 63:8748-8756.

Oikawa, H., Yokota, T., Sakano, C., Suzuki, Y., Naya, A., and Ichihara, A. 1998b. Solanapyrones, phytotoxins produced by Alternaria solani: Biosynthesis and isolation of minor components. Bioscience, Biotechnology, and Biochemistry 62:2016-2022.

Payne, G.A., Nystrom, G.J., Bhatnagar, D., Cleveland, T.E., and Woloshuk, C.P. 1993. Cloning of the afl-2 gene involved in aflatoxin biosynthesis from Aspergillus flavus. Applied and Environmental Microbiology 59:156-162.

Peever, T.L., Chen, W., Abdo, Z., and Kaiser, W.J. 2012. Genetics of virulence in Ascochyta rabiei. Plant Pathology 61:754-760.

Peplow, A.W., Tag, A.G., Garifullina, G.F., and Beremand, M.N. 2003. Identification of new genes positively regulated by tri10 and a regulatory network for trichothecene mycotoxin production. Applied and Environmental Microbiology 69:2731-2736.

Rouxel, T., Grandaubert, J., Hane, J.K. et al. 2011. Effector diversification within compartments of the Leptosphaeria maculans genome affected by Repeat-Induced Point mutations. Nature Communications 2:202.

Roy, S., Singh, S.K., Choudhury, S.R., and Sengupta, D.N. 2009. An insight into the biological functions of family X-DNA polymerase in DNA replication and repair of plant genome. Plant Signaling & Behavior 4:678-681.

Santra, D.K., Tekeoglu, M., Ratnaparkhe, M., Kaiser, W.J., and Muehlbauer, F.J. 2000. Identification and mapping of QTLs conferring resistance to Ascochyta blight in chickpea. Crop Science 40:1606-1612.

Schmidt, L.E., Gloer, J.B., and Wicklow, D.T. 2007. Solanapyrone analogues from a Hawaiian fungicolous fungus. Journal of Natural Products 70:1317-1320.

Schmittgen, T.D., and Livak, K.J. 2008. Analyzing real-time PCR data by the comparative CT method. Nature Protocols 3:1101-1108.

Siewers, V., Viaud, M., Jimenez-Teja, D., Collado, I.G., Gronover, C.S., Pradier, J.-M., Tudzynsk, B., and Tudzynski, P. 2005. Functional analysis of the cytochrome P450 monooxygenase gene bcbot1 of Botrytis cinerea indicates that botrydial is a strain-specific

virulence factor. Molecular Plant-Microbe Interactions 18:602-612.

Singh, K.B., Reddy, M.V., and Nene, Y.L. 1984. International testing of chickpeas for resistance to Ascochyta blight. Plant Disease 68:782-784.

Slot, J.C., and Rokas, A. 2011. Horizontal transfer of a large and highly toxic secondary metabolic gene cluster between fungi. Current Biology 21:134-139.

Stone, J.M., Heard, J.E., Asai, T., and Ausubel, F.M. 2000. Simulation of fungal-mediated cell death by fumonisin B1 and selection of fumonisin B1-resistant (fbr) Arabidopsis mutants. Plant Cell 12:1811-1822.

Tsuge, T., Harimoto, Y., Akimitsu, K., Ohtani, K., Kodama, M., Akagi, Y., Egusa, M., Yamamoto, M., and Otani, H. 2013. Host-selective toxins produced by the plant pathogenic fungus Alternaria alternata. FEMS Microbiology Reviews 37:44-66.

Turgeon, B.G., and Lu, S.-W. 2000. Evolution of host specific virulence in Cochliobolus heterostrophus. Pages 93-126 in Fungal Pathology. (J. W. Kronstad ed.) The Netherlands, Springer.

Vail, S., and Banniza, S. 2008. Structure and pathogenic variability in Ascochyta rabiei populations on chickpea in the Canadian prairies. Plant Pathology 57:665-673.

Walton, J.D. 1996. Host-selective toxins: Agents of compatibility. Plant Cell 8:1723-1733.

Waring, P., Eichner, R.D., Müllbacher, A., and Sjaarda, A. 1988. Gliotoxin induces apoptosis in macrophages unrelated to its antiphagocytic properties. Journal of Biological Chemistry 263:18493-18499.

Wilson, A.D., and Kaiser, W.J. 1995. Cytology and genetics of sexual incompatibility in Didymella rabiei. Mycologia 87:795-804.

Wu, D., Oide, S., Zhang, N., Choi, M.Y., and Turgeon, B.G. 2012. ChLae1 and ChVel1 regulate

T-toxin production, virulence, oxidative stress response, and development of the maize pathogen Cochliobolus heterostrophus. PLoS Pathogens. 8:e1002542.

Yamtich, J., and Sweasy, J.B. 2010. DNA polymerase family X: Function, structure, and cellular roles. Biochimica et Biophysica Acta - Proteins and Proteomics 1804:1136-1150.

Yoder, O.C. 1973. A selective toxin produced by Phyllosticta maydis. Phytopathology 63:1361-1366.

Yu, J.-H., Hamari, Z., Han, K.-H., Seo, J.-A., Reyes-Domínguez, Y., and Scazzocchio, C. 2004. Double-joint PCR: A PCR-based molecular tool for gene manipulations in filamentous fungi. Fungal Genetics and Biology 41:973-981.

CHAPTER THREE

GENOMIC ANALYSIS OF ASCOCHYTA RABIEI IDENTIFIES DYNAMIC GENOME

ENVIRONMENTS OF SOLANAPYRONE BIOSYNTHESIS GENE CLUSTER

AND A NOVEL TYPE OF PATHWAY-SPECIFIC RAGULATOR

ABSTRACT

Secondary metabolite genes are often clustered together and situated in particular genomic regions like subtelomere that can facilitate niche adaptation in fungi. Solanapyrones are toxic secondary metabolites produced by fungi occupying different ecological niches. Full genome sequencing of the ascomycete, Ascochyta rabiei, revealed a solanapyrone biosynthesis gene cluster embedded in an AT-rich isochore-like region proximal to a telomere end and surrounded by Tc1/Mariner-type transposable elements. The highly AT-rich environment of the solanapyrone cluster is likely the product of repeat-induced point mutations. Several secondary metabolism-related genes are found in the flanking regions of the solanapyrone cluster. Although the solanapyrone cluster appears to be resistant to repeat-induced point mutations, a P450 monooxygenase gene adjacent to the cluster has been degraded by such mutations. Among the solanapyrone cluster genes, sol4 encodes a novel type of Zn(II)2Cys6 zinc cluster transcription factor. Deletion of sol4 caused a total loss of solanapyrone production, but did not compromise growth and development. Gene expression studies with the sol4-deletion mutant delimited the

boundaries of the solanapyrone gene cluster and revealed that sol4 is a specific regulator for solanapyrone biosynthesis and is necessary and sufficient for induction of the solanapyrone cluster genes. These results showed that the solanapyrone gene cluster was housed within a dynamic genome environment and likely horizontally transferred. Despite the high degeneracy of the surrounding genomic regions, the gene cluster appears to have maintained its integrity, suggesting important roles of solanapyrones in fungal biology.

INTRODUCTION

Subtelomere, the region upstream of telomere on a linear chromosome, is characterized by a richness in repetitive sequences and frequent chromosomal rearrangement, lacking synteny among closely-related species and even within the same species in fungi (Barry et al., 2003;

Galagan et al., 2005; Fairhead and Dujon, 2006; Fedorova et al., 2008). Subtelomeric regions are also enriched with duplicate, transolcated and horizontally transferred genes and thought to be evolutionarily labile and important for niche adaptation (Berriman et al., 2002; Fairhead and

Dujon, 2006; Coelho et al., 2013). These plastic chromosomal regions often include

‘contingency genes’, such as genes involved in membrane transport (Fairhead and Dujon, 2006;

Coelho et al., 2013), nutrient assimilation (Naumov et al., 1995; Fairhead and Dujon, 2006), anaerobiosis (Rachidi et al., 2000), and virulence (Hernandez-Rivas et al., 1997; Barry et al.,

2003; Verstrepen et al., 2005). Secondary metabolite gene clusters also exhibit a subtelomeric bias (Maiya et al., 2007; Perrin et al., 2007; McDonagh et al., 2008; Palmer and Keller, 2011).

Secondary metabolites (SMs) in fungi play a diverse role in host and niche specialization, serving as virulence factors, antibiotics and metal ion-chelating agents (Michael and Pattenden,

1993; Turgeon and Bushley, 2010). Solanapyrones are polyketide-derived SMs that were originally found in the plant pathogenic fungi, Ascochyta rabiei and Alternaria solani (Ichihara et al., 1983; Alam et al., 1989), as well as in ascomycetous fungi occupying different ecological niches (Schmidt et al., 2007; Trisuwan et al., 2009; Wu et al., 2009b). The fact that solanapyrones are produced by distantly related species like Nigrospora (Sordariomycetes), but not by closely related Ascochyta or Alternaria (Dothideomycetes) suggested the acquisition of the entire gene cluster by horizontal gene transfer (HGT) events. Recently the solanapyrone biosynthesis gene cluster was identified from Alternaria solani (Kasahara et al., 2010). However, lack of the genomic information on solanapyrone gene cluster in other fungi hinders further examination of the HGT hypothesis. A. rabiei is a host-specific parasite with a narrow host range, infecting domesticated chickpea (Cicer arietinum L.) and their wild relatives (Chen et al., 2004;

Frenkel et al., 2010). Among solanapyrones, solanapyrone A was proven to be the most toxic to chickpea, and it has been claimed to be a key virulence factor in A. rabiei (Kaur, 1995; Hamid and Strange, 2000). However a recent functional study indicated that the phytotoxic SMs are not required for causing disease in the host plants of A. rabiei and Alternaria solani (Kim et al.,

2015).

Genomes of filamentous fungi are larger and more complex than prokaryotic genomes because of the presence of AT-rich isochores containing large noncoding regions and

degenerated genes (i.e., pseudogenes). These AT-rich isochores are also rich in transposable elements (TEs) and repetitive DNA sequences, which scatter along the genome alternating with large GC-equilibrated regions in which most of the predicted genes reside (Eyre-Walker and

Hurst, 2001). Only a few TEs are currently active. However, TEs are thought to have greatly impacted on gene function and genome structure during genome evolution (Kazazian, 2004;

Ayarpadikannan and Kim, 2014). TEs can be broadly categorized into two major groups, retrotransposons and DNA transposons (Wicker et al., 2007). Although the majority of TEs in genomes of the fungal class Dothideomycetes are retrotransposon types, the Tc1/Mariner superfamily of DNA transposons are frequently found in certain members of the fungal class

(Ohm et al., 2012; Manning et al., 2013; Santana et al., 2014). The Tc1/Mariner–type TEs are highly variable in size, but generally delimited by terminal inverted repeats and encode a single transposase (Marini et al., 2010). A large number of Tc1/Mariner-type TEs have been degenerated by extensive mutations and a few functional copies can be found across the genome of Cochliobolus heterostrophus (Santana et al., 2014).

In most ascomycetous fungi, TEs are targets for repeat-induced point mutation (RIP) that involves transitions of dinucleotide CpA to TpA (and TpG to TpA for the complementary strand) in pairs of duplicated DNA sequences during the dikaryotic phase between meiosis (Selker et al.,

1987; Margolin et al., 1998). RIP is considered a genome surveillance mechanism that mutates repetitive sequences to limit the accumulation of TEs (Galagan and Selker, 2004). Recent comparative genomic studies suggested that RIP also plays a significant role in the evolution of

pathogenicity-related effector genes that populate repeat-rich regions (Rep and Kistler, 2010;

Rouxel et al., 2011). The higher mutation rate associated with the RIP process could introduce new functional alleles with premature stop codons or non-synonymous substitutions into the genes adjacent to repetitive sequences.

The boundaries of SM gene clusters can be defined through identifying their respective pathway-specific regulators that coordinate the biosynthesis of SMs (Keller and Hohn, 1997).

Solanapyrone gene clusters in A. rabiei and Alternaria solani are known to contain a transcription factor similar to fungal-specific Zn(II)2Cys6 zinc cluster transcription factor (C6 zinc cluster TF) family (Kasahara et al., 2010; Kim et al., 2015). The C6 zinc cluster TF family represents one of the largest classes of transcriptional regulators in fungi, as exemplified by Gal4 from Saccharomyces cerevisiae (Todd and Andrianopoulos, 1997). C6 zinc cluster TFs have been characterized as regulators for primary and secondary metabolism and many different cellular processes (MacPherson et al., 2006; Näär and Thakur, 2009). This TF family contains a highly conserved N-terminal DNA-binding domain consisting of six cysteines in the pattern

CX2CX6CX6CX2CX6C which complex two Zn (II) ions (Pan and Coleman, 1990). Immediately following the DNA-binding domain is the coiled-coil region, a very short stretch of amino acids

(15 residues in the case of Gal4), which provides an interface for homo- or hetero-dimerization

(Schjerling and Holmberg, 1996; Näär and Thakur, 2009). Many fungal specific C6 zinc cluster

TFs have been reported to dimerize through the coiled-coil region and bind to the inverted repeat sequence like CGG triplets (Reece and Ptashne, 1993; MacPherson et al., 2006). Another

conserved feature of C6 zinc cluster TFs is the presence of weak homology region located in the middle of the protein, termed the middle homology region (MHR) (Schjerling and Holmberg,

1996). The exact role of MHR is not well understood, however, it is thought to regulate the transcriptional activity of the C6 zinc cluster TFs upon cognate ligand binding (Näär and Thakur,

2009).

Here we describe the solanapyrone gene cluster and its genomic environment in A. rabiei and functionally characterized a novel type C6 zinc cluster transcription factor for the cluster genes. We will discuss horizontal transfer of the solanapyrone gene cluster, potential impacts of

RIP on the genomic environment, and regulation of solanapyrone biosynthesis.

MATERIALS AND METHODS

Fungal Strains and Whole Genome Sequencing

A. rabiei strains AR628 (ATCC 201622) and AR21 (ATCC 76502) were obtained from the worldwide A. rabiei collection maintained at the USDA Western Regional Plant Introduction

Station. Genomic DNA (10 µg) of AR628 strain was sequenced at the Washington State

University Laboratory for Biotechnology and Bioanalysis on a PacBio RS platform (Pacific

Biosciences, Menlo Park, CA). The raw data from 15 SMRT cells were assembled using the

HGAP protocol of the SMRT Analysis v2.0.0 software (Pacific Biosciences).

Sliding-window Analysis of RIP Indices and GC Content

To determine how the nucleotide compositions of the 100 kb region from the telomeric repeats change as a function of the position relative to the telomere, GC content and RIP indices were calculated in a “window” of 200 nucleotides. The window was slid in 50 bp increments

(from centromere to telomere direction) and values were recalculated for each position. This was reiterated until the right-hand edge of the window met the 3′ end of the contig. GC content was measured as the percentage of G or C nucleotides in each window. RIP index I was calculated as

ApT/TpA, and RIP index II was CpA + TpG/ApC + GpT (Margolin et al., 1998). These operations were performed automatically using a perl script (RIPindex.pl, kindly provided by Dr.

Mark Farman, University of Kentucky). To analyze GC content and RIP indices for transposable elements and ORF10, the values were calculated in a “window” of 200 nucleotides with 20 bp increments in the 5′ to 3′ direction.

Generation of Gene Replacement Constructs and sol4-deletion Mutants

The sol4 gene was deleted using the split marker method (Catlett et al., 2003). DNA fragments of the gene replacement constructs overlapping within the hygromycin B phosphotransferase (hph) gene were amplified using a double-joint PCR method (Yu et al., 2004) with a minor modification. In the first round PCR from genomic DNA of the AR628 isolate, an

838 bp upstream of the sol4 coding sequence was amplified with primer 4-1/primer 4-2 pair, and a 1,094 bp downstream of the sol4 coding sequence was amplified with primer 4-3/primer 4-4

pair. A 1,372 bp hph cassette was amplified from pDWJ5 using HYG-F/HYG-R primer pair. The primer 4-2 and primer 4-3 carried 27-bp sequence tails that overlapped with the 5′ and 3′ ends of the hph cassette, respectively. In the second round PCR, each sol4 flanking DNA fragment was fused to the hph cassette through PCR by overlap extension. The resulting PCR products were cloned into the pGEM-T Easy vector (Promega, USA). DNA sequences of the split-marker constructs were verified by sequencing with T7 or SP6 primer. The vectors cloned with the upstream flanking DNA construct and the downstream flanking DNA construct were used as templates for the third round PCR with nested primer pairs, primer 4-7/HY-R pair for the upstream flanking DNA construct and primer 4-8/YG-F pair for the downstream flanking DNA construct. The upstream and downstream split-marker constructs were used for transformation of isolates AR628 and AR21.

Fungal Transformation and Confirmation of Transformants.

Preparation of fungal protoplasts and transformation was conducted as described in

Chapter 2. Briefly, one to 5 μg of each upstream and downstream split-marker DNA construct or

10 μg of the sol4-overexpression plasmid construct were used for genetic transformation of A. rabiei. Putative transformants were subcultured on V8 agar containing hygromycin B (200

μg/mL), and two rounds of single spore isolation were conducted to obtain homokaryotic transformants. Selected transformants were screened by a diagnostic PCR with the primer

4-11/primer 4-12 pair to amplify the full length DNA sequence of sol4 gene in wild-types and hph cassette in putative Δsol4 mutants (Fig. 3-1). Homologous integration of the split-marker

DNA constructs into the sol4 gene locus was verified by PCR with an upstream flanking primer, primer 4-1, and HYG-R.

Solanapyrone Extraction and Detection

For extraction of solanapyrones from liquid cultures of the WT strain (AR628 and AR21) and their Δsol4 mutants, thirty milliliters of Czapek-dox media supplemented with cations (Chen and Strange, 1991) in 250-mL Erlenmeyer flasks were inoculated with 100 μL of conidial suspension of the strains (1 × 107 spores mL-1), and incubated at 20°C without shaking. After 18 days, cultures were filtered through four layers of Miracloth (Calbiochem, USA) in a vacuum to remove mycelium. Solanapyrones were extracted using Sep-Pak® Vac 6 cc (1g) tC18 cartridges

(Waters Corp., Milford, MA, USA). Individual cartridges were eluted with 2 mL of MeOH and the eluents were subjected to LC-MS analysis. Chromatographic separation was achieved using an ACQUITY UPLC system as described in Chapter 2. The presence or absence of peaks corresponding to solanapyrones in the culture extracts of the WT and Δsol4 mutants was confirmed by comparing retention times, and mass and UV spectra of solanapyrone A standard

(kindly provided by Dr. Hideaki Oikawa, Hokkaido University) and purified solanapyrone C

(Kim et al., 2015) in separate LC runs.

Growth Measurement

Agar plugs (3 mm diameter) containing actively growing mycelia on PDA were transferred to fresh PDA and the cultures were incubated under 12 h fluorescent light at 20 ± 2°C.

100

Colony diameters in orthogonal directions in five replicate plates (13.5 cm diameter) for each strain were measured in 3-day intervals for up to 36 days.

Quantitative Real-time RT-PCR Analyses

RNA was extracted from PDA cultures of AR628 WT strain, Δsol5 and Δsol4 mutants at

9 days after growth, using the RNeasy Plant Mini Kit (Qiagen) and adopting the method described in (Schumann et al., 2013). On-column digestion of possible genomic DNA contamination was performed using the RNase-Free DNase Set (Qiagen). First-strand cDNA synthesis was prepared from total RNA using the iScript cDNA Synthesis Kit (Bio-Rad) according to manufacturer’s instructions. Real-time RT-PCR conditions are as described in

Chapter 2. The quantification of the relative transcript levels was performed with the comparative CT method normalization (Schmittgen and Livak, 2008). The transcript levels of target genes were normalized against the Actin1 gene transcript levels and the relative gene expressions levels were log2-transformed. The means of the three biological replicates and standard deviations of the relative expression values were presented. Primers used in this study and their sequences are listed in Appendix A.

Generation of sol4-overexpression Strains and RT-PCR Analysis

For sol4-overexpression strain generation, full length DNA sequences of sol4 (1,741 bp) and its downstream UTR region (276 bp) was amplified from genomic DNA of AR628 isolate with 5′ and 3′ primers tailed with BamHI and HindIII restriction sites, respectively. The resulting

101

PCR product was partially digested with BamHI restriction enzyme as one BamHI site is present in the coding sequence, and subsequently digested by HindIII enzyme. The larger DNA fragmenet (2,029 bp) was eluted and then ligated to vector pHNU3PelA (Yang et al., 1994) precut with the same restriction enzymes. The vector contains the pelA promoter from

Aspergillus nidulans that can be induced by polygalacturonic acid (PGA) and repressed by the presence of glucose in the media (Yang et al., 1994). The vector carrying the sol4 gene was transformed into the AR628 strain as described previously (Akamatsu et al., 2010). PCR analysis was conducted as described earlier (Wu et al., 2012), to verify integration of the plasmid by single crossover at the native sol4 genomic locus (Fig. S3-1). The resulting transformants carried two copies of the sol4 gene, one driven by the inducible pelA promoter, and the other by the native sol4 promoter. Two sol4-overexpression strains (4OE2-2 and 4OE6-2) were tested to see if sol4 overexpression activates other sol genes under PGA induction. AR628 or sol4-overexpression strains were first grown in half-strength PDB for 2 days by shaking at 120 rpm. Mycelia were harvested, washed twice with distilled water and transferred to CM broth

(Leach et al., 1982) containing 1% PGA. The cultures were further incubated for 4 days, and total RNA was extracted from mycelia after washing with distilled water 2 times. RT-PCR conditions are as described in Chapter 2.

Synteny Block Analysis

Synteny blocks between the draft genomes of A. rabiei, Didymella exigua, and

Leptosphaeria muculans were computed by SyMAP (Soderlund et al., 2011). The repeat-masked

102

draft genome sequence of Leptosphaeria muculans (Rouxel et al., 2011) and Didymella exigua were obtained from JGI website (http://www.jgi.doe.gov/). Genomic sequences were first aligned using promer/MUMmer (Kurtz et al., 2004). Raw anchors resulting from MUMmer were clustered into gene anchors. Minimum number of anchors for synteny block detection was set to seven as a default value. The A. rabiei contig (835 kb) harboring solanapyrone gene cluster was pairwise compared with L. muculans supercontig 1 and D. exigua scaffold 6.

RESULTS

Solanapyrone Biosynthesis Gene Cluster is Situated in a Subtelomeric Region

The genome of AR628 strain of A. rabiei was sequenced to 72× coverage with an N50 of

1.1 Mb, using the PacBio RS platform. The genome assembly had a total size of 40.89 Mb, scaffolded into 144 contigs. The solanapyrone biosynthesis gene cluster was found in a contig with the size of 835 kb. The contig had a tandem array of telomeric repeats (TTAGGG)9 at the 3´ end and also included a telomere-linked helicase (TLH) gene within ≈ 15 kb from the repeats

(Fig. 3-2; Table 3-1). The TLH gene appears to be mutated by point mutations and indels, showing a partial sequence match with a full-length TLH gene. In addition, the GC-equilibrated region within ≈ 10 kb from the repeats was filled with a mosaic of gene relics, most of which showed sequence similarity to TLH genes in Pyrenophora teres f. teres (not shown).

The solanapyrone gene cluster was situated ≈ 45 kb from the telomere end (Fig. 3-2). The gene cluster is comprised of six genes (sol1–sol6) with the same order and orientation of the

103

solanapyrone gene cluster found in Alternaria solani (Kasahara et al., 2010), showing a remarkably high level of amino acid sequence identity between the homologous genes (96–98%;

Table 3-1). However, unlike the solanapyrone gene cluster in Alternaria solani, the gene cluster in A. rabiei is interrupted by an AT-rich isochore-like region (≈ 13 kb) between sol2 and sol3 genes, separating the gene cluster into halves (Fig. 3-2).

We identified other putative SM-related genes in the flanking regions of the solanapyrone gene cluster. Three open reading frames (ORF1, ORF2, and ORF3) found in the flanking region distal to the telomere showed homology to membrane transporter, non-ribosomal peptide synthetase, and dioxygenase, respectively (Fig. 3-2; Table 3-1). Also found were some pseudogenes with no apparent ORF, however, they showed significant blast hits with known protein families (> 77% identity at DNA sequence level), such as ubiquitin carboxyl-terminal hydrolase and aminotransferase. Within a region (≈ 45 kb) between the gene cluster and the telomere end, only a single ORF (ORF10) can be found, which encodes a P450 monooxygenase.

ORF10 includes only a partial P450 catalytic domain, and thus it is likely non-functional.

RIP Has Affected TEs and Adjacent Genes, but Not Solanapyrone Gene Cluster

In the subtelomeric region (100 kb, GenBank accession: KR139658), transposable elements (TEs) are scattered along AT-rich regions (Fig. 3-2). The five TEs found in the genomic locus are all DNA transposons with characteristics of the Tc1/Mariner superfamily; four are with high DNA sequence similarity to Molly transposon from Stagonospora nodorum,

104

one is similar to the transposase-like gene found in the AM-toxin gene cluster from Alternaria alternata (Table 3-1).

In fungi (especially filamentous ascomycetes), the presence of AT-rich regions is often indicative of the action of RIP, a defense mechanism that likely acts to protect the genome from invasive repeated sequences such as viruses and TEs (Cambareri et al., 1989). To survey whether the AT-rich regions were caused by RIP, we conducted a sliding-window analysis to calculate two RIP indices (I and II) across the 100-kb genomic locus. Regions with a RIP index I >1 and a

RIP index II <1 are likely to have undergone RIP (Margolin et al., 1998; Wu et al., 2009a). As indicated by both indices, there is a strong sign of RIP in the AT-rich regions within the solanapyrone gene cluster and between the cluster and telomere (Fig. 3-2). However, the regions in which the solanapyrone cluster genes are located appear to be unaffected by RIP.

The TEs located close to the solanapyrone cluster genes exhibited a high degree of structural conservation with the Molly transposon found in S. nodorum, having nearly identical length of terminal inverted repeat (TIR) and transposase domain in the elements (Fig. S3-2).

However, the overall GC content of the proximal TEs (39–41%) was lower than that of the Molly transposon (52%). RIP analyses indicated that the low GC content of the proximal TEs has arisen from RIP, suggesting that the TEs are degenerated copies of Molly transposons (Fig. 3-3 A).

Concomitantly, several premature stop codons in the coding regions of a putative transposase were identified in the degenerated TEs, which may have been introduced directly by RIP (Fig.

105

S3-2). It is likely that the AT-rich regions found in the subtelomeric region have resulted from

RIP that mutated these TEs and other repeated DNA sequences.

SM-related genes near TEs are often affected by RIP and the degree of RIP on genes proximal to a repeat sequence was inversely proportional to distance from the repeated sequence

(Ohm et al., 2012). ORF10 encoding a partial P450 enzyme was embedded in the middle of

AT-rich regions proximal to one of the degenerated TEs (Fig. 3-2), raising the possibility that the upstream and downstream sequences of ORF10 that perhaps encoded an intact functional P450 domain has been affected by RIP. The deduced amino acid sequence of ORF10 was best matched with a full-length P450 enzyme from Coniosporium apollinis that belongs to the class

Eurotiomycetes, sharing 75% identity (Fig. S3-3). The DNA sequence similarity between upstream and downstream regions of ORF10 and corresponding regions of the full-length P450 was lower but still showed 52 and 54% identity at DNA level, respectively, indicating that the flanking regions have been diverged to a greater extent (Fig. S3-3). This led us to question whether the sequence divergence might be the consequences of the RIP processes. Although not prominent in RIP responses, there was indication of potential RIP occurrence in upstream and downstream regions of ORF10 especially when compared to the functional P450 homolog (Fig.

3-3 B). The result suggests that the abolished N- and C-termini of the otherwise functional P450 domain of ORF10 is affected by RIP to some extent.

106

sol4 Gene Encodes a Novel Type of Zn(II)2Cys6 Transcription Factor

The genomic environment around the solanapyrone gene cluster revealed that additional

SM-related genes such as non-ribosomal peptide synthetase (ORF2) and dioxygenase (ORF3) were neighbored closely by the solanapyrone gene cluster. It is uncertain that these genes are involved in solanapyrone biosynthesis or coordinately regulated with other solanapyrone cluster genes. In the 100-kb subtelomeric region, we could not identify other transcription factor

(TF)-encoding genes than the sol4 gene that was previously predicted to encode a putative TF

(Kasahara et al., 2010). However, the deduced amino acid sequence of the sol4 gene lacked a

DNA-binding domain, only containing the middle homology region (MHR) commonly found in typical C6 zinc cluster TFs like yeast Gal4 (Fig. 3-4). The two best blast hits using the protein sequence as query were with hypothetical proteins in Podospora anserina and Aspergillus niger, sharing 41 and 40% sequence identity, respectively. Unlike the Sol4 protein, those two hypothetical proteins contained the conserved DNA-binding and MHR, the hallmarks of the C6 zinc cluster TF family (Fig. S3-4). sol4 Gene is Required for Solanapyrone Production, but Not for Fungal Development

To examine if the sol4 gene encodes a bona fide TF and specifically regulates the solanapyrone cluster genes, we generated sol4-deletion (Δsol4) mutants from two A. rabiei strains (AR628 and AR21), using the split-marker gene replacement strategy (Catlett et al., 2003) illustrated in Figure 3-1. LC-MS analysis of 18-d old static culture of AR628 strain and its Δsol4

107

mutant confirmed that the mutant culture was devoid of solanapyrones that were detected in the wild-type (WT) culture (solanapyrone A, eluted at 4.19 min; solanapyrone C, eluted at 4.03 min)

(Fig. 3-5 A). Another Δsol4 (AR21 background) mutant also completely lost the ability to produce solanapyrones (not shown). Growth of the Δsol4 mutant in the culture was similar to that of the WT strain, as indicated by dry weight of mycelial mats (Fig. 3-5 B).

The growth arrest of the WT strains was conspicuous at later stages of incubation on PDA media, with a restricted colony morphology, presumably due to toxicity of solanapyrones (Fig.

3-6 A). In contrast, Δsol4 mutants were able to grow further at a constant rate until the colony margin reached the edge of the agar plates. The growth of Δsol4 mutants appears to be unlimited as their growth rates were constant at least up to 36 days after incubation (Fig. 3-6 C). Close examination of hyphal tip growth of the WT strains and Δsol4 mutants showed that Δsol4 mutants were actively growing as indicated by highly branched hyphae at 3 weeks after incubation, while the WT strains seemed to attenuate or cease the growth showing aggregate hyphae at the colony margin (Fig. 3-6 B, right panels). No defects were observed in sporulation and conidial morphology of Δsol4 mutants, although the spore mass oozing out from pycnidia differed in color (Fig. 3-6 B, left panels; Fig. 3-7). sol4 Gene is a Pathway-specific Regulator for Solanapyrone Biosynthesis

To investigate the effect of sol4 gene in the control of solanapyrone cluster genes, transcript levels of the sol genes (sol1–sol6) were compared between the AR628 wild-type (WT)

108

strain and its Δsol4 mutant. Additionally, the Δsol5 mutant that lacks the final step enzyme for solanapyrone biosynthesis was included in the gene expression study because it was reported that the sol4 gene was autonomously overexpressed in the mutant, possibly due to a positive feedback loop (i.e. accumulation of intermediates triggers the entire metabolic pathway) (Kim et al., 2015).

Real-time RT-PCR analysis showed that expression of all of the sol genes decreased in the Δsol4 mutant (Fig. 3-8 A). In particular, the expression level of the sol1 gene encoding a polyketide synthase that initiates the solanapyrone biosynthetic pathway was dramatically reduced in the

Δsol4 mutant, showing a ≈ 700-fold decrease when compared to the WT strain. In contrast, the expression levels of sol1, sol2, sol4, and sol6 were significantly elevated in the Δsol5 mutant, presumably due to the inherently elevated level of sol4 gene expression (Fig. 3-8 A).

To check if the sol4 gene affects expressions of genes outside of the cluster and to define boundaries of the solanapyrone gene cluster, transcript levels of the flanking ORFs (ORF2,

ORF3, and ORF10) and two known polyketide synthase genes (PKS1 and PKS2) were compared between the WT strain and the Δsol4 and Δsol5 mutants. The genes outside of the cluster remained unaffected by deletion of the sol4 or sol5 gene (Fig. 3-8 A). Student’s t-test showed that there was no difference in the transcript levels between the strains. It is currently unknown whether these genes are functional, except for the PKS1 gene which is involved in melanin biosynthesis (Akamatsu et al., 2010). Nevertheless, significant amounts of the transcripts of

SM-related genes were detected from all tested strains. We also attempted to compare, if transcribed, transcript levels of one of the degenerated TEs (pseudogene 6) next to ORF10. No

109

transcripts could be amplified from all the tested strains, further supporting the fact that this element has been pseudogenized. sol4 Gene is Solely Sufficient for Induction of Solanapyrone Biosynthetic Genes

The upregulation of the cluster genes in the Δsol5 mutant may result directly from the elevated expression of the sol4 gene. To verify this, we generated sol4-overexpression (4OE) strains carrying an additional copy of the sol4 gene under the control of the pelA promoter from the WT strain. The second copy of the sol4 gene can be inducible when the 4OE strains are grown on polygalacturonic acid (PGA)-containing media, which induces the pelA promoter.

After induction with PGA, expression levels of the cluster genes (sol1–sol6) were compared between the WT strain and 4OE strains. RT-PCR analysis showed that the sol4 gene was highly overexpressed in the 4OE strains and the expression levels of all of the cluster genes except for the sol3 gene were increased in the 4OE strains compared to the WT strain (Fig. 3-8 B). A similar trend was observed in the Δsol5 mutant, where the expression level of the sol3 gene was not significantly different from the WT strain (Fig. 3-8 A), suggesting requirement of another factor for sol3 gene induction.

DISCUSSION

We found that the solanapyrone gene cluster is housed within 80 kb from a telomere end in A. rabiei (Fig. 3-2). In fungi, conditionally dispensable chromosomes often accommodate genes involved in the production of SMs, host-selective toxins, and effectors that have a role in

110

host range delineation and specialization (Hatta et al., 2002; Ma et al., 2010; Mehrabi et al.,

2011; Balesdent et al., 2013). Synteny analyses of the A. rabiei draft genome data with genomes of fungal species within the same fungal order Pleosporales, available on the JGI website

(Grigoriev et al., 2014), showed that a large portion of the contig harboring the solanapyrone gene cluster was syntenic (mesosyntenic sensu stricto) to the largest scaffold of related species

(3.38-Mb scaffold of Leptosphaeria maculans and 4.03-Mb scaffold of Setosphaeria turcica), indicating that the solanapyrone gene cluster resides in one of the core chromosomes of A. rabiei.

Accordingly, despite a high degree of karyotype variations observed between A. rabiei strains

(Akamatsu et al., 2012), all strains examined so far are known to produce solanapyrones (Höhl et al., 1991; Kaur, 1995; Kim et al., 2015).

The synteny block analysis of the contig harboring the solanapyrone gene cluster with syntenic scaffolds of the two most related Dothideomycetes Didymella exigua and Leptosphaeria maculans genomes available in the JGI website showed that, although GC-equilibrated regions of the contig shares many synteny blocks with regions of the scaffolds of the two related species, the subtelomeric region (100-kb from telomeric repeats) harboring the solanapyrone gene cluster was unique to A. rabiei (Fig. 3-9). We also attempted to compare genomic environments of the solanapyrone gene cluster between A. rabiei and Alternaria solani. Unfortunately, we could obtain only a short contig (20 kb) encompassing the sol1 and sol2 genes and ≈ 10-kb flanking region of sol1 gene from the genome sequence data of Alternaria solani. The gene profile in the

10-kb flanking region of the Alternaria solani sol1 gene was totally different from that of A.

111

rabiei, containing genes involved in nutrient assimilation and antibiotic resistance (Fig. 3-10).

Unlike A. rabiei, the genes found in the flaking region of Alternaria solani sol1 gene are highly similar to related Dothideomycetes such as Cochliobolus miyabeanus and Setosphaeria turcica.

Although a species-specific genomic region can only be properly assessed through a more extensive survey of closely related species like legume-infecting Ascochyta species, the subtelomeric region is likely a species-specific region at the end of a core chromosome in A. rabiei.

Solanapyrone Gene Cluster Resides in a Species-specific “Genomic Island”

Horizontal transfer event of a gene (or an entire gene cluster) can be supported by circumstantial evidence such as patchy phyletic distribution of the gene within closely related species, phylogenetic incongruence of the gene with the known species phylogeny, and variations in codon usage or GC content (Fitzpatrick, 2012). Ascochyta fungi infecting different legumes form a monophyletic group (Barve et al., 2003; Peever et al., 2007; Chilvers et al.,

2009). However, sol genes were not found in the draft genome data of closely related Ascochyta species infecting different legumes (personal communication with Dr. J. Lichtenzveig). Also, the exceptionally high sequence similarity of the cluster genes between A. rabiei and Alternaria solani suggests its acquisition by a HGT event between the two species, or from a third organism independently.

112

The solanapyrone gene cluster in A. rabiei is situated in a subtelomeric region. AT-rich regions near telomere with repetitive DNA sequence are prone to undergo ectopic recombination at a much higher rate than expected for homologous recombination (Freitas-Junior et al., 2000;

Dean et al., 2005). Moreover, several copies of degenerated Tc1/Mariner-type TEs were found around the gene cluster (Fig. 3-2). Tc1/Mariner-type TEs are widely found in diverse taxa as fungi, rotifers, insects and mammals (Robertson, 1993; Plasterk et al., 1999; Arkhipova and

Meselson, 2005), and known to be horizontally transferred between distant taxa in mammals or insects (Robertson et al., 1998; Lampe et al., 2003). The finding that the solanapyrone gene cluster in A. rabiei is placed in such a dynamic genomic environment further supports the likelihood of HGT of the gene cluster to A. rabiei. It is also notable that, although

Dothideomycetes is one of the most extensively sequenced fungal taxa (Grigoriev et al., 2014), the pseudogenes and ORFs found within the subtelomeric region of A. rabiei were more similar to genes found in species that belong to other fungal classes (Table 3-1), suggesting that the

100-kb subtelomeric region is a reservoir of foreign DNA materials that may contribute to niche adaptation.

RIP May Play a Role in Diversification of Secondary Metabolites

RIP normally operates on repeated sequences but it can leak to regions flanking the repeated sequence (Ohm et al., 2012). There is increasing evidence that RIP leakage to adjacent

DNA sequences and coding regions contributes to the diversification of pathogenicity-related effector genes (Rep and Kistler, 2010; Van de Wouw et al., 2010).

113

ORF10 encoding a truncated P450 domain is located proximal to a degenerated TE. Some putative homologs encoding full-length P450 enzymes were found from species in classes

Eurotiomycetes and Sordariomycetes but not in Dothideomycetes, implying either its foreign origin or fast divergence. Given the taxonomic distance, the degree of sequence identity (75% at amino acid level) was considerably high between ORF10 and the putative homolog P450 gene from Coniosporium apollinis. The upstream and downstream of ORF10 was mildly affected by

RIP that may have been originated from surrounding AT-rich isochore-like regions containing repeated sequences and TEs. However, despite the discovery of extended homology and evidence for RIP, there were some indels and point mutations not relevant to RIP in the regions upstream and downstream of ORF10, which did not align very well with the homologous P450 gene (Fig. S3-3), suggesting that RIP is not the sole source of mutation acting on the genomic loci.

Solanapyrone analogs with more complex chemical structures than solanapyrones produced by A. rabiei and Alternaria solani have been identified from distant fungal species

(Schmidt et al., 2007; Trisuwan et al., 2009; Wu et al., 2009b). Considering the structural diversity of solanapyrone analogs, there may be additional genes involved in their production in other solanapyrone-producing fungi. It is currently unknown whether ORF10 was part of the solanapyrone gene cluster. However, it is plausible that RIP-mediated degradation of biosynthetic genes for a metabolite could be one of the mechanisms to achieve such a diverse chemical structure of SMs sharing the common backbone, analogous to the effector

114

diversification processes in plant pathogenic fungi during the genome evolution. A future direction of the study is to identify solanapyrone gene clusters and neighboring genes from other solanapyone-producing fungi in order to compare gene contents, their degeneracy and metabolic outputs.

Solanapyrone Gene Cluster is Controlled by a Novel Type of Transcriptional Factor

We identified the sol4 gene as a pathway-specific regulator for solanapyrone biosynthesis.

This unusual type of TF containing only MHR specifically regulates the solanapyrone cluster genes. Although several SM-related genes were found adjacent to the solanapyrone gene cluster, deletion of sol4 did not affect their gene expression (Fig. 3-8 A). The sharp demarcation of transcriptional control by sol4 also helped us to delimit the borders of solanapyrone gene cluster in A. rabiei.

As far as we are aware, there are no reports on regulatory roles of this type of TF in biological processes. Notably, in the gene cluster of cercosporin (a polyketide produced by

Cercospora spp.), three putative TF-encoding genes were found. One encodes a conventional C6 zinc cluster TF (CTB8) that is known to coordinately regulate genes in that cluster, but the other two are truncated versions of the C6 zinc cluster TF with unknown function, lacking the conserved DNA-binding domain (Chen et al., 2007). We could not identify any other

TF-encoding genes around the solanapyrone gene cluster. However, given the fact that the gene expression was not completely shut down by sol4-deletion and that the Sol4 protein contains a

115

coiled coil region and MHR through which it interacts with other proteins and regulates transcriptional activity, it is conceivable that the sol4 gene activates the gene cluster, in conjunction with other TFs.

C6 zinc cluster TFs are known to bind to an inverted repeat sequence like

[CGG(N6-11)CCG] with some variations in the triplet sequence (Todd and Andrianopoulos, 1997;

MacPherson et al., 2006). Using the MEME motif discovery tool (Bailey et al., 2009), consensus inverted repeat sequences such as [CCG(N8)CGG] or [GCC(N10)GGC] were found upstream of the coding regions of sol genes (Table 3-2), implying that there is a C6 zinc cluster TF for the regulation of solanapyrone biosynthesis. C6 zinc cluster TFs usually repress the target genes under non-inducing conditions, as is well illustrated in galactose catabolism by Gal4 in yeast

(Traven et al., 2006). Therefore, Sol4 may have a role in derepression of the cluster genes by interacting with an as-yet-unknown C6 zinc cluster TF.

It is notable that pre-mRNA species of sol1 sol2 and sol5 genes accumulated in the WT strain, while the degrees of the pre-mRNA accumulation were reduced in 4OE strains (Fig. 3-8

B). The elevated expression of sol genes in 4OE strains is likely attributable to more efficient processing of pre-mRNA to mRNA. In the same context, considerable amounts of pre-mRNA species of sol genes were still transcribed in Δsol4 mutants (Fig. 3-11). These results suggest that the Sol4 protein may facilitate mRNA processing directly or possibly via interaction with splicing factors, recruiting them to the transcriptional machinery. The identification of specific

116

interacting partners of the Sol4 protein will shed light on a novel regulatory mechanism for SM biosynthesis.

Solanapyrones Play an Important Role in the Ecology of A. rabiei

Solanapyrone has been claimed to be a virulence factor as it exhibits strong phytotoxicity

(Kaur, 1995; Hamid and Strange, 2000). However, Δsol4 mutants generated from two WT strains (AR628 and AR21) differing in pathotype showed a virulence level similar to their respective WT progenitor on susceptible chickpea cultivars (Fig. 3-12), supporting the previous finding that solanapyrones are not required for virulence (Kim et al., 2015). Why then does the fungus produce the auxiliary metabolite which may impose a metabolic cost to the organism?

The universal production of solanapyrones among A. rabiei strains suggested that it may play an important role in the pathogen biology other than parasitism. In fact, solanapyrones A and C were recently shown to be selectively antibiotic against gram positive bacteria such as Bacillus subtilis, a common soil bacterium (Wang et al., 2014). We also observed that A. rabiei strains exhibited a restricted colony growth but Δsol4 mutants were able to grow almost indeterminately on agar plates (Fig. 3-6). This restricted growth may be a consequence of antibiotic activity of solanapyrones that accumulated in the media.

The genomic environment of the solanapyrone gene cluster is marked by abundance in

TEs and SM-related genes, many of which underwent RIP-mediated degradation and subsequent pseudogenization. Nevertheless, the gene cluster has remained intact in such a mutation hotspot

117

region. Alternatively, it may be that, if the HGT hypothesis is valid, the gene cluster was recently transferred to the AT-rich region. It is difficult to conclude which is the case, however, either case implies the importance in maintaining the capacity to produce solanapyrone for the pathogen biology; otherwise the gene cluster would have been pseudogenized or, if ho rizontally transferred, it would not have been fixed in the existing population of A. rabiei. Therefore, solanapyrones may confer a fitness advantage to the organism by acting as antibiotic agents.

ACKNOWLEDGMENTS

I am grateful to Drs. Turgeon Gillian and Dongliang Wu (Cornell University, USA) for kindly providing pHNU3PelA and Dr. Mark Farman (University of Kentucky, USA) for sharing a perl script used in the RIP indices analyses. The genome sequence data of Didymella exigua

(CBS 183.55) were produced by the US Department of Energy Joint Genome Institute

(http://www.jgi.doe.gov/) in collaboration with the user community.

118

Table 3-1. The solanapyrone biosynthesis gene cluster and flanking open reading frames and pseudogenes in Ascochyta rabiei Genea Relative positionb Length (bp) Predicted function Closest blast match (accession number) I.Dc E-value

P1 6,363 – 7,942 1,580 transposase TLS1 in AM-toxin region (AB525200) 68 5e-92

P2 9,046 – 9,783 738 ubiquitin C-terminal hydrolase ubiquitin hydrolase (XM_001259501) 78 6e-166

O1 12,384 – 10,712 1,673 major fascilitator super family hypothetical protein (XP_008030938) 71 0

P3 13,063 – 14,396 1,334 aminotransferase hypothetical protein (XM_007837320) 77 0

O2 15,169 – 19,287 4,121 non-ribosomal peptide synthetase hypothetical protein (XP_007835510) 69 0

O3 20,805 – 21,410 606 Fe(II)-dioxygenase hypothetical protein (XP_008030935) 84 0

O4 22,905 – 30,959 8,055 polyketide synthase sol1, polyketide synthase (BAJ09789) 97 0

O5 32,264 – 33,753 1,490 O-methyltransferase sol2, O-methyltransferase (BAJ09788) 98 0

P4 34,253 – 36,118 1,866 transposase S. nodorum transposon molly (AJ488502) 69 8e-157

119

P5 42,611 – 44,476 1,866 transposase S. nodorum transposon molly (AJ488502) 70 0

O6 47,917 – 47,009 909 dehydrogenase sol3, dehydrogenase (BAJ09787) 96 0

O7 50,140 – 48,400 1,741 transcription factor sol4, transcription factor (BAJ09786) 96 0

O8 52,989 – 51,240 1,750 Diels-Alderase sol5, oxidase/Diels-Alderase (BAJ09785) 98 0

O9 53,583 – 55,242 1,660 P450 monooxygenase sol6, P450 monooxygenase (BAJ09784) 97 0

O10 62,240 – 61,527 714 P450 monooxygenase hypothetical protein (XP_007784573) 75 4e-127

P6 63,499 – 65,364 1,867 transposase S. nodorum transposon molly (AJ488502) 68 2e-132

P7 82,019 – 80,154 1,866 transposase S. nodorum transposon molly (AJ488502) 68 4e-135

P8 85,301 – 88,831 3,531 helicase telomere-linked helicase (XM_001942430) 64 6e-164 a Gene designation is according to Figure 1. P and O indicate pseudogene and open reading frame, respectively. b Relative position

(5′ – 3′) of genes found within 100 kb region from the telomere end. Telomeric repeats (TTAGGG)9 were positioned at 99,948–100,001. c Percent amino acid sequence identity. For pseudogenes, percent sequence identity at DNA level was presented.

Table 3-2. The locations of putative binding motifs for a Zn(II)2Cys6 zinc cluster transcription factor in the promoter regions of biosynthetic genes for solanapyrone production.

Genea Binding site Ib Binding site IIc

d AR sol1 285 ,[GCA(N10)GGC] None

AR sol2 320, [ACC(N10)GGC] 321, [CCG(N8)CGG]

AR sol3 252, [GCC(N10)GGC] 253, [CCG(N8)CGG]

AR sol5 355, [GCC(N10)GGC] 356, [CCG(N8)CGG]

AR sol6 224, [GCC(N10)GGC] 225, [CCG(N8)CGG]

AS sol1 438, [GAC(N10)GGC] 439, [ACG(N8)CGG]

AS sol2 320, [ACC(N10)GGC] 321, [CCG(N8)CGG]

AS sol3 259, [GCC(N10)GGC] 260, [CCG(N8)CGG]

AS sol5 357, [GCC(N10)GGC] 358, [CCG(N8)CGG]

AS sol6 310, [GCC(N10)GGC] 311, [CCG(N8)CGG] aAR and AS prefixes indicate Ascochyta rabiei and Alternaria solani, respectively. b Perfect sites contain the 5′-GCC(N10)GGC-3′ consensus motif. Imperfect sites contain a single base pair mismatch in one of the two triplet half sites (GCC or GGC), highlighted in bold. c Perfect sites contain the 5′-CCG(N8)CGG-3′ consensus motif. Imperfect sites contain a single base pair mismatch in one of the two triplet half sites (CCG or CGG), highlighted in bold. dThe sites represent those found within 500 bp upstream of the ATG start codon. The numbers indicate relative position of the binding site from ATG start codon.

120

Figure 3-1. Gene replacement via split-marker strategy and PCR analysis of sol4-deletion. A,

Schematic diagram of targeted gene replacement strategy. B, PCR analysis of A. rabiei wild-type isolates and their Δsol4 mutants. Replacement of the sol4 gene by the hph cassette was verified in two independent mutants from each isolate using a primer pair, primer 4-11 and primer 4-12

(wild-types, 1,950 bp; Δsol4 mutants, 1,581 bp). C, PCR verification for homologous integration of the replacement fragment to the correct genomic site with a primer pair, primer 4-1 and HY-R

(1,604 bp).

121

Figure 3-2. Solanapyrone biosynthesis gene cluster and its genomic environment.

Sliding-window analysis of nucleotide composition was conducted with a chromosome end harboring solanapyrone biosynthesis cluster genes. GC content (%, black line), RIP index I

(TpA/ApT, blue line) and RIP index II (CpA+TpG/ApC+ GpT, red line) were calculated in a

200-bp window, which was slid in 50-bp increments across the terminal 100 kb from the telomere end. The x-axis shows the distance from the left edge of the window. At the rightmost region, the telomere repeats (TTAGGG)9 were found. Schematic diagram of arrangement and orientation of open reading frames (O1–O10, red arrows) and pseudogenes (P1–P8, blue arrows) were shown in the middle. The shaded area indicates the region harboring the solanapyrone cluster genes (sol1–sol6). Relative position and details of the gene content in the 100 kb region is listed in Table 1. Horizontal solid line are provided to show the 50% GC mark, while dashed line shows a RIP index value of 1.

122

123

Figure 3-3. Degeneracy of transposable elements and ORF10 located near the solanapyrone gene cluster. A, Sliding-window analyses of nucleotide composition were conducted for the full-length functional Molly transposon from Stagonospora nodorum (1,866 bp, GenBank accession: AJ488502) and pseudogenes (P4–P6) found in the subtelomeric region (P4 and P5:

1,866 bp, P6: 1,867 bp), including 5′ and 3′ TA insertion sites, terminal inverted repeats and putative transposase. Values for GC content (%), black line; RIP index I (TpA/ApT), blue line; and RIP index II (CpA + TpG / ApC + GpT), red line. These values were calculated across a

200-bp window, which was slid in 20-bp increments across the full-length elements, Horizontal solid line are provided to show the 50% GC mark, while dashed line shows a RIP index value of

1. B, Sliding-window analyses of nucleotide composition of ORF10 plus its 5′ and 3′ flanking regions (2029 bp, in total) and the homologous P450 (2020 bp, GenBank accession:

XP_007784573). Black box below each graph indicates the position of coding regions of the sequences. ORF10 and the corresponding homologous region in the P450 are shaded for comparison.

124

Figure 3-4. Structural organization of typical C6 zinc cluster transcription factor (Gal4) and Sol4 protein. C6 zinc cluster TFs consist of a highly conserved N-terminal zinc cluster DNA-binding domain consisting of six cysteines coordinating two zinc ions (Zn, purple), immediately followed by a coiled coil region (C, green) that involves protein-protein interactions, a large C-terminal domain dubbed middle homology region (MHR, blue) that regulates transcriptional activities, and followed by an acidic activation domain (Ac, red) in the distal C-terminal end. Note that

Sol4 protein lacks the characteristic N-terminal zinc cluster DNA-binding domain of C6 zinc cluster TF family but contains MHR, followed by a coiled-coil region.

125

Figure 3-5. Lack of solanapyrone production in Δsol4 mutant culture. A, Overlaid base peak ion chromatograms of equivalent quantities of culture extracts of AR628 wild-type (WT) strain

(black line) and the respective Δsol4 mutant (red line), indicated by observed m/z values with their respective retention times (in parenthesis) for each compound at the peak. Peaks corresponding to solanapyrone A and C (eluted at 4.19 and 4.03 min, respectively, in the WT culture extract) are not found in the Δsol4 mutant. B, Mean dry weight of the WT strain and

Δsol4 mutant. Mycelial mats were dried and weighed when the cultures were harvested and analyzed for solanapyrone production. Error bars are standard deviation (n = 28).

126

127

Figure 3-6. Indeterminate in vitro growth of Δsol4 mutants. A, Representative pictures of colony morphology of wild-type (WT) strains, AR628 and AR21, and their corresponding Δsol4 mutants at 2, 4, and 8 weeks after incubation on PDA. For 8 week-old culture, strains were grown from the left-hand edge of plates. B, Sporulation of the WT strains and Δsol4 mutants at 3 weeks after incubation on PDA (left panels). Spore mass of both strains oozing out from pycnidia was visible on the colony surface. Note different colors of spore mass oozing out from pycnidia of the WT strains (ochre) and Δsol4 mutants (pink). Growth arrest of the WT strains was evident from aggregate hyphae on the colony margin seen from the back sides of PDA plates

(right panels), while actively growing hyphae were observed in Δsol4 mutants. C, Growth curves of the WT strains and their corresponding Δsol4 mutants on PDA. Colony diameter was measured at 3-day intervals. Error bars are standard deviation (n = 5).

128

Figure 3-7. Conidial morphology of wild-type strains (AR628WT and AR21WT) and their corresponding Δsol4 mutants. No difference in size and morphology of conidia was observed between WT strains and Δsol4 mutants.

129

130

Figure 3-8. Pathway-specific control of sol4 gene for solanapyrone cluster genes. A, Expression levels of solanapyrone cluster genes (sol1–sol6), flanking open reading frames (ORF2, ORF3 and ORF10), and other secondary metabolites-related genes (PKS1 and PKS2) in AR628 wild-type (WT) strain, its Δsol5 and Δsol4 mutant. Total RNA was extracted from a 9-d old colony grown on potato dextrose agar. Quantitative RT-PCR was performed and the gene expressions were normalized by Actin1 expression levels in respective samples. The relative gene expressions levels were log2-transformed. Error bars are standard deviation of three biological replicates. Asterisks indicate that no transcripts were detected from the samples due to the specific deletion of respective genes. B, Expression levels of solanapyrone cluster genes were analyzed by RT-PCR in the WT strain and two independent sol4-overexpression strains (4OE2-2 and 4OE6-2). Total RNA was extracted 4 days after polygalacturonic acid induction. Primer sets used for the analysis were designed to amplify neighboring exons such that larger bands (when two bands are present) are amplicons of pre-mRNAs with an intron. Similar results were obtained from three independent experiments. M = 100 bp DNA ladder.

131

Figure 3-9. Synteny blocks shared between the A. rabiei contig harboring solanapyrone biosynthesis gene cluster and scaffolds of related Dothideomycetes. Didymella exigua scaffold 6

(upper) and Leptosphaeria maculans supercontig 1 (lower) was pairwise compared to the contig of A. rabiei, and only 835-kb regions showing synteny with the A. rabiei contig are presented.

For the A. rabiei contig, GC content is provided across the entire contig (835 kb) and the subtelomeric region (100 kb) harboring solanapyrone biosynthesis gene cluster is highlighted in grey box. Note that the L. maculans supercontig 1 showed mesosynteny with the A. rabiei contig, while the D. exigua is largely macrosyntenic to the A. rabiei contig owing to its close taxonomic relationship with A. rabiei (teleomorph Didymella rabiei).

132

Figure 3-10. Gene content of the flanking region of sol1 gene in Alternaria solani. A, GC content (%, black line), RIP index I (TpA/ApT, blue line) and RIP index II (CpA+TpG/ApC+

GpT, red line) were calculated in a 200-bp window, which was slid in 50-bp increments across ≈

10 kb of the flanking sequence. Horizontal solid line are provided to show the 50% GC mark, while dashed line shows a RIP index value of 1. B, Schematic diagram of arrangement and orientation of three open reading frames (red boxes) and one pseudogene (blue box). Genes found were pectin lyase (similar to XM_001796354, 67% identity, E-value = 4e-155), hexokinase (similar to XP_008029266, 94% identity, E-value = 0), hypothetical protein (similar to XP_007692661, 88% identity, E-value = 1e-119), and aminoglycoside 3′-phosphotransferase

(similar to XP_007692660, 85% identity, E-value = 0). Note that the partial coding region of pectin lyase gene showed strong RIP response.

133

Figure 3-11. Comparison of expression levels of solanapyrone cluster genes (sol1–sol6) between

AR628 wild-type (WT) and its Δsol4 mutant via RT-PCR analysis. Total RNA was extracted from 9-d old colony grown on PDA. M = 100 bp DNA ladder marker. Total genomic DNA

(gDNA) of the wild-type was used to compare PCR products from DNA. The size differences were due to presence of an intron in genomic DNA.

134

Figure 3-12. Comparison of virulence between Δsol4 mutants and their wild-type progenitors.

Mean disease severity of two chickpea cultivars Spanish White and Dwelley caused by wild-type strains AR628 (pathotype II), AR21 (pathotype I), and their corresponding Δsol4 mutants.

Disease score was assessed 2 weeks after inoculation, based on the 1–9 rating scale (1 = no symptom, 9 = dead plants), according to (Chen et al., 2004). Student’s t-test showed that the disease severity caused by the Δsol4 mutants was not different from that caused by their respective wild-type progenitors.

135

REFERENCES

Arkhipova, I.R., and Meselson, M. 2005. Diverse DNA transposons in rotifers of the class Bdelloidea. Proceedings of the National Academy of Sciences of the United States of America 102:11781-11786.

Ayarpadikannan, S., and Kim, H.-S. 2014. The impact of transposable elements in genome evolution and genetic instability and their implications in various diseases. Genomics & Informatics 12:98-104.

Bailey, T.L., Boden, M., Buske, F.A., Frith, M., Grant, C.E., Clementi, L., Ren, J., Li, W.W., and Noble, W.S. 2009. MEME Suite: Tools for motif discovery and searching. Nucleic Acids Research 37:W202-W208.

Balesdent, M.-H., Fudal, I., Ollivier, B., Bally, P., Grandaubert, J., Eber, F., Chèvre, A.-M., Leflon, M., and Rouxel, T. 2013. The dispensable chromosome of Leptosphaeria maculans shelters an effector gene conferring avirulence towards Brassica rapa. New Phytologist 198:887-898.

Barry, J.D., Ginger, M.L., Burton, P., and McCulloch, R. 2003. Why are parasite contingency genes often associated with telomeres? International Journal for Parasitology 33:29-45.

136

Berriman, M., Hall, N., Sheader, K., Bringaud, F., Tiwari, B., Isobe, T., Bowman, S., Corton, C., Clark, L., Cross, G.A.M., Hoek, M., Zanders, T., Berberof, M., Borst, P., and Rudenko, G. 2002. The architecture of variant surface glycoprotein gene expression sites in Trypanosoma brucei. Molecular and Biochemical Parasitology 122:131-140.

Catlett, N.L., Lee, B.N., Yoder, O.C., and Turgeon, B.G. 2003. Split-marker recombination for efficient targeted deletion of fungal genes. Fungal Genetics Newsletter 50:9-11.

Chen, H., Lee, M.-H., Daub, M.E., and Chung, K.-R. 2007. Molecular analysis of the cercosporin biosynthetic gene cluster in Cercospora nicotianae. Molecular Microbiology 64:755-770.

Chilvers, M.I., Rogers, J.D., Dugan, F.M., Stewart, J.E., Chen, W., and Peever, T.L. 2009. Didymella pisi sp. nov., the teleomorph of Ascochyta pisi. Mycological Research 113:391-400.

Coelho, M.A., Gonçalves, C., Sampaio, J.P., and Gonçalves, P. 2013. Extensive intra-kingdom horizontal gene transfer converging on a fungal fructose transporter gene. PLoS Genetics 9:e1003587.

Dean, R.A., Talbot, N.J., Ebbole, D.J., Farman, M.L., Mitchell, T.K., Orbach, M.J., Thon, M., Kulkarni, R., Xu, J.-R., Pan, H., Read, N.D., Lee, Y.-H., Carbone, I., Brown, D., Oh, Y.Y., Donofrio, N., Jeong, J.S., Soanes, D.M., Djonovic, S., Kolomiets, E., Rehmeyer, C., Li, W., Harding, M., Kim, S., Lebrun, M.-H., Bohnert, H., Coughlan, S., Butler, J., Calvo, S., Ma, L.-J., Nicol, R., Purcell, S., Nusbaum, C., Galagan, J.E., and Birren, B.W. 2005. The genome sequence of the rice blast fungus Magnaporthe grisea. Nature 434:980-986.

Eyre-Walker, A., and Hurst, L.D. 2001. The evolution of isochores. Nature Reviews Genetics 2:549-555.

Fairhead, C., and Dujon, B. 2006. Structure of Kluyveromyces lactis subtelomeres: Duplications and gene content. FEMS Yeast Research 6:428-441.

137

Fedorova, N.D., Khaldi, N., Joardar, V.S. et al. 2008. Genomic islands in the pathogenic filamentous fungus Aspergillus fumigatus. PLoS Genetics 4:e1000046.

Fitzpatrick, D.A. 2012. Horizontal gene transfer in fungi. FEMS Microbiology Letters 329:1-8.

Freitas-Junior, L.H., Bottius, E., Pirrit, L.A., Deitsch, K.W., Scheidig, C., Guinet, F., Nehrbass, U., Wellems, T.E., and Scherf, A. 2000. Frequent ectopic recombination of virulence factor genes in telomeric chromosome clusters of P. falciparum. Nature 407:1018-1022.

Frenkel, O., Peever, T.L., Chilvers, M.I., Ozkilinc, H., Can, C., Abbo, S., Shtienberg, D., and Sherman, A. 2010. Ecological genetic divergence of the fungal pathogen Didymella rabiei on sympatric wild and domesticated Cicer spp. (Chickpea). Applied and Environmental Microbiology 76:30-39.

Galagan, J.E., and Selker, E.U. 2004. RIP: The evolutionary cost of genome defense. Trends In Genetics 20:417-423.

Galagan, J.E., Calvo, S.E., Cuomo, C. et al. 2005. Sequencing of Aspergillus nidulans and comparative analysis with A. fumigatus and A. oryzae. Nature 438:1105-1115.

Hatta, R., Ito, K., Hosaki, Y., Tanaka, T., Tanaka, A., Yamamoto, M., Akimitsu, K., and Tsuge, T. 2002. A conditionally dispensable chromosome controls host-specific pathogenicity in the fungal plant pathogen Alternaria alternata. Genetics 161:59-70.

Hernandez-Rivas, R., Mattei, D., Sterkers, Y., Peterson, D.S., Wellems, T.E., and Scherf, A. 1997. Expressed var genes are found in Plasmodium falciparum subtelomeric regions. Molecular

138

and Cellular Biology 17:604-611.

Ichihara, A., Tazaki, H., and Sakamura, S. 1983. Solanapyrones A, B and C, phytotoxic metabolites from the fungus Alternaria solani. Tetrahedron Letters 24:5373-5376.

Kaur, S. 1995. Phytotoxicity of solanapyrones produced by the fungus Ascochyta rabiei and their possible role in blight of chickpea (Cicer arietinum). Plant Science 109:23-29.

Kazazian, H.H. 2004. Mobile elements: Drivers of genome evolution. Science 303:1626-1632.

Keller, N.P., and Hohn, T.M. 1997. Metabolic pathway gene clusters in filamentous fungi. Fungal Genetics and Biology 21:17-29.

Kim, W., Park, C.-M., Park, J.-J., Akamatsu, H.O., Peever, T.L., Xian, M., Gang, D.R., Vandemark, G., and Chen, W. 2015. Functional analyses of the Diels-Alderase gene sol5 of Ascochyta rabiei and Alternaria solani indicate that the solanapyrone phytotoxins are not required for pathogenicity. Molecular Plant-Microbe Interactions. 28:482-496.

Kurtz, S., Phillippy, A., Delcher, A.L., Smoot, M., Shumway, M., Antonescu, C., and Salzberg, S.L. 2004. Versatile and open software for comparing large genomes. Genome Biology 5:R12-R12.

Lampe, D.J., Witherspoon, D.J., Soto-Adames, F.N., and Robertson, H.M. 2003. Recent horizontal transfer of mellifera subfamily mariner transposons into insect lineages representing four different orders shows that selection acts only during horizontal transfer. Molecular Biology and Evolution 20:554-562.

Leach, J., Lang, B.R., and Yoder, O.C. 1982. Methods for selection of mutants and in vitro culture of Cochliobolus heterostrophus. Journal of General Microbiology 128:1719-1729.

Lupas, A., Van Dyke, M., and Stock, J. 1991. Predicting coiled coils from protein sequences. Science 252:1162-1164.

Ma, L., van der Does, H., Borkovich, K., Coleman, J., Daboussi, M., Di Pietro, A., Dufresne, M., Freitag, M., Grabherr, M., Henrissat, B., Houterman, P., Kang, S., Shim, W., Woloshulk, C.,

139

Xie, X., Hu, J., Antoniw, J., Baker, S., Bluhm, B., Breakspear, A., Brown, D., Butchko, R., Chapman, S., Coulson, R., Coutinho, P., Danchin, E., Diener, A., Gale, L., Gardiner, D., and Goff, S. 2010. Comparative genomics reveals mobile pathogenicity chromosomes in Fusarium oxysporum. Nature 464:367 - 373.

MacPherson, S., Larochelle, M., and Turcotte, B. 2006. A fungal family of transcriptional regulators: The zinc cluster proteins. Microbiology and Molecular Biology Reviews 70:583-604.

Manning, V., Pandelova, I., Dhillon, B., Wilhelm, L., Goodwim, S., Berlin, A., Figueroa, M., Freitag, M., Hane, J., Henrissat, B., Holman, W., Kodira, C., Martin, J., Oliver, R., Robbertse, B., Schackwitz, W., Schwartz, D., Spatafora, J., Turgeon, B., Yandava, C., Young, S., Zhou, S., Zeng, Q., Grigoriev, I., Ma, L., and Ciuffetti, L. 2013. Comparative genomics of a plant-pathogenic fungus, Pyrenophora tritici-repentis, reveals transduplication and the impact of repeat elements on pathogenicity and population divergence. Genes Genomes Genetics 3:41 - 63.

Margolin, B.S., Garrett-Engele, P.W., Stevens, J.N., Fritz, D.Y., Garrett-Engele, C., Metzenberg, R.L., and Selker, E.U. 1998. A methylated Neurospora 5S rRNA pseudogene contains a transposable element inactivated by repeat-induced point mutation. Genetics 149:1787-1797.

Marini, M., Zanforlin, T., Santos, P., Barros, R., Guerra, A., Puccia, R., Felipe, M., Brigido, M., Soares, C., Ruiz, J., Silveira, J., and Cisalpino, P. 2010. Identification and characterization of Tc1/mariner-like DNA transposons in genomes of the pathogenic fungi of the Paracoccidioides species complex. BMC Genomics 11:130.

McDonagh, A., Fedorova, N.D., Crabtree, J., Yu, Y., Kim, S., Chen, D., Loss, O., Cairns, T., Goldman, G., Armstrong-James, D., Haynes, K., Haas, H., Schrettl, M., May, G., Nierman, W.C., and Bignell, E. 2008. Sub-telomere directed gene expression during initiation of invasive aspergillosis. PLoS Pathogens 4:e1000154.

Mehrabi, R., Bahkali, A.H., Abd-Elsalam, K.A., Moslem, M., Ben, M., Barek, S., Gohari, A.M., Jashni, M.K., Stergiopoulos, I., Kema, G.H.J., and de Wit, P.J.G.M. 2011. Horizontal gene and chromosome transfer in plant pathogenic fungi affecting host range. FEMS Microbiology Reviews 35:542-554.

140

Michael, J.P., and Pattenden, G. 1993. Marine metabolites and metal ion chelation: The facts and the fantasies. Angewandte Chemie 105:1-23.

Näär, A.M., and Thakur, J.K. 2009. Nuclear receptor-like transcription factors in fungi. Genes & Development 23:419-432.

Naumov, G.I., Naumova, E.S., and Louis, E.J. 1995. Genetic mapping of the alpha-galactosidase MEL gene family on right and left telomeres of Saccharomyces cerevisiae. Yeast 11:481-483.

Ohm, R.A., Feau, N., Henrissat, B., Schoch, C.L., Horwitz, B.A., Barry, K.W., Condon, B.J., Copeland, A.C., Dhillon, B., Glaser, F., Hesse, C.N., Kosti, I., LaButti, K., Lindquist, E.A., Lucas, S., Salamov, A.A., Bradshaw, R.E., Ciuffetti, L., Hamelin, R.C., Kema, G.H.J., Lawrence, C., Scott, J.A., Spatafora, J.W., Turgeon, B.G., de Wit, P.J.G.M., Zhong, S., Goodwin, S.B., and Grigoriev, I.V. 2012. Diverse lifestyles and strategies of plant pathogenesis encoded in the genomes of eighteen Dothideomycetes fungi. PLoS Pathogens 8:e1003037.

Palmer, J.M., and Keller, N.P. 2011. Secondary metabolism in fungi: Does chromosomal location matter? Current Opinion in Microbiology 13:431-436.

Pan, T., and Coleman, J.E. 1990. GAL4 transcription factor is not a "zinc finger" but forms a Zn(II)2Cys6 binuclear cluster. Proceedings of the National Academy of Sciences of the United States of America 87:2077-2081.

Perrin, R.M., Fedorova, N.D., Bok, J.W., Cramer, R.A., Wortman, J.R., Kim, H.S., Nierman, W.C., and Keller, N.P. 2007. Transcriptional regulation of chemical diversity in Aspergillus fumigatus by LaeA. PLoS Pathogens 3:e50.

Plasterk, R.H.A., Izsvák, Z., and Ivics, Z. 1999. Resident aliens: The Tc1/mariner superfamily of transposable elements. Trends In Genetics 15:326-332.

Rachidi, N., Martinez, M.-J., Barre, P., and Blondin, B. 2000. Saccharomyces cerevisiae PAU genes are induced by anaerobiosis. Molecular Microbiology 35:1421-1430.

Reece, R., and Ptashne, M. 1993. Determinants of binding-site specificity among yeast C6 zinc cluster proteins. Science 261:909-911.

141

Rep, M., and Kistler, H.C. 2010. The genomic organization of plant pathogenicity in Fusarium species. Current Opinion in Plant Biology 13:420-426.

Robertson, H.M. 1993. The mariner transposable element is widespread in insects. Nature 362:241-245.

Robertson, H.M., Soto-Adames, F.N., Walden, K.K.O., Avancini, R.M.P., and Lampe, D.J. 1998. The mariner transposons of animals: Horizontally jumping genes. Pages 268-284 in Horizontal Gene Transfer (Syvanen, M. and Kado, C., eds.) London, Chapman and Hall.

Rouxel, T., Grandaubert, J., Hane, J.K. et al. 2011. Effector diversification within compartments of the Leptosphaeria maculans genome affected by repeat-induced point mutations. Nature Communications 2:202.

Santana, M., Silva, J., Mizubuti, E., Araujo, E., Condon, B., Turgeon, B., and Queiroz, M. 2014. Characterization and potential evolutionary impact of transposable elements in the genome of Cochliobolus heterostrophus. BMC Genomics 15:536.

Schjerling, P., and Holmberg, S. 1996. Comparative amino acid sequence analysis of the C6 zinc cluster family of transcriptional regulators. Nucleic Acids Research 24:4599-4607.

Schmidt, L.E., Gloer, J.B., and Wicklow, D.T. 2007. Solanapyrone analogues from a Hawaiian fungicolous fungus. Journal of Natural Products 70:1317-1320.

Schmittgen, T.D., and Livak, K.J. 2008. Analyzing real-time PCR data by the comparative CT method. Nature Protocols 3:1101-1108.

Schumann, U., Smith, N., and Wang, M.-B. 2013. A fast and efficient method for preparation of high-quality RNA from fungal mycelia. BMC Research Notes 6:71.

Selker, E.U., Cambareri, E.B., Jensen, B.C., and Haack, K.R. 1987. Rearrangement of duplicated DNA in specialized cells of Neurospora. Cell 51:741-752.

Soderlund, C., Bomhoff, M., and Nelson, W.M. 2011. SyMAP v3.4: A turnkey synteny system with application to plant genomes. Nucleic Acids Research 39:e68-e68.

Todd, R.B., and Andrianopoulos, A. 1997. Evolution of a fungal regulatory gene family: The Zn(II)2Cys6 binuclear cluster DNA binding motif. Fungal Genetics and Biology 21:388-405.

Traven, A., Jelicic, B., and Sopta, M. 2006. Yeast Gal4: A transcriptional paradigm revisited.

142

EMBO Reports 7:496-499.

Turgeon, B.G., and Bushley, K.E. 2010. Secondary metabolism. Pages 376-395 in Cellular and Molecular Biology of Filamentous Fungi (Borkovich, K.A. and Ebbole, D.J., eds.). Washington DC, The United States of America: ASM Press.

Van de Wouw, A.P., Cozijnsen, A.J., Hane, J.K., Brunner, P.C., McDonald, B.A., Oliver, R.P., and Howlett, B.J. 2010. Evolution of linked avirulence effectors in Leptosphaeria maculans is affected by genomic environment and exposure to resistance genes in host plants. PLoS Pathogens 6:e1001180.

Verstrepen, K.J., Jansen, A., Lewitter, F., and Fink, G.R. 2005. Intragenic tandem repeats generate functional variability. Nature Genetics 37:986-990.

Wicker, T., Sabot, F., Hua-Van, A., Bennetzen, J.L., Capy, P., Chalhoub, B., Flavell, A., Leroy, P., Morgante, M., Panaud, O., Paux, E., SanMiguel, P., and Schulman, A.H. 2007. A unified classification system for eukaryotic transposable elements. Nature Review Genetics 8:973-982.

Wu, C., Kim, Y.-S., Smith, K.M., Li, W., Hood, H.M., Staben, C., Selker, E.U., Sachs, M.S., and Farman, M.L. 2009a. Characterization of chromosome ends in the filamentous fungus Neurospora crassa. Genetics 181:1129-1145.

Wu, D., Oide, S., Zhang, N., Choi, M.Y., and Turgeon, B.G. 2012. ChLae1 and ChVel1 regulate T-toxin production, virulence, oxidative stress response, and development of the maize pathogen Cochliobolus heterostrophus. PLoS Pathogens 8:e1002542.

Wu, S.H., Chen, Y.W., Shao, S.C., Wang, L.D., Yu, Y., Li, Z.Y., Yang, L.Y., Li, S.L., and Huang, R. 2009b. Two new solanapyrone analogues from the endophytic fungus Nigrospora sp. YB-141 of Azadirachta indica. Chemistry & Biodiversity 6:79-85.

Yang, G., Turgeon, B.G., and Yoder, O.C. 1994. Toxin-deficient mutants from a toxin-sensitive

143

transformant of Cochliobolus heterostrophus. Genetics 137:751-757.

144

CHAPTER FOUR

GROWTH AND TISSUE-SPECIFIC PRODUCTION OF SOLANAPYRONE A

AND ITS ECOLOGICAL ROLE DURING SAPROBIC GROWTH OF

THE FUNGAL PATHOGEN, ASCOCHYTA RABIEI

ABSTRACT

Solanapyrone A is a phytotoxic secondary metabolite produced by two distinct plant pathogenic fungi, Ascochyta rabiei and Alternaria solani. We previously demonstrated that this toxin is not required for causing diseases by comparing virulence of solanapyrone-minus mutants to that of wild-type progenitors. Although the toxin is dispensable for causing diseases, the fact that all A. rabiei strains examined so far produce solanapyrone A prompted us to examine its possible ecological roles in the fungal biology. Solanapyrone A was specifically produced during the saprobic growth of A. rabiei. Concomitantly, the expressions of key biosynthetic genes were high during the growth on chickpea straws, but negligible in infected plant tissues. Interestingly, the expression of the gene encoding the last step enzyme for solanapyrone biosynthesis was induced during pycnidia formation. In agar plate confrontation assays with fungal saprobes that were commonly found in chickpea debris, the wild-type strains of A. rabiei effectively suppressed the growth of the competing fungi, while the solanapyrone-minus mutants did not.

With purified solanapyrone A, the half maximal inhibitory concentrations against the fungal

145

competitors were estimated, ranging from 53 to 82 μM. Taken together, the results suggest that solanapyrone A is a tissue-specific secondary metabolite produced in saprobic growth phase of A. rabiei, and plays an antagonistic role against a wide range of saprobes found in nature.

INTRODUCTION

Ascochyta rabiei (Pass.) Labrousse is a host-specific fungus infecting chickpea (Cicer arietinum L.) and its wild relatives (Hernandez-Bello et al., 2006; Frenkel et al., 2010). The pathogen can devastate entire chickpea fields and lead to total crop loss when environmental conditions are conducive to the disease (Nene and Reddy, 1987). For plant pathogenic fungi, studies on their biology and genetics have been focused largely on infection processes and interactions with their host organisms. Although survival of the facultative plant pathogens between cropping seasons in crucial for the disease recurrence, little is known about their saprobic growth phase after the cropping season and interactions with other microorganisms.

A. rabiei is a necrotrophic pathogen which derives nutrients from dead or dying plant tissues, and often kills host cells in advance of colonization (Höhl et al., 1990; Rea et al., 2002;

Chen et al., 2004). The fungus is known to produce a set of phytotoxic secondary metabolites

(SMs), solanapyrones A, B, and C (Alam et al., 1989). Solanapyrones were originally isolated from culture of Al. solani, the causal agent of early blight of potato (Ichihara et al., 1983).

Recently, the solanapyrone biosynthesis gene cluster was identified in Al. solani (Kasahara et al.,

2010). The cluster comprises six genes (sol1–sol6) encoding all the necessary biosynthetic

146

enzymes as well as a putative pathway-specific regulator (sol4 gene) for solanapyrone production. Among the cluster genes, the sol1 gene encodes a polyketide synthase that initiates the solanapyrone biosynthesis pathway. The sol5 gene encodes a unique oxidoreductase that catalyzes the last step of the biosynthetic pathway, involving in both oxidation and the subsequent cyclization of the precursor compound via Diels-Alder reaction (Oikawa et al., 1998;

Kasahara et al., 2010). Despite the amount of documented phytotoxicity of solanapyrones (Alam et al., 1989; Höhl et al., 1991; Kaur, 1995; Hamid and Strange, 2000), we have demonstrated that solanapyrones are not required for virulence either in A. rabiei or Al. solani (Kim et al., 2015).

Deletion of sol5 or sol4 gene led to no solanapyrone production, but the resulting mutants exhibited a similar degree of virulence to their respective wild-type progenitors (Chapters 2 and

3).

A. rabiei is an obligate outcrossing species, requiring two strains with opposite mating types to develop its sexual stage during the overwintering period (Wilson and Kaiser, 1995). A high degree of genotypic variation and presence of both mating types at similar frequencies within the A. rabiei populations are indicative of frequent mating events in the field (Kaiser and

Küsmenoglu, 1997; Navas-Cortés et al., 1998; Armstrong et al., 2001; Peever et al., 2004).

Although either sexual (ascospores) or asexual (conidia) spores can serve as initial inoculum, air-borne ascospores resulted from the sexual reproduction are thought to be important for the long-distance dissemination (Trapero-Casas et al., 1996; Kaiser, 1997; Milgroom and Peever,

2003), and to have a stronger capacity in germination and penetration to the host tissues than

147

conidia under harsh environmental conditions (Trapero-Casas and Kaiser, 2007). During the growing season, secondary infections occur by dispersion of conidia produced in the asexual fruiting bodies (pycnidia) to nearby tissues or neighboring plants (Trapero-Casas and Kaiser,

1992b, a). After the host decays the fungus in the infected chickpea debris undergo the overwintering period during which the fungus may encounter various microbial competitors including bacteria, fungi, and nematodes. A survival study showed that A. rabiei managed to survive in chickpea debris and remain viable for more than 2 years on soil surface at a low relative humidity (< 30%), however the survival rate was decreased at higher humidity levels

(Kaiser, 1973).

Dugan et al. (2005) examined diversity of fungal species residing in chickpea debris and their possible usage as biological control agents. Several saprobic fungi such as Alternaria tenuissima, Al. infectoria, Ulocladium consortiale, Epicoccum purpurascens, and U. atrum (in the order of abundance) were predominantly isolated from chickpea debris left in a field in 2003 and 2004 (Dugan et al., 2005). These fungal species are also found as common saprobes growing on different organic substrata (Runa et al., 2009). A. rabiei may need to compete with those fungi and other microorganisms in order to proliferate during the saprobic growth unlike the relatively constant host environment.

Fungal SMs have diverse roles in nature, such as antibiosis, development, metal chelation, pathogenicity, and communication (Turgeon and Bushley, 2010). Fungi survive and disseminate as mycelia or spores. In some instances, these fungal propagules are protected by SMs with

148

cytotoxic or antifeedant activities, as well as by melanized dense mycelial network like sclerotia and ascocarp walls (Gloer et al., 1988; Wang et al., 1995; Joshi et al., 1999; Degenkolb et al.,

2008). Also, some basidiocarp-forming fungi (i.e. mushrooms) are known to produce terpene-derived SMs to deter fungivorous animals from feeding on the fruiting bodies or to eliminate bacterial predators from the basidiocarps (Donnelly et al., 1982; Camazine, 1983).

These SMs may serve survival functions for the organisms producing them.

Solanapyrones are produced by all reliably identified A. rabiei strains in culture media

(Höhl et al., 1991; Kaur, 1995; Kim et al., 2015). The universal production of the toxin prompted us to examine possible ecological roles of solanapyrones. Here we describe growth and tissue-specific production of solanapyrones. In addition, we used the solanapyrone-minus mutants to elucidate their role in competitive interactions with other microorganisms that likely be encountered in nature.

MATERIALS AND METHODS

Fungal Strains and Growth Conditions

A. rabiei strains AR628 (ATCC 201622) and AR21 (ATCC 76502) were obtained from the A. rabiei collection maintained at the USDA Western Regional Plant Introduction Station.

AR628 strain expressing GFP under the control of a translation elongation factor promoter

(TEFprom::GFP) (Akamatsu et al., 2006; Kim et al., 2015) was used to monitor the parasitic and saprobic growth. Seeds (3 per pot) of chickpea cultivar Spanish White susceptible to the

149

Ascochyta blight disease were sown in one-litre plastic pot containing Sunshine mix #4 (Sun Gro

Horticulture, Canada) and maintained in a growth chamber (Conviron Model PGR15, Canada) for 2 weeks under a 12 h photoperiod with 20°C day and 18°C night at > 95% relative humidity.

For inoculum preparation, TEFprom::GFP strain was grown on V8 agar media (200 mL V8 juice,

5 3 g CaCO3, 20 g agar in 1 L distilled water) for 2 weeks. The spore suspension (2 × 10 spores mL-1) of the strain was spray-inoculated until run-off, and the parasitic growth on stem tissues was monitored by fluorescence microscopy between 1 and 9 dpi. Fungal infection was tracked by following the constitutively expressed GFP signal. Fluorescence microscopy was carried with the

Zeiss Axioskop 40 FL microscope equipped with the Zeiss AxioCam HRc camera. Filter sets 38 and 43 were used for GFP and dsRed detection, respectively. For monitoring the saprobic growth, senescencing chickpea plants were collected and air-dried, and dried chickpea stems (straws) were cut in pieces and autoclaved. Straws with similar dimensions (2 ~ 3 mm in diameter; ≈ 30 mm in length) were selected and soaked in the spore suspension (2 × 105 spores mL-1) for 30 min.

Inoculated straws were placed on petri dishes with three layers of wet filter papers in order to maintain high relative humidity. The plates were kept in a growth chamber (12/12 hr, light/dark cycle at 20oC).

Solanapyrone Extraction

To detect solanapyrones, chickpea stems or straws inoculated with TEFprom::GFP strain were collected at 3, 5, 7 and 9 dpi. Different parts of stems sprayed with distilled water were also sampled as a control treatment. At each time point, three stem cuts or three straws (≈ 3 cm in

150

length) showing a similar degree of the fungal growth were homogenized in 1 mL of ethyl acetate (EtOAc) in a 2 mL screw-cap tube, and placed in 4 °C for 16 h. The tubes were centrifuged at 2,200 × g for 1 min, and the supernatants were collected and evaporated to dryness on a vacufuge® (Eppendorf, Germany). Pellets were reconstituted with 200 μL of methanol, and the solutions were subjected to LC-MS/MS analysis. Chromatographic separation and mass spectrometric analysis were achieved using an ACQUITY UPLC system coupled with Synapt

G2-S (Waters Corp.), as described earlier (Kim et al., 2015). Isoflavonoid standard compounds

(biochanin A, formononetin, calycosin, pseudobaptigenin, maackiain and medicarpin) were purchased from ChromaDex® (Irvine, CA, USA). Solanapyrone A standard was kindly provided by Dr. Hideaki Oikawa. These chemicals were used for identification of the metabolites in the extracts by comparing the retention times and mass spectra of the standards with those of the detected peaks. Isoflavone glucoside conjugates found in the extracts were identified by determining the elemental compositions of the precursor ions, and by examining mass fragmentation patterns of the precursor ions with the fragments generated from the chemical structures of the proposed candidates, using the MassFragment software (Waters Corp.).

Quantitative Real-time RT-PCR Analyses

Total RNAs were extracted from chickpea stems or straws inoculated with A. rabiei strain

AR628 at 3, 5, 7, 9 and 11 dpi, using RNeasy Plant Mini Kit (Qiagen). At each time point, 15 to

20 stem cuts were harvested. For straws, samples were first frozen with liquid N 2, and the fungal mycelia growing on the surface were scraped out from the straws with a single-edge razor blade.

151

On-column digestion of possible genomic DNA contamination was performed using RNase-Free

DNase Set (Qiagen). Real-time RT-PCR analyses were performed to monitor expression levels of sol1, sol4, sol5, and AGT genes during the course of the parasitic and saprobic growth of A. rabiei. First-strand cDNA synthesis was prepared from total RNA using the iScript cDNA

Synthesis Kit (Bio-Rad) according to manufacturer’s instructions. Real-time RT-PCR conditions are as described in Chapter 2. The quantification of the relative transcript levels was performed with the comparative CT method normalization (Schmittgen and Livak, 2008). The transcript levels of target genes were normalized against the Actin1 gene transcript levels. The averages of the three biological replicates and standard deviations of the relative expression values were presented. Forward primer for the amplification of sol genes were designed to bind to exon-exon junction thereby specifically detecting mRNA species. Primers used in this study and their sequences are listed in Appendix A.

Effects of carbon source on solanapyrone production

A. rabiei strain AR628 was grown on yeast nitrogen base (YNB; BD Diagnostic Systems) agar media containing a single carbon source in plastic petri dishes (50 × 15 mm). Distilled water containing a single carbon source and bacto agar (BD Diagnostic Systems) were autoclaved after being buffered to pH 5.8 with 100 mM of 2-(N-morpholino)ethanesulfonic acid (MES), and 10×

YNB stock solution was added to the solution. The final concentration of a single carbon source and agar was 1% and 2%, respectively. The carbon substrates used were all purchased from

Sigma (St. Louis, MO, USA). Cellulose (catalog no.C6288), xylan from birchwood (X4252),

152

soluble starch (S9765), pectin from apple (P8471), D-cellobiose (C7252), D-maltose (M9171),

D-glucose (D9434), D-xylose (X1500), L-arabinose (A3256), D-galacturonic acid (48280) were used. The culture was incubated at 20°C under a 12/12 light regime. The colonies growing on each carbon source were considered ellipses, and their longer and shorter diameters were measured. The area of the colonies was calculated as area = Π × r1 × r2, where r1 and r2 are the long and short radii of the colony, respectively. Colony diameter and area were measured at 2 weeks after incubation, and solanapyrone A was extracted by soaking the whole agar with fungal colony taken out from the petri dish (5 replicates for each carbon source) into 8 mL of EtOAc in

50 mL falcon tube. After homogenization at 4°C for 16 h, 400 µL of solution was transferred to microcentrifuge tube, evaporated to dryness, reconstituted with 50 µL of MeOH, and subjected to HPLC quantitation.

Generation and Verification of the Reporter Strain Expressing GFP under the Native sol5

Gene Promoter

The plasmids pTRI5::GFP and pII99dsRed were kindly provided by Dr. Wilhelm Schäfer

(Ilgen et al., 2009). No restriction enzyme sites were available between the TRI5 promoter and the GFP gene in the pTRI5::GFP. Therefore, to replace the TRI5 promoter with a promoter region plus 132-bp N-terminus coding region of the sol5 gene, the GFP gene and the sol5 promoter region were amplified, put together in the pGEM T-easy vector (Promega), and cloned to the pTRI5::GFP plasmid precut with KpnI and XbaI restriction enzymes. Details on the multiple cloning steps for the replacement were illustrated in Figure S4-1. The resulting plasmid

153

was designated pSol5::GFP. The plasmid was linearized in the promoter sequence with Bsu36I and integrated into the genome of A. rabiei strain AR628 by homologous recombination (Fig.

S4-2), via protoplast-mediated transformation (Akamatsu et al., 2010). A verified transformant from the first transformation was transformed with plasmid pII99dsRed that carries the dsRed gene under the control of the glycerol-3-phosphate dehydrogenase promoter. To target this plasmid into the backbone of the already integrated pSol5::GFP, the pII99dsRed plasmid was digested with PvuI in the ampicillin resistance gene prior to transformation. The resulting reporter strain was named Sol5prom::GFP. The strategy to generate the reporter strain and verification of the homologous integration of the two plasmids into the reporter strain were presented in Figure S4-2.

Confrontation Assays

In vitro confrontation assays between A. rabiei strains (wild-type strains, AR628 and

AR21, and their respective Δsol5 and Δsol4 mutants) and the fungal competitors (Al. tenuissima,

E. purpurascens, B. cinerea, and U. atrum) were established as follows. Agar plugs excised from growing front of colonies of each fungus were placed 7 cm apart in parallel on PDA plates and incubated at room temperature under a 12/12 light regime. A. rabiei strains were allowed to grow in advance 3 days before the placement of Al. tenuissima, E. purpurascens, and U. atrum, and 7 days before the placement of B. cinerea. Photographs were taken after 3 weeks. The confrontation assays were performed in triplicate plates, and single cultures of A. rabiei strains were used as controls. For multiple-confrontation assays with S. sclerotiorum (WMA1 isolate)

154

(Xu and Chen, 2013), the WT strain and solanapyrone-minus mutants (Δsol5 and Δsol4) were placed on corners of the plate, and then S. sclerotiorum was placed on the center of the plate 10 days after the growth of A. rabiei strains. The plates were photographed 6 days and 8 days after S. sclerotiorum growth had begun.

Antifungal Assays

Solanapyrone A was purified from culture extract of AR628 isolate as described previously (Kim et al., 2015). Petri dishes (50 × 15 mm) containing 8 mL of half-strength PDA media and the indicated concentration of solanapyrone A (0, 25, 50, 100, or 200 μM) were inoculated at the center with agar plugs harboring actively growing hyphae of each fungal species (A. rabiei, Al. solani Al. tenuissima, E. purpurascens, B. cinerea, and U. atrum). At 5 days after incubation, colony diameters of fungal species at the different concentrations of solanapyrone A were measured for determining dose-response curves. For each treatment, colony diameter on three replicate plates was measured (two replicate plates for 200 μM of solanapyrone

A). To estimate the half maximal inhibitory concentrations (IC50) of solanapyrone A for each fungal species, the dose-response curve was fitted with R software and the "drc" package (Ritz and Streibig, 2005). Two non-linear regression models were used depending on the features of the dose-response curve. For A. rabiei, a hormesis parameter was estimated to obtain the IC50 of solanapyrone A as a weak hormetic effect, although statistically non-significant, was observed at the low doses, showing an inverted J-shape dose-response curve (Fig. S4-3). For the other fungal species, a log-logistic model was applied to estimate the IC50 of solanapyrone A.

155

RESULTS

Saprobic Growth-dependent Production of Solanapyrone A

A. rabiei undergoes two different growth phases during its life cycle: parasitic phase during host infection and saprobic phase on dead or decaying matters such as plant debris during overwintering period. To address whether the fungal development and solanapyrone production differ by the source of plant materials on which they grow, a GFP-labeled A. rabiei strain

(TEFprom::GFP) was inoculated to a susceptible chickpea plant (parasitic phase) and to autoclaved chickpea straw (saprobic phase). The growth of TEFprom::GFP was monitored under the fluorescence microscope as well as its solanapyrone production using a high resolution

LC-MS system.

In both growth phases, spores germinated as early as 1 day post inoculation (dpi), and the resulting mycelia produced pycnidia equally well on the living and dead plant materials (Fig.

4-1). No major difference was observed between the two growth phases in terms of growth rate and development. Therefore, we compared metabolic profiles of stem lesions and chickpea straws colonized by TEFprom::GFP strain. The main chemical components extracted from stem lesions of living plants at 5 dpi were isoflavonoids including known chickpea phytoalexins, medicarpin and maackiain, in addition to formononetin and biochanin A that were detected in uninoculated plants as constitutive isoflavonoids of chickpea (Fig. 4-2 A; Table 4-1). We could not detect solanapyrone A or its known derivatives at earlier or later stage of infection (3, 7 or 9

156

dpi; Fig. S4-4). Notably, a significant amount of solanapyrone A was detected from chickpea straws colonized by the fungus as early as 5 dpi (Fig. 4-2 A), approximately at initiation of pycnidia formation (Fig. 4-1). The amount of solanapyrone production gradually increased over time during the growth on chickpea straws (Fig. 4-2 B).

Suppression of Solanapyrone Cluster Genes during Plant Infection

To ascertain the saprobic phase-dependent production of solanapyrone A, we investigated gene expression levels of the key solanapyrone biosynthesis enzymes (sol1 and sol5) and the pathway-specific regulator (sol4) during both parasitic and saprobic growth of A. rabiei strain

AR628. In addition, acetylglutamate kinase (AGT) gene was included in the analysis to ensure the quality of fungal RNA extracted from infected plant tissues as the AGT gene was reportedly induced during the infection process (Singh et al., 2012). The gene expression levels were monitored over the course of plant infection processes and during the growth on chickpea straws for 11 dpi at 2-day intervals. The sol1, sol4 and sol5 genes were highly expressed during the saprobic growth on chickpea straws (Fig. 4-3). In contrast, only a trace amount of transcripts for the solanapyrone cluster genes were amplified from stem lesions, while the expression levels of

AGT gene did not show significant difference between the two growth phases (Fig. 4-3). The fungal genes were not amplified from uninoculated host tissues, confirming the specificity of primer sets used in the study.

157

The Effect of Carbon Sources on Solanapyrone Production

We hypothesized that carbon sources readily accessible during saprobic growth on chickpea straw, such as cellulose and hemicellulose, better support solanapyrone production. To address this, we used several mono-, di-, poly-saccharides as sole carbon sources for A. rabiei growth and solanapyrone production. In preliminary studies, the sporulation and solanapyrone production of A. rabiei were poor in static liquid cultures containing polysaccharide as a sole carbon source. Also, it was observed that, during the growth of A. rabiei, difference in pH changes in the agar media each containing different carbon source seems to override the effect of the carbon source for solanapyrone production (Fig. 4-4). Therefore, the agar media buffered to pH 5.8 was used in this study. Although there was a variation in growth rate, A. rabiei was able to develop pycnidia and produce spores on the pH-buffered agar media each containing different simple and complex sugars (Fig. 4-5; Fig. S4-5). The tested monosaccharides all supported high solanapyrone production (Fig. 4-5). Interestingly, solanapyrone production differed by disaccharides made up of two glucose molecules; higher on cellobiose with β (1→4) glycosidic bond than on maltose with α (1→4) bond (Fig. 4-5). For polysaccharides, solanapyrone production was significantly higher on xylan than on pectin or starch (Fig. 4-5). A. rabiei poorly grew on cellulose and, hence, did produce only a small amount of solanapyrone A (not shown).

158

Tissue-specific Localization of Solanapyrone

Knowing the spatial location and timing of SM production is crucial for understanding its biological or ecological roles. Thus, we generated a reporter strain (Sol5prom::GFP) expressing a green fluorescent protein (GFP) under the control of the native sol5 promoter. In addition, the reporter strain was further manipulated to express a red fluorescent protein (dsRed) driven by a constitutive promoter in order to visualize the fungal growth during microscopic examination.

Since the sol5 gene encodes the final step enzyme for solanapyrone biosynthesis, the reporter strain enabled us to pinpoint the solanapyrone production during the fungal growth and development in real time with cellular resolution. The expression levels of sol5 gene and the introduced GFP gene both driven by the endogenous sol5 gene promoter were highly correlated over the growth on PDA media (Pearson’s r = 0.9933; Table 4-2).

To localize solanapyrone production, GFP signal of the reporter strain was monitored throughout the fungal growth on PDA media. We could not observe GFP signal in actively growing mycelia of young colonies. It was not until the fungal colonies get mature and initiate conidiation that GFP signals can be observed from mycelia. Notably, GFP signals were enriched around young developing pycnidia (Fig. 4-6 A). Solanapyrone production appears to be stimulated in swollen and interwoven hyphae around mature pycnidia (Figs. 4-6 B and C).

Mycelia on the growing front of aged colonies did not show GFP signals. This specific GFP localization pattern indicated that solanapyrone production was associated with the conidiation

159

process. However, pycnidia and adjacent hyphae developing on chickpea stems and leaves were in large part free of GFP (Fig. 4-7).

It was reported that the N terminal region of Sol5 protein includes a signal peptide of 20 amino acids (a.a.) for secretion (Kasahara et al., 2010). The GFP gene in the Sol5prom::GFP strain was fused to the 3´ end of the first 132 nucleotides (44 a.a.) of the sol5 coding sequence including the start codon (see Materials and Methods). Thus, the translated GFP protein includes the signal peptide at its N terminus. Indeed, the GFP protein was secreted to outside of the fungal cells and accumulated on the media in the aged culture of Sol5prom::GFP strain (Fig. 4-8). In contrast, TEFprom::GFP strain that produces a GFP protein with no signal peptide showed constitutive GFP expressions confined to mycelia and pycnidia (Fig. 4-8).

Antagonistic Effect of Solanapyrone against Saprobic Fungal Species

Fast-growing saprobic fungi are likely the most challenging competitors for limited sources of nutrients during overwintering periods of A. rabiei. Previously, we identified that sol5-deletion (Δsol5) mutants led to the accumulation of precursor compounds (Kim et al., 2015) and sol4-deletion (Δsol4) mutants were devoid of solanapyrones and precursor compounds (ref).

Using these solanapyrone-minus mutants, we investigated a potential role of solanapyrones in competitive interactions with saprobic fungal species. Several fungal species frequently isolated from chickpea debris after harvest (Dugan et al., 2005) were selected for in vitro agar plate confrontation assay. Tested were Alternaria tenuissima, Epicoccum purpurascens, Botrytis

160

cinerea, and Ulocladium atrum. The saprobic fungi were grown in dual culture either with the

WT strain, Δsol5 or Δsol4 mutant on PDA media. Notably, the WT strain inhibited the growth of all the tested saprobic fungi (Fig. 4-9 A). The colonies of WT strains and the saprobic fungi never came in contact each other after prolonged incubation. In contrast, the saprobic fungi were able to grow toward the colonies of Δsol5 and Δsol4 mutants, contacting and often invading the colony boundaries of the solanapyrone-minus mutants (Fig. 4-9 A). A similar result was obtained with another WT strain (AR21) and its Δsol5, Δsol4 mutants in the competitive interactions with the same set of saprobic fungi (Fig. 4-10).

We also examined if solanapyrones inhibit not only saprobic fungi but plant pathogenic fungi. Sclerotinia sclerotiorum is a pathogen of various legume plants including chickpea, and produces sclerotia on dead or decaying plant matters as an overwintering structure (Chen et al.,

2011). S. sclerotiorum was co-cultured with A. rabiei WT strain, Δsol5 and Δsol4 mutants. A clear inhibition zone was visible between the colonies of the WT strain and S. sclerotiorum, whereas the colonies of the solanapyrone-minus mutants were soon invaded by S. sclerotiorum

(Fig. 4-9 B). Moreover, S. sclerotiorum produced sclerotia on the colonies of solanapyrone-minus mutants after further incubation (Fig. 4-9 B).

Antifungal Activities of Solanapyrone A

To precisely examine the antifungal activity of solanapyrone A, the saprobic fungi used in the confrontation assays were grown on PDA media amended with purified solanapyrone A.

161

At a concentration of 200 μM, solanapyrone A completely inhibited the growth of the saprobic fungi, although Al. tenuissima showed a marginal growth (Fig. 4-11). The half maximal inhibitory concentrations (IC50) of solanapyrone A against the fungal species ranged from 53 to

82 μM. Interestingly, solanapyrone A was even inhibitory to the producing strains, but A. rabiei and Al. solani appeared to be more tolerant than other fungal species. Solanapyrone A did not show any toxicity to A. rabiei at least up to 100 μM, while the growth of other tested fungi was considerably suppressed (Fig. 4-11). Also, Al. solani exhibited a higher IC50 value than other tested saprobic fungi.

DISCUSSION

Although humankind have exploited fungi-derived SMs to their own benefits in many ways (pharmaceutical, industrial or agricultural), information on their actual role in the biology and ecology of the producing organism is limited. Solanapyrone A was reported to selectively inhibit X-family DNA polymerases that exert their function exclusively on DNA repair process and thereby also considered to have a pharmaceutical potential for cancer therapy (Mizushina et al., 2002). Interestingly, solanapyrone A was found in two distinct fungal plant pathogens and proposed to be a virulence factor (Höhl et al., 1991; Kaur, 1995; Hamid and Strange, 2000).

However, we recently demonstrated that solanapyrones are not required for causing diseases on the host plants of A. rabiei and Al. solani (Kim et al., 2015). Nevertheless, the fact that all the tested A. rabiei strains isolated from different regions of the world produce solanapyrone A

162

(Höhl et al., 1991; Kaur, 1995; Kim et al., 2015) suggested that it has an indispensable role in the pathogen biology. In this study we provided evidence for ecological roles of solanapyrone A in antagonistic interactions with potential saprobic fungal competitors.

Filamentous fungi vary in their life strategies in relation to nutrient availability, to environmental factors, and in their relation to interacting microorganisms. Most ruderal

(R-selected) fungi (e.g. species in Chytridiomycota and Zygomycota) have evolved to rapidly colonize newly available resources and reproduce. This ecological group of fungi appears to be generally poor producers of SMs (Frisvad et al., 2008). In contrast, A. rabiei is primarily a plant parasite and seems to have a poor saprobic capacity, as the fungus requires more than one week to fully develop its asexual fruiting bodies on dead organic matters like autoclaved chickpea straws (Fig. 4-1). On the other hand, saprobic fungi naturally inhabiting chickpea debris, such as

Al. tenuissima and U. atrum, were able to grow and sporulate within 2–3 days on chickpea straws

(data not shown). Therefore, A. rabiei may need a potent antibiotic agent for competition and survival to secure its territory and nutrients therein till maturity from ruderal fungi and combative

(C-selected) fungi like the “fungi resident” in chickpea debris. The agar confrontation assays showed that A. rabiei WT strains can suppress the growth of all the tested saprobic fungal species isolated from chickpea debris, while the solanapyrone-minus mutants cannot (Figs. 4-9 and 4-10). We tested additional fungal species isolated from chickpea debris, such as

Cladosporium herbarum, and a species in Al. infectoria group (Dugan et al., 2005), as well as an unidentified fungal species that occurred as an air-borne contaminant during this study. None of

163

them could come in contact with the colony of the WT strain in the agar plate (Fig. 4-12), indicating a broad range of toxicity of solanapyrones. It is notable that ascochitine, another polyketide-derived SM produced by the closely related species, A. pisi and A. fabae (Bertini,

1956; Oku and Nakanishi, 1963), showed significant antifungal activities against several fungal species. However, some fungi were insensitive to the compound as they metabolized it to a reduced form (Oku and Nakanishi, 1964; Nakanishi and Oku, 1969).

Filamentous fungi follow a programmed developmental sequence from spore germination to vegetative hyphae, which can differentiate into either sexual or asexual fruiting bodies. In many instances, SMs appear only after conidiation process has commenced (Sekiguchi and

Gaucher, 1977; Bird et al., 1981; Reiss, 1982; Hicks et al., 1997; Calvo et al., 2002). Using the reporter strain (Sol5prom::GFP), we demonstrated that solanapyrones were produced in a tissue-specific manner. The GFP signals were largely confined to hyphae around pycnidia (Fig.

4-6). This is in line with the previous observations that solanapyrone production was associated with the onset of sporulation (Höhl et al., 1991; Kim et al., 2015). Also, the high accumulation of

GFP signals in the agar media indicated that the last step of solanapyrones biosynthetic pathway takes place outside of the fungal cell (Fig. 4-8). This compartmentalization of the key biosynthesis enzyme and the precursor substrate may allow for the sequestration of the toxic end product and thereby prevent self-toxicity, as with many cytotoxic SMs produced by fungi

(Chanda et al., 2009; Imazaki et al., 2009; Roze et al., 2011; Lim and Keller, 2014). Intriguingly, although solanapyrone production was specifically associated with the fruiting body formation,

164

solanapyrones were not produced when the fungus massively sporulates on the host tissues (Fig.

4-2 A), suggesting that solanapyrone production is not only linked to the conidiation process but also depends on other environmental factors, nutrient availability, or competing microorganisms.

Pycnidia are formed on necrotic lesions of infected chickpea tissues. After the plants are killed by the fungus, conidia may disperse and grow on nearby decaying plants or dead organic matters, during which the obligate outcrossing fungus may have better chance to find strains of opposite mating type and undergo the asexual and sexual stage development and overwintering

(Kaiser and Küsmenoglu, 1997; Barve et al., 2003). Chickpea straw has been used as a substrate for in vitro sexual stage induction of A. rabiei and is likely one of the major natural substrates for the saprobic growth (Wilson and Kaiser, 1995; Trapero-Casas et al., 2012). We showed that chickpea straw alone as a substrate supported an appreciable amount of solanapyrone A production (Fig. 4-2). During the saprobic growth, A. rabiei may require more arsenals for competing with other potential microbes than in the relatively protected host niches that is nearly free of other microorganisms that otherwise might compete for the same substrates. Recent comparative genomics studies highlighted that obligate parasitic fungal plant pathogens such as powdery mildews have extensively lost genes involved in secondary metabolism (Spanu et al.,

2010). In addition, a stress-tolerant (S-selected) fungus Baudoinia compniacensis that occupies a harsh ecological niche (Scott et al., 2007) has only a few SM-related genes, but other necrotrophic plant pathogens in its sister taxa retain a plethora of SM-related genes in their genomes (Ohm et al., 2012). This suggests that fungal species living in a relatively relaxed

165

environment (i.e., with less microbial competitors) have lost their SM-producing capacity, possibly to reduce fitness loads that may otherwise be imposed.

It is currently unknown how the growth stage-specific production of solanapyrone A can be established at a molecular level. Many studies have indicated that global transcriptional regulators mediate SM production in fungi, responding to environmental cues such as nutrient availability and ambient pH (Hynes, 1975; Dowzer and Kelly, 1991; Tilburn et al., 1995; Yin and Keller, 2011). In A. rabiei, xylan and cellobiose supported higher solanapyrone production, compared to the other complex carbons (Fig. 4-4). This suggests that these cellulose and hemicellulose components are likely the preferable carbon sources for the production of solanapyrone A during the saprobic growth. We observed that the key biosynthetic genes (sol1 and sol5) and the pathway-specific regulator (sol4) for solanapyrone biosynthesis are highly expressed during the growth on chickpea straw rich in cellulose and hemicellulose materials, but not during the host infection (Fig. 4-3). For plant pathogenic fungi, little information is available on differential gene expressions between their parasitic and saprobic growth. A recent transcriptomic study on human pathogenic fungi Coccidioides, the causal agents of coccidiodomycosis revealed that a large number of genes were upregulated in the saprobic growth phase compared to the parasitic growth phase, and vice versa (Whiston et al., 2012). In multicellular eukaryotes, changes in chromatin structure are required for time- and tissue-specific gene regulation. There is increasing evidence that epigenetic controls are involved in fungal SM biosynthesis (Reyes-Dominguez et al., 2010; Connolly et al., 2013). It was reported that

166

heterochromatin formations on SM gene clusters predominantly located near telomere have a role in repression of the cluster genes (Connolly et al., 2013). The solanapyrone biosynthesis gene cluster is also situated in a subtelomeric region of the A. rabiei genome (Chapter 3).

Therefore, on a speculative note, during the parasitic growth of A. rabiei, the cluster gene may have been silenced by such epigenetic effects. Genome-wide transcriptome and methylome analyses in A. rabiei during its saprobic and parasitic growth would reveal novel features on the growth stage-specific regulation of SM biosynthesis and on the global changes in gene expressions depending on different life strategies

Solanapyrone A exhibited significant antifungal activities at micromolar range against several saprobic fungi isolated from chickpea debris (Fig. 4-11). Moreover, it was recently reported that solanapyrone A was selectively toxic to gram positive bacteria such as Bacillus subtilis commonly found in soil (Wang et al., 2014). A. rabiei is a host-specific parasite and have to survive and reproduce sexually during the saprobic stage in order to reiterate its disease cycle

(Kaiser, 1997; Milgroom and Peever, 2003). Given the broad spectrum of the antibiotic activities and the growth stage-specific production, we suggest that solanapyrone A plays a crucial role in protection of their propagules from a wide variety of microorganisms in nature.

ACKNOWLEDGMENTS

I am grateful to Dr. Wilhelm Schäfer (University of Hamburg, Germany) for providing pTRI5::GFP and pII99dsRed plasmids.

167

Table 4-1. Peak assignments of compounds detected in extracts of chickpea stems and autoclaved chickpea straws inoculated with A. rabiei strain AR628.

m/z [M + H]+ m/z [M + H]+ No Rt (min) Compounds calculated observed a

1 2.40 Calycosinb 285.08 285.10 2 2.49 Pseudobaptigenin (glu-mal)c 531.11 531.12 → 283.08 3 2.56 Formononetin (glu-mal)c 517.14 517.14 → 269.10 4 2.75 Maackiain (glu-mal)c 533.13 533.13 → 285.09 5 2.86 Medicarpin (glu-mal)c 519.15 519.16 → 271.11 6 2.97 Biochanin A (glu-mal)c 533.13 533.15 → 285.10 7 3.18 Pseudobaptigeninb 283.06 283.08 8 3.26 Formononetinb 269.08 269.10 9 3.50 Maackiainb 285.08 285.09 10 3.66 Medicarpinb 271.10 271.11 11 3.92 Biochanin Ab 285.08 285.10 12 3.94 Unknown – 287.16 13 4.06 Unknown – 376.28 14 4.19 Solanapyrone A 303.16 303.16 a Values for m/z of the precursor ion [M + H]+ and major fragment ion detected in tandem mass analyses. b Identities of isoflavone aglycons were confirmed with the retention time and the mass spectra of standards. c O-glucoside malonylated conjugates (glu-mal) of isoflavones were defined on basis of elemental composition of the precursor [M + H]+ ions and their mass fragmentation patterns.

168

Table 4-2. Correlation of the expression levels between GFP and solanapyrone cluster genes in the sol5prom::GFP reporter strain.

PDAa sol1 sol4 sol5 dsRed GFP

3 days 0.033b 0.001 0.121 7.414 0.072

6 days 0.473 0.003 1.076 3.454 0.633

9 days 0.956 0.005 2.060 2.993 1.134

12 days 1.994 0.007 2.811 3.229 1.807

15 days 2.540 0.010 4.161 2.256 2.610

18 days 1.404 0.012 4.074 2.478 2.281

r c 0.930 0.960 0.993 – 0.808 – a Total RNA was extracted from the reporter strain (sol5prom::GFP) grown on potato dextrose agar media in 18 days with 3 day-intervals. b ΔCq expressions of sol1, sol4, sol5, dsRed, and GFP genes normalized by Actin1 reference gene. c Pearson correlation coefficient (r) of the sol gene (or dsRed) expression levels with that of the

GFP gene.

169

Figure 4-1. Visualization of A. rabiei during parasitic and saprobic growth. Growth of the

TEFprom::GFP strain on infected chickpea stems (cultivar Spanish White) over the course of infection (upper panels) and on autoclaved chickpea straw (lower panels). The parasitic and saprobic growth phases were visualized under a fluorescence microscope with GFP filter from 1 day post inoculation (dpi) to 9 dpi. Structures in round or oval shapes are pycnidia, the asexual fruiting bodies of the fungus. The images of mature pycnidia at 15 dpi were taken in the brightfield. Arrowheads indicate germinating spores. Arrows point out spore mass exudating from ostioles. Bar = 100 μm.

170

171

Figure 4-2. Growth phase-dependent solanapyrone production. A, Base peak intensity chromatograms of extracts of uninoculated chickpea stems (a), stem lesions (b), and autoclaved chickpea straws (c) inoculated with A. rabiei strain AR628 at 5 dpi. Most of the detected peaks were identified as chickpea isoflavonoids and their glucoside conjugates. Note that solanapyrone

A was detected during the saprobic growth as verified by comparison with the standard compound (d). The numbers above the peaks indicate the following: 1, calycosin; 2, pseudobaptigenin glucoside malonylated; 3, formononetin glucoside malonylated; 4, maackiain glucoside malonylated; 5, medicarpin glucoside malonylated; 6, biochanin A glucoside malonylated; 7, pseudobaptigenin; 8, formononetin; 9, maackiain; 10, medicarpin; 11, biochanin

A; 12, unknown (286 Da); 13, unknown (375 Da); 14, solanapyrone A. The observed m/z values of precursor and product ions for the peaks are presented in Table 1. B, Quantification of solanapyrone A production during the saprobic growth. Chickpea straw (≈ 3 cm in length) colonized by the A. rabiei strain were extracted at the indicated time points, and the amount of solanapyrone A was quantified by HPLC. Means of solanapyrone A production per straw are shown. Bars indicate standard deviation (n = 6–8).

172

Figure 4-3. Time-course expression of solanapyrone biosysnthesis cluster genes during the parasitic and saprobic growth of A. rabiei. Expression levels of solanapyrone cluster genes (sol1, sol4, and sol5) were compared between parasitic (dashed lines) and saprobic growth (solid lines).

Chickpea cultivar Spanish White and autoclaved chickpea straws were inoculated with A. rabiei strain AR628, and total RNA was extracted from stem lesions or chickpea straws colonized by the A. rabiei strain between 3 and 11 dpi at 2 days intervals. Quantitative RT-PCR was performed and gene expression levels of sol1, sol4, sol5 and AGT (acetylglutamate kinase) were normalized by Actin1 expression levels in respective samples. Means of the relative gene expressions levels (log2-transformed) are presented. Bars indicate standard deviation of three biological replicates.

173

Figure 4-4. pH shifts and solanapyrone A production of A. rabiei growing on different carbon sources. A. rabiei AR628 strain was grown on minimal media (set to pH 5.8) containing bacto agar (2%) and a single carbon source (1%). A, pH changes of the media were monitored before the start of culture (filled bar) and after two weeks of incubation (open bar). The pH of colony surface was measured at three different parts of the colony and averaged out to be one measurement. Means of six replicate plates were shown. Bars indicate standard deviation. B,

Solanapyrone A production were measured after 2 weeks of incubation on media containing polysaccharide (cellulose, xylan, starch, and pectin), disaccharide (cellobiose and maltose), or monosaccharide (glucose and xylose). Values for solanapyrone A quantification are presented as total yield of solanapyrone A per one plate. Means of five replicate plates were shown. Bars indicate standard deviation.

174

Figure 4-5. Effects of carbon sources on production of solanapyrone A. A. rabiei AR628 strain was cultured on minimal media (buffered to pH 5.8) containing Bacto agar (2%) and a single carbon source (1%). Colony diameter (filled bars) and solanapyrone A production (open bars) were measured after 2 weeks of incubation on media containing polysaccharide (xylan, starch, and pectin), disaccharide (cellobiose and maltose), or monosaccharide (glucose, xylose, arabinose, and galacturonic acid). Values for solanapyrone A quantification are presented as total yield of solanapyrone A divided by colony surface area (μg/cm2). Means of five replicate plates were shown. Bars indicate standard deviation.

175

Figure 4-6. Tissue-specific production of solanapyrone A. Sol5prom::GFP strain exhibiting constitutive dsRed expression as well as GFP expression under the control of the endogenous sol5 promoter was grown on PDA media. The cultures at different ages were scanned in brightfield and UV modes. Overlay images of photos taken with the GFP and dsRed filters individually were combined. A, Strong GFP signals took place around young developing pycnidia. B–D, Induction of sol5 gene was confined to amorphous hyphae (arrows) around mature pycnidia. Bar = 50 μm.

176

Figure 4-7. No GFP induction during plant infection. Sol5prom::GFP strain was inoculated to a susceptible chickpea cultivar Spanish White. Necrotic lesions formed on leaves and stems were scanned in brightfield and UV modes. The fungal growth and infections are evident with strong dsRed signals. Structures in round or oval shapes are pycnidia. Note that GFP fluorescence was not observed in the fungal mycelia or pycnidia colonizing the host tissues. Bar = 200 μm.

177

Figure 4-8. Secretion of solanapyrones outside of fungal cells. Sol5prom::GFP and

TEFprom::GFP strains were grown on PDA media. The Sol5prom::GFP strain expresses GFP under the control of the endogenous sol5 promoter. The GFP coding sequence in the

Sol5prom::GFP strain was fused to 3′ of the first 132 nucleotides of the sol5 coding sequence that reportedly encodes a signal peptide for secretion. GFP signals from the Sol5prom::GFP strain were enriched in the agar media as well as in hyphae near pycnidia, whereas

TEFprom::GFP strain expressing GFP under the control of the translation elongation factor promoter showed a constitutive expression pattern of GFP in all the hyphae and pycnidia. Bar =

100 μm.

178

179

Figure 4-9. Antagonistic potentials of A. rabiei (AR628 strain) against fungal competitors. A,

Front sides of confrontation plates, with A. rabiei AR628 strain (WT) or solanapyrone-minus mutants (Δsol5 and Δsol4) on the left side of the plates, and with saprobic fungi isolated from chickpea debris on the right side. At, Ep, Bc, and Ua mark Alternaria tenuissima, Epicoccum purpurascens, Botrytis cinerea, and Ulocladium atrum, respectively. B, Multiple-confrontation plates with Sclerotinia sclerotiorum growing from the center, and with A. rabiei WT, Δsol5 and

Δsol4 on corners of the plate. A. rabiei strains were allowed to grow in advance 10 days before the placement of S. sclerotiorum. Arrows indicate sclerotia of S. sclerotiorum formed on the colonies of the solanapyrone-minus mutants. The plates were photographed 6 days (left panel) and 8 days (right panel) after inoculation of S. sclerotiorum.

180

Figure 4-10. Antagonistic potentials of A. rabiei (AR21 strain) against fungal competitors. Front sides of confrontation plates with A. rabiei AR21 strain (WT) or solanapyrone-minus mutants

(Δsol5 and Δsol4) on the left side of the plates, and with saprobic fungi isolated from chickpea debris on the right side. At, Ep, Bc, and Ua mark Alternaria tenuissima, Epicoccum purpurascens, Botrytis cinerea, and Ulocladium atrum, respectively.

181

Figure 4-11. Antifungal activities of solanapyrone A. The colony diameter was measured on half-strength PDA media at 5 days after the growth of different fungal species in the presence of indicated concentrations of solanapyrone A (0, 25, 50, 100, or 200 μM). Ar, As, At, Ep, Bc, and

Ua mark Ascochyta rabiei, Alternaria soliani, Alternaria tenuissima, Epicoccum purpurascens,

Botrytis cinerea, and Ulocladium atrum, respectively. The half maximal inhibitory concentrations (IC50) of solanapyrone A against the fungal competitors (At, Ep, Bc, and Uc) or solanapyrone-producing fungi (Ar and As) were indicated in parenthesis. To estimate the IC50 values, colony diameter measurement for three replicate plates were averaged out at each concentration of solanapyrone A (two replicates for 200 μM). Values indicate mean ± standard error.

182

Figure 4-12. Front sides of confrontation plates with A. rabiei AR628 strain (WT) or solanapyrone-minus mutants (Δsol5 and Δsol4) on the left side of the plates. Cladosporium herbarum, a species in Al. infectoria group, and an unidentified fungal species were placed on the right side of the plates.

183

REFERENCES

Akamatsu, H.O., Chilvers, M.I., Stewart, J.E., and Peever, T.L. 2010. Identification and functio n of a polyketide synthase gene responsible for 1,8-dihydroxynaphthalene-melanin pigment biosynthesis in Ascochyta rabiei. Current Genetics 56:349-360.

Armstrong, C.L., Chongo, G., Gossen, B.D., and Duczek, L.J. 2001. Mating type distribution and incidence of the teleomorph of Ascochyta rabiei (Didymella rabiei) in Canada. Canadian Journal of Plant Pathology 23:110-113.

Bertini, S. 1956. Su di un composto ad antibiotica prodotto da Ascochyta pisi Lib. Annali Sperimentaz Agraria (Roma) 11:545-556.

Bird, B.A., Remaley, A.T., and Campbell, I.M. 1981. Brevianamides A and B are formed only after conidiation has begun in solid cultures of Penicillium brevicompactum. Applied and Environmental Microbiology 42:521-525.

Calvo, A.M., Wilson, R.A., Bok, J.W., and Keller, N.P. 2002. Relationship between secondary metabolism and fungal development. Microbiology and Molecular Biology Reviews 66:447-459.

Camazine, S. 1983. Mushroom chemical defense: Food aversion learning induced by hallucinogenic toxin, muscimol. Journal of Chemical Ecology 9:1473-1481.

Chanda, A., Roze, L.V., Kang, S., Artymovich, K.A., Hicks, G.R., Raikhel, N.V., Calvo, A.M.,

184

and Linz, J.E. 2009. A key role for vesicles in fungal secondary metabolism. Proceedings of the National Academy of Sciences of the United States of America 106:19533-19538.

Chen, W., Sharma, H.C., and Muehlbauer, F.J. 2011. Compendium of Chickpea and Lentil Diseases and Pests, St. Paul, MN, APS Press.

Connolly, L.R., Smith, K.M., and Freitag, M. 2013. The Fusarium graminearum histone H3 K27 methyltransferase KMT6 regulates development and expression of secondary metabolite gene clusters. PLoS Genetics 9:e1003916.

Degenkolb, T., von Döhren, H., Fog Nielsen, K., Samuels, G.J., and Brückner, H. 2008. Recent advances and future prospects in peptaibiotics, hydrophobin, and mycotoxin research, and their importance for chemotaxonomy of Trichoderma and Hypocrea. Chemistry & Biodiversity 5:671-680.

Donnelly, D., Sanada, S., O'Reilly, J., Polonky, J., Prange, T., and Pascard, C. 1982. Isolation and structure (X-ray analysis) of the orsellinate of armillol, a new antibacterial metabolite from Armillaria mellea. Journal of the Chemical Society, Chemical Communications:135-137.

Dowzer, C.E., and Kelly, J.M. 1991. Analysis of the creA gene, a regulator of carbon catabolite repression in Aspergillus nidulans. Molecular and Cellular Biology 11:5701-5709.

Dugan, F.M., Lupien, S.L., Hernandez-Bello, M., Peever, T.L., and Chen, W. 2005. Fungi resident in chickpea debris and their suppression of growth and reproduction of Didymella rabiei under laboratory conditions. Journal of Phytopathology 153:431-439.

Frisvad, J.C., Andersen, B., and Thrane, U. 2008. The use of secondary metabolite profiling in chemotaxonomy of filamentous fungi. Mycological Research 112:231-240.

Gloer, J.B., TePaske, M.R., Sima, J.S., Wicklow, D.T., and Dowd, P.F. 1988. Antiinsectan

185

aflavinine derivatives from the sclerotia of Aspergillus flavus. Journal of Organic Chemistry 53:5457-5460.

Hicks, J.K., Yu, J.H., Keller, N.P., and Adams, T.H. 1997. Aspergillus sporulation and mycotoxin production both require inactivation of the FadA G alpha protein-dependent signaling pathway. EMBO Journal 16:4916-4923.

Hynes, M.J. 1975. Studies on the role of the areA gene in the regulation of nitrogen catabolism in Aspergillus nidulans. Australian Journal of Biological Sciences 28:301-313.

Ichihara, A., Tazaki, H., and Sakamura, S. 1983. Solanapyrones A, B and C, phytotoxic metabolites from the fungus Alternaria solani. Tetrahedron Letters 24:5373-5376.

Ilgen, P., Hadeler, B., Maier, F.J., and Schäfer, W. 2009. Developing kernel and rachis node induce the trichothecene pathway of Fusarium graminearum during wheat head infection. Molecular Plant-Microbe Interactions 22:899-908.

Imazaki, A., Tanaka, A., Harimoto, Y., Yamamoto, M., Akimitsu, K., Park, P., and Tsuge, T. 2009. Contribution of peroxisomes to secondary metabolism and pathogenicity in the fungal plant pathogen Alternaria alternata. Eukaryotic Cell 9:682-694.

Joshi, B.K., Gloer, J.B., Wicklow, D.T., and Dowd, P.F. 1999. Sclerotigenin: A new antiinsectan benzodiazepine from the sclerotia of Penicillium sclerotigenum. Journal of Natural Products

186

62:650-652.

Kaiser, W.J. 1973. Factors affecting growth, sporulation, pathogenicity, and survival of Ascochyta rabiei. Mycologia 65:444.

Kaiser, W.J. 1997. Inter- and intranational spread of ascochyta pathogens of chickpea, faba bean, and lentil. Canadian Journal of Plant Pathology 19:215-224.

Kaiser, W.J., and Küsmenoglu, I. 1997. Distribution of mating types and the teleomorph of Ascochyta rabiei on chickpea in Turkey. Plant Disease 81:1284-1287.

Kaur, S. 1995. Phytotoxicity of solanapyrones produced by the fungus Ascochyta rabiei and their possible role in blight of chickpea (Cicer arietinum). Plant Science 109:23-29.

Lim, F.Y., and Keller, N.P. 2014. Spatial and temporal control of fungal natural product synthesis. Natural Product Reports 31:1277-1286.

Milgroom, M.G., and Peever, T.L. 2003. Population biology of plant pathogens: The synthesis of plant disease epidemiology and population genetics. Plant Disease 87:608-617.

Nakanishi, T., and Oku, H. 1969. Mechanism of selective toxicity: Absorption and detoxication of an antibiotic, ascochitine, by sensitive and insensitive fungi. Phytopathology 59:1563-1565.

Navas-Cortés, J.A., Pérez-Artés, E., Jiménez-Díaz, R.M., Llobell A, Bainbridge, B.W., and Heale, J.B. 1998. Mating type, pathotype and RAPDs analysis in Didymella rabiei, the agent

187

of Ascochyta blight of chickpea. Phytoparasitica 26:199-212.

Nene, Y.L., and Reddy, M.V. 1987. Chickpea diseases, their control. Pages 233-270 in The chickpea (Saxena, M.C., and Singh, K.B., eds.). Wallingford, UK: CAB International.

Ohm, R.A., Feau, N., Henrissat, B. et al. 2012. Diverse lifestyles and strategies of plant pathogenesis encoded in the genomes of eighteen Dothideomycetes fungi. PLoS Pathogens 8:e1003037.

Oku, H., and Nakanishi, T. 1963. A toxic metabolite from Ascochyta fabae having antibiotic activity. Phytopathology 53:1321-1325.

Oku, H., and Nakanishi, T. 1964. Reductive detoxification of an antibiotic, ascochitine, by an insensitive fungus, Fusarium lycopersici Sacc. Naturwissenschaften 51:538-538.

Rea, G., Metoui, O., Infantino, A., Federico, R., and Angelini, R. 2002. Copper amine oxidase expression in defense responses to wounding and Ascochyta rabiei invasion. Plant Physiology 128:865-875.

Reiss, J. 1982. Development of Aspergillus parasiticus and formation of aflatoxin B1 under the influence of conidiogenesis affecting compounds. Archives of Microbiology 133:236-268.

Reyes-Dominguez, Y., Bok, J.W., Berger, H., Shwab, E.K., Basheer, A., Gallmetzer, A., Scazzocchio, C., Keller, N., and Strauss, J. 2010. Heterochromatic marks are associated with the repression of secondary metabolism clusters in Aspergillus nidulans. Molecular Microbiology 76:1376-1386.

Ritz, C., and Streibig, J.C. 2005. Bioassay analysis using R. Journal of Statistical Software 12 Issue 5.

Roze, L.V., Chanda, A., and Linz, J.E. 2011. Compartmentalization and molecular traffic in secondary metabolism: A new understanding of established cellular processes. Fungal

188

genetics and biology 48:35-48.

Runa, F., Park, M., and Pryor, B. 2009. Ulocladium systematics revisited: Phylogeny and taxonomic status. Mycological Progress 8:35-47.

Schmittgen, T.D., and Livak, K.J. 2008. Analyzing real-time PCR data by the comparative CT method. Nature Protocols 3:1101-1108.

Scott, J.A., Untereiner, W.A., Ewaze, J.O., Wong, B., and Doyle, D. 2007. Baudoinia, a new genus to accommodate Torula compniacensis. Mycologia 99:592-601.

Sekiguchi, J., and Gaucher, G.M. 1977. Conidiogenesis and secondary metabolism in Penicillium urticae. Applied and Environmental Microbiology 33:147-158.

Singh, K., Nizam, S., Sinha, M., and Verma, P.K. 2012. Comparative transcriptome analysis of the necrotrophic fungus Ascochyta rabiei during oxidative stress: Insight for fungal survival in the host plant. PLoS ONE 7:e33128.

Spanu, P.D., Abbott, J.C., Amselem, J. et al. 2010. Genome expansion and gene loss in powdery mildew fungi reveal tradeoffs in extreme parasitism. Science 330:1543-1546.

Tilburn, J., Sarkar, S., Widdick, D.A., Espeso, E.A., Orejas, M., Mungroo, J., Peñalva, M.A., and Arst, H.N. 1995. The Aspergillus PacC zinc finger transcription factor mediates regulation of both acid- and alkaline-expressed genes by ambient pH. EMBO Journal 14:779-790.

Trapero-Casas, A., and Kaiser, W.J. 1992a. Development of Didymella rabiei, the teleomorph of Ascochyta rabiei, on chickpea straw. Phytopathology 82:1261-1266.

Trapero-Casas, A., and Kaiser, W.J. 1992b. Influence of temperature, wetness period, plant age, and inoculum concentration on infection and development of Ascochyta blight of chickpea. Phytopathology 82:589-596.

Trapero-Casas, A., and Kaiser, W.J. 2007. Differences between ascospores and conidia of Didymella rabiei in spore germination and infection of chickpea. Phytopathology 97:1600-1607.

189

Trapero-Casas, A., Luque-Marquez, F., and Kaiser, W. 2012. Development of the teleomorph of Ascochyta rabiei on culture media. European Journal of Plant Pathology 134:773-782.

Wang, H.J., Gloer, J.B., Wicklow, D.T., and Dowd, P.F. 1995. Aflavinines and other antiinsectan metabolites from the ascostromata of Eupenicillium crustaceum and related species. Applied and Environmental Microbiology 61:4429-4435.

Whiston, E., Zhang Wise, H., Sharpton, T.J., Jui, G., Cole, G.T., and Taylor, J.W. 2012. Comparative transcriptomics of the saprobic and parasitic growth phases in Coccidioides spp. PLoS ONE 7:e41034.

Wilson, A.D., and Kaiser, W.J. 1995. Cytology and genetics of sexual incompatibility in Didymella rabiei. Mycologia 87:795-804.

Xu, L., and Chen, W. 2013. Random T-DNA mutagenesis identifies a Cu/Zn superoxide dismutase gene as a virulence factor of Sclerotinia sclerotiorum. Molecular Plant-Microbe Interactions 26:431-441.

Yin, W., and Keller, N.P. 2011. Transcriptional regulatory elements in fungal secondary metabolism. Journal of Microbiology 49:329-339.

190

CHAPTER FIVE

USE OF NON-TARGETED METABOLIC PROFILING AS A TOOL FOR

THE CHEMOTAXONOMY OF LEGUME-ASSOCIATED

ASCOCHYTA AND ALLIED ANAMORPHIC GENERA

ABSTRACT

Chemotaxonomy and the comparative analysis of metabolic features in related fungal species would provide valuable information relating to ecology and evolution. Here, we investigated the chemical diversity of legume-associated Ascochyta and Phoma species and the possible use of the non-targeted metabolic profiling approach for their taxonomic classification using liquid chromatography-mass spectrometry (LC-MS)-based metabolomics. The metabolic features of 45 strains including 11 known species isolated from various legumes were extracted, and the datasets were analyzed by chemometrics methods such as principal component and hierarchical clustering analyses to compare their metabolic features. We found a high degree of intra-species consistency in metabolic profile, but inter-species diversity was high. Molecular phylogenies of legume-associated Ascochyta/Phoma species were reconstructed using three protein-coding genes, and the five main chemical groups that were detected in the hierarchical clustering analysis were mapped onto the phylogeny. Patterns and groupings of chemical features largely conformed to the species phylogeny. These results indicated that the distinct fungal

191

lineages have diversified their metabolic capacities as they have evolved independently.

Chemotaxonomy has not been applied to the genera Ascochyta and Phoma, mainly due to the lack of information on their metabolic content. The non-targeted metabolic profiling can be an effective tool for a polyphasic taxonomy of the largest and most complex fungal taxa.

INTRODUCTION

Many species in the genera Ascochyta and Phoma are recognized as primary or opportunistic pathogens on different plants of agricultural and economical importance, including some notorious pathogens with quarantine status (Sutton 1980, Boerema et al. 2004, Aveskamp et al. 2008). Some species are known to produce mycotoxins, which could directly threaten animal and human health (Rabie et al. 1975, Bennett 1983, Pedras & Biesenthal 2000, Sørensen et al. 2010). The genera Ascochyta and Phoma both are polyphyletic, being embedded in the order Pleosporales within the class Dothideomycetes (Boerema & Bollen 1975, Aveskamp et al.

2010). Species of Ascochyta and Phoma are highly similar in morphology, physiology, and in the disease symptoms they produce (Aveskamp et al. 2010 and references therein). In the

Saccardoan system, Ascochyta and Phoma are distinguished only by conidial morphology; two-celled conidia in the former and one-celled conidia in the latter (Boerema & Bollen 1975,

Boerema et al. 2004). However, the two-celled conidial morphology appears to have evolved independently multiple times during evolution of several lineages in the Pleosporales. Also, the type of conidiogenous cells has been used as an ultra-structural character to distinguish the two

192

genera (Boerema & Bollen 1975). This character is however not consistent with the conidial ontogeny of Phoma species in some instances (Sutton & Sandhu 1969). Due to lack of reliable morphological characters to distinguish the two genera as well as a high degree of environmental variation, systematics of these genera has never been fully resolved (Aveskamp et al. 2008). This has resulted in unclear taxonomic placement and many species having both Ascochyta and

Phoma synonyms (Aveskamp et al. 2010).

The internal transcribed spacer (ITS) regions of the nuclear rDNA operon have been proposed as a universal DNA barcode that can be used to circumscribe species boundaries of several taxonomic groups of fungi (Druzhinina et al. 2005, Summerbell et al. 2007, Seifert 2009).

However, the locus is conserved and thus sometimes not suitable for species delimitation of closely related taxa (Nilsson et al. 2008). The genus Phoma is one of the largest and the most complex fungal genera, with more than 3,000 infrageneric taxa described (Monte et al. 1991). To resolve the phylogeny of Phoma at species level or even below, several protein-coding genes such as cytochrome c oxidase subunit I (COI), actin and β-tubulin, were used (Aveskamp et al.

2009, Habibi et al. 2014). Although the COI was successfully applied to Penicillium taxonomy

(Seifert et al. 2007), the locus does not exhibit taxon-specific conserved single nucleotide polymorphism (SNP) in a subset of Phoma taxa (Aveskamp et al. 2009). The actin locus showed a high degree of SNPs enough to resolve phylogeny of a Phoma taxon below species level

(Aveskamp et al. 2009). However, interspecific variations are shown to be very high and, as a result, the obtained sequences are not aligned properly, resulting in many losses of informative

193

sites (Aveskamp et al. 2009). In addition, significant topological incongruence has been observed between the phylogeny estimated from the β-tubulin locus and other housekeeping genes among legume-associated Ascochyta species, possibly due to incomplete lineage sorting and/or gene flow and hybridization (Habibi et al. 2014).

A polyphasic approach integrating morphological, physiological features and DNA sequences would be necessary in fungal systematics to classify diverse Ascochyta and Phoma spp. (Aveskamp et al. 2010). Chemotaxonomy is a method of biological classification based on the similarity of chemical compounds such as carbohydrate, lipid, amino acid and secondary metabolites (SM) (Benedict 1970, Tosch et al. 2006, Velázquez et al. 2006). Filamentous fungi

(especially ascomycetes) are known to produce a vast array of SMs such as terpenes, polyketides, non-ribosomal peptides, as well as many other small organic compounds of mixed biosynthetic origin. Extensive genome sequencing efforts revealed that genomic differences between fungal species are often related to the number and similarity of genes involved in SM biosynthesis like polyketide synthases and non-ribosomal peptide synthetases, which result in the production of a unique set of SMs (Galagan et al. 2005, Nierman et al. 2005, Pel et al. 2007). Therefore, it is tempting to use SM profiles as a taxonomic diagnostic tool, and the chemotaxonomy is proven to be effective in Aspergillus and Penicillium classification (Smedsgaard & Nielsen 2005, Frisvad et al. 2007, Frisvad et al. 2008).

Several Ascochyta species infect cool season food legumes in a host-specific manner. A. pisi Lib., A. fabae Speg., A. lentis Vassiljevsky, A. viciae-villosae Ondrej, and A. rabiei (Pass)

194

Labr. are pathogens of pea (Pisum sativum L.), faba bean (Vicia faba L.), lentil (Lens culinaris

Medik.), hairy vetch (Vicia villosa Roth), and chickpea (Cicer arietinum L.), respectively (Nene et al. 1988, Kaiser 1997, Chen et al. 2011). These legume-associated Ascochyta fungi form a monophyletic group that also includes some other Ascochyta and Phoma spp. sampled from non-legume hosts. The species infecting different legume hosts can be recognized by phylogenetic analyses with DNA sequences of multiple protein-coding genes, each being grouped into separate, well-supported clades (Peever et al. 2007, Chilvers et al. 2009). In addition to the phylogenetic species recognition, the biological species concept has been employed to describe these closely related species (Kaiser et al. 1997, Hernandez-Bello et al.

2006). In vitro genetic crosses were made between strains of A. rabiei, A. fabae and A. lentis

(Kaiser et al. 1997). A. rabiei was able to mate neither with A. fabae nor A. lentis. In contrast, the genetic crosses between A. fabae and A. lentis strains produced sexual fruiting bodies with viable ascospores, although intrinsic postzygotic mating barriers were observed, such as nonstandard numbers of ascospores in asci and variable size of ascospores, suggesting that the three species are biologically distinct (Kaiser et al. 1997). In a separate study, genetic crosses were made between A. pisi and A. fabae and between A. lentis and A. viciae-villosae (Hernandez-Bello et al.

2006). In both combinations, normal asci and viable ascospores were produced and molecular markers segregated normally, indicating no intrinsic mating barriers. However, the pathogenic abilities of the progenies derived from the cross between A. pisi and A. fabae were greatly impaired, indicating that genes controlling host specificity likely represent extrinsic postzygotic

195

mating barriers acting to prevent hybrid formation and thereby maintain species integrity. These results revealed that the taxa used in these crosses were closely related yet still biologically distinct (Hernandez-Bello et al. 2006).

Legume-associated Ascochyta and Phoma species are phylogenetically and biologically well described fungi, which allows us to examine the utility of chemotaxonomy in their classification. To date, however, a major problem for the application of chemotaxonomy is that only a few SMs have been identified from Ascochyta and Phoma species. Metabolomics is the comprehensive analysis of the biochemical content of cells, tissues or entire organisms, usually from analysis of extracts (Oliver et al. 1998). Liquid chromatography, coupled with time-of-flight mass spectrometry (LC-TOF-MS), is a powerful tool for metabolic profiling studies in many fields of biology (Hodson et al. 2007, Krug et al. 2008, Dunn et al. 2011). The non-targeted approach that analyzes and compares whole metabolomes of fungal extracts is effective especially when limited information is available on metabolite production of species of interest.

We conducted the metabolic profiling of 11 Ascochyta and Phoma spp. and many ecological strains provisionally identified as Ascochyta that were isolated from different legume hosts. The primary objective of this research was to assess the feasibility that the 11 Ascochyta and Phoma species can be distinguished and recognized by their metabolic profiles, supporting the previous species delimitations using both phylogenetic and/or biological species concepts.

We also were interested in testing the application of chemotaxonomy for classifying species in

196

highly diverse taxa in general. A secondary objective was to determine the evolutionary relationships between known Ascochyta/Phoma spp. and closely related strains associated with cultivated legumes or related wild plants by performing multivariate chemometrics analyses with their metabolic profiles, in conjunction with phylogenetic analysis. A subsidiary objective was to identify SMs from newly identified species such as P. koolunga.

MATERIALS AND METHODS

Fungal Strains

Forty-five single-conidial isolates of Ascochyta spp., Phoma spp., and Alternaria solani were obtained from a culture collection maintained by the USDA Western Region Plant

Introduction Station, Pullman, Washington. Strains of A. rabiei, A. lentis, A. pisi, A. fabae, A. pinodes, and A. pinodella used in the previous study (Peever et al. 2007) were selected for chemotaxonomic analysis. We attempted to sample at least 3 strains per species, where possible, and also preferentially selected strains that were deposited in the American Type Culture

Collection. Also, among strains isolated from various wild vetches (Vicia spp.) in the Republic of

Georgia in 2004 (Peever et al. 2007), we selected strains that formed well supported subclades distinct from the above-mentioned Ascochyta spp. in previous phylogenetic analyses (Peever et al. 2007) to compare their metabolic features with those of the known Ascochyta spp. Additional strains used in this study included strains from Medicago sativa (Akamatsu et al. 2008),

Astragalus lentiginosus (Habibi et al. 2015) and Lupinus sp. collected by T.L. Peever and W.J.

197

Kaiser in Washington and Idaho, USA in 2001–2005 and strains from Lathyrus sativus collected by A. Infantino in Italy in 2007–2013. Six strains of the recently described species Phoma koolunga isolated from Pisum sativum (Davidson et al. 2009) were also included. In a preliminary study, the six strains showed identical metabolic profiles, and therefore the FT07010 strain (recoded as PK4 in this study) was chosen in the current study. Finally, we included representative strains of P. herbarum (CBS615.75) and Didymella exigua (CBS183.55), the type species of genus Phoma and Didymella, respectively. The details on the strains used in this study are described in Table 5-1.

Fungal Culture and Sample Preparation

To obtain the metabolic profiles of fungal strains for the chemometrics analyses, all the strains were grown on V8 agar media (200 mL V8 juice, 3 g CaCO3, 20 g agar in 1 L distilled water) for 10 days, then 20–30 agar plugs (5 mm diameter) containing actively growing mycelia were inoculated to the oat kernel solid media. Strains were cultured in 150-mL Erlenmeyer flasks containing oat kernels (50 mL in volume) that were soaked in water overnight and then autoclaved. The cultures were allowed to grow at 20°C under an alternating 12 hr photoperiod for 14–18 days until pycnidia formation indicated by black pigmentation was observed on the oat kernels. Ethyl acetate (EtOAc) was used as a solvent for extraction. This organic solvent with medium polarity is capable of extracting a diverse array of metabolites other than those that are extremely polar or non-polar. This solvent is useful for SM extraction because it has the properties to denature and subsequently precipitate macromolecules such as proteins, as well as

198

to cause partition of highly polar compounds to the aqueous phase (Bose et al. 2014). To extract metabolites from the inoculated oat kernels, 50 mL of EtOAc was added to the culture and gently shaken for 20 minutes at room temperature. Extracts were dried over anhydrous MgSO4 and filtered through four layers of cheesecloth to remove oat kernel debris. One and half milliliter of the extracts were transferred to 2-mL screw-cap tubes and centrifuged at 2,200 × g for 5 min.

One milliliter of the supernatants were collected and evaporated to dryness on a vacufuge®

(Eppendorf, Germany). The same procedure was conducted for autoclaved oat kernels incubated without fungal inoculation to obtain a “blank” extract. Pellets were reconstituted with 200 μL of methanol, and the solutions were subjected to LC-MS analysis.

Structural Identification of Ascochitine and Pinolidoxin

Ascochitine and pinolidoxin were purified from the extracts of Phoma koolunga strain

PK4 and Ascochyta pinodella strain PMP3, respectively. The strains were cultured on 200 g of autoclaved oat kernels in a mason jar. Upon completion of the culture period (2 weeks), oat kernels colonized by either PK4 or PMP3 strain were extracted twice with 150 mL of EtOAc.

The combined extracts were dried over MgSO4 and filtered through four layers of cheesecloth to remove oat kernel debris and MgSO4. The filtrate was concentrated by rotary evaporator under reduced pressure to give a crude yellow culture extract. Ascochitine was semi-purified by following an acid-base extraction procedure described in literature (Foremska et al. 1992). The resulting residue was further purified by preparative TLC on silica gel using a mixture of dichloromethane/methanol/formic acid (96:3:1) to afford ascochitine as yellow oil (2.6 mg). The

199

compound was confirmed by LC-MS and 1H NMR (Figs. S5-1 and S5-2). For pinolidoxin purification, flash column chromatography was performed using a mixture of EtOAc/hexanes

(2:3) on a silica gel to give a product as a white solid (5.8 mg). Fractions containing pinolidoxin were recognized by its mass spectra (ESI MS (+) m/z 361.1 [M + Na]+; Cimmino et al. 2012) and collected to perform 1H NMR analyses (Fig. S5-2). The spectral characterization of pinolidoxin was compared with literature 1H NMR data (García-Fortanet et al. 2005). Identity of solanapyrone A produced by A. rabiei strains was confirmed with the retention time and the mass spectra of the reference compound provided by Dr. H. Oikawa (Hokkaido University,

Japan).

Mass Spectrometric Analyses

Chromatographic separation was performed using an ACQUITY UPLC system (Waters

Corp., USA) on a 2.1 mm × 50 mm, 1.7 µm ACQUITY UPLC BEH C18 column (Waters Corp.).

The mobile phase A was water with 0.1% formic acid, and mobile phase B was acetonitrile with

30°C, and the flow rate was 0.45 mL min-1. MS analysis was performed on an inline Synapt

G2-S HDMS (Waters Corp.) time of flight mass spectrometer. Positive and negative ion electrospray modes were used for data collection. In positive mode, the desolvation gas was nitrogen (800 L h-1), and the collision gas was argon (2.0 mL min-1). The data acquisition range

200

was m/z 50–1,000. The source temperature was 120°C, and the desolvation temperature was

250°C. The cone voltage was 30 V. The lock mass compound used was leucine encephalin with a reference mass at m/z 556.2771. In negative mode, ESI conditions were: capillary voltage 2.5 kV; source temperature 90 °C; cone voltage 40 V; desolvation temperature 150 °C; desolvation gas flow 500 L h-1 (nitrogen gas); and scan rate 0.1 s/scan with 0.015 s interscan delay. The lock mass compound used was leucine encephalin with a reference mass at m/z 554.2615. For the

TOF experiments, data were acquired in the MSE mode in which two separate scan functions were programmed for the MS acquisition method. One scan function was set at low collisio n energy (trap at 4 eV and transfer at 2 eV), and the other scan function was set at high collision energy (trap ramped from 15–50 eV and transfer at 2 eV). The mass spectrometer switched rapidly between the two functions during data acquisition. As a result, information on intact precursor ions and on product ions was obtained from a single LC run.

Chemometrics Analyses

We performed two independent chemometrics analyses using LC-MS datasets. In the first experiment, the LC-MS dataset of fungal extracts of 21 strains was acquired with positive mode.

In the second experiment, to reproduce the first experiment and analyze more diverse taxa, the

LC-MS dataset of fungal extracts of 40 strains including most of strains used in the first experiment was acquired with both positive and negative modes. The LC-MS datasets were further processed to extract and align peaks from the chromatograms of strains, using Progenesis

QI software (Waters Corp.). To discard peaks corresponding to extremely polar and nonpolar

201

compounds as well as noise peaks, the early- and late-eluting peaks (< 0.5 min and > 7.0 min in the chromatograms) and the peaks with absolute ion intensity less than 100 were excluded from the analyses, resulting in a total of 1,733 and 4,610 putative metabolic features (variables) for the first and second datasets, respectively. To aid the metabolic profiling process, the LC-MS data of the “blank” extract (oat kernel only) for each dataset were also analyzed to extract metabolic features and to use as a background reference. These reference ion peaks were removed from all fungal extracts data in the matrices, resulting in a total of 1,075 and 2,770.

The resulting data matrices (21 observations with 1,075 variables and 40 observations with 2,770 variables) were exported as comma-separated value files and imported to Statistica v.12 (StatSoft, Inc., USA). All data were normalized prior to multivariate analysis, using the statistical method implemented in Progenesis QI (Waters Corp.). Unsupervised analysis using principal component analysis (PCA) was initially performed to assess any grouping or trends among A. rabiei, A. pisi, and A. pinodes strains in the first dataset, using the Statistica software

(StatSoft, Inc., USA). To achieve better clustering of data points, supervised partial least square

(PLS) analyses were performed, in which predetermined groupings by known species were set to classify the datasets. For hierarchical clustering analyses, matrices of Pearson correlation coefficients for the datasets were constructed using the MultiExperiment Viewer (MeV) software

(Institute for Genomic Research, CA, USA). The robustness of the clusters produced was determined by looking at the consensus of the two clustering methods, complete linkage and unweighted pair group method with arithmetic mean (UPGMA). Both methods agreed in

202

dividing strains into five main clusters and several singletons in the second dataset.

DNA Sequencing

For the phylogenetic analysis, partial DNA sequences of three protein-coding genes, glyceraldehyde 3-phosphate dehydrogenase (G3PD), chitin synthase (CHS), and translation elongation factor alpha (EF) were used. To obtain the DNA sequences from the strains, primers

(CHS-79 and CHS-354 for CHS gene; EF1-728F and EF1-986R for EF gene) were used for PCR amplification (Carbone & Kohn 1999). Also, G3PD gene was amplified from the strains with gpd-1 and gpd-2 primers (Berbee et al. 1999). PCR condition consisted of 95 C for 3 min followed by 32 cycles of 95 °C for 30 s, 52 °C for 30 s, and 72 °C for 45 s. Amplicons were direct sequenced for each strand with sequence reactions containing 10–50 ng DNA, 320 nM each of forward and reverse primers, and 4 μL BigDye Terminator Cycle Sequencing Ready

Reaction Mix (Applied Biosystems, USA), in 10 μL total volumes. Reaction conditions consisted of 35 cycles of 15 s at 96 °C, 15 s at 50 °C, and 4 min at 60 °C. PCR Products were purified with

Centriflex Gel Filtration Cartridges (Edge Biosystems, USA), dried in a rotary evaporator and sequences were read in a PE Biosystems Model 3730 Automated DNA Sequencer (Applera

Corp., USA). All sequencing was performed at the Laboratory for Biotechnology and

Bioanalysis, School of Molecular Biosciences, Washington State University. The obtained G3PD,

CHS, and EF sequences were deposited under GenBank accession numbers.

KR184153–KR184188. Additional DNA sequences were retrieved from GenBank for datasets used in previous phylogenetic studies (Peever et al. 2007).

203

Phylogenetic Analysis

A maximum likelihood phylogeny was estimated from the combined G3PD, CHS, and EF datasets in a Bayesian framework using Markov chain Monte Carlo (MCMC) sampling in

MrBayes version 3.2 (Ronquist et al. 2012). Models of sequence evolution and model parameter estimates for each dataset were as described previously (Peever et al. 2007). In brief, the TIM+G model with unequal base frequencies (A = 0.262, C = 0.314, G = 0.230, T = 0.193), four substitution rate parameters (1.000, 4.4137, 2.0745, 2.7045, 7.7086, 1.000) and gamma distributed rates (shape parameter = 0.318) was chosen for the G3PD dataset. For the CHS dataset, the HKY+G model with unequal base frequencies (A = 0.2992, C = 0.2244, G = 0.2816,

T = 0.1949), a transition:transversion ratio of 2.364 (two substitution rate parameters) and gamma distributed rates (shape parameter = 0.1355) was selected. For the EF dataset, the HKY model with unequal base frequencies (A = 0.2052, C = 0.3213, G = 0.2339, T = 0.2396) and a transition:transversion ratio of 2.2724 was selected. For analysis of the combined CHS, EF, and

G3PD dataset, a partitioned model was implemented in MrBayes with evolutionary models and parameters estimated independently for each genomic region as described above. Each run of the sampler consisted of 2,000,000 generations of the Markov chain. Four chains were run in each analysis (one heated and three cold) with the temperature parameter set at 0.1 and random chains swapped one time per generation. Two independent analyses (each of 2,000,000 generations) each were started from a random tree. Trees were sampled every 500 generations and the first

500,000 generations (1,000 trees) of each analysis were discarded as burn-in. Average posterior

204

probabilities were estimated for each node of the phylogeny across runs (6,000 trees from

2,000,000 total generations of the MCMC). The combined phylogeny was rooted by Alternaria solani. Clades were inferred based on posterior probabilities greater than or equal to 95%.

Sequence alignments of concatenated G3PD, CHS, and EF genes were deposited in TreeBASE

(Study 17,485, www.treebase.org).

RESULTS

Metabolic Profiling of Legume-associated Ascochyta and Phoma

To profile the metabolite content of legume-associated Ascochyta and Phoma spp., a total of 45 strains were cultivated on autoclaved oat kernels, and culture extracts were subjected to

LC-MS analyses. Metabolic profiles were further processed and analyzed using multivariate chemometrics methods to assess intra-species consistency and inter-species differences in metabolite production. A preliminary study confirmed that oat kernel cultures of the tested strains yielded a diverse array of metabolites, including previously known SMs of the tested strains, such as ascochitine, solanapyrone A, and pinolidoxin produced by A. pisi, A. rabiei, and

A. pinodes, respectively (Bertini 1956, Alam et al. 1989, Evidente et al. 1993).

Two independent experiments were carried out for the metabolic profiling of the

Ascochyta/Phoma strains. In the first experiment, we were interested in assessing the feasibility that legume-associated Ascochyta species can be distinguished and recognized by their metabolic profiles, we analyzed 21 fungal strains including well-described species such as A. pisi, A. rabiei,

205

and A. pinodes whose teleomorphs (sexual stage) have been connected to Didymella (Petrak

1924, Trapero-Casas & Kaiser 1992, Peever et al. 2007, Chilvers et al. 2009). In the second study, we included more diverse strains that were less well characterized but provisionally identified as Ascochyta species based on morphology and phylogenetic analyses. Also, included were the type species of Ascochya, Phoma, and Didymella. The details on the strains used in the analyses are described in Table 5-1, and their colony morphology is shown in Figure 5-1.

Chemoconsistency

Initially, we chose six Ascochyta and Phoma species; A. pisi (strain codes: APs), A. pinodes (MPs), P. koolunga (PK4), A. pinodella (PMPs), A. rabiei (ARs, G10, and M305), and P. medicaginis (ASs). Strains representing the same species from different geographical origins were used, when possible (Table 5-1) to assess the consistency of metabolite production across regions. Strains (ID1A, ID3A, ID4A, and G11) that were morphologically and phylogenetically similar to Ascochyta species (Chilvers et al. 2006, Peever et al. 2007, Chilvers et al. 2009, Habibi et al. 2015), were also included to see if their metabolic features are similar to or distinct from the known species.

Metabolic profiles of the 21 strains were obtained using LC-MS analyses of culture extracts. A total of 1,075 unique ions (after data processing, see Materials and methods) were detected from the extracts. Principal Component Analysis (PCA) was conducted to assess any groupings or trends among the strains. Plots of the PCA scores revealed differences in the

206

metabolic profiles of the known Ascochyta/Phoma species. The results showed four distinct clusters; A. pisi (including P. koolunga and the Ascochyta-like strains), A. rabiei, P. medicaginis and A. pinodes (together with A. pinodella) (Fig. 5-2 A). A supervised multivariate analysis using partial least square (PLS) statistics, whereby strains within the same species were grouped together, showed that A. pisi strains were separated from other strains that were grouped together in the PCA score plot. In the PLS score plot, however, A. pinodes and A. pinodella strains remained grouped together (Fig. 5-2 A).

A hierarchical clustering analysis showed that the grouping patterns of strains were consistent with that of PLS analysis (Fig. 5-2 B). The Ascochyta-like strains (ID1A, ID3A, and

G11) were clustered together in the analysis and their metabolic features were distinct from others. Also, P. koolunga strain and ID4A strain (isolated from Lupinus sp.) were not grouped with A. pisi in the hierarchical clustering analysis, but seem to be closer to A. pisi than any other species in terms of chemical similarity. The metabolic features of A. pinodes and A. pinodella were almost identical. Pairwise comparison of total ion chromatograms of representative strains of A. pisi, A. rabiei, and A. pinodes highlighted consistency of the metabolic profile between strains of different geographical origins, as well as the reproducibility of data acquisition (Fig.

5-2 C). Also, we detected ascochitine, solanapyrone A, and pinolidoxin, as a major metabolite of

A. pisi, A. rabiei, and A. pinodes, respectively (Fig. 5-2 C).

207

Identification of Ascochitine and Pinolidoxin

The grouping by PCA and hierarchical clustering analysis showed that P. koolunga PK4 strain and ID4A strain are chemically similar to A. pisi (Fig. 5-2 B). Visual inspection of their chromatograms found that the strains also produce ascochitine (Fig.5-3 A). A. pinodes is phylogenetically close to A. pinodella (syn. P. medicaginis var. pinodella) (Chilvers et al., 2009;

Peever et al., 2007). We also detected pinolidoxin as a major metabolite in all tested A. pinodella strains (PMPs) (Fig. 5-3 B). Ascochitine and pinolidoxin were purified from the extract of PK4 and PMP3 strains, respectively, and the identities of the purified compounds were confirmed by

1H NMR analyses (Fig. S5-2).

Chemodiversity

For the second experiments, we analyzed more diverse Ascochyta/Phoma spp. and ecological strains, including most of the strains used in the first chemometrics analysis. Strains of

A. fabae, A. lentis, and A. viciae-villosae and strains isolated from different wild vetches (Vicia spp.) in Republic of Georgia were included (Table 5-1). It was previously reported that strains found in wild vetches (strain codes starting with ‘G’) were morphologically indistinguishable from and phylogenetically related to A. pisi, A. fabae, A. lentis, and A. viciae-villosae (Peever

2007, Peever et al. 2007). In addition, we analyzed metabolic profiles of strains collected from grasspea (Lathyrus sativus) in Italy (strain codes starting with ‘ER’), which are phylogenetically

208

close to A. lentis (T.L.Peever, unpublished). Also included were P. herbarum and Didymella exigua, the type species for the genera Phoma and Didymella.

Hierarchical clustering analysis of the LC-MS dataset of 40 strains detected five major chemically similar groups and strains of P. koolunga, P. medicaginis, P. herbarum, and D. exigua as singletons at a 0.75 distance threshold (Fig.5-4 A). The ‘chemical group 1’ is comprised of A. pisi, A. fabae, A. viciae-villosae and strains isolated from different wild vetches (Vicia spp.).

Although the metabolic profiles of strains of A. pisi and A. fabae were highly similar to each other, the two closely related species were still distinguishable by their metabolic features (Fig.

5-4 A). The strains isolated from grasspea (ER1415, ER1478, and ER1813) were clustered together with A. lentis strains within the ‘chemical group 2’ (Fig. 5-4 A), as expected from their phylogenetic similarity. However, the metabolic features of the A. lentis strains were highly variable and one strain (AL6) was placed within the ‘chemical group 1’.

Consistent with the first analysis, A. rabiei, A. pinodes (and A. pinodella), P. medicaginis,

P. koolunga, and the Ascochyta-like strains (ID1A, ID3A, and G11) each represented independent chemical groups. In addition, P. herbarum (PH) and Didymella exigua (DE) strains showed distinct metabolic profiles and were not grouped with any of the legume-associated

Ascochyta/Phoma spp. (Fig. 5-4 A). It has been reported that a distantly related plant pathogenic fungus, Alternaria solani, produces the same set of SMs, solanapyrones, as does A. rabiei

(Ichihara et al. 1983). We included two Al. solani strains (ALS1 and ALS2). Interestingly, the Al. solani strains were grouped together with A. rabiei strains, rather than representing an

209

independent group (Fig. 5-4 A). A supervised multivariate analysis using partial least square

(PLS) statistics, whereby predetermined groupings by known species were set to classify the dataset, also showed a distinction of metabolic features of the five major chemical groups and the singletons (Fig. 5-4 B).

Chemotaxonomy

To put the observed metabolic profiles in an evolutionary framework, a Bayesian phylogeny of the strains was estimated from three protein-coding genes, glyceraldehyde

3-phosphate dehydrogenase, chitin synthase, and translation elongation factor alpha (Fig. 5-5), and the five major chemical groups observed in the hierarchical clustering analysis (Fig. 5-4 A) were traced onto the combined phylogeny. The Bayesian phylogeny revealed four major clades.

Within clade A, two subclades, A. pisi and A. fabae (including strains isolated from wild vetches) were evident. Within clade B, two subclades, A. lentis (including strains isolated from grasspea) and A. viciae-villosae (including strains isolated from other wild vetches) were evident. Within clade C, three subclades, A. rabiei, P. medicaginis, and the Ascochyta-like strains (ID1A, ID3A, and G11) were evident. Clade D included A. pinodes and A. pinodella. Some putative recombination events in the protein-coding genes appear to have occurred given that an A. pinodes strain (MP19) was closer to A. pinodella (PMPs) than other A. pinodes strains in the phylogeny, as has been observed previously (Peever et al. 2007).

210

Tracing the five major chemical groups onto the Ascochyta/Phoma phylogeny showed that the topology of the dendrogram of Ascochyta/Phoma spp. based on their chemical similarity is largely congruent with the topology of the species phylogeny. Previously, strain AV11 isolated from big-flower vetch (Vicia grandiflora) received the formae specials designation (A. fabae f. sp. vicia; Leath 1994). Given the phylogenetic and chemical distinctiveness (Figs. 5-4 and 5-5), strains AV11, G13, and G16 seem to be conspecifics. Within clade B, the two subclades were chemically distinct (Fig.5-5). The metabolic features of the A. lentis subclade seem to have been diverged from the sister subclade (A. viciae-villosae) and clade A (A. pisi and A. fabae). Within clade C, the metabolic features of each subclade were distinct from one another and appear to have been diversified as these lineages evolve independently. Strains G11, ID1A, and ID3A strains were phylogenetically and chemically distinct from others and may represent an independent evolutionary lineage. The Bayesian phylogeny showed that P. koolunga (PK4) was distinct from other species, so was an unidentified Ascochyta sp. (ID4A), but their relative relationship with other species was uncertain (Fig. 5-5). The lower posterior probability values

(88% for clade C and 92% for clades A, B, and C) were mainly because the two strains were sometimes placed external to the clades containing A. pisi, A. fabae, A. lentis, A. rabiei and P. medicaginis. Their chemical similarity to A. pisi and the production of ascochitine in common

(Figs. 5-2 and 5-3) suggested their closer relationship to clade A.

The Bayesian phylogeny indicated that A. pinodes/A. pinodella, P. herbarum and D. exigua are distinct from the other legume-associated Ascochyta/Phoma spp. (clades A, B, and C).

211

Recently, A. pinodes and A. pinodella (syn. P. medicaginis var. pinodella) were transferred to the new genus Peyronellae that was elevated from Phoma sections (Aveskamp et al. 2010). The distinct metabolic features and the observed phylogenetic distance from P. herbarum, the type species for the genus Phoma, and the other Ascochyta/Phoma spp. including P. medicaginis justify the reclassification.

DISCUSSION

The legume-associated Ascochyta family is a group of phylogenetically closely related fungi. Due to paucity of morphological characters, their species boundaries have been delimited using phylogenetic and biological species recognition (Kaiser et al. 1997, Hernandez-Bello et al.

2006, Peever et al. 2007, Chilvers et al. 2009). However, ecological divergence such as host specificity appears to have played significant roles in speciation of these legume-associated

Ascochyta species (Peever 2007, Peever et al. 2007). The levels of prezygotic isolation barrier between phylogenetically closely related species in the fungal family are low and therefore some of the recognized species are still interfertile producing viable progeny (Hernandez-Bello et al.

2006). Non-targeted metabolic profiling and its application to chemotaxonomy could aid in classification of fungal species because distinct metabolic features may reflect lack of extensive gene flows between distinct yet interfertile species, albeit without prior knowledge on their metabolite production. In this study, although potentially recombining taxa (e.g. A. pinodes and A. pinodella) were chemically indistinguishable, we were able to recognize previously described

212

Ascochyta and Phoma spp. by their metabolic features. The classification of the legume-associated Ascochyta/Phoma spp. based on chemical similarity results in a grouping of the strains that is in good agreement with the phylogenetic analysis (Figs. 5). Each evolutionary lineage appears to diversify its metabolic capacity and the known Ascochyta/Phoma spp. exhibited characteristic metabolic profiles.

Clade A

One of the aims of this study was to resolve evolutionary relationships of ecological strains collected from wild vetches (Vicia spp.) and the known Ascochyta spp. within clades A and B. Although A. pisi, A. fabae and A. fabae f. sp. vicia each exhibited specific metabolic profiles, the phylogenetically well supported subclades including strains from wild vetches were not clustered by their chemical similarity. It is likely that subclades within clade A may not have chemically diverged from each other much yet. Alternatively, a history of hybrid formations may have jumbled the metabolic profiles of fungal strains infecting wild plants. In fact, some putative recombination events were detected among strains infecting wild pea (Pisum elatius) in A. pisi and A. fabae clades (Peever et al. 2007). Hybridization may benefit fungal species infecting wild plants by modulating a combination of genes related to host specificity, thereby enabling host jump especially when the previous host became no longer available or extinct in wild.

Strains that belong to the ‘chemical group 1’ produced ascochitine as a major SM (data not shown). Ascochitine was initially identified as a selective antifungal agent (Nakanishi & Oku

213

1969). Interestingly, ascochitine is widely found in species in genera Ascochyta and Phoma, including non-legume plant pathogens, A. hyalospora and P. clematidina (Venkatasubbaiah &

Chilton 1992, Smith et al. 1994), and a marine-derived fungus A. salicorniae (syn.

Stagonosporopsis salicorniae) (Seibert et al. 2006). Therefore, ascochitine may be an ancestral

SM (a plesiomorphic trait) of legume-associated Ascochyta/Phoma spp., but have been lost in some lineages in clades B and C.

Clade B

Strains collected from grasspea (Lathyrus sativus) are phylogenetically and chemically close to A. lentis, and likely the conspecifics. However, ER1415 and AL1 strains isolated from grasspea and lentil, respectively, showed strong host specialization in reciprocal inoculation experiments on their hosts (T.L. Peever, unpublished), suggesting that an ecological speciation may be underway in these sympatric taxa. Also, this study showed that the metabolic profiles of strains from grasspea were more similar to each other (Fig. 5-4). Ascochitine, the signature SM of the ‘chemical group 1’, was not detected in the A. lentis subclade (chemical group 2) and the metabolic feature of clade B was significantly different from those of clade A, despite the close phylogenetic relationship. It was recently reported that an A. lentis strain produce anthraquinones and some other SMs (Andolfi et al. 2013), which have not been found in other Ascochyta and

Phoma spp. so far. In vitro genetic cross experiments showed that interspecific hybrids between

A. lentis and A. viciae-villosae (between the subclades within clade B) can be formed in laboratory conditions (Hernandez-Bello et al. 2006). However, clear distinction in metabolic

214

profiles between A. lentis and A. viciae-villosae suggested that some sorts of isolation barriers have prevented the hybrid formation in nature (Hernandez-Bello et al. 2006).

Clade C

Metabolic profiles of legume-associated Ascochyta/Phoma strains were species-specific and consistent among strains of the same species isolated from different geographical regions and host plants, as best shown in A. rabiei strains. In the first chemometrics analysis, we included A. rabiei strains isolated from cultivated chickpea (Cicer arietinum), annual wild chickpea (C. judaicum) and perennial wild chickpeas (C. montbretti and C. ervoides) (Table 5-1).

Significant genetic differentiation was previously reported between A. rabiei strains isolated from C. arietinum and C. judaicum (Frenkel et al. 2010). Nevertheless, the metabolic profiles of

AR628 and M305 strains isolated from C. arietinum and C. judaicum, respectively, were nearly identical (Fig. 5-2). This indicated that the metabolic feature is conserved among strains in genetically differentiated populations and can be reproducibly used as a species-specific taxonomic character.

Horizontal gene transfer (HGT) is defined as the movement of stable genetic material between different strains or species (Doolittle 1999). Although rare, several HGT events of SM biosynthesis gene clusters between fungal species have been reported (Khaldi et al. 2008, Akagi et al. 2009, Khaldi & Wolfe 2011, Slot & Rokas 2011). It was proposed that solanapyrone biosynthesis gene cluster in A. rabiei was horizontally transferred from Al. solani or a distant

215

yet-unknown-species as solanapyrone production is unique to A. rabiei among closely related

Ascochyta spp. and the homologous cluster genes between A. rabiei and Al. solani share a high degree of DNA sequence similarity (> 97%) (Kim et al. 2015). Such HGT of SM biosynthesis gene cluster between species may blur species boundaries in chemotaxonomic classification.

However, the derived SMs as autapomorphic traits in a cladistic sense seem to make A. rabiei strains even more distinguishable from closely related species by metabolic feature.

Strains ID1A and ID3A isolated from spotted locoweed (Astragalus lentiginosus) and

G11 strain from tiny vetch (Vicia hirsuta) appear to be conspecific, given the close phylogenetic and chemotaxonomic relationship. The two strains from spotted locoweed appear to be host-specific only causing disease on spotted locoweed but not on other legumes (Habibi et al.

2015). It needs to be confirmed that whether the strains can cause disease on tiny vetch, and conversely, whether strain G11 can cause disease on spotted locoweed. The distinct metabolic feature and DNA sequences may indicate that the strains represent a novel Ascochyta or Phoma species. In the two independent chemometrics analyses, the metabolic features of these strains were always clustered with those of P. medicaginis strains (AS1 and AS4), consistent with their closer phylogenetic relationship. These two lineages both produce one-celled conidia, while A. rabiei strains produce a mixture of one-celled and two-celled conidia, showing an example of the polyphyletic nature of this morphological character.

216

Clade D

Based on DNA sequences, differentiating A. pinodes from A. pinodella has been difficult (Onfroy et al. 1999, Barve et al. 2003); however, the former is homothallic but the latter is heterothallic (Bowen et al. 1997, Onfroy et al. 1999). The metabolic features of the two closely related species were also highly similar to each other (Figs. 5-2 and 5-4). The high level of chemical similarity between two species might be resulted from gene exchange possibly via hybrid formation. A putative hybridization or introgression status of strain MP19 needs to be confirmed in future studies. Pinolidoxin is a major SM produced by both species and known to be phytotoxic (Evidente et al. 1993). Structurally similar compounds including herbarumin II are also produced by P. herbarum (Fausto Rivero-Cruz et al. 2000). The dienoate group of pinolidoxin at the lactone ring (Fig. 5-2) can undergo enzymatic hydrolysis, yielding the alcohol derivative, which is herbarumin II. Both pinolidoxin and herbarumin II were found in culture of

A. pinodes (Cimmino et al. 2012). However, pinolidoxin has not yet been found in P. herbarum, suggesting a divergence of the shared biosynthetic pathway in these two evolutionary lineages.

Concluding Remarks

Chemotaxonomy in conjunction with phylogenetic analysis may provide novel insights into species and chemical ecology (Becerra 1997). The genome speaks of how many and what kinds of metabolites could potentially be produced in an organism. However, it is the phenotypic metabolites actually being produced and used during fungal life cycle that are biologically

217

informative and useful for chemotaxonomy. The legume-associated Ascochyta and Phoma spp. are primarily plant pathogens, but at the same time, good saprobes as with most of fungi. The oat kernel cultures supported a uniform growth of the strains as well as consistent production of SMs that were previously described in legume-associated Ascochyta and Phoma species. The non-targeted metabolite profiling employs simple, fast, and inexpensive growth and extraction methods and relatively short LC-MS analysis time per sample (< 15 min), which enables high-throughput chemotaxonomic studies on highly diverse fungal families with minimum available information on their metabolite production.

ACKNOWLEDGMENTS

I am grateful to Drs. J. Davidson (South Australian Research and Development Institute,

Australia) and A. Infantino (Roma Tre University, Italy) for providing Ascochyta/Phoma strains.

218

Table 5-1. List of Ascochyta and Phoma strains used in this study Host Fungal Species Isolate Code a Location Collector Year Reference

Cicer arietinum Ascochyta rabiei AR628 (201622) Malikiye, Syria J. Geistlinger 1995 Peever et al., 2007 – – AR21 (76502) Idaho.USA W.J. Kaiser 1996 – Cicer montbretti – AR738 (201625) Bulgaria W.J. Kaiser 1996 – Cicer ervoides – Georgia-10 Ateni, Georgia W.J. Kaiser 2004 – Cicer judaicum – M305 Ramot Menashe, Israel O. Frenkel 2005 Frenkel et al., 2010 Lens culinaris A. lentis AL1 (96419) Australia W.J. Kaiser N/A Peever et al., 2007 – – AL2 (96420) Brazil W.J. Kaiser N/A – – – AL3 (46979) Canada W.J. Kaiser N/A – – – AL6 (46982) Russia W.J. Kaiser N/A – – – AL11 (46980) India W.J. Kaiser N/A –

219 Pisum sativum A. pisi AP1 (201617) Bulgaria W.J. Kaiser 1992 – –

– – AP4 (210620) Saskatoon, Canada B. Gossen N/A – – AP5 Saskatoon, Canada B. Gossen N/A – Vicia faba A. fabae AF1 (96418) Saskatoon, Canada B. Vandenberg 1992 – – – AF4 (96409) Iran C. Bernier 1992 – – – AF8 (201613) Cambridge, UK R.A.A. Morrall 1992 – Pisum sativum A. pinodes MP1 (201628) Oregon, USA J. Baggett N/A – – – MP2 (201629) Ireland J. Kraft N/A – – – MP19 (201632) Argentina D. Webster 1996 – Lens culinaris A. pinodella PMP1 N/A W.J. Kaiser N/A Chilvers et al., 2009 – – PMP3 (58660) Washington, USA W.J. Kaiser 1996 Peever et al., 2007 Cicer arietinum – PMP4 (58662) Washington, USA W.J. Kaiser 1996 – Vicia villosa A. viciae-villosae. AV1 Washington, USA W.J. Kaiser 1983 – Vicia lathyroides Ascochyta sp. AV8 Washington, USA W.J. Kaiser 1994 – Vicia grandiflora – AV11 Connecticut, USA K.T. Leath 1994 –

Table 5-1. Continued Host Fungal Species Is olate Code a Location Collector Year Reference

Vicia grandiflora Ascochyta sp. Georgia-2 Dedoplistskaro, Georgia W.J. Kaiser 2004 Peever et al., 2007 – – Georgia-3 Dedoplistskaro, Georgia W.J. Kaiser 2004 – Vicia sepium – Georgia-4 Manglesi, Georgia W.J. Kaiser 2004 – Vicia grandiflora – Georgia-8 Ateni, Georgia W.J. Kaiser 2004 – – – Georgia-13 Ateni, Georgia W.J. Kaiser 2004 – Vicia cordata – Georgia-16 Tserovani, Georgia W.J. Kaiser 2004 – Lathyrus sativus – ER1415 Salerno, Italy A. Infantino 2007 N/A – – ER1478 Salerno, Italy A. Infantino 2008 – – – ER1813 Salerno, Italy A. Infantino 2013 – Vicia hirsuta – Georgia-11 Ateni, Georgia W.J. Kaiser 2004 Chilvers et al., 2006

220 Astragalus sp. – ID1A Idaho, USA W.J. Kaiser 2005 Habibi et al., 2015 Chilvers et al., 2009

– – ID3A Idaho, USA W.J. Kaiser 2005 Lupinus sp. – ID4A Idaho, USA W.J. Kaiser 2005 N/A Medicago sativa Phoma medicaginis AS1 Washington, USA T.L. Peever 2001 Chilvers et al., 2009 – – AS4 Washington, USA T.L. Peever 2001 Akamatsu et al., 2008 Pisum sativum P. koolunga PK4 (FT07010)b Australia J.A. Davidson 2007 Davidson et al., 2009 Rosa multiflora P. herbarum PH (CBS615.75) Netherlands G.H. Boerema 1973 Chilvers et al., 2009 Rumex arifolius Didymella exigua DE (CBS183.55) France E. Müller 1953 – Solanum Alternaria solani ALS1 Washington, USA N/A 2011 Kim et al., 2015 tuberosum – ALS2 Idaho, USA N/A 2010 – a American Type Culture Collection accession numbers (or CBS codes for Phoma herbarum and Didymella exigua) in parentheses, where applicable, and the strain codes for Georgia-XX strains were shortened as G-XX in the text. b Phoma koolunga FT07010 strain used in the previous study (Davidson et al., 2009) was recoded as PK4 in this study. N/A = not available. Dashe indicate the same as the above row.

Figure 5-1. Colony morphology of Ascochyta and Phoma species. Strains were indicated by their respective codes. Detailed information of the strains is in Table 5-1. Two Alternaria solani strains (ALS1 and ALS2) were also included in this study as an outgroup for the

Ascochyta/Phoma taxonomy. Photos were taken 1 week after incubation on PDA.

221

222

Figure 5-2. Consistent metabolic profiles between strains within the same species. A, PCA scores plot (upper), PC 2 versus PC 3 showing the variation in the metabolic profiles from 21 fungal strains: AP1, AP4, and AP5 (closed circles); ID4A (cross); PK4 (asterisk); ID1A, ID3A, and G11 (open circles); AS1 and AS4 (diamonds); MP1, MP2, and MP19 (open squares); PMP1,

PMP3, and PMP4 (closed squares); AR628, AR21, AR738, G10, and M305 (triangles). PLS scores plot (lower), showing the supervised separation of the strains. B, Dendrogram of the 21 strains based on chemical similarity, based on UPGMA clustering method. The heatmap of respective metabolites corresponding to each strain is presented. C, Overlaid total ion chromatograms (TICs) of A. pisi strains [AP1 (black) and AP5 (red)] in upper panel, A. rabiei strains [AR628 (black) and M305 (red)] in middle panel, and A. pinodes strains [MP1 (black) and MP2 (red)] in lower panel. Chemical structure of ascochitine (upper), solanapyrone A

(middle), or pinolidoxin (lower) was shown within each panel.

223

Figure 5-3. Chromatograms relating to the detection of ascochitine and pinolidoxin. A, Base peak ion chromatograms (BPIs) of AP5, PK4, and ID4A strains. Peaks corresponding to ascochitine were indicated by observed m/z values and retention times (in parenthesis). B, BPIs of MP1 and PMP3 strains. Peaks corresponding to pinolidoxin were indicated by observed m/z values and retention times (in parenthesis).

224

Figure 5-4. Chemical diversity and grouping observed in Ascochyta/Phoma species. A,

Dendrogram of 40 fungal strains based on chemical similarity. Numbers and color coding indicate the five main chemical groups according to UPGMA clustering method at a 0.75 distance threshold. B, PLS scores plot, PC 1 versus PC 2, showing the supervised separation of the major chemical groups and singletons (grey open circles).

225

226

Figure 5-5. Bayesian phylogeny of Ascochyta/Phoma species. The phylogeny was estimated from the combined dataset of glyceraldehyde 3-phosphate dehydrogenase, chitin synthase, and translation elongation factor alpha for Ascochyta and Phoma spp. isolated from various legume plants. The phylogeny was rooted by Alternaria solani strains. Numbers at major nodes indicate

Bayesian posterior probabilities (PP) of sampling the node among 6,000 trees (2 million total generations of the MCMC chains). Clades inferred based on PP greater than or equal to 95%.

The five major chemical groups detected in the hierarchical clustering analysis are traced onto the phylogeny, showing distinct metabolic features in each evolutionary lineage. Gray lines at the terminal clades indicate ungrouped strains (singletons). Branch lengths are proportional to the inferred amount of evolutionary change and the scale represents 0.05 nucleotide substitutions per site.

227

REFERENCES

Akagi, Y., Akamatsu, H., Otani, H., and Kodama, M. 2009. Horizontal chromosome transfer, a mechanism for the evolution and differentiation of a plant-pathogenic fungus. Eukaryotic Cell 8:1732-1738.

Akamatsu, H.O., Chilvers, M.I., and Peever, T.L. 2008. First report of spring black stem and leaf spot of alfalfa in Washington state caused by Phoma medicaginis. Plant Disease 92:833-833.

Aveskamp, M.M., Gruyter, J.d., and Crous, P.W. 2008. Biology and recent developments in the systematics of Phoma, a complex genus of major quarantine significance. Fungal Diversity 31:1-18.

Aveskamp, M.M., Woudenberg, J.H.C., De Gruyter, J., Turco, E., Groenewald, J.Z., and Crous, P.W. 2009. Development of taxon-specific sequence characterized amplified region (SCAR) markers based on actin sequences and DNA amplification fingerprinting (DAF): A case study in the Phoma exigua species complex. Molecular Plant Pathology 10:403-414.

Becerra, J.X. 1997. Insects on plants: Macroevolutionary chemical trends in host use. Science 276:253-256.

228

Benedict, R.G. 1970. Chemotaxonomic relationships among basidiomycetes. Advances in Applied Microbiology 13:1-23.

Bennett, J.W. 1983. Secondary metabolism and differentiation in mycelia fungi. Pages 1-32 in Secondary metabolism and differentiation in fungi (Bennett J.W., and Ciegler A., eds.). New York, USA: Marcel Dekker Inc.

Berbee, M.L., Pirseyedi, M., and Hubbard, S. 1999. Cochliobolus phylogenetics and the origin of known, highly virulent pathogens, inferred from ITS and glycealdehyde-3-phosphate dehydrogenase gene sequences. Mycologia 91:964-977.

Bertini, S. 1956. Su di un composto ad antibiotica prodotto da Ascochyta pisi Lib. Annali Sperimentaz Agraria (Roma) 11:545-556.

Boerema, G.H., and Bollen, G.J. 1975a. Conidiogenesis and conidial septation as differentiating criteria between Phoma and Ascochyta. Persoonia 8:111-144.

Boerema, G.H., and Bollen, G.J. 1975b. Conidiogenesis and conidial septation ads differentiating criteria between Phoma and Ascochyta. Persoonia 8:111-144.

Bose, U., Hewavitharana, A.K., Vidgen, M.E., Ng, Y.K., Shaw, P.N., Fuerst, J.A., and Hodson, M.P. 2014. Discovering the recondite secondary metabolome spectrum of Salinispora species: A study of inter-species diversity. PLoS ONE 9:e91488.

Bowen, J.K., Lewis, B.G., and Matthews, P. 1997. Discovery of the teleomorph of Phoma medicaginis var. pinodella in culture. Mycological Research 101:80-84.

Carbone, I., and Kohn, L.M. 1999. A method for designing primer sets for speciation studies in filamentous ascomycetes. Mycologia 91:553-556.

Chen, W., Sharma, H.C., and Muehlbauer, F.J. 2011. Compendium of Chickpea and Lentil Diseases and Pests, St. Paul, MN, APS Press.

229

Chilvers, M.I., Horton, T.L., Peever, T.L., Kaiser, W.J., and Muehlbauer, F.J. 2006. First report of Ascochyta blight of Vicia hirsuta (hairy tare) in the Republic of Georgia caused by Ascochyta sp. Plant Disease 90:1555-1555.

Chilvers, M.I., Rogers, J.D., Dugan, F.M., Stewart, J.E., Chen, W., and Peever, T.L. 2009. Didymella pisi sp. nov., the teleomorph of Ascochyta pisi. Mycological Research 113:391-400.

Cimmino, A., Andolfi, A., Fondevilla, S., Abouzeid, M.A., Rubiales, D., and Evidente, A. 2012. Pinolide, a new nonenolide produced by Didymella pinodes, the causal agent of Ascochyta Blight on Pisum sativum. Journal of Agricultural and Food Chemistry 60:5273-5278.

Davidson, J.A., Hartley, D., Priest, M., Krysinska-Kaczmarek, M., Herdina, McKay, A., and Scott, E.S. 2009. A new species of Phoma causes ascochyta blight symptoms on field peas (Pisum sativum) in South Australia. Mycologia 101:120-128.

Doolittle, W. 1999. Lateral genomics. Trends Cell Biol 9:M5-M8.

Druzhinina, I.S., Kopchinskiy, A.G., Komoń, M., Bissett, J., Szakacs, G., and Kubicek, C.P. 2005. An oligonucleotide barcode for species identification in Trichoderma and Hypocrea. Fungal Genetics and Biology 42:813-828.

Dunn, W.B., Goodacre, R., Neyses, L., and Mamas, M. 2011. Integration of metabolomics in heart disease and diabetes research: Current achievements and future outlook. Bioanalysis 3:2205-2222.

Evidente, A., Lanzetta, R., Capasso, R., Vurro, M., and Botralico, A. 1993. Pinolidoxin, a phytotoxic nonenolide from Ascochyta pinodes. Phytochemistry 34:999-1003.

Frisvad, J.C., Andersen, B., and Thrane, U. 2008. The use of secondary metabolite profiling in chemotaxonomy of filamentous fungi. Mycological Research 112:231-240.

230

Frisvad, J.C., Larsen, T.O., de Vries, R., Meijer, M., Houbraken, J., Cabañes, F.J., Ehrlich, K., and Samson, R.A. 2007. Secondary metabolite profiling, growth profiles and other tools for species recognition and important Aspergillus mycotoxins. Studies in Mycology 59:31-37.

Galagan, J.E., Calvo, S.E., Cuomo, C. et al. 2005. Sequencing of Aspergillus nidulans and comparative analysis with A. fumigatus and A. oryzae. Nature 438:1105-1115.

Habibi, A., Peever, T.L., Kim, W., Chilvers, M.I., Chen, W., Kaiser, W.J., and Muehlbauer, F.J. 2015. First report of Ascochyta blight of spotted locoweed (Astragalus lentiginosus) in Idaho caused by Ascochyta sp. Plant Disease DOI: 10.1094/pdis-02-15-0174-pdn.

Habibi, A., Frederick, Z.A., Koberstein, S.L., Liu, L., Nazarov, T., Salamone, A.L., Wheeler, D.L., Carris, L.M., and Peever, T.L. How many species of Ascochyta cause Ascochyta blights of cultivated legumes and their wild relatives? Poster session presented at: Mycological Society of America; 2014 Jun 8-12; East Lansing, MI.

Hodson, M.P., Dear, G.J., Roberts, A.D., Haylock, C.L., Ball, R.J., Plumb, R.S., Stumpf, C.L., Griffin, J.L., and Haselden, J.N. 2007. A gender-specific discriminator in Sprague-Dawley rat urine: The deployment of a metabolic profiling strategy for biomarker discovery and identification. Analytical Biochemistry 362:182-192.

Ichihara, A., Tazaki, H., and Sakamura, S. 1983. Solanapyrones A, B and C, phytotoxic metabolites from the fungus Alternaria solani. Tetrahedron Letters 24:5373-5376.

Kaiser, W.J. 1997. Inter- and intranational spread of ascochyta pathogens of chickpea, faba bean, and lentil. Canadian Journal of Plant Pathology 19:215-224.

Kaiser, W.J., Wang, B.C., and Rogers, J.D. 1997. Ascochyta fabae and A. lentis: Host specificity, teleomorphs (Didymella), hybrid analysis, and taxonomic status. Plant Disease 81:809-816.

Khaldi, N., Collemare, J., Lebrun, M., and Wolfe, K. 2008. Evidence for horizontal transfer of a secondary metabolite gene cluster between fungi. Genome Biology 9:R18.

231

Krug, D., Zurek, G., Schneider, B., Garcia, R., and Müller, R. 2008. Efficient mining of myxobacterial metabolite profiles enabled by liquid chromatography-electrospray ionisation-time-of-flight mass spectrometry and compound-based principal component analysis. Analytica Chimica Acta 624:97-106.

Leath, K.T. 1994. Foliar blight of big-flower vetch caused by Ascochyta fabae f. sp. vicia. Plant Disease 78:637-639.

Monte, E., Bridge, P.D., and Sutton, B.C. 1991. An integrated approach to Phoma systematics. Mycopathologia 115:59-103.

Nakanishi, T., and Oku, H. 1969. Mechanism of selective toxicity: Absorption and detoxicationof an antibiotic, ascochitine, by sensitive and insensitive fungi. Phytopathology 59:1563-1565.

Nierman, W.C., Pain, A., Anderson, M.J. et al. 2005. Genomic sequence of the pathogenic and allergenic filamentous fungus Aspergillus fumigatus. Nature 438:1151-1156.

Nilsson, R.H., Kristiansson, E., Ryberg, M., Hallenberg, N., and Larsson, K.-H. 2008. Intraspecific ITS variability in the Kingdom Fungi as expressed in the international sequence databases and its implications for molecular species identification. Evolutionary Bioinformatics Online 4:193-201.

Oliver, S.G., Winson, M.K., Kell, D.B., and Baganz, F. 1998. Systematic functional analysis of the yeast genome. Trends in Biotechnology 16:373-378.

Onfroy, C., Tivoli, B., Corbiere, R., and Bouznad, Z. 1999. Cultural, molecular and pathogenic variability of Mycosphaerella pinodes and Phoma medicaginis var. pinodella isolates from dried pea (Pisum sativum) in France. Plant Pathology 48:218-229.

Pedras, M.C., and Biesenthal, C.J. 2000. HPLC analyses of cultures of Phoma spp.:

232

Differentiation among groups and species through secondary metabolite profiles. Canadian Journal of Microbiology 46:685-691.

Peever, T.L. 2007. Role of host specificity in the speciation of Ascochyta pathogens of cool season food legumes. European Journal of Plant Pathology 119:119-126.

Pel, H.J., de Winde, J.H., Archer, D.B. et al. 2007. Genome sequencing and analysis of the versatile cell factory Aspergillus niger CBS 513.88. Nat Biotech 25:221-231.

Petrak, F. 1924. Mykologische Notizen VII. Annales Mycologici 22.

Rabie, C.J., Rensburg, S.J.v., Watt, J.J.v.d., and Lubben, A. 1975. Onyalai - the possible involvement of a mycotoxin produced by Phoma sorghina in the aetiology. South African Medical Journal 57:1647-1650.

Ronquist, F., Teslenko, M., van der Mark, P., Ayres, D.L., Darling, A., Höhna, S., Larget, B., Liu, L., Suchard, M.A., and Huelsenbeck, J.P. 2012. MrBayes 3.2: Efficient Bayesian phylogenetic inference and model choice across a large model space. Systematic Biology 61:539-542.

Sørensen, J.L., Aveskamp, M.M., Thrane, U., and Andersen, B. 2010. Chemical characterization of Phoma pomorum isolated from Danish maize. International Journal of Food Microbiology 136:310-317.

Seifert, K.A. 2009. Progress towards DNA barcoding of fungi. Molecular Ecology Resources 9:83-89.

Seifert, K.A., Samson, R.A., deWaard, J.R., Houbraken, J., Lévesque, C.A., Moncalvo, J.-M., Louis-Seize, G., and Hebert, P.D.N. 2007. Prospects for fungus identification using CO1 DNA barcodes, with Penicillium as a test case. Proceedings of the National Academy of Sciences of the United States of America 104:3901-3906.

233

Slot, J.C., and Rokas, A. 2011. Horizontal transfer of a large and highly toxic secondary metabolic gene cluster between fungi. Current biology 21:134-139.

Smedsgaard, J., and Nielsen, J. 2005. Metabolite profiling of fungi and yeast: from phenotype to metabolome by MS and informatics. Journal of Experimental Botany 56:273-286.

Summerbell, R.C., Moore, M.K., Starink-Willemse, M., and Van Iperen, A. 2007. ITS barcodes for Trichophyton tonsurans and T. equinum. Medical Mycology 45:193-200.

Sutton, B.C. 1980. The Coelomycetes, Fungi imperfecti with pycnidia, acervuli and stromata. 1st edn. Commonwealth Mycological Institute, U.K.

Sutton, B.C., and Sandhu, D.K. 1969. Electron microscopy of conidium development and secession in Cryptosporiopsis sp., Phoma fumosa, Melanconium bicolor and M. apiocarpum. Canadian Journal of Botany 47:745-749.

Tosch, W., Lanthaler, K., Boote, V., Stretz, D., Robson, G.D., Geiger, E., and Drucker, D.B. 2006. Molecular species of phosphatidylethanolamine from continuous cultures of Saccharomyces pastorianus syn. carlsbergensis strains. Yeast 23:75-82.

Trapero-Casas, A., and Kaiser, W.J. 1992. Development of Didymella rabiei, the teleomorph of Ascochyta rabiei, on chickpea straw. Phytopathology 82:1261-1266.

Velázquez, E., del Villar, M., Grondona, I., Monte, E., and González-Villa, T. 2006. Ultrastructural and chemotaxonomic analysis of a xylanolytic strain of Cryptococcus adeliensis isolated from sheep droppings in Spain. Archives of Microbiology 186:195-202.

Venkatasubbaiah, P., and Chilton, W.S. 1992. Phytotoxins of Ascochyta hyalospora, causal agent of lambsquarters leaf spot. Journal of Natural Products 55:461-467.

234

CHAPTER SIX

SUMMARY

Ascochyta blight of chickpea caused by Ascochyta rabiei is the most devastating disease of chickpea especially in areas where cool, humid weather persists during the cropping season.

Although cultivars with improved resistance to the disease have been developed in breeding programs worldwide, the newly introduced resistance of cultivars can be overcome by emergence of a new pathotype or strains with enhanced virulence. Little is known about the mechanism by which the pathogen initiates and establishes infection in host tissues. Until this study solanapyrone produced by A. rabiei was the solely known potential virulence factor as it exhibits profound phytotoxicity to chickpea.

Under the present research the role of solanapyrone in infection processes of A. rabiei was examined. Solanapyrone A exhibited a broad spectrum of phytotoxicity to various legumes, whereas solanapyrone C and pro-solanapyrone II-diol (the immediate precursor of solanapyrones) showed a marginal and zero toxicity, respectively. In A. rabiei, the solanapyrone gene cluster was identified, which comprises six genes (sol1–sol6). Among the cluster genes, sol5 and sol4 genes were found to encode the last step enzyme and a pathway-specific regulator for solanapyrone biosynthesis, respectively. Solanapyrone-minus mutants (sol5- and sol4-deletion mutants) were generated from three A. rabiei strains of different pathotypes and geographic origins. No defect in growth and development was observed in the resulting mutants.

However, these mutants were equally virulent as wild-type progenitors to chickpea cultivars with

235

varying degree of resistance to the disease. These results indicate that solanapyrones are neither a host-selective toxin nor a virulence factor of A. rabiei.

Genome sequence data of A. rabiei revealed that the solanapyrone biosynthesis cluster is placed in a subtelomeric region. The solanapyrone gene cluster is situated in a highly dynamic genomic environment where many pseudogenes and transposable elements were found. The flanking regions of the solanapyrone gene cluster were degenerated presumably by repeat-induced point mutations, but the regions in which the solanapyrone gene cluster resides appear to be resistant to such mutations, suggesting that solanapyrones are required for the pathogen biology and may confer fitness advantages during the pathogen life cycle other than parasitic phase.

To understand the role of solanapyrones in pathogen biology, the spatial and temporal patterns of solanapyrone production were investigated. LC-MS analyses and gene expression studies of key biosynthetic enzymes confirmed that the production of solanapyrones was highly growth stage-dependent. Transcripts of sol1, sol4 and sol5 genes were detected in small quantity during infection processes, but abundant during growth on chickpea straws (saprobic growth).

Concomitantly, solanapyrone A can be detected from chickpea straws colonized by A. rabiei. In addition, a GFP-expressing reporter strain that reflects solanapyrone biosynthesis exhibited a localized GFP expression pattern around pycnidia, which is in consistent with the previous observations that solanapyrone production is associated with conidiation. The antifungal activities of solanapyrone A were evident against several saprobic fungi that were isolated from

236

chickpea debris left in a field. In addition to the saprobic stage-dependent production and significant antifungal activities, the specific association of solanapyrone production with pycnidia formation is suggestive of its protective role against competing microorganisms during saprobic growth of A. rabiei. Indeed, co-culturing of the saprobic fungi either with A. rabiei wild-type strains or solanapyrone-minus mutants showed that solanapyrones can suppress the growth of the potential fungal competitors during saprobic growth, but solanapyrone-minus mutants cannot. These results suggest that solanapyrone A plays an important role in competition, and presumably in survival of the pathogen in nature.

The initial objective of this study was to prove the involvement of solanapyrones in pathogenicity or virulence and to use the toxins as a screening tool for breeding lines.

Solanapyrone A has been considered a key virulence factor for decades without direct evidence, but in this study, is proven to be not. Exonerating soalanpyrone A as a virulence factor will fascilitate research toward identification of bona fide virulence factors in the chickpea-A. rabiei interactions. Although the initial objective cannot be accomplished, this series of studies yielded a better understanding of survival strategy of the pathogen in nature. The information would be potentially useful to prevent recurrence of the disease in the following crop seasons.

237

APPENDIX A

Supplementary Table. Primers used in this study. Primer Sequence 5′ to 3′ Description PCR a primer 5-1 GCTTGTGTAACCTCTGAA flanking region upstream of 800 TCTCA sol5 FPb primer 5-2 TTGACCTCCACTAGCTCC flanking region upstream of AGCCAAGCCCTTGTCAG sol5 RPc with overhang ATTGGCGTGATGTCA sequence of 5′end of hph primer 5-3 GGGCAAAGGAATAGAGT flanking region downstream of 1,354 AGATGCCGACCGTCTAG sol5 RP with overhang GCTATCGGATTCAAGGAA sequence of 3′end of hph primer 5-4 GCGGATGGACGATCTGG flanking region downstream of TTGATC sol5 FP primer 5-7 AATTCCAATACCACTCGT flanking region upstream of ACCATG sol5 FP nested primer primer 5-8 GATGTCTTCGAAGAAGT flanking region downstream of ACTG sol5 RP nested primer primer 5-11 AAAAGAACACACGAGA FP upstream of sol5, 2,330 GCTCATCC verification primer (in WT) primer 5-12 CTGATCTATGTCCCTCTTT RP downstream of sol5, 1952 GATCC verification primer (in Δsol5 ) primer 4-1 GCGTCGATGCAGATTCTT flanking region upstream of 838 ACTTA sol4 FPb primer 4-2 TTGACCTCCACTAGCTCC flanking region upstream of AGCCAAGCCCATCGTGT sol4 RPc with overhang TGTACTCATGTGGTT sequence of 5′end of hph primer 4-3 GCAAAGGAATAGAGTAG flanking region downstream of 1,094 ATGCCGACCGTTAGGCGA sol4 RP with overhang TGGAGTAGAGAGAAG sequence of 3′end of hph primer 4-4 CATCCAAGCGAGCGAGA flanking region downstream of ATGGAG sol4 FP primer 4-7 AGAGTAAGCGGAGCGGA flanking region upstream of TCAAAG sol4 FP nested primer primer 4-8 CCAGAAGTGGACCTCTG flanking region downstream of ATGCGA sol4 RP nested primer primer 4-11 GTGTGTCCACCTGCCTAT FP upstream of sol4, 1,950 GTATC verification primer (in wild primer 4-12 CACGCTTCTTCCCGTGCA RP downstream of sol4, type) TCTG verification primer 1,581 (in sol4)

239

Supplementary Table. Continued Primer Sequence 5′ to 3′ Description PCR a HYG-F GGCTTGGCTGGAGCTAG FP to amplify hph from 1,372 TGGAG pDWJ5 HYG-R CGGTCGGCATCTACTCTA RP to amplify hph from TTCCTT pDWJ5 YG-F CGATGTAGGAGGGCGTG FP coding region of hph for GATATGTCC overlapping extension, pair with primer-8 HY-R GTATTGACCGATTCCTTG RP coding region of hph for CGGTCCGAA overlapping extension, pair with primer-7 POLL-F TATCGTTGTCACCCATCC Real-time RT-PCR for 179 TG At1g10520 POLL-R GCTCCTGACCAGGATAA GTG EF1a-F CACCACTGGAGGTTTTG Real-time RT-PCR for AGG At5g60390 EF1a-R TGGAGTATTTGGGGGTG GT Sol4BamH5 GGATCCATGATGCCGTCC FP at ATG with BamHI cut 2,029 ACCCTCATC site for 4OE construction Sol4Hind3 AAGCTTTCGCAAGGTCT RP at +276 bp from stop GGGAAATTC codon with HindIII cut site for 4OE construction Sol4OE5 GTTGTGTGTCCACCTGCC FP at +92 bp from ATG for 2,187 (with TATG confirmation of 4OE strains Sol4OE5 and Sol4OE3 CGGGAAGCTGCGAGAAG 4OE RP at +354 bp from stop Sol4OE3 ATAAG codon for confirmation of primer pair) 4OE strains DW38 AGATGGTCAACGCTGCTT pelA promoter primer for ≈ 2,200 (with AC confirmation of 4OE strains, DW38 and 141 bp away from PvuII site Sol4OE3 primer pair) Sol1RT-F TTGGTATTGGTTCGCTCG RT-PCR for sol1 530 AGGT (615, Sol1RT-R TCAACAGCGGTTGACAT for gDNA or CCTCT pre-mRNA)

240

Supplementary Table. Continued Primer Sequence 5′ to 3′ Description PCR a Sol2RT-F CACATCTCCATGGCTTTG RT-PCR for sol2 390 GCTC (470, Sol2RT-R GTTCGCTGCTTAGCACCC for gDNA or AGAA pre-mRNA) Sol3RT-F GTTCGCCTTGATGGCAA RT-PCR for sol3 572 GACTG (622, Sol3RT-R CGCGCATCCAGAGGATG for gDNA or TTCAA pre-mRNA) Sol4RT-F TGGATCAACCAGATCGTC RT-PCR for sol4 522 CATC (588, Sol4RT-R CTTACGGTGCAGTACGCA for gDNA or TCTA pre-mRNA) Sol5RT-F AGAACCCAGCGTGCATC RT-PCR for sol5 534 TATAC (631, Sol5RT-R GATCATGGAACCTCCCCA for gDNA or TATC pre-mRNA) Sol6RT-F GCAAAGTGCTAACACCC RT-PCR for sol6 498 GCTCT (606, Sol6RT-R CGTTTAGCTGTTCTAGGC for gDNA or TTGG pre-mRNA) Actin1RT-F CAATGGTTCGGGTATGTG RT-PCR for Actin1 482 CAAG Actin1RT-R GAAGAGCGAAACCCTCG TAGAT Sol1realt-F GTTGGCATGGGCTGTAGA Real-time RT-PCR for sol1 129 TG Sol1realt-R ATGGTGGAATCCCTTGCG AG Sol2realt-F CACTACATGCTCGATGAA Real-time RT-PCR for sol2 123 TGC Sol2realt-R GAACATTGGCACACCGA AG Sol3realt-F CGTTCAGGTTAATCACCT Real-time RT-PCR for sol3 126 GG Sol3realt-R CAGAAGTGGACCTCTGA TG

241

Supplementary Table. Continued Primer Sequence 5′ to 3′ Description PCR a Sol4realt-F CAACCTTCGCCTTGCAA Real-time RT-PCR for sol4 140 AAG Sol4realt-R GAAGTACTCTCTCGGTG AAC Sol5realt-F CGCGAACAATTTTGGCAT Real-time RT-PCR for sol5 162 TG Sol5realt-R ATGGCAGACTTCTTGTCC AG Sol6realt-F CAAGCAATACGGACCGG Real-time RT-PCR for sol6 135 TG Sol6realt-R GGAAACGAGATGCATCG ATAG ORF2realt-F ATTGGACCCGCACCAAAT Real-time RT-PCR for ORF2 140 AC ORF2realt-R GTCTTCATGGGATCTCCA AGG ORF3realt-F CCGCTCTAGCGATAAGA Real-time RT-PCR for ORF3 136 AGG ORF3realt-R CGTAGCAACCTGATGCA AC ORF10realt- CACTGCCATCCTTGAGAC Real-time RT-PCR for ORF10 127 F AG ORF10realt- GAGACTTCGCTGTTCTTG R CC Peudo6 GCACTTTGACAGGCATAC Real-time RT-PCR for 144 realt-F AAAG pseudogene 6 Peudo6 GAGTGTGGAGGCATGCA realt-R TAG PKS1realt-F CAACATGTCTCCACGTG Real-time RT-PCR for PKS1 130 AAG PKS1realt-R AATGCGGTTCAGCTTTGT GG PKS2realt-F GAACCTTGCTGGTGCATC Real-time RT-PCR for PKS2 123 G PKS2realt-R GCTTTGGACAGCGACTT GAG

242

Supplementary Table. Continued Primer Sequence 5′ to 3′ Description PCR a Actin1realt-F GTATCATGATCGGTATGG Real-time RT-PCR for Actin1 134 GACAG Actin1realt-R CCAGATCTTCTCCATGTC GTCC GFP5-BamH GGATCCGTGAGCAAGGG FP b to amplify GFP gene from 1,108 I CGAGGAGCTGTT pTRI5::GFP GFP3-XbaI TCTAGAAGTCCTCGTGTA RP c to amplify GFP gene from CTGTGTAAGC pTRI5::GFP Sol5P5-KpnI GGTACCGCTTGTGTAACC FP to amplify a sol5 promoter 951 TCTGAATCTC region Sol5P3-Bam GGATCCTGCTGCAGGGT RP to amplify a sol5 promoter H TTATTGCATTC region 5F-GFP5 GTCTATGCTGAGATCGTT FP for 5′ flanking region 1,192 CACG confirmation of Sol5prom::GFP 5F-GFP3 GGTGGTGCAAATGAACT RP at GFP gene, 5′ flanking TCAG region confirmation of Sol5prom::GFP 3F-AMP5 GTACTGAGAGTGCACCAT FP at amp gene, 3′ flanking 1,242 ATGC region confirmation of Sol5prom::GFP 3F-AMP3 GGAAACACGACTGCAGT RP for 3′ flanking region CTG confirmation of Sol5prom::GFP AGTreal-F CGCTGTTTGGTTCCAGG Real-time RT-PCR for acetyl- GTAT glutamatekinase (AGT) gene AGTreal-R CGGCGTAGCGGTCTTGAT A GFPreal-F CTACGGCAAGCTGACCC Real-time RT-PCR for GFP 133 TGAAG gene GFPreal-R GTCGTGCTGCTTCATGTG GTCG DsRedreal-F CGAGGACGTCATCAAGG Real-time RT-PCR for dsRed 140 AGT gene DsRedreal-R CCTTGGTCACCTTCAGCT TG a PCR = PCR product length in basepair; b FP = forward primer; c RP = reverse primer

243

APPENDIX B

245

Figure S2-1. Pairwise alignment of DNA sequences (A), and deduced amino acid sequences (B) of sol5 genes between Al. solani

(As) and A. rabiei (Ar). Identical nucleotides and amino acids are shaded black.

246

Figure S2-2. The correlation details among protons and between protons and carbons of prosolanapyrone II-diol by 300 MHz

1H NMR (A) and 75 MHz 13C NMR (B), analyses. The presence and substitution pattern of the purified compound were

compared with those of the previously described precursor compound, prosolanapyrone II. 1H NMR analysis revealed that the

compound was still keeping prosolanapyrone II skeleton. However, one set of signals corresponding to a double bond

disappeared and two new multiple signals at 4.01 ppm and 3.85 ppm were observed in the 1H NMR spectrum. A peak

corresponding to the terminal methyl group also shifted upfield. This comparison suggested that the terminal double bond of

prosolanapyrone II was modified in the Δsol5 mutant.

Figure S2-3. 300 MHz COSY NMR of prosolanapyrone II-diol. Based on 1H-1H COSY spectrum exhibiting 10 interesting correlations, H9 and H10 can be clearly assigned. This COSY data indicate that the terminal double bond of prosolanapyrone II is oxidized.

247

Figure S2-4. 300 MHz HMBC NMR of prosolanapyrone II-diol. To clarify the structure, the connectivity between protons and carbons was confirmed by interpretation of HMBC data. Key correlations from H-9 to C-9 and from H-9 to C-10 clearly suggested that the terminal double bond of prosolanapyrone II was oxidized and resulted in formation of a vicinal diol.

248

Figure S2-5. Verification of Arabidopsis polymerase lamda (POLL) mutant genotypes. Relative

POLL gene expression in wild-type (Col-0), POLL overexpression transgenic plants (POLLOE), and POLL knockdown mutants (POLLKD), normalized by translation elongation factor (EF1α,

At5g60390) expression levels in respective samples. Error bars are standard deviation of two biological replicates.

249

250

Figure S3-1. Verification of integration of pHNU3PelA plasmid carrying sol4 open reading frame in sol4-overexpression

(4OE) strains. A, Strategy to overexpress the sol4 gene and positioning of the primers used for a diagnostic PCR. B, PCR

analysis of AR628 wild-type (WT) strain and two independent 4OE strains confirmed a single crossover event at the sol4

genomic locus in 4OE strains, using two primer pairs, Sol4OE5 and Sol4OE3 (P1) and DW38 and Sol4OE3 (P2). M1 = 100

bp DNA ladder, M2 = λ DNA cut with HindIII.

251

Figure S3-2. Multiple DNA sequence alignment of Molly transposon from Stagonospora nodorum and degenerated transposons (pseudogene 4–6) located proximal to the solanapyrone gene cluster in A. rabiei. Identical nucleotides are shaded in black. The regions marked with purple line indicate terminal inverted repeat. The coding regions of putative transposase are indicated by arrows. The first putative RIP mutation (CpA to TpA) which introduced premature stop codon in the transposase domain of pseudogene 4, 5, or 6 are indicated by red boxes in comparison to the functional Molly transposon. Note that the four elements are of nearly identical length and include 5′ and 3′ TA insertion sites.

252

253

Figure S3-3. Pairwise sequence alignment of ORF10 and a P450 homologous gene found in

Coniosporium apollinis. A, Pairwise comparison of deduced amino acid sequence of ORF10 and the homologous P450 gene from Coniosporium apollinis (GenBank accession: XP_007784573).

Only a region conserved with ORF10 (240 out of 535 residues) was shown for the P450.

Alignment shaded to indicate similarity with black corresponding to blocks of identical residues, and with grey corresponding to conservative substitutions. Pairwise DNA sequence comparison between upstream of ORF10 and the corresponding region of the hypothetical protein (B), between the coding region of ORF10 and the corresponding region of the hypothetical protein

(D). Note that the coding region of ORF10 has no indels and is more conserved with the hypothetical protein compared with the upstream or downstream regions.

254

255

Figure S3-4. Multiple amino acid sequence alignment of A. rabiei Sol4 (KM244525) and two hypothetical proteins from Aspergillus niger (XP_001397039) and Podospora anserina

(XP_001905776). Alignment was shaded to indicate similarity with black corresponding to blocks of identical residues, and with grey corresponding to conservative substitutions.

Zn(II)2Cys6 (C6) zinc cluster DNA-binding domain is boxed in purple and middle homology region (MHR) is marked with blue line. The coiled coil region in Sol4 amino acid sequence detected by COILS program (Lupas et al., 1991) is boxed in green (a region with prediction values greater than 0.8 in scanning windows of 14). No coiled coil region was predicted for the two hypothetical proteins. Note that Sol4 lacks the C6 zinc cluster DNA-binding domain.

256

257

Figure S4-1. Strategy to construct the pSol5::GFP plasmid. The plasmid pTRI5::GFP was kindly provided by Dr. Wilhelm Schäfer (University of Hamburg) (Ilgen et al., 2009). 1) The

GFP-coding region was derived from plasmid pTRI5::GFP, with primers GFP5-BamHI and

GFP3-XbaI, introducing a BamHI restriction site into the 5′ end of the PCR product. The PCR product (1,108 bp) was cloned into pGEM-T easy cloning vector (Promega). 2) A 951-bp fragment of the sol5 gene locus was amplified with primers Sol5P5-KpnI binding 807 bp upstream and Sol5P3-BamHI binding 132 bp downstream of the sol5 start codon, introducing

KpnI and BamHI restriction sites into the 5′ and 3′ ends of the PCR product, respectively. The

PCR product was cloned into pGEM-T easy cloning vector. 3) The pGEM-T easy vectors containing the GFP-coding region or the 951-bp fragment were double-digested with NdeI and

BamHI, and the 951-bp fragment was isolated and ligated to the upstream of the GFP-coding region in the pGEM-T easy vector. 4) The sol5 promoter fragment fused with the GFP fragment was isolated from the pGEM-T backbone by KpnI-XbaI double digestion. 5) Finally, the insert was ligated to pTRI5::GFP precut with the same restriction enzymes. The resulting vector was designated pSol5::GFP.

258

259

Figure S4-2. Verification of integration of pSol5::GFP and pII99dsRed plasmids into the sol5 genomic locus. A, Schematic illustration for the generation of Sol5prom::GFP strain. The plasmid pSol5::GFP carrying the sol5 promoter–GFP fusion construct was linearized with restriction enzyme Bsu36I in the sol5 promoter sequence. Transformation of the plasmid into

AR628 wild-type strain (WT) results in a single crossover with the targeted genomic locus. Both the sol5 gene and the integrated GFP gene are under the control of the native sol5 promoter. The pII99dsRed was linearized in the amp sequence with PvuI, resulting in a targeted integration into the backbone of pSol5::GFP already present in the genome. These transformations led to the reporter strain with GFP expression driven by the endogenous sol5 promoter and constitutive dsRed expression. B, PCR analysis of Sol5prom::GFP strain confirmed a single crossover event at the targeted genomic locus. Diagnostic PCRs with primer pairs 1 (P1) and primer pair 2 (P2) verified the homologous integrations of the plasmids at the sol5 gene locus, amplifying 1,192-bp and 1,242-bp DNA fragments, respectively. A primer pair (P3: 5F-GFP5 and 3F-AMP3) amplified a 1,122-bp DNA fragment only from the WT strain, due to the long insertion in the reporter strain between the primer pair. M = 200 bp DNA ladder.

260

Figure S4-3. Dose response curves of solanapyrone A. The colony diameter was measured on half-strength PDA media at 5 days after the growth of different fungal species in the presence of solanapyrone A (0, 25, 50, 100, or 200 μM). Ar, As, At, Ep, Bc, and Ua mark Ascochyta rabiei,

Alternaria soliani, Alternaria tenuissima, Epicoccum purpurascens, Botrytis cinerea, and

Ulocladium atrum, respectively. The half maximal inhibitory concentrations (IC50) of solanapyrone A against the fungal species were estimated by fitting the dose-response curve with

R software and the "drc" package (Ritz and Streibig, 2005).

261

262

Figure S4-4. Growth phase-dependent solanapyrone production. Base peak intensity chromatograms of extracts of uninoculated

chickpea stems (upper), inoculated chickpea stems (middle), and autoclaved chickpea straws colonized by A. rabiei strain AR628

(lower) at 3, 5, 7 and 9 days post inoculation (dpi). Most of the detected peaks were identified as chickpea isoflavonoids and their

glucoside conjugates. Note that solanapyrone A was detected only from chickpea straws colonized by A. rabiei as early as 5 dpi.

The numbers above the peaks indicate the following: 1, calycosin; 2, pseudobaptigenin glucoside malonylated; 3, formononetin

glucoside malonylated; 4, maackiain glucoside malonylated; 5, medicarpin glucoside malonylated; 6, biochanin A glucoside

malonylated; 7, pseudobaptigenin; 8, formononetin; 9, maackiain; 10, medicarpin; 11, biochanin A; 12, unknown (286 Da); 13,

unknown (375 Da); 14, solanapyrone A.

263

Figure S4-5. pH shifts and colony morphology of A. rabiei growing on different carbon sources. A. rabiei AR628 strain was grown

on minimal media (buffered to pH 5.8) containing a single carbon source (1%). A, pH changes of the media were monitored before

the start of culture (filled bar) and after two weeks of incubation (open bar). Means of five replicate plates were shown. Bars indicate

standard deviation. B, Representative pictures of AR628 strain growing on different carbon sources. C, Enlarged views of plates

shown in B to highlight pycnidia formation.

Figure S5-1. Mass spectra of peaks corresponding to ascochitine and pinolidoxin. A, Major adduct ions for ascochitine in culture extracts of AP5, PK4, and ID4A strains were indicated by observed m/z values. B, Major adduct ions for pinolidoxin in culture extracts of MP1 and PMP3 strains were indicated by observed m/z values.

264

Figure S5-2. Structural identification of ascochitine and pinolidoxin. A, 300 MHz 1H NMR in

CDCl3 (ppm) for ascochitine purified from culture of Phoma koolunga (strain PK4). B, 300 MHz

1 H NMR in CDCl3 (ppm) for pinolidoxin purified from culture of Ascochyta pinodella (strain

PMP3).